Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The amygdala and conditioned taste aversion
(USC Thesis Other)
The amygdala and conditioned taste aversion
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
THE AMYGDALA AND CONDITIONED TASTE AVERSION
by
Nicholas N. Foster
__________________________________________________________________
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
May 2011
Copyright 2011 Nicholas N. Foster
ii
TABLE OF CONTENTS
List of Figures
Abbreviations
Abstract
Chapter 1 The Amygdala and Learning
1.0 Aurora
1.1 Classical conditioning
1.1.1 The conditioned emotional response
The fear state
Lesion studies
Interim evaluation: the role of the amygdala
Demonstrating criterion “a”
Demonstrating criterion “b”
Resolving the alternative hypotheses
1.1.2 Contextual fear conditioning
Contextual memory
1.1.3 Others: eyeblink, reward, conditioned attention
Classical eyeblink conditioning
Reward and attention
Context associated with reward
1.2 Instrumental conditioning
1.2.1 Aversive conditioning
Inhibitory avoidance
Role of the amygdala: modulation?
Role of the amygdala: US representation?
Active avoidance
1.2.2 Appetitive conditioning
1.3 Recapitulation and conclusions
Chapter 2 The Amygdala in Conditioned Taste Aversion
2.0 Conditioned taste aversion: Aversive memory of a
special kind
2.1 Some brain areas involved in CTA
The CS Pathway
The US Pathway
CS-US Association
The CR Pathway
2.2 Lesion studies demonstrating participation of the
amygdala in CTA
2.3 The role of the amygdala in CTA
v
vi
vii
1
1
5
7
7
10
15
16
18
23
31
36
38
38
39
41
42
43
44
48
59
63
67
75
80
80
87
87
91
98
100
105
111
iii
Table 2.0 Periods of CTA when a given role must
participate in CTA
2.3.1 What the evidence shows
CS Pathway
US Pathway
CR Pathway
CS Novelty
Modulation
Engram
Summary of the role of the amygdala
Chapter 3 Reconsolidation
3.0 Consolidation and reconsolidation
3.1 Some features and parameters of reconsolidation
Reconsolidation can be induced in a number of
vertebrate and invertebrate species using a
variety of learning tasks
Older memories appear to be more stable and harder
to replasticize than recent memories
Well-trained memories are more stable
Increasing CS re-exposure time can replasticize
more stable memories
The CS re-exposure duration affects whether
reconsolidation or extinction is initiated
The learned behavior is still exhibited up to several
hours after the amnestic treatment is given and
re-exposure to CS occurs
Reconsolidation is similar, but not identical to
consolidation, requiring in some instances
different neural areas and different biochemical
underpinnings
Reconsolidation can be used to enhance memory
too
Summary of the main features
3.2 Ruling out alternative explanations
The amnesia observed after ‘disruption of
reconsolidation’ is due to failure to retrieve an
intact memory
The amnesia is due to extinction, perhaps even a
drug-induced accelerated extinction
The effect is due to neurotoxicity or some other
non-specific effect of the reconsolidation-
disrupting drug
3.3 What is reconsolidation and why does it exist?
113
116
120
124
127
132
139
143
150
152
152
153
153
154
156
157
159
161
162
165
166
167
167
169
172
173
iv
3.4 How is the amygdala involved in reconsolidation?
3.5 What is known about reconsolidation and CTAs?
3.6 Summary
Chapter 4 Methodology
4.0 Overview
4.1 Materials and methods
Animals and husbandry
Fistulation
Behavioral procedures
Histological procedures
Statistical analyses
4.2 Experimental designs
Experiment 1
Experiment 2
Experiment 3
Experiment 4
Chapter 5 Results
5.0 Experiment 1
5.1 Experiment 2
5.2 Experiment 3
5.3 Experiment 4
Chapter 6 Discussion
6.0 Summary of results
6.1 Interpretation of the behavioral results
6.2 Interpretation of the histological results
C nucleus
L nucleus
B nucleus
Assessment of the amygdala model
6.3 Conclusions
References
178
180
184
185
185
186
186
187
188
192
194
195
195
195
195
197
198
198
199
200
201
210
210
212
218
218
219
221
226
229
231
v
LIST OF FIGURES
Figure 1.0 The amygdala of a male Sprague-Dawley rat
Figure 1.1 Model of the neuroanatomical pathways in the
classical fear circuit
Figure 1.2 A possible neuroanatomical model for the
contextual fear conditioning circuit
Figure 1.3 A model of the path of neuromodulator actions
in the B amygdala
Figure 1.4 An overall model of the role of the amygdala in
fear conditioning
Figure 2.0 The CS pathway
Figure 2.1 The US pathway
Figure 2.2 Possible sites of CS-US association
Figure 2.3 Loci of CS-US pathway convergence within the
amygdala
Figure 2.4 The CR pathway
Figure 2.5 A theoretical model of the CS novelty pathway
Figure 2.6 Illustration of a hypothetical model of the
modulatory pathway in the CTA neural circuit
Figure 2.7 A hypothesized model of the engram
Figure 4.0 The orofacial observation chamber
Figure 4.1 Schematics of the behavioral procedures used in
the experiments
Figure 5.0 The results of Experiment 1
Figure 5.1 The results of Experiment 2
Figure 5.2 Behavioral and histological results of
Experiment 3
Figure 5.3 Histological data from Experiment 3
Figure 5.4 The behavioral and histological results of
Experiment 4
Figure 5.5 Histological data from Experiment 4
Figure 6.0 Neuropsychological model of the
reconsolidation-extinction effect
4
20
37
53
78
92
97
101
102
107
134
143
145
189
196
204
205
206
207
208
209
225
vi
ABBREVIATIONS
AB accessory basal nucleus of the
amygdala
AMPA alpha-amino-3-hydroxy-5-
methyl-4-isoxazolepropionic acid
AP area postrema
B basal nucleus of the amygdala
BLA basolateral complex: the basal,
lateral, and accessory basal nuclei
C central nucleus of the amygdala
cAMP cyclic adenosine
monophosphate
Co cortical nucleus of the amygdala
CR conditioned response
CREB cAMP response element binding
protein, a transcription factor
CS conditioned stimulus
CTA conditioned taste aversion
GABA gamma amino butyric acid
IC insular cortex
i.p. intraperitoneal (intra-abdominal)
L lateral nucleus of the amygdala
LiCl lithium chloride
LTP long-term potentiation
M medial nucleus of the amygdala
MITC midline and intralaminar
thalamic complex
NMDA N-methyl-D-aspartate
NST nucleus of the solitary tract
PBN parabrachial nucleus
REX reconsolidation-extinction group
SEX standard extinction group
SUC sucrose-only group
TTX tetrodotoxin
UR unconditioned response
US unconditioned stimulus
VPMpc parvocellular
ventroposteromedial thalamic
nucleus
5-HT 5-hydroxytryptamine (serotonin)
vii
ABSTRACT
Reconsolidation is a temporary plasticity that memories undergo when recalled,
and these memories can be disrupted, usually through pharmacological intervention.
Recently, a purely behavioral treatment was devised that was shown to disrupt
reconsolidation of fear memories (Monfils, Cowansage, Klann & LeDoux, 2009). We
applied a modified form of this treatment to the classically conditioned taste aversion
paradigm, wherein rats received intraoral infusion of sucrose followed by injection of
lithium chloride. During extinction each day, treatment rats received reconsolidation-
extinction (REX), consisting of a brief reminder infusion of sucrose one hour before a
one minute extinction infusion. Control rats received standard extinction (SEX), wherein
they received only the one minute infusion. Aversive orofacial behaviors were tallied
and compared by percentile bootstrap tests. Each rat was extinguished until it had
recovered, and then 5 days later sucrose was given again to test spontaneous recovery,
measured by comparing aversive behaviors on the last extinction day to behaviors
displayed on the spontaneous recovery test. The REX group recovered faster, showing
significantly fewer aversive behaviors by extinction day 4. Both groups showed
spontaneous recovery, but the SEX group exhibited significantly more aversive behaviors
during the spontaneous recovery test. These results indicate that the REX treatment
disrupted the taste aversion memory. A second experiment, conducted similarly to the
first but with an additional unconditioned sucrose-only group, examined expression of the
immediate early gene c-Fos in the subnuclei of the lateral, central, and basal nuclei of the
amygdala. Neural tissue was collected one hour after the first extinction test, allowing
viii
assessment of amygdala participation in reconsolidation. The REX group contained more
c-Fos-positive neurons than the other two groups in the dorsal lateral subnucleus of the
lateral nucleus and the parvocellular subnucleus of the basal nucleus. This activity in the
REX rats may reflect reconsolidation of the CTA memory and activation of an
amygdalocortical pathway that suppresses extinction memory formation, respectively.
1
1
THE AMYGDALA AND LEARNING
1.0 Aurora
In the 1930s, two scientists at the University of Chicago, Heinrich Klüver and
Paul Bucy, were attempting to localize neural targets of the psychedelic drug mescaline
(Neylan, 1997). Using the gross lesion technique of lobectomy, surgical extirpation of an
entire lobe of the brain, they proceeded to remove bilaterally the frontal, parietal, or
occipital lobes of various monkey species. The effects of these lesions were
unremarkable, to the extent that Klüver and Bucy (1939) believed that monkeys with
such lesions would be able to survive in the wild, provided that sections of primary visual
cortex were left in place. However, in one rhesus monkey, Aurora, they removed one,
and then two months later the other of her temporal lobes, the parts of the brain that
underlie the temples. Upon excision of the second temporal lobe, there were dramatic
changes in her behavior. She was possessed by an insatiable curiosity for any object
within her reach or range, a compulsion the famous neurologist Wernicke mysteriously
coined “hypermetamorphosis.” For example, if presented with a screw, she would seize
the screw and immediately bring it to her mouth, where she would gently bite, lick, sniff,
and probe it with her lips. If presented serially with 100 screws, all 100 would be
similarly examined. If she was offered an array of various items, she would orally
investigate each of them in the same manner, regardless of their identity: nails, a lump of
2
wax, a banana, cigarette butts, a glass, and feces. Anything edible so examined was
immediately ingested, even foods normally rejected by intact monkeys such as bacon and
smoked fish. These “oral tendencies” to palpate objects buccally not only applied to
inanimate objects, but to living things as well, such as the fingers of the investigators,
rats, mice, and snakes, which would all be placed into her mouth.
The fact that Aurora would approach a snake, much less put it into her mouth and
gently bite it, was remarkable to Klüver and Bucy. Intact monkeys are terrified of
snakes, and all the monkeys the two scientists studied, including Aurora, were wild
monkeys. They had presumably encountered snakes in the wild, and perhaps some had
even seen another animal fall prey to a snake. This loss of fear towards a natural predator
was actually part of a greater deficit in fear and aggression. The temporal lobectomized
monkeys were much more docile, allowing the experimenters to approach, stroke, and
carry them without protest. When other monkeys bit or attacked them, or made an
aggressive display such as bared teeth, the lobectomized animals were not deterred, nor
did they screech or flee, but instead continued to approach their attackers, trying to
examine them. Indeed, the only time these animals showed any agitation was when an
interesting object lay outside of their reach through the bars of their cages (Klüver &
Bucy, 1939).
This body of signs, including hypermetamorphosis, oral tendencies, and loss of
fear (along with playful, acrobatic “antics” and hypersexuality), now collectively known
as KLÜVER-BUCY SYNDROME, inspired other scientists to investigate the precise neural
geography underlying these characteristics. The temporal lobes contain the amygdala
3
and much of the hippocampus, both of which are large nuclear complexes that lie
ensconced in a mantle of cerebral cortex; the syndrome could result from the loss of any
or all of these structures. Although not all of the symptoms of Klüver-Bucy syndrome
are due to amygdala damage, after some delay (Weiskrantz, 1956), the amygdala (Figure
1.0) was identified as the structure underpinning the fear deficit in Klüver-Bucy
syndrome in monkeys (Aggleton & Passingham, 1981; Meunier, Bachevalier, Murray,
Malkova & Mishkin, 1999; Kalin, Shelton, Davidson & Kelley, 2001). This finding was
later verified in humans, and replicated in cats, pigeons, mice, rats, and rabbits (Ursin,
1965; Blanchard & Blanchard, 1972; Slotnick, 1973; Cohen, 1975; Kapp, Frysinger,
Gallagher & Haselton, 1979; LaBar, LeDoux, Spencer & Phelps, 1995), illustrating the
conservation of this important brain structure across the evolution of a number of species.
Many of the fear deficits identified in these species are disruptions of learned
fear. The Weiskrantz study (1956) implemented two fear learning paradigms, the
conditioned emotional response and an active avoidance task, both of which were
acquired with difficulty by monkeys with amygdala lesions. In humans, temporal lobe
damage disrupted acquisition of a fear-conditioned skin conductance response (LaBar, et
al., 1995). Even the loss of ophidiophobia in Aurora and her cohort may be seen as an
impairment of learning and memory, since monkeys do not fear snakes innately.
Laboratory-bred monkeys will eagerly reach over the head of a live boa constrictor to
obtain a food reward lying just beyond; wild-reared monkeys will immediately withdraw
as far as possible from the snake and never reach for the food. However, if the captive-
bred monkeys view a video of another monkey reacting with distress to the presence of a
4
snake, they will thenceforth display similar fear reactions to snakes (Cook & Mineka,
1990), demonstrating observational fear learning.
Figure 1.0 The amygdala of a male Sprague-Dawley rat. Tissue has been
cut in the coronal plane and stained with neuron-specific NeuN antibody.
The structures labeled are the central (C), lateral (L), basal (B), accessory
basal (AB), medial (M), and cortical (Co) nuclei. The dark phalanx of
neurons along the lower right edge of the section is the piriform cortex.
The Klüver and Bucy monkeys also demonstrated learning deficits in a non-fear
paradigm, Klüver’s FORM BOARD TASK (Klüver & Bucy, 1939). This procedure involves
presenting the monkey a tray with two wells in it; each well is covered by a distinctive
card, one circular, the other square. The monkey is supposed to push aside the circular
card (for instance) and retrieve a food reward from the well below; the square card covers
an empty well, and if it is selected, the whole tray is removed before the monkey can take
the food from the other well. This task takes intact monkeys hardly any time to learn,
and if given a few training trials with a one-well tray, they make zero errors. The
5
lobectomized monkeys selected stimuli randomly at first, but eventually, after hundreds
of trials, they learned to choose the circle, achieving errorless performance. Thus
temporal lobectomy seems to have a widespread effect on learning, impacting both fear
learning and positively reinforced learning as well.
After decades of research following these initial forays, much has been revealed
about the function of the amygdala. The amygdala is indisputably involved in learning
and memory. In particular, the amygdala associates stimuli, such as auditory or visual
cues, with events that are significant to the animal, such as dangerous, nourishing, or
sexual encounters. Later, it responds to those cues and prepares the animal to deal with
the events that those stimuli predict. There are two major ways it prepares the animal,
classically conditioned responses and instrumental responses. Correspondingly, there are
two major output pathways from the amygdala, one from the central nucleus to the
hindbrain mediating classically conditioned responses and one from the basolateral
complex to the forebrain mediating instrumental responses. As a final note, most of the
studies in this chapter involve rats; those that do not will be identified.
1.1 Classical conditioning
Classical conditioning was serendipitously discovered in Russia by the
physiologist Ivan Pavlov, who was studying the secretion of digestive juices in dogs.
Pavlov noticed that his dogs would salivate copiously at the sight of food, before it was
actually placed into their mouths; he observed that the salivary reflex could also be
elicited in his dogs by the odor of the food, the sight of the food dish, the sound of the
6
approaching footsteps of the caretaker who fed them, or any other arbitrary stimulus that
was paired with the presentation of food. Identifying this curiosity as a phenomenon that
lends itself extremely well to the physiological investigation of psychology, he went on to
elaborate many of the properties of classical conditioning in great detail (Pavlov, 1927).
PAVLOVIAN or CLASSICAL CONDITIONING is a form of learning in which a CONDITIONED
STIMULUS (CS), a “neutral” stimulus or cue, is presented in close temporal proximity to an
UNCONDITIONED STIMULUS (US), a stimulus that naturally elicits a reflexive response.
After one or more such CS-US pairings (termed “conditioning” or “acquisition”), the CS
attains the ability itself to evoke the reflexive response of the US. As an example, a
rabbit, when puffed in the eye with a jet of air, will contract its nictitating membrane to
protect its eye. The air puff (the US) automatically elicits the eyeblink response, a
reaction called the UNCONDITIONED RESPONSE (UR). In a classical conditioning paradigm,
a CS such as a tone (or sometimes a light) is presented to the rabbit shortly before the air
puff is given. After a number of CS-US pairings, the tone CS becomes able to trigger the
eyeblink response (named the CONDITIONED RESPONSE, CR), even in the absence of the
US. Furthermore, if during training the CS-US delay interval is consistent, say 300 ms,
the rabbit learns to time its eyeblink to block the air puff (Koekkoek, et al., 2003). Yet
classical conditioning applies not only to such adaptive and overt behaviors as blinking
and drooling, but to less obvious reflexes such as heart rate changes, endocrine responses
like insulin or adrenaline release, and even emotional responses like fear.
7
1.1.1 The conditioned emotional response
In CLASSICAL FEAR CONDITIONING, a neutral CS is paired with a fear-inducing US,
which allows the CS to invoke a fear state. Typically, the animal is placed into a
conditioning box that contains a light bulb (or perhaps a speaker) and a wire grid floor.
Several seconds after the light turns on, an electric shock is delivered to the animal
through the electrified floor grid. The shock evokes two main unconditioned responses:
the first is a jump response, i.e., the animal leaps off of the electrified floor; the second is
a state of fear. After a few CS-US pairings, the light CS, which when presented alone
would merely rouse the animal’s attention, now invokes a state of fear, a CONDITIONED
EMOTIONAL RESPONSE. Testing for the conditioned emotional response involves turning
on the light and quantifying the behavioral response that the animal makes.
Testing typically occurs in either the home cage or a test box that is dissimilar to
the training box. After fear conditioning takes place, the CR consists of many of the
same responses as the UR, although there are some differences. In some cases the animal
may display bradycardia, whereas the UR entails increased heart rate. Interestingly, the
CS does not elicit a jump response, which was one of the URs and which could possibly
allow the animal to avoid the painful shock if properly timed; instead, the CS causes the
animal to freeze, which is an element of the fear response. Thus the animal learns that
the CS predicts an unavoidable painful stimulus, and it responds by timorously awaiting
the impending electric shock.
The fear state. Given that so many of the studies examining the role of the
amygdala in learning utilize fear-learning paradigms, a description of the fear state is
8
warranted. Fear in rats can be evoked in an unconditioned manner by many stimuli, like
pain stimuli such as electric shock, the presence of certain predators (e.g., cats), and
environmental contexts such as bright open spaces (Davis, 1992). The fear state is a
constellation of responses (Davis, 2000), with autonomic, behavioral, and cognitive
components. The autonomic responses are reflective of activation of the sympathetic
nervous system, commonly known as the FIGHT-OR-FLIGHT RESPONSE. These include the
release of the stress hormones corticosterone (or cortisol), epinephrine, and
norepinephrine; cardiovascular changes like increased blood pressure and tachycardia,
constriction of visceral blood vessels and dilation of vessels in somatic musculature;
respiratory changes; perspiration, urination, defecation, piloerection, and pupillary
dilation. These changes in concert can prepare the organism for intense physical activity
(Smith & DeVito, 1984; LeDoux, 1987).
Behavioral fear responses can vary with the circumstances, showing flexibility
and something of an environmental dependency (Bolles & Fanselow, 1980). For instance,
one common behavioral response is a passive defensive reaction called freezing, in which
the rat suppresses all movements except for those necessary for respiration; such a
response could help evade detection by a nearby predator. However, if there is an avenue
of flight from the dangerous stimulus or situation, the rat instead may make an escape or
avoidance response. If a predator is in close proximity and escape is impossible, the rat
may make an attack response; an aggressive response may also occur if the rat stands a
reasonable chance of victory, such as when fighting a conspecific. Thus behavioral fear
9
responses can vary in ways that are adapted to surviving the particular dangerous
circumstance.
Cognitive changes also accompany fear induction. One critical change is
increased vigilance, a state of heightened generalized attention (or “alerting” attention) in
which an animal’s senses can become more responsive. Electroencephalographic
recordings show greater cortical desynchrony during a fear state, indicative of greater
neural processing pursuant to increased attention (Whalen, Kapp & Pascoe, 1994).
Related to this is the FEAR-POTENTIATED STARTLE RESPONSE. Rats and other animals show
a sensitivity to sudden loud noises by exhibiting an involuntary somatic start, e.g., the
sound of a balloon bursting can cause people to flinch. Brief electric shock and fear-
inducing stimuli have been shown to enhance the amplitude of flinch in response to
unexpected loud noise bursts (Davis, 1989; Hitchcock, Sananes & Davis, 1989; Sananes
& Davis, 1992). Thus, activation of a lightbulb that has previously been paired with
shock will cause an animal to show a bigger jolt to a sudden loud noise than it would in
darkness. Another cognitive change is hypoalgesia, wherein activation of the fear state
reduces pain perception (Terman, Shavit, Lewis, Cannon & Liebeskind, 1984). This is
believed to have survival value, since hypoalgesia could prevent pain from an injury from
interfering with behaviors necessary for dealing with a life-threatening situation, such as
running from a predator on a broken foot. And finally, fear is of course characterized by
feelings of doom.
The multifaceted fear state, comprised of almost all of these responses, can be
evoked when a previously-conditioned CS is presented. Thus almost any of the
10
individual fear responses can be used to assess fear conditioning in animals (except for
feelings of doom, which must be inferred based upon the behaviors they motivate) and
many of them have been utilized in both classical and instrumental conditioning
(reviewed in Davis, 2000). Also, note the fact that most of the descriptions above are of
the unconditioned fear response, evoked by unconditioned stimuli, although conditioned
fear responses are for the most part exactly the same. While lesion of the amygdala can
interfere with some expressions of unconditioned fear
1
, conditioned fear responses, since
they are learned, will be the focus of this paper.
Lesion studies. The most straightforward and oft-used technique for assessing
the involvement of a particular brain area in a particular behavior is the lesion method. If
the amygdala is an integral part of the neural circuit governing fear learning, then its
destruction should affect performance in a fear learning paradigm, probably in a
deleterious way.
Lesions of the amygdala have, in fact, been shown to impair classical
conditioning of fear, with demonstrations of deficits in autonomic, behavioral, and
cognitive components of fear responses. Several autonomic features of fear have been
assessed after conditioning. In humans, conditioning of the galvanic skin response (an
index of perspiration) is impaired after unilateral or bilateral amygdalectomy to treat
seizure disorders (LaBar, et al., 1995), and after ischemia-induced bilateral amygdala
1
In a classic fear study by Blanchard and Blanchard (1972), amygdalectomized and control rats were
exposed to a Thorazine-sedated yet still-conscious cat, an animal that unconditionally elicits fear in rats.
While control rats spent much of the test session freezing and avoiding the cat, which sat in serene repose
in the corner of the test chamber, amygdaloid rats approached and contacted the cat. One rat climbed onto
the cat’s back, then onto its head, then proceeded to nibble on the cat’s ear. At this point, the cat seized the
rat in its jaws, briefly shook it, and then released it. The rat, unperturbed, scaled back onto the cat.
11
lesion (Bechara, et al., 1995). In rats, extensive neurotoxic lesions (made using neuron-
specific poisons, which do far less collateral damage than the electric lesion technique) of
the amygdala, encompassing the central (C), basolateral complex (BLA), and medial (M)
groups more or less, completely blocked conditioned corticosterone release in response to
white noise bursts that had been previously paired with footshock (Goldstein,
Rasmusson, Bunney & Roth, 1996), while baseline corticosterone levels were not
statistically different from controls. The number of fecal boluses excreted by these
lesioned animals during a 30-min test session was also fewer. These autonomic deficits
occurred whether surgical intervention was before or after training. Lesion of the dorsal
part of the amygdala (consisting of the C and lateral [L] nuclei) approximately two weeks
before fear conditioning impaired conditioned blood pressure increases when a tone CS
was sounded (LeDoux, Sakaguchi, Iwata & Reis, 1986). Their responses were similar to
a group of intact pseudoconditioned control rats, who received tone and shock stimuli in
a random pattern, and thus did not show pressor reflexes to the tone.
Heart rate conditioning too is affected by amygdala lesion, although the nature of
the conditioned cardiac response varies across species. In rabbits, like many other
species, the natural, unconditioned cardiac reflex to electric shock is accelerated heart
rate, but the conditioned cardiac response is decreased heart rate. In one study (Kapp, et
al., 1979), groups of rabbits with and without C nucleus lesions received multiple
pairings of a tone CS and paraorbital shock (i.e., shock to the eyelids). Partial or total
electric ablation of the C nucleus substantially reduced (but did not eliminate) the
bradycardic CR. Presentation of a novel stimulus alone (without a subsequent shock)
12
elicits an unconditioned bradycardic response, called the HEART RATE ORIENTING RESPONSE,
which attenuates as the rabbit habituates to the stimulus; C nucleus lesion had no effect
upon this phenomenon, demonstrating the preservation of at least one kind of heart rate
reactivity. Thus, C lesion apparently causes a specific deficit in fear conditioned
bradycardia. Pigeons (Cohen, 1975) and rats (Sananes & Campbell, 1989) exhibit fear
conditioned tachycardia, but the deleterious effect of amygdala lesion on conditioning
this response is the same as in rabbits. In the study by Sananes and Campbell (1989), rat
pups received pairings of a grape odor CS and subdermal electrocution. Animals that had
received C nucleus lesions had no elevation in heart rate in response to the grape odor. In
pigeons, electric obliteration of the archistriatum (the avian homologue of the amygdala)
prevented acquisition of a tachycardic CR to a light CS paired with footshock (Cohen,
1975). In all, a number of autonomic conditioned fear responses are impaired by large
amygdala lesions and by more restricted C nucleus-specific lesions.
Behavioral conditioned responses are also affected by amygdala lesion, and one
of the most widely examined behaviors is freezing (probably because it is tractable to
experimentation). Large NMDA lesions
2
engrossing the C, BLA, and parts of the M
nucleus completely blocked freezing responses to a brief white noise CS (Goldstein,
Rasmusson, Bunney & Roth, 1996). Electrolysis of the C and L nuclei of the dorsal
amygdala (LeDoux, Sakaguchi, Iwata & Reis, 1986) or just the L nucleus (LeDoux,
2
N-methyl-D-aspartate (NMDA) is an amino acid that selectively binds to and activates a particular
glutamate receptor subtype called the NMDA receptor. The activation of this receptor opens a channel in
the neuronal membrane that allows Na
+
and Ca
2+
to enter the cell, directly causing an excitatory
postsynaptic potential. Infusion of a high concentration of NMDA into a part of the brain causes
overexcitement of neurons in that area, leading to excitotoxic death. Axons passing through the area are
spared. Ibotenic acid, a less specific glutamate receptor agonist, is likewise used as a neurotoxin.
13
Cicchetti, Xagoraris & Romanski, 1990) roughly two weeks prior to fear conditioning
impaired conditioned freezing responses; responses were assessed both during
exploration of a new cage and during drinking by parched rats, situations that motivate a
rat to move. Their behavioral responses did not differ from a group of pseudoconditioned
control rats, demonstrating an absence of the freezing CR in the lesion groups.
Neurotoxic lesion of the BLA either prior to or following auditory fear conditioning also
severely curtails time spent freezing during subsequent tone-only testing (Maren,
Aharonov & Fanselow, 1996). Assessment of locomotor activity in these rats revealed
no difference from sham-lesioned controls, ruling out a motor hyperactivity disorder that
interfered with freezing (confirmed by Goosens & Maren, 2001). Several investigations
have more finely localized the amygdalar nuclei responsible for this form of learning,
finding that the C and L nuclei are indispensable. Separate lesions of the C and L nuclei,
made with electrolysis or neurotoxins, found that ablation of either nucleus disrupts
conditioning of the freezing response (Killcross, Robbins & Everitt, 1997; Amorapanth,
LeDoux & Nader, 2000; Goosens & Maren, 2001; Nader, Majidishad, Amorapanth &
LeDoux, 2001). These studies found no disruptive effect on learning when lesions
targeted the basal (B), accessory basal (AB), both B and AB, or M nucleus (though see
Selden, Everitt, Jarrard & Robbins, 1991). Together, these studies elucidate a specific
amygdalar substrate, the C and L nuclei, for conditioned fear-induced freezing.
Amygdala damage is also disruptive to fear conditioning when using cognitive
measures, such as fear-potentiated startle. Excitotoxic NMDA lesions encompassing the
L and B nuclei completely blocked conditioned fear-induced enhancement of the startle
14
reflex with a visual CS, effects occurring whether lesions were made before or after
training (Sananes & Davis, 1992). This was probably not due to decreased shock
sensitivity, since lesioned and control animals responded to footshock with jump
reactions of comparable amplitudes. This study also shows that an intact C nucleus alone
is insufficient to mediate normal expression when the B and L are destroyed after
training. Yet elimination of the C nucleus can also block fear-potentiated startle.
Specifically, loss of the C nucleus virtually erased fear conditioning acquired prior to
surgery (Hitchcock & Davis, 1986; Kim & Davis, 1993), but did not prevent a modest,
impaired relearning; if, however, the C was expunged before any fear training took place,
no potentiation of startle could be exacted, even when the shock current was nearly
doubled during training. Incidentally, the unconditioned jump reaction to footshock was
unaffected in these lesioned animals, and in fact the magnitude of the jump response
increased monotonically across pre- and post-lesion reactivity tests; this suggests that the
foot pain sensory pathway was intact in these animals, as was the jump motor pathway,
and that some non-associative learning even took place in this US-UR pathway.
Furthermore, the deficit expressed by these animals was not due to damage of the visual
CS pathway, which would have prevented the animals from seeing or responding to the
fear-inducing light stimulus; visual pathways were intact, because a flash of light
suppressed startle responses normally in a visual PREPULSE INHIBITION paradigm
3
(Hitchcock & Davis, 1986). Potentiation of the startle reflex by conditioned fear to a
tone CS is also disrupted by lesion of the C nucleus (Hitchcock & Davis, 1987). Thus the
3
Prepulse inhibition: Attenuation of unconditioned startle amplitude by a visible light flicker that leads the
loud noise by a short interval (60 ms, more or less).
15
amygdala participates in cognitive aspects of conditioned fear. In total, the studies
presented in this section provide compelling evidence that the amygdala contributes to
fear conditioning in some way.
Interim evaluation: The role of the amygdala. As the preceding section
illustrates, amygdala lesions appear to corrupt conditioned fear with virtually every type
of CR measured, cutting across the autonomic, behavioral, and cognitive categories of
fear responses. This suggests that the amygdala may be at the center of the fear
responding circuit. Furthermore, amygdaloid fear deficits are obtained no matter what
CS is used, be it tone, light, or odor (as well as environmental context, see Section 1.1.2).
This implies that there is a confluence of sensory information from many modalities in
the amygdala. It is unlikely that all of these myriad sensory and motor pathways merely
pass through the amygdala on their way to some other all-important neural hotspot,
particularly so given that the neurotoxic lesion data indicate that neurons intrinsic to the
amygdala constitute part of the fear circuit. Hence a picture is beginning to emerge of the
amygdala’s role in classical fear conditioning: engram, a place where a memory is
encoded. According to Lavond, Kim, and Thompson (1993), to qualify as the repository
of a Pavlovian engram, a brain area “must a) be a locus of convergence for the
conditioned and unconditioned stimulus systems and b) produce a unique output that is a
function of proper (contingent) pairing of these systems.” The foregoing evidence
suggests that the amygdala may satisfy these two conditions, but a convergence of data
from a variety of methodologies is necessary to make this assertion with more
confidence.
16
Demonstrating criterion “a”. In seeking to establish the amygdala as a place
where CS and US sensory pathways merge, neuroanatomical data proves invaluable.
First of all, in accordance with the first half of the requirement, sensory information
regarding the CS must reach the amygdala. In fact, visual and auditory information do
target the amygdala, specifically the L nucleus (Doron & LeDoux, 1999). The thalamus
is the primary sensory relay nucleus that modulates and distributes ascending sensory
information of both internal and external
4
origin to the cerebral hemispheres
(Groenewegen & Witter, 2004). Several auditory thalamic nuclei provide afferent
projections to the L amygdala, namely the medial region of the medial geniculate
nucleus, posterior intralaminar nucleus, and suprageniculate nucleus (LeDoux, Farb &
Ruggiero, 1990). Thalamic visual areas also innervate the L amygdala, with afferents
arising from the lateral posterior nucleus (the rodent homologue of the pulvinar), but not
the lateral geniculate nucleus (Turner & Herkenham, 1991; Pitkänen, 2000). These
auditory and visual areas do not directly target the B nucleus (Turner & Herkenham,
1991) or the C nucleus (LeDoux, et al., 1990), suggesting that the L nucleus is the
primary amygdalar target of these sensory paths. Plus, auditory and visual cortices also
project heavily to the L as well as the C nucleus (McDonald, 1998; Pitkänen, 2000),
although in the case of auditory fear conditioning these afferents may provide a
somewhat redundant sensory input to the fear conditioning circuit, since auditory fear
conditioning does not require the auditory cortex
5
(LeDoux, Sakaguchi & Reis, 1984).
4
Except for the olfactory senses which by and large are extrathalamic.
5
Still, the auditory cortical pathway is probably not completely redundant, since cortex is capable of more
complex processing. Hence cortex may be critical for relaying complex CSs. On the other hand, simple
17
Olfactory sensory information may reach the L indirectly through a projection from the
piriform cortex, the primary olfactory cortex (Ottersen, 1982). The L may also receive
olfactory information via intraamygdaloid afferents from the medial (M) and cortical
(Co) nuclei (Pitkänen, 2000), which are amygdaloid nuclei that process olfactory input
(Swanson & Petrovich, 1998). In sum, the L nucleus is the primary recipient of
ascending visual and auditory sensory information in the amygdala, and it has inputs
from important olfactory-related structures as well (Figure 1.1).
The second half of condition “a” is that sensory information regarding the noxious
US must be conveyed to the amygdala. The main thalamic areas for nociceptive
information are the ventral posterolateral nucleus, central lateral nucleus, nucleus
submedius, and the posterior thalamic nuclear group (Willis, Westlund & Carlton, 2004).
In the posterior thalamic nuclear group, pain-responsive fibers of the spinothalamic tract
terminate in the suprageniculate nucleus, medial portion of the medial geniculate nucleus,
posterior intralaminar nucleus, posterior nucleus, and posterior triangular nucleus (Zhang
& Giesler, 2005). As was already established in the preceding paragraph, many of these
nuclei project to the L nucleus of the amygdala, so fulfilling the requirement that the L
nucleus receive US pathway afferents (Figure 1.1). Note that many of these same
thalamic nuclei also receive auditory and visual afferents; verily, the posterior thalamus
could potentially accommodate a Pavlovian fear engram. This hypothesis is evaluated
and rejected below. Additionally, some pain afferents from the thalamus and other
structures reach the C nucleus of the amygdala. The lateral parabrachial nucleus projects
CSs like pure tones can be effectively mediated by either the thalamo-amygdala or cortico-amygdala
auditory pathways (Romanski & LeDoux, 1992).
18
to the C amygdala (Bernard, Alden & Besson, 1993), as do some offshoots of the
spinohypothalamic pathway (Cliffer, Burstein & Giesler, 1991), and fibers from the
posterior nucleus of the thalamus (LeDoux, et al., 1990). Altogether, the L amygdala is
in an excellent position to form an association between conditioned and unconditioned
stimuli, since this nucleus receives branches from external sensory and nociceptive
pathways.
Of course, just because both CS and US pathways reach the L amygdala does not
necessarily mean that they converge. However, electrophysiological data reveals that in
fact there is a population of neurons in the L that is the target of both stimulus pathways.
Single unit recording of neurons in the L amygdala in anesthetized rats detected neurons
that responded with short latency (15 ms) to a simple auditory stimulus, clicks. Many
units in the L also responded to footshocks (17 ms latency). Importantly, a high
proportion of units (23 of 26) that responded to auditory stimulation also responded to
noxious stimulation (Romanski, Clugnet, Bordi & LeDoux, 1993), and the authors’
statistical analysis of the recorded waveforms indicated that the responsive units were
single neurons. Since a set of neurons in the L nucleus is postsynaptic to both CS and US
pathways, condition “a” is satisfied.
Demonstrating criterion “b”. Criterion “b” of Lavond et al. (1993) stipulates
that, should the L amygdala be the seat of the fear engram, then its activity should change
in a way that reflects the independent and dependent parameters of conditioning. We
interpret this to mean that over the course of conditioning, the L should become
19
responsive to the CS, and it should activate (at least a subset of) the behavioral,
autonomic, and cognitive components of fear.
Electrophysiological data reveals that the L amygdala does seem to fulfill the
condition that its reactivity must change as a result of conditioning. Recording of unit
activity in the L nucleus found that a subset of auditory responsive units demonstrated
potentiated electrical responses during and after tone-shock fear training (Quirk, Repa &
LeDoux, 1995; Rogan, Stäubli & LeDoux, 1997). Additionally, some initially
unresponsive units became tone-sensitive during conditioning, suggesting that some
synaptic activity that was below the detection threshold became strengthened across
conditioning. It was observed in many cases during paired training that unit responses
began potentiating within the first few tone-shock pairings, culminating in a five-fold
increase in stimulus-elicited activity measured during the first block of tone-only
EXTINCTION
6
trials (Quirk, et al., 1995). Over the course of tone-only extinction trials, the
potentiated unit responses returned to their pre-conditioning baseline, in a manner that
corresponded to the extinction of tone-evoked freezing (Rogan, et al., 1997). Unpaired
presentation of the CS and US does not alter auditory-evoked potentials. These studies
demonstrate the first half of criterion “b”, that the L amygdala becomes responsive to the
CS after contingent CS-US pairings, but not after non-contingent pairings.
6
Extinction: After training, the CS is administered without the US, to assess the strength of acquisition.
With repeated CS exposure in this manner, the strength of the CR it elicits will decay.
20
Figure 1.1 Model of the neuroanatomical pathways in the classical fear circuit. Olfactory projections to the
amygdala are not depicted. PBN, parabrachial nucleus; Thal, thalamus.
However, activities measured throughout the BLA also show conditioning-
induced responses to a tone CS (Muramoto, Ono, Nishijo & Fukuda, 1993), although the
average response latency was approximately 50 ms after tone onset, much longer than the
mean 20 ms response latency recorded from neurons of the dorsal L nucleus (Quirk, et
21
al., 1995). Similarly, in rabbits, C amygdala response latencies to the CS are on the order
of 30-50 ms (Pascoe & Kapp, 1985). This could indicate that the B, AB, and C nuclei are
downstream of L nucleus activity. In fact, anatomical data supports this conception.
Neurons of the L nucleus project to the B, AB, and to the C nucleus (Pitkänen, et al.,
1995). The L projection to the C nucleus is not reciprocated, i.e., it is one-way
(Jolkkonen & Pitkänen, 1998), confirming the downstream relationship of the C to the L.
The C nucleus has traditionally been considered the main output station of the
amygdala, and it has descending projections to numerous areas in the brainstem involved
in many of the various autonomic, behavioral, and cognitive components of the fear state
(Pitkänen, 2000; Sah, Faber, de Armentia & Power, 2003). For autonomic control, the C
nucleus innervates both the lateral hypothalamus (Rosen, et al., 1991), through which it
can affect blood pressure (LeDoux, Iwata, Cicchetti & Reis, 1988), and the motor nucleus
of the vagus nerve (Rosen, et al., 1991), which can modulate heart rate (in rabbit
bradycardia conditioning: Kapp, Pascoe & Bixler, 1984). The paraventricular
hypothalamus also receives an input from the C nucleus (Rosen et al., 1991), which gives
the C the ability to activate the hypothalamic-pituitary-adrenal axis, thereby stimulating
corticosterone release (Senba & Ueyama, 1997). For behavioral control, the C nucleus
projects to the periaqueductal grey (Rosen, et al., 1991), an area known to mediate the
freezing response (LeDoux, Iwata, Cicchetti & Reis, 1988). For cognitive control, the
projections of the C nucleus to the noradrenergic locus coeruleus (Danielson, Magnuson
& Gray, 1989) and the cholinergic basal forebrain (Grove, 1988) may stimulate fear-
induced vigilance (Whalen, Kapp & Pascoe, 1994). The C also sends an efferent
22
projection to the nucleus reticularis pontis caudalis, a brainstem nucleus forming a link in
the acoustic startle pathway (Rosen, et al., 1991); this connection allows the C to
potentiate the startle reflex (Hitchcock & Davis, 1991). These descending fiber tracts
(summarized in Figure 1.1) put the C amygdala in an excellent position to activate all the
components of fear in a coordinated fashion. In fact, electrical stimulation of the C has
been demonstrated to potentiate startle (Rosen & Davis, 1988), trigger pressor responses
(Iwata, Chida & LeDoux, 1987), and elicit freezing, increased respiratory rate, pupillary
dilation, and cortical desynchrony (in rabbits, Applegate, Kapp, Underwood & McNall,
1983; Kapp, Supple & Whalen, 1994). Similarly, in humans, amygdala stimulation
7
has
been reported to increase blood pressure, accelerate respiratory rate, increase pupillary
dilation, and importantly it has caused feelings of fear and anxiety (Chapman, et al.,
1954). Given this body of anatomical and activational evidence, it is logical to conclude
that the C amygdala constitutes the head of the fear circuit CR pathway.
To recap, there are data demonstrating conditioning-induced changes in L nucleus
activity, including responsivity to the CS that precedes activity in other amygdalar nuclei;
one-way L nucleus projections to the C nucleus, the nexus of the CR pathway; and lesion
studies described previously which illustrate the necessity of the L nucleus in simple fear
conditioning. Although a demonstration that stimulation of the L amygdala activates the
fear response would complete the picture (to our knowledge no such study has been
published), from the studies laid out here it can be deduced that the L amygdala can
7
Electrodes were implanted in the amygdalae of 5 epileptics in order to locate their seizure foci. Multiple
electrodes were implanted, but because of a lack of both neuroimaging technology and, of course,
histological verification, the exact location of stimulation could not be determined.
23
activate the CR pathway. Therefore it is reasonable to consider the second proviso of
condition “b” to be satisfied. All together, then, the two criteria that must be fulfilled in
order to consider a brain area to be the locus of an engram, as specified by Lavond et al.
(1993), have been met.
Still, some have been left with vexing doubts (Cahill, Weinberger, Roozendaal &
McGaugh, 1999). Perhaps the L amygdala meets the engram criteria, but acts merely to
consolidate long-term memory elsewhere, in the actual locus of the engram, while having
itself no endogenous memory storage. Or perchance some other brain area, the thalamus
for instance, which also meets the two criteria for an engram (this evidence was not fully
elaborated above), is the genuine site of the engram, and thalamic activity during and
after conditioning is speciously reflected in the L amygdala, which is downstream of the
thalamus. However, as shall be illustrated next, both of these plausible alternatives are
refutable.
Resolving the alternative hypotheses. The first alternative hypothesis is that the
amygdala consolidates classical fear memory elsewhere. If true, then the participation of
the amygdala would be critical during and for some time after conditioning, and its
necessity would wane over time as the fear memory became consolidated in the distally
located engram. At some point after conditioning, then, the amygdala would become
completely dispensable, and its destruction would not hinder expression of the fear
memory. However, post-training lesions of the amygdala virtually eliminate fear-
conditioning no matter the interval between training and surgery. Excitotoxic lesions of
the BLA complex made 1, 14, or 28 days after auditory fear conditioning (Maren, et al.,
24
1996), and NMDA lesions encompassing the L and B nuclei made 3 days following
visual fear conditioning (Sananes & Davis, 1992), completely blocked expressions of fear
at all intervals. Even post-conditioning BLA lesions made 16 months following
acquisition disrupted expression of a freezing CR, with no apparent abatement of the
memory deficit when compared to animals lesioned just one day following acquisition
(Gale, et al., 2004). It strains credulity that the amygdala so critically modulates memory
formation 16 months after the learning event.
Further evidence from a functional (reversible) lesion study also leads to the same
conclusion. In said study (Wilensky, Schafe & LeDoux, 2000), pre-training inhibition of
neural activity in the L amygdala (and likely the B as well) with the inhibitory GABA
A
receptor agonist muscimol prevented freezing responses to a tone CS when tested 24 h
after single trial training. This ability was dose dependent, with no impairment caused by
0.001 nmol muscimol, a partial effect from 0.088 nmol, and near complete blockade of
learning with 0.44 nmol. Thus the efficacy of the drug in this area was demonstrated.
However, if the amygdala acts only to modulate memory formation, then inactivation of
the amygdala just after training should also produce profound deficits, for non-
consolidated memory lasts only several hours (Davis & Squire, 1984), so modulation of
memory would likely take place in that time frame. Yet muscimol had no effect on
conditioning when administered just after the training trial (Wilensky, et al., 2000). Thus
the notion that the role of the amygdala is restricted to modulation of memory storage
elsewhere is not supported by these data.
25
The second alternative hypothesis is that the associative memory for classical fear
conditioning is stored in another brain area, such as the thalamus, and that thalamic
connections with the L amygdala have led to the mistaken conclusion that the amygdala
contains the engram. Rebuttal of this hypothesis will take two approaches: most effort
will go towards presenting evidence that memory is stored in the L, insofar as current
understanding of physiological substrates of memory allows; but also, the thalamus
specifically will be ruled out as the hold of the engram. In addressing the issue of
memory storage in the L amygdala itself, first recall that many neurons in the L nucleus
respond to both the CS and US, and these cells change their responsiveness to the CS
after conditioning. The changes in the activity or responsiveness of these neurons would
likely be due to the induction of LONG-TERM POTENTIATION, the superstar of physiological
models of learning and memory (Malenka & Nicoll, 1997). Therefore, establishing that
fear conditioning-induced long-term potentiation occurs in the L amygdala, and that
disruption of this process impairs classical fear memory, will leave little room to argue
against a role for the L in fear memory storage.
In fact, long-term potentiation induced by fear conditioning has been reasonably
demonstrated in the L amygdala. As previously described, an experiment conducted by
Rogan, Stäubli, and LeDoux (1997) entailed auditory fear conditioning in rats with
recording electrodes chronically implanted in the L amygdala, yielding the detection 24 h
later of increased field potentials that were CS-evoked and were proportional to the
incidence of freezing behavior. In a dovetailing study by McKernan and Shinnick-
Gallagher (1997), it was shown that this potentiation results at least in part from
26
strengthening of the thalamo-amygdala pathway. Rats underwent either standard
auditory fear conditioning or pseudoconditioning, followed 24 h later by extraction of
neural tissue samples containing the auditory thalamus and the amygdala. In vitro
stimulation of the thalamo-amygdala pathway elicited potentiated current responses in
neurons of the L nuclei of conditioned but not pseudoconditioned rats. These potentiated
currents were shown to be mediated in large part by AMPA glutamate receptors, a
property that is consistent with long-term potentiation. Furthermore, a different pathway,
the piriform-amygdala pathway, was not potentiated in these samples, indicating that the
fear circuitry was specifically potentiated—as per the input specificity property of long-
term potentiation. Together, these two experiments are believed to constitute one of the
first demonstrations of learning-induced long-term potentiation in vivo (Malenka &
Nicoll, 1997).
However suggestive the apparent correlation between fear conditioning and long-
term potentiation in the L amygdala, it is insufficient to establish a causal relationship. If,
on the other hand, disruption of long-term potentiation similarly disrupted classical fear
memory, then that would complement the correlational findings and verify the necessity
of amygdalar long-term potentiation in fear memory. A variety of data substantiate this
necessity. For instance, infusion of AP5, an NMDA glutamate receptor antagonist that
prevents the induction of long-term potentiation, into the BLA complex before fear
conditioning prevented acquisition of conditioned freezing (Miserendino, Sananes, Melia
& Davis, 1990; Lee & Kim, 1998; Davis, 2000). When administration of a high dose of
AP5 (25 nmol) was interposed between conditioning and testing (roughly one week after
27
training, one week before testing), startle responses were potentiated normally during the
test phase, showing that AP5 had no lingering toxic effects (Miserendino, et al., 1990).
Furthermore, when acquisition was carried out normally, injection of an effective dose of
AP5 (12.5 nmol) just prior to testing had no effect upon the expression of fear-potentiated
startle. Thus NMDA receptor activation in the BLA is essential for the acquisition, but
not the expression, of conditioned fear. This is consistent with a necessary role for long-
term potentiation in conditioned fear.
Protein synthesis is another essential substrate of long-term potentiation. More
correctly, there are two general phases of long-term potentiation, early and late; it is the
late phase that requires new protein synthesis. Similarly, long-term memory formation
requires protein synthesis, while short-term memory (lasting several hours) does not
require new protein manufacture (Davis & Squire, 1984). As such, the L amygdala
should be sensitive to blockade of protein synthesis, particularly when tested at long- but
not short-term intervals. This is the case (Schafe & LeDoux, 2000). Immediately
following a one-trial auditory fear conditioning session, rats were infused through
cannulae aimed at the L and B nuclei with various doses of anisomycin, an antibiotic that
prevents protein synthesis. Twenty-four hours later they were placed into a testing
chamber and given tone-only tests. Anisomycin attenuated freezing in an apparently
dose-dependent manner, with the highest dose (62.5 µg) significantly reducing time spent
freezing, the moderate dose (6.2 µg) having a lesser, non-significant effect, and the
lowest dose (0.62 µg) having an effect comparable to vehicle infusion. In a second
experiment, rats were trained as above, then infused with anisomycin (62.5 µg) or
28
vehicle, and then subjected to both short-term memory tests 4 h later and long-term
memory tests 24 h later. At the early test, all rats exhibited strong freezing responses
when the tone CS was replayed; during the later test, the control rats maintained their
high level of freezing, while the drug-infused rats froze significantly less. Thus long-
term memory was impaired while short-term memory was intact. This was not due to
drug-induced state-dependent learning, because when a group of rats was administered
anisomycin both immediately after the training trial and 4 h before the late test (i.e., early
and late testing occurred in the same drug state), they behaved exactly the same as the
single-infusion anisomycin group, demonstrating poor recall. Additionally, when drug-
treated rats were subjected to drug-free retraining a week later, fear conditioning was
acquired normally, proving that anisomycin had no lasting deleterious effect in these rats.
In sum, the results indicate that protein synthesis in the L (and/or B) nucleus of the
amygdala serves an important transient function in the formation of classical fear
memory, findings that are consonant with an obligatory role for long-term potentiation in
the L amygdala.
While these findings support the notion that long-term potentiation in the L
amygdala is the physiological foundation of classical fear memory, they may also support
another possibility. Specifically, Cahill and colleagues (1999) posited that it is the
medial geniculate nucleus of the thalamus that learns, and the amygdala merely reads-out
its activity; moreover, any manipulation of the amygdala simply acts to interfere with
learning in the thalamus. This hypothesis was rebutted in an excellent study by Schafe,
Doyere, and LeDoux (2005). Rats were implanted with cannulae in the L amygdala and
29
recording electrodes in the L amygdala and the medial geniculate nucleus of the thalamus
(the auditory thalamus). The authors’ design was to impair consolidation of long-term
potentiation in the L nucleus while simultaneously recording neural activity in the L and
the medial geniculate. Shortly before tone-shock training, rats received intraamygdala
infusion of either vehicle or an inhibitor of MAPK, a kinase that is critical to the
consolidation of both long-term memory and long-term potentiation (Thomas & Huganir,
2004). All rats were given tone-only testing at 3 h and 24 h after training, which allows
assessment of short-term and long-term memory, respectively. During the early test, both
groups demonstrated freezing to the tone CS, and augmented field potentials were
recorded in the L nucleus, indicating that learning and neuroplastic change had occurred
in both groups. At the late test 24 h later, control animals retained their high levels of
freezing and large field potential responses; drug-infused rats showed significantly less
freezing and diminished field potential responses when compared to controls and to their
own early test levels. Additionally, there was a significant correlation in the drug-infused
group between the magnitude of the field potential retained in the late test and the degree
of freezing exhibited (r=0.92). These data strongly suggest that long-term potentiation in
the L amygdala is responsible for the behavioral changes seen after fear conditioning.
The medial geniculate showed increased field potentials at the early and late tests, but L
amygdala MAPK inhibition had no effect at either time point on potentiation in the
medial geniculate. Furthermore, there was no correlation between the net field potential
across tests and time spent freezing (r=-0.12). If potentiation in the medial geniculate
were causally related to fear-induced freezing, then those two variables would necessarily
30
correlate. Altogether, the experimental design allowed long-term memory formation to
occur everywhere other than the L amygdala, as well as allowing the fear circuit to
remain intact, and still fear responses were attenuated. Short-term memory and early
long-term potentiation were unaffected by these procedures, demonstrating that normal
acquisition did occur, however transiently. Long-term potentiation in the L amygdala
correlated with the expression of fear behavior, whereas long-term potentiation in the
medial geniculate nucleus, though enhanced through training, did not. Additionally,
manipulation of the L amygdala was without effect upon long-term potentiation in the
thalamus, showing that functional lesion of the L nucleus does not simply disrupt
learning elsewhere. The results of this study and the other discussed above (Schafe &
LeDoux, 2000) effectively rule out the second alternative hypothesis, that the fear engram
lies outside of the amygdala in the thalamus. In conclusion, consolidation of long-term
potentiation in the L amygdala appears necessary for long-term fear memory: the L
nucleus is the seat of the engram in associative fear learning.
To hedge just a bit, the L nucleus may not be the only place where classical fear
memory is stored. Recent evidence indicates that the classical fear engram may be
distributed across the L and C nuclei (Wilensky, Schafe, Kristensen & LeDoux, 2006).
In this study, rats were surgically implanted with guide cannulae aimed at either the C or
L nuclei. Just prior to tone-shock fear conditioning, these rats were infused with either
muscimol or vehicle. Testing 24 h later showed that pretraining functional inactivation of
either the C or L nuclei disrupted fear memory retention. In fact, inactivation of the C
induced significantly greater amnesia than L nucleus inactivation. In a follow-up
31
experiment, muscimol was infused into the C nucleus of another group of rats just after
fear conditioning. The subsequent test trial revealed that posttraining inactivation of the
C did not disrupt fear memory, indicating its participation is germane to acquisition
processes and not merely post-acquisition modulatory processes. The authors then asked
whether the role of the C nucleus in the acquisition process was similar to that of the L
nucleus, memory storage. To assess this, they infused the protein synthesis inhibitor
anisomycin into the C nucleus just after fear conditioning. During short-term testing 4 h
later, these rats displayed normal fear reactions, indicating normal short-term retention of
conditioned fear. However, during the test of long-term memory 24 h later, these animals
were significantly less fearful (as assessed by freezing) when compared to vehicle-
infused rats and to their own short-term test performance. Thus the C nucleus is not just
the head of the CR pathway or the output station of the amygdala; it also appears to bear
part of the classical fear engram.
1.1.2 Contextual fear conditioning. CONTEXTUAL FEAR CONDITIONING, also a form of
classical fear conditioning, entails learning to fear an environment in which an aversive
stimulus is encountered. In a typical contextual fear conditioning experiment, a rat is
placed into a conditioning box for several minutes and then briefly electrocuted. Later,
when the rat is returned to the conditioning box, a fear reaction is evoked and the rat
freezes. As one may surmise, classical fear conditioning that utilizes a light or tone CS
(termed “cued” fear conditioning) also entails conditioning a fear reaction to the context,
which acts essentially as another CS, albeit typically one of lesser salience than the
explicit cue. In cued fear conditioning, testing for a fear CR usually takes place in the
32
rat’s home cage or a test box dissimilar to the conditioning box in order to avoid
assessing contextual fear conditioning. If a rat tested for cued fear is already freezing due
to contextual fear, it may lead to spurious results.
Naturally, contextual fear conditioning utilizes much of the same neural circuitry
as cued fear conditioning, including the amygdala. Electrolytic lesion of the amygdala
prior to contextual fear conditioning attenuates freezing behavior. In a single training
session consisting of 3 one-second shocks separated by one minute, lesion of the C (with
variable damage to L and B) significantly reduced time spent freezing in the 60 s
following each shock. When replaced into the shock chamber 24 h later, lesioned
animals spent as little time freezing as they did prior to conditioning (Kim, Rison &
Fanselow, 1993). These findings have been corroborated by others (Helmstetter, 1992b;
Phillips & LeDoux, 1992; Goosens & Maren, 2001). Neurotoxic ablation of the BLA
also substantially attenuated subsequent contextual fear learning as assessed by freezing,
even with an extensive training protocol consisting of 25 shocks (Maren, 1998). An
additional training session (25 more shocks) for these rats yielded a moderate increase in
freezing that was nonetheless significantly below the level of freezing exhibited by non-
lesioned rats. Post-conditioning excitotoxic lesion of the BLA also dramatically reduces
freezing responses to frightening contexts, even with post-operative retraining (Maren, et
al., 1996). Furthermore, extensive training prior to BLA lesion did not lessen this deficit,
nor did it facilitate retraining afterwards (Maren, 1998). Lidocaine inactivation of the
amygdala (cannula placements fell within the C, L, and B nuclei) during testing of
33
contextual fear conditioning disrupted the expression of freezing (Helmstetter, 1992a), as
did GABA receptor stimulation (Holahan & White, 2004).
Electric lesion of either the C or BLA also considerably reduced conditioned
hypoalgesia (Helmstetter, 1992b). Rats were given a subcutaneous injection of
formaldehyde in their rear paw, then placed into a conditioning box that they had
previously been trained to fear. Non-lesioned rats exhibited far fewer signs of foot pain
(specifically, licking the wounded paw and holding the paw close to the body) than rats
with C or BLA lesions. Large excitotoxic lesions encompassing portions of both C and
BLA yielded the same pattern of results, suggesting that neurons intrinsic to these cell
groups are responsible for the hypoalgesia deficit. Thus lesions of the C or BLA complex
profoundly affect contextual fear conditioning.
However, unlike cued fear conditioning, which does not require the B nucleus, the
contextual fear conditioning circuit depends upon the both the L and B amygdala. Using
an interesting surgical methodology, Goosens and Maren (2001) made gross, nonspecific
electrolytic lesions of the amygdala in one hemisphere, and individual neurotoxic lesions
of the C, L, or B nuclei in the contralateral hemisphere. After recovery, rats were
transported to the conditioning chambers where they received 15 tone-shock pairings;
remember that cued fear conditioning can also entail conditioning to context. Twenty-
four hours later, an 8-min re-exposure to the conditioning boxes revealed that loss of the
L or B (or C, as above) significantly reduced freezing behavior compared to rats with
sham lesions or unilateral electric lesions only. Functional lesion studies have provided
further evidence of a role for the L and B nuclei in contextual fear conditioning. In a
34
study utilizing mice, lidocaine infusion into the B nucleus 5 min before training disrupted
context conditioning (Calandreau, Desmedt, Decorte & Jaffard, 2005). Lidocaine
inactivation of L nucleus also disrupted context conditioning. Likewise, infusion of the
NMDA antagonist AP5 into the B nucleus prior to conditioning completely disrupted
freezing behavior during the test session 24 (Fanselow & Kim, 1994) or 48 h later
(Savonenko, Werka, Nikolaev, Zieliñski & Kaczmarek, 2003). Thus the B nucleus seems
to be required for contextual fear conditioning. The question that follows is in what
capacity is the B required?
The B nucleus may be the site of associative memory storage for contextual fear,
similar to the role of the L nucleus in cued fear conditioning. If so, then the B would
need to be a place of confluence of the CS and US pathways, and it would need to
activate the fear response pathway after contextual fear conditioning. A neural center
important for the processing of polymodal contextual and environmental CS information
is the hippocampus, a structure that has long been known to form spatial memories
(Kandel, 2000). Lesions of the hippocampus prevent contextual fear conditioning, but
not cued fear conditioning (Selden, et al., 1991; Phillips & LeDoux, 1992). Hippocampal
projections to the amygdala target the B nucleus in particular (Ottersen, 1982; Pitkänen,
2000), thus this connection could constitute the CS pathway. The US pathway may reach
the B nucleus from a number of sources. The L nucleus, which of course receives US
information, may relay the US pathway to the B nucleus, since the L nucleus projects
heavily to the B nucleus (Pitkänen, 2000; Sah, et al., 2003). Other potential sources of
US information are somatosensory inputs from the parabrachial nucleus, the insular
35
cortex, or the cuneiform nucleus, but interestingly not from the nociceptive portions of
the thalamus (LeDoux, et al., 1990). Ergo the potential for CS-US association in the B
nucleus exists.
There is also evidence demonstrating that the B nucleus is capable of LTP (long-
term potentiation), a likely prerequisite for learning to occur in this area. In fact, the
hippocampus-B nucleus pathway in particular can be potentiated (Maren & Fanselow,
1995). In vivo stimulation of the ventral hippocampus while recording from the B
nucleus revealed a couple of features of this pathway. First, the short latency and
correlation of field potentials recorded with stimulus trains delivered indicated that this
pathway is a direct, monosynaptic projection. Second, this pathway exhibited tetanus-
induced LTP, and infusion of AP5 into the B nucleus before the stimulation protocol
prevented LTP induction. As mentioned above, injection of AP5 into the B nucleus prior
to contextual fear conditioning blocked learning (Fanselow & Kim, 1994), suggesting
that B nucleus LTP may be critical for contextual fear learning. Finally, in rats with
lesions of either the B nucleus or the ventral hippocampus, contextual fear conditioning
was significantly impaired relative to sham operated rats (Maren & Fanselow, 1995).
Thus the hippocampus-B nucleus pathway is capable of LTP, and the integrity of this
pathway is necessary for normal contextual fear conditioning.
With these data, a model of the contextual fear circuit can be supposed. The
hippocampus, a nuclear complex sensitive to environmental and spatial stimuli, detects
contextual CS features. Its output targets the B nucleus, where it joins fibers of the
shock-sensitive US pathway. When the CS and US pathways are activated
36
simultaneously, LTP occurs and strengthens the synaptic connections of the CS pathway
in the B nucleus. Re-exposure to the conditioned context would stimulate the
hippocampus, which, with its recently strengthened output tract, could then activate the B
nucleus. The B nucleus has connections with the C nucleus, the nexus of the fear CR
pathway, providing a way for the CS pathway to activate the fear response system.
Although plausible, the strength of this model is undermined by evidence demonstrating
that ablation of the entire BLA complex weakens, but does not prevent contextual fear
conditioning (Maren, 1998; Cahill, Vazdarjanova & Setlow, 2000), suggesting that
perhaps the BLA may not be the only area involved in contextual fear learning. The C
nucleus, a locus of memory storage in cued fear conditioning, may sustain continued
contextual fear performance in the absence of the BLA, since it too is a recipient of
hippocampal efferents (Pitkänen, 2000).
Contextual memory. Some evidence implicates the B amygdala in the
modulation of pure contextual memory (as opposed to contextual fear memory). A
paradigm known as the CONTEXT PREEXPOSURE FACILITATION EFFECT has been utilized to
examine the role of the amygdala in contextual memory. This paradigm is based on the
IMMEDIATE SHOCK DEFICIT EFFECT, the fact that a rat that is placed into a novel context and
immediately shocked will fail to show fear when replaced into the context 24 h later; it is
postulated that this deficit is due to inadequate context memory formation in the
hippocampus (Rudy, Barrientos & O’Reilly, 2002; Rudy, Huff & Matus-Amat, 2004), a
CS deficit. However, if one preexposure to the context is given 24 h before the fear
conditioning trial, then immediate shock contextual fear conditioning is effective,
37
presumably because the brief preexposure allows sufficient formation of a context
memory. Temporary inactivation of the B nucleus with the GABA
A
agonist muscimol
either before or just after preexposure attenuated the context preexposure facilitation
effect, but inactivation 3 h after preexposure was without effect (Huff & Rudy, 2004).
Inhibition of protein synthesis in the B nucleus just after preexposure left the context
preexposure facilitation effect intact, meaning that actual storage of the context memory
is probably not in the B. Thus the B amygdala may modulate context memory formation
in the hippocampus, a function that presumably has relevance for contextual fear
conditioning as well.
Figure 1.2 A possible neuroanatomical model of the contextual fear
conditioning circuit.
38
1.1.3 Others: eyeblink, reward, conditioned attention.
Classical eyeblink conditioning. The amygdala has been shown to affect
classical eyeblink conditioning as well (see Section 1.1 for description). During
conditioning, a phenomenon occurs called REFLEX FACILITATION, wherein a neutral tone or
the CS increases the magnitude or shortens the latency of the air puff-elicited eyeblink
US. Thus when CS-US pairs are presented, the eyeblink that is evoked is stronger than
the eyeblink elicited by an unsignalled, unexpected air puff. Electrolytic lesions targeting
the C nucleus in New Zealand rabbits did not affect the initial presentation of reflex
facilitation, but eliminated the sustained display of this response seen in control rabbits
(Weisz, Harden & Xiang, 1992). Since reflex facilitation is exhibited to even the first
CS-US pair, the initial facilitation is thought to be an unconditioned facilitative response,
but that subsequent reflex facilitation reflects a rapidly acquired conditioned facilitation.
The amygdala may mediate this facile acquisition of conditioned reflex facilitation.
Lesions of the C nucleus also retarded acquisition of the conditioned eyeblink
response when the auditory CS intensity was suboptimal (65 dB), but not when the CS
parameters promoted optimal conditioning (85 dB). The authors believed that the
optimal conditioning procedure occluded the lesion-induced deficit, and that the role of
the amygdala is to modulate eyeblink conditioning by enhancing the efficacy of CSs;
optimal CSs would need no enhancing, hence the lack of a deficit when the CS was
delivered at 85 dB (ibid.). The enhancing effect of the amygdala on suboptimal CSs may
in fact be due to a kind of aversive conditioning, since the air puff is an unpleasant
stimulus, and the conditioned fear response includes a state of vigilance and increased
39
attention. Thus the amygdala would quickly associate the tone with the aversive air puff,
such that the tone in short order would elicit a conditioned attention response. Attention
in turn can enhance stimulus processing and speed learning in the parts of the brain that
mediate eyeblink conditioning.
Reward and attention. Electrophysiological recordings of neurons throughout
the amygdala, but with a slight majority the BLA, have shown a relationship between
neuronal activity, reward stimuli, and conditioned stimuli that predict reward. Of the
neurons sampled in one study (Muramoto, Ono, Nishijo & Fukuda, 1993), 44% of BLA
units responded to rewarding stimuli, i.e., to glucose, intracranial self-stimulation of the
medial forebrain bundle, or both. During conditioning, which entailed pairing each of
glucose or stimulation with a distinctive pure-tone CS, a subset of the reward-responsive
units began responding to the associated CS. Irrelevant tone stimuli that were not paired
with reward were either not responded to, or were habituated to. When reward was
withheld during the tone-only extinction trials, the units eventually ceased to respond to
the CSs. This seems to suggest that, as with aversive stimuli in fear conditioning, neutral
stimuli can become associated with rewarding USs.
Consistent with this idea are studies of conditioned attention, a reward-based
classical learning paradigm mediated by the amygdala. In one procedure used to study
conditioned attention, a food-deprived rat is placed into a conditioning box containing a
food hopper and a lightbulb. The light is turned on for 10 s prior to the delivery of a food
pellet by the hopper. Over several trials, the rat learns that the light indicates a coming
food reward, and begins responding to the light with an orienting response called rearing,
40
in which a rat raises itself up onto its hind legs and looks at the light. Typically, the rat
rears for the first few seconds of the light stimulus, and then it approaches the food
hopper to await the pellet (Baxter & Murray, 2002). Rats with lesions of the C nucleus
fail to show this conditioned attentive response, even though unconditioned orienting
responses to novel light stimuli are intact (Gallagher, Graham & Holland, 1990; Han,
McMahan, Holland & Gallagher, 1997). Furthermore, lesioned rats will still reliably
retrieve the food pellets from the hopper. Lesions of the B nucleus do not affect
conditioned attention (Baxter & Murray, 2002). Thus it seems that, just as with classical
fear conditioning, a neutral CS paired with a rewarding US can acquire the ability to
evoke a CR, and that this process is consistent with the amygdalar contribution to the
classical fear conditioning circuit.
Another similar kind of behavior is AUTOSHAPING, which, despite its operant-like
name, is a form of appetitive classical conditioning quite similar to conditioned attention.
It involves an animal learning to approach a visual CS that precedes food presentation in
a distal location, a behavior that is Pavlovian, not instrumental (Bussey, Everitt &
Robbins, 1997). In this paradigm too, neurotoxic lesions of the C amygdala eliminate
conditioned approach responses, while lesions centered in the B nucleus had no effect
upon this conditioning (Parkinson, Robbins & Everitt, 2000). In another study with a
similar design (Hitchcott & Phillips, 1998), rats were conditioned to associate a visual CS
with access to a sucrose solution. Just after each training session, rats were infused in
either the C or B nucleus with R(+) 7-hydroxy-DPAT, a D3 dopamine receptor agonist.
41
Approach responses were measured
8
, revealing that dopamine agonists in the C amygdala
enhanced approach responding, while infusions in the B nucleus were without effect.
Thus, lesion of the C nucleus disrupts conditioned approach, and activation of the C with
a dopamine agonist facilitates conditioned approach, which suggests that the C nucleus
may mediate reward-based classical conditioning.
Context associated with reward. Amygdala lesions also disrupt the motivating
impact of a rewarding environment. Dehydrated groups of B nucleus- and sham-lesioned
rats were placed into a test chamber containing a water spout, and drinking behavior was
measured with a lickometer during a 1-minute test. On the first test, both groups had a
mean latency of about 16 seconds to initiate drinking. Several hours later, a second 1-
minute test was given to the still dehydrated rats. The sham lesioned group displayed a
shorter latency to begin drinking than before (on the order of 6 seconds), while the B-
lesioned group maintained its 16 second latency. Both groups drank at the same rate, and
made approximately the same number of licks during the test, indicating that B-lesioned
rats were able to respond appropriately to a primary reinforcer. The deficit, the authors
concluded, was in the rats’ ability to respond to conditioned reward (Seldon, et al., 1991).
In this case the CS could be the context of the test chamber, a context indicating access to
water and which should stimulate approach to the drinking spout (although the authors
did not explicitly propose this hypothesis). Just as in contextual fear conditioning,
lesions of the B nucleus seem to disrupt the ability of a context to become associated with
8
Unfortunately the stimulus lightbulb was located on the same wall of the training box as the feeding
aperture, which could have confounded the measurement of CS-elicited approach responses with US-
elicited approach responses, since the two stimuli were in close proximity. However, US-elicited approach
responses are unaffected by C nucleus lesions (Gallagher, Graham & Holland, 1990), so alterations in the
approach responses reported here probably reflect the manipulation of CS-elicited approach.
42
reward, since the B nucleus is the amygdalar recipient of hippocampal contextual sensory
information.
1.2 Instrumental conditioning
INSTRUMENTAL CONDITIONING is a learning and memory system in which the
frequency of a behavior is changed contingent upon the consequences (punishing or
rewarding) that follow it. Said another way, the animal uses its behavior as a tool in
order to attain a desired consequence, by either avoiding punishers or approaching
rewards (Dickinson & Balleine, 1994). The behavioral responses are not emitted
reflexively, as in classical conditioning, but are instead performed more as the result of a
cost-benefit analysis or hedonic calculus, i.e., the animal makes a decision about whether
or not to respond. There are four instrumental conditioning paradigms: positive
reinforcement, negative reinforcement, positive punishment, and negative punishment.
The reinforcement paradigms increase behavioral frequency, while the punishment
paradigms decrease behavioral responses. POSITIVE REINFORCEMENT is the administration
of a rewarding stimulus, to the effect of increasing the behavior the animal was exhibiting
shortly before reward delivery. NEGATIVE REINFORCEMENT is the removal of an aversive
stimulus, which is a desired and rewarding consequence that likewise promotes the
behavior that preceded it. POSITIVE PUNISHMENT (or just “punishment”) is the delivery of
an aversive stimulus that attenuates occurrences of the foregoing behavior. NEGATIVE
PUNISHMENT (sometimes termed “penalty”) is the removal of a rewarding stimulus, which
is also aversive and will diminish the behavior it follows.
43
1.2.1 Aversive conditioning. Passive and active avoidance are the two forms of
instrumental fear conditioning, corresponding to the instrumental conditioning paradigms
of positive punishment and negative reinforcement, respectively. In the PASSIVE
AVOIDANCE task (also called “inhibitory avoidance”), a rat is placed into a conditioning
box with two chambers separated by a partition. One side is white-floored and well-lit,
which rats do not enjoy, preferring instead dark, cloistered dens. A door in the partition
opens into the other chamber, which is dark and thus inviting to the rat. When the rat
enters the dark room, an electrified floor grid delivers a brief shock to the rat, which is
then removed to its home cage. Later (typically 48 h), the rat is replaced into the light
chamber of the test apparatus, the door in the partition is opened, and the latency for the
rat to enter the dark chamber is recorded. It is called passive avoidance (or inhibitory
avoidance) because the rat must refrain from entering the dark chamber. The task is
instrumental in nature, a punishment paradigm specifically, because the rat uses its
behavior of remaining in the light room to avoid the shock it expects to receive in the
dark room.
ACTIVE AVOIDANCE is an instrumental learning task quite similar to inhibitory
avoidance, except that in this case the animal must perform a behavior in order to escape
from an aversive stimulus. A typical active avoidance training procedure employs an
apparatus which for all intents and purposes may be identical to an inhibitory avoidance
training box. During training the animal is placed into the dark shock chamber and
electrocuted. The aperture leading to the lighted safe chamber is left open, and the
animal will usually escape through it hastily after shock delivery. Upon replacement of
44
the animal into the dark chamber, the electrical grid floor is turned on after several
seconds, giving the animal a brief time to flee in order to avoid shock and to escape the
fear-inducing context. In this sense, the active avoidance paradigm could be conceived
as a form of negative reinforcement, since the animal is placed into an environment that
constitutes a continuous aversive stimulus from which the animal can work to be
relieved.
Inhibitory avoidance. The description in the introductory paragraph of Section
1.2.1 is an example of inhibitory avoidance conditioning, where an animal can avoid an
aversive consequence by abstaining from the performance of a behavior. Large lesions of
the amygdala have a deleterious effect upon this form of learning. Using a one-trial
passive avoidance training session, Dunn and Everitt (1988) found that rats with gross
excitotoxic lesions of the amygdala (encompassing the BLA complex and parts of the C
and Co nuclei) exhibited a much shorter latency to enter the shock chamber during the
testing session than control rats, even though both groups had increased their entry
latency relative to the training session. Thus, while both groups seemed hesitant to re-
enter the shock chamber, amygdala-lesioned rats were significantly less so.
Cahill and McGaugh (1990) confirmed and extended these findings, using rats
with similarly placed lesions in a somewhat different passive avoidance procedure.
Severely water-deprived rats were trained in a Y-MAZE TASK, in which they ran down a
corridor and then had to choose between the right (dimly backlit) and the left (dark) arms
of a Y-shaped maze in order to receive water, which was in the backlit arm. The next
time the animals were placed into the maze, they rapidly went down the backlit corridor,
45
where instead of water, they received an electric shock until they retreated from that arm.
During the following test in the Y-maze, all rats were reluctant to enter the backlit arm,
but amygdala-lesioned rats displayed a significantly shorter latency to enter than control
rats, similar to the findings of Dunn and Everitt (1988). In a follow-up experiment, they
implemented an odor-shock passive avoidance procedure, in which rats were exposed to
amyl acetate vapor (a banana-apple scent) followed by a brief shock. Testing took place
in a two-chamber test box with a door between the chambers, with one chamber bearing
the odor and the other odorless. The time each rat spent in each chamber was recorded in
a 5-min test. Compared to groups that had been exposed to the odor only (no shock),
non-lesioned shock-trained animals spent significantly less time in the scented chamber,
i.e., they avoided the odor; amygdala-lesioned animals, on the other hand, were not
significantly different from the no-shock rats. In fact, the authors reported that the
lesioned rats went right up to the odor-producing apertures in the scented chamber, a
behavior the non-lesioned rats never showed. Thus whether context or odor is the target
stimulus, amygdala lesions attenuate aversive instrumental conditioning. The consistent
finding by these laboratories and others (Pellegrino, 1968; Nagel & Kemble, 1976;
Jellestad & Bakke, 1985; Antoniadis & McDonald, 2000) is that lesions of the amygdala
reduce the efficiency of aversive instrumental conditioning.
Other studies have attempted to define more precisely the part of the amygdala
exerting an influence over aversive instrumental conditioning. In another Y-maze
inhibitory avoidance experiment (Vazdarjanova & McGaugh, 1998), BLA-lesioned and
sham-lesioned rats were shocked several times in one arm of the maze. When re-
46
introduced to the Y-maze the following day and allowed to roam freely for 8 min, the
sham rats spent most of the time freezing in one of the non-shocked arms, and most failed
to enter the arm where shock was received (judging by the 7.9 min average latency to
enter the shock arm). The BLA rats spent slightly (but significantly) more time in the
shock arm than the sham rats (roughly 25 s versus 0 s, respectively), with an average
latency of 3.75 min to enter the shock arm. Still, the BLA rats nonetheless spent most of
their time in the other two non-shocked arms of the maze. Thus the lesions seemed to
attenuate but not eliminate the avoidance behavior; this diminution was not attributable to
reduced shock sensitivity, as BLA rats had the same flinch and jump responses across
several shock amplitudes as sham rats. The histology indicated that in the BLA group,
the BLA complex was consistently damaged, while the C nucleus was entirely avoided.
Another study corroborated this lesion effect: when using a standard two-chamber
inhibitory avoidance training apparatus, retention of an avoidance response was
eliminated by BLA lesion, although its facile reacquisition was not significantly hindered
(Berlau & McGaugh, 2003). These indicate that BLA lesions impair avoidance responses
to aversive stimuli, even though the rats’ behavior seems to indicate at least some
appropriate sensibility about the danger of the shock chamber, perhaps due to retained
classical fear conditioning in these animals.
Other studies have used specific nuclear lesions to refine further the critical
circuitry, to no definitive end. In a single-trial inhibitory avoidance training setup,
pretraining neurotoxic lesions of the C amygdala virtually eliminated avoidance of the
shock chamber (Roozendaal & McGaugh, 1996), while in a similar study from the same
47
laboratory, C nucleus lesions merely attenuated avoidance responses (Tomaz, Dickinson-
Anson & McGaugh, 1992). In yet another study, pretraining excitotoxic lesions of the C
nucleus did not affect rats’ ability to avoid a punished bar-press response, while rats with
lesions encompassing the B and portions of the L nucleus failed to avoid the punished bar
(Killcross, et al., 1997). Electrolytic lesions of the L nucleus in mice also disrupted
performance in a passive avoidance training protocol, and greater damage to the L was
associated with poorer performance (Slotnick, 1973). Avoidance responses in rats, too,
were adversely affected by L nucleus excitotoxic lesion (Tomaz, et al., 1992). Rats with
neurotoxic lesions of the B nucleus subsequently exhibited normal avoidance reactions
when trained in an inhibitory avoidance procedure (Seldon, et al., 1991; Tomaz, et al.,
1992; Roozendaal & McGaugh, 1996). Yet other studies found that lesion of the B
nucleus significantly retarded avoidance performance (Lorenzini, Bucherelli, Giachetti,
Mugnai & Tassoni, 1991; Killcross, et al., 1997). The inconsistency of this body of
results is puzzling.
A number of possible reasons for this variation come to mind. Perhaps the actual
neuroanatomical substrate of inhibitory avoidance is not one of the prominent nuclei (see
Figure 1.0), but one of the intercalated nuclei, a number of small, aciniform or laminar
bodies of neurons scattered throughout the amygdala, positioned between the major
nuclei. Focused lesions, like those used in the studies above, would likely not reliably
include one of the outlying intercalated nuclei. Or perhaps multiple associations are
learned in parallel during passive avoidance conditioning, and the loss of one subsystem
is obscured by the emergence of another that acts as a decent substitute. Another
48
possibility is that little differences in the inhibitory avoidance training paradigms could
engage the amygdala differently. For instance, retention of single-trial training tends to
resist amygdala lesion, but multi-trial training seems susceptible. Or, learning seems fine
when tested immediately (intact short-term memory), but when tested at a delay ( ≥24 h)
it degrades (disrupted consolidation).
Role of the amygdala: modulation? McGaugh has long been the standard
bearer of the theory that the role of the amygdala during inhibitory avoidance learning is
to modulate memory formation in other brain areas (McGaugh, Ferry, Vazdarjanova &
Roozendaal, 2000; McGaugh, 2002; McGaugh, 2004). His arguments advocating a
modulatory role for the amygdala rest mainly on studies demonstrating that memory-
modulating hormones and drugs exert their effects through the amygdala. That evidence
will be described anon, but first relevant lesion data will be reviewed. The posttraining
lesion technique is useful in evaluating possible modulatory contributions a brain area
may make to learning. This procedure allows the conditioning phase to take place in
unoperated animals, so that learning occurs normally, and then after a short or long
interval the lesions are produced. One would expect that, should the amygdala modulate
memory persistence, its destruction would have ever less of an effect on performance as
more time passed between the training session and lesion induction, since at these later
time points the aversive memory would be successfully converted into a long-term
storage form and would no longer be in need of modulation.
Posttraining amygdala lesions are seen to impair recall of inhibitory avoidance
training at all intervals tested. Large excitotoxic amygdala lesions made 7 d after training
49
were strongly disruptive to rats’ ability to avoid the shock chamber of a passive
avoidance apparatus relative to non-lesioned rats (Parent, Quirarte, Cahill & McGaugh,
1995; Bermudez-Rattoni, Introini-Collison, Coleman-Mesches & McGaugh, 1997). Yet
these lesioned rats delayed entry into the shock chamber marginally but significantly
longer than non-shocked (i.e., non-trained) control rats (Parent, et al., 1995), indicating a
minor retention of aversive memory. This disruption in memory retention is nominally
ameliorated by increasing the number of preoperative training trials to 20 (Parent, Tomaz
& McGaugh, 1992). In another inhibitory avoidance experiment, the effects of
pretraining electrolytic lesions and posttraining lesions made at 2, 5, or 10 d intervals
were compared (Liang, et al., 1982). The 2 d lesion group performed similarly to the
pretraining lesion group, poorly. The 5 and 10 d delayed lesion groups fared a mite better
at the avoidance task than groups receiving pretraining lesions. Thus as more time
passed between the training session and lesion induction (2, 5, or 10 d), the better the
animals seemed to perform; yet all lesion groups performed dramatically worse than the
unoperated control groups, who avoided the shock chamber unwaveringly. In a
subsequent experiment in the same laboratory (Parent, West & McGaugh, 1994), animals
were given 0, 1, or 10 inhibitory avoidance training trials. Surgery was delayed for 30 d
after training, then large excitotoxic amygdala lesions (engrossing the C and BLA) or
sham surgeries were performed. Destruction of the amygdala virtually eliminated
retention of the avoidance task, as lesioned animals showed almost no hesitation to enter
the shock chamber; those lesioned rats receiving 10 training trials appeared to have
somewhat of a longer entrance latency than the 0- or 1-trial lesioned rats (who were
50
indistinguishable behaviorally), but this tendency was not statistically validated. In
assessing the results of these posttraining lesion experiments as a group, they do not
provide compelling evidence that the function of the amygdala is merely to modulate
aversive learning in other brain regions. The fact that rats with amygdala lesions perform
so poorly on the inhibitory avoidance task, with at best only a minimal mitigating effect
of increased preoperative training or up to 30 d of posttraining respite, indicates more
than just a modulatory role for the amygdala in inhibitory avoidance learning and
memory.
Functional lesions, however, give some indication that the amygdala may be
involved in modulatory processes related to inhibitory avoidance learning. Temporary
inactivation of the BLA with lidocaine infusions administered immediately after single-
trial avoidance training shortened rats’ latency to enter the shock chamber and truncated
the time they spent in the safe chamber 48 h later (Parent & McGaugh, 1994); rats
infused in the C nucleus behaved similarly to control rats infused with saline. Retention
of the avoidance memory was also disrupted when BLA infusions were made 6 h (but not
24 h) after training. Infusion of TTX (tetrodotoxin
9
) throughout the amygdala disrupted
retention when inhibition was initiated 15 or 90 min after a moderately intense training
protocol (15 shocks), but inactivation at 6 or 24 h was without effect (Bucherelli, Tassoni
& Bures, 1992). Inhibition of the BLA with GABA
A
receptor stimulation immediately
after single-trial passive avoidance training blocked retention of the aversive memory
when tested 48 h later (Wilensky, Schafe & LeDoux, 2000). Together, these studies
9
Tetrodotoxin: a potent inhibitor of voltage-gated Na
+
channels that effectively prevents neural activity.
51
indicate that normal avoidance conditioning requires posttraining activity in the BLA
complex of the amygdala, a widely accepted hallmark of modulatory activity.
Yet other functional lesion data reveal a more integral role for the amygdala in
retention of avoidance memory. When a GABA agonist was administered into the BLA
of rats just prior to retention testing (48 h after normal training), these animals too failed
to avoid the shock chamber, entering it as unhesitatingly as non-shocked control animals
(Berlau & McGaugh, 2003). This suggests a requisite role for the amygdala in memory
retrieval, since BLA activity during expression of avoidance behavior was necessary.
Cycloheximide-induced inhibition of protein synthesis in the amygdala 30 min before or
immediately after single-trial training strongly attenuated retention of the avoidance
response 48 h later (Berman, Kesner & Partlow, 1978). Retention was normal when
tested just 30 min after training, demonstrating that short-term memory was intact, but
that long-term retention was impaired. Recordings of electrical activity in the amygdala
after bilateral cycloheximide injections in control rats revealed no untoward effects of the
drug on amygdala activity for 6 h after injection or during periodic sampling during the
following 24 h; thus protein synthesis inhibition did not affect memory acquisition or
amygdala activity-dependent modulation of distal brain areas, but instead seemed to
antagonize actual memory storage in the amygdala. Both of these results provide reason
to suspect a more important role for the amygdala in inhibitory avoidance memory than
modulation.
Much of the basis for the belief that the amygdala modulates the formation of
long-term inhibitory avoidance memory is the fact that a number of memory-modulating
52
drugs and hormones affect the amygdala. Of signal importance to this belief, the central
nervous system neuromodulator and adrenal stress hormone norepinephrine can modulate
inhibitory avoidance memory crystallization, and this effect is mediated by the BLA.
Intraamygdala infusions of norepinephrine can facilitate training efficacy. When low
doses of norepinephrine (0.1, 0.2, or 0.3 µg) are injected immediately after single-trial
training, the latency to enter the shock chamber is greatly extended relative to vehicle-
injected controls in a retention test 24 h later, whereas high doses (1, 4.2, or 5 µg) are
without effect (Gallagher, Kapp, Musty & Driscoll, 1977; Liang, Juler & McGaugh,
1986; Liang, McGaugh & Yao, 1990). Norepinephrine does not exert this augmentative
effect by acting as an additional punisher, because animals that were given training
without footshock, and then infused, were unafraid of the shock chamber (Liang, et al.,
1986). Infusions of either of the norepinephrine β receptor antagonists propranolol and
alprenolol into the amygdala immediately after single-trial inhibitory avoidance training
disrupted retention, while infusions delayed for 6 h were without significant effect
(Gallagher, et al., 1977). In addition, low doses of these antagonists or high doses of
their dextral enantiomers were without effect on retention, indicating that interference
with noradrenergic neurotransmission was responsible for the amnesia, as opposed to
some non-specific effect. Moreover, retention deficits were reversed by simultaneous
infusion of norepinephrine and propranolol (in a “cocktail”), again indicating
noradrenergic specificity, since their competitive actions cancelled out each other. Both
the α and β adrenoreceptors participate in this modulatory effect, and the BLA
specifically seems to be an important locus of action of the noradrenergic system. The β
53
receptor agonist clenbuterol enhances the avoidance response when infused into the BLA
shortly after single-trial training (Ferry & McGaugh, 1999), indicating that the β
adrenoreceptor is involved in memory modulation. Co-application of an α
adrenoreceptor antagonist with the clenbuterol blocked the memory-aiding effect of the β
agonist (Ferry, Roozendaal & McGaugh, 1999), revealing that α and β adrenoreceptors
interact. Similar infusion of an analog of cyclic adenosine monophosphate (cAMP
10
) in
conjunction with the α antagonist revealed no detrimental effects of the antagonist on the
ability of cAMP to potentiate avoidance memory. This suggests that α adrenoreceptor
stimulation normally acts to facilitate the β adrenoreceptors’ creation of cAMP, somehow
(McGaugh, 2004). This complicated pathway is summarized in Figure 1.3.
Figure 1.3 A model of the path of neuromodulator actions in the B amygdala. Unless
norepinephrine (NE) release is inhibited by opioids (flathead arrow), it acts upon its α
and β adrenoreceptors. Stimulation of the β receptor induces the generation of cAMP
(cyclic adenosine monophosphate), while α receptor stimulation acts through
glucocorticoids to facilitate cAMP creation. At some point downstream, ACh
(acetylcholine) exerts its effect. See text for a complete description.
Delay of infusion of an effective dose to 3 h after training annulled the amplifying
effect of norepinephrine (Liang, et al., 1986), as did delaying the infusion of its agonist
10
cAMP: the postsynaptic second messenger generated by β adrenoreceptor stimulation.
54
clenbuterol by merely 10 or 20 min (anecdotally reported in Introini-Collison, Miyazaki
& McGaugh, 1991). The brief window of opportunity afforded to norepinephrine
suggests it may have some specific effect on consolidation. Since norepinephrine is
administered just after training, it does not affect acquisition processes, per se. Rather,
studies of LTP in other model systems (the rat hippocampus and Aplysia nervous system)
have shown that conversion of LTP from its early to its late form requires coincident
activation of glutamate receptors with modulatory metabotropic receptors (i.e., those
receptors binding dopamine and serotonin in the systems above, respectively). Since
norepinephrine is a modulatory neurotransmitter, it is possible that it can convert early
LTP into late LTP, as long as it is applied in a timely enough fashion to effect this
conversion.
Another class of adrenal stress hormone, the corticosteroids
11
, also modulates
memory formation by acting upon the amygdala. Application of a glucocorticoid agonist
in the B nucleus shortly after one-trial training enhanced avoidance of the shock chamber
during the test session, while the same treatment in the C nucleus was without effect on
retention performance (Roozendaal & McGaugh, 1997). Lesion of the B nucleus did not
disrupt single-trial conditioning but did prevent systemic corticosterone enhancement of
learning (Roozendaal & McGaugh, 1996), whereas lesion of the C nucleus did not
obstruct corticosterone memory modulation. This corticosteroid effect upon avoidance
learning is actually mediated by the noradrenergic system in the amygdala. A
posttraining systemically administered glucocorticoid agonist was shown to facilitate
11
Corticosteroids: steroid hormones comprised of the glucocorticoid and mineralocorticoid subclasses.
55
avoidance responses, yet this effect was completely obviated by pretraining inoculation
with any of several β adrenoreceptor antagonists infused into the B amygdala (Quirarte,
Roozendaal & McGaugh, 1997). Furthermore, simultaneous co-infusion of a
glucocorticoid agonist and a β adrenoreceptor antagonist into the B nucleus similarly
eliminated the avoidance memory augmentation seen with glucocorticoid agonists alone,
while an α adrenoreceptor antagonist was without effect on the glucocorticoid agonist
(ibid.; Roozendaal, Quirarte & McGaugh, 2002). In another study, rats received an intra-
basal pretraining infusion of a glucocorticoid antagonist, followed after training by
clenbuterol, the β agonist; clenbuterol was still seen to improve avoidance memory, but a
much higher dose was required to achieve the same effect as that attained in the absence
of the glucocorticoid antagonist (Roozendaal, et al., 2002). In other words, the essential
effect of β adrenoreceptor stimulation to facilitate inhibitory avoidance memory was
intact, but it was modulated by glucocorticoids. Thus, the memory modulatory effects of
corticosteroids seem to occur via amygdalar β adrenoreceptors. Further experimentation
found that corticosteroids actually appear to mediate the ability of α adrenoreceptors to
boost β adrenoreceptor function (described previously). Intra-basal infusion of a
glucocorticoid antagonist followed by an α
1
adrenoreceptor agonist averted the enhanced
memory formation elicited by the α
1
agonist alone; higher doses of the α
1
agonist could
not overcome this occlusion (ibid.). Altogether, glucocorticoids seem to mediate the
modulatory effect that α adrenoreceptor stimulation has upon β adrenoreceptors (Figure
1.3).
56
The endogenous opioid system also modulates aversive memory consolidation,
and it too is mediated by the noradrenergic innervation of the amygdala. Intraamygdala
injection of the opioid agonist β-endorphin immediately after one-trial inhibitory
avoidance training eliminated expression of avoidance responses 48 h later (Introini-
Collison, Ford & McGaugh, 1995). Simultaneous administration of the β-adrenoreceptor
agonist clenbuterol overcomes the opioid-induced amnesia in a dose-dependent manner.
The opioid antagonist naloxone facilitates memory formation when infused into the
amygdala (McGaugh, Introini-Collison & Nagahara, 1988; Introini-Collison, Nagahara &
McGaugh, 1989), an effect that is completely reversed by concurrent co-localized
infusion of the β blocker propranolol (Introini-Collison, et al., 1989). A microdialysis
study has revealed that systemic β-endorphin actually inhibits the release of
norepinephrine in the amygdala, while naloxone potentiates release (Quirarte, Galvez,
Roozendaal & McGaugh, 1998). Hence, like corticosteroids, endogenous opioids exert a
modulatory effect on inhibitory avoidance memory, and they do so by restraining the
secretion of norepinephrine in the amygdala (Figure 1.3).
The neurotransmitter acetylcholine is another neurochemical that can modulate
inhibitory avoidance memory, when acting upon the metabotropic muscarinic
acetylcholine receptor. Both systemic (whole body) and intraamygdala infusions of a
muscarinic receptor agonist can enhance avoidance learning when administered just after
single-trial training (Introini-Collison, Dalmaz & McGaugh, 1996). The memory
improvement from systemic muscarinic stimulation was not blocked by simultaneous
blockade of β adrenoreceptors in the amygdala, suggesting that unlike the
57
neuromodulators above, acetylcholine does not act upon the amygdala noradrenergic
system. In fact, evidence indicates that the modulatory effects of norepinephrine on
avoidance memory are mediated by the cholinergic system of the amygdala. As already
discussed, the β norepinephrine agonist clenbuterol can improve avoidance memory
formation when infused into the amygdala (Ferry & McGaugh, 1999). Simultaneous
intraamygdala infusion of the muscarinic acetylcholine antagonist atropine with
clenbuterol prevented any avoidance memory enhancement during the retention test
(Introini-Collison, et al., 1996). Furthermore, simultaneous stimulation of amygdalar
muscarinic receptors and blockade of amygdalar β adrenoreceptors with propranolol
immediately after single-trial avoidance training resulted in enhanced avoidance
performance in the retention test 48 h later, i.e., the propranolol failed to disrupt the
mnemonic action of muscarinic stimulation (ibid.). Thus, it appears that cholinergic
activity is later in the chain of events than noradrenergic activity during the modulation of
inhibitory avoidance memory. Overall, then, neuromodulators like corticosteroids and
opioids act in the amygdala to modify noradrenergic activity, which in turn acts upon the
cholinergic system to modify memory consolidation (Figure 1.3).
Thus, certainly, drugs affecting the consolidation of inhibitory avoidance memory
converge upon and act within the BLA. Yet while the amygdala may modulate memory
formation in other brain areas (and it certainly does affect LTP in the hippocampus
[McGaugh, 2004], and pure contextual memory [see section 1.1.2, Contextual Memory]),
these modulatory effects are also consistent with the idea that it is the BLA itself that is
being modulated. Convincing evidence that the amygdala modulates inhibitory
58
avoidance memory formation in other brain areas of the rat is surprisingly scarce, given
all of the ink that has been devoted to the topic. Among the most compelling is a study
that explored the effects of severing major amygdalar afferent/efferent fiber tracts on the
ability of intraamygdala norepinephrine infusions to modulate avoidance memory (Liang,
et al., 1990). Rats received pretraining sham surgery or lesions of the stria terminalis, a
fiber path connecting the amygdala with the front half of the brain. Later, they received
one-trial training followed immediately by intraamygdala injections of vehicle or low
(0.2), moderate (1) or high (5) doses of norepinephrine (in µg). In the non-lesioned
animals the norepinephrine was seen to have a facilitating effect (low dose), no effect
(moderate dose), or an impairing effect (high dose) upon avoidance memory formation,
relative to vehicle. However, stria terminalis lesions eliminated any effects of
norepinephrine, causing all animals to have moderate retention just like the non-lesioned,
vehicle-infused controls and the 1 µg group. If the amygdala was a place of memory
storage, and norepinephrine was directly infused there, it does not make sense why stria
terminalis lesions would disrupt only modulatory processes, rather than blocking
avoidance memory entirely. Furthermore, the high dose of norepinephrine severely
impaired memory, so if memory was stored in the amygdala and the stria terminalis was
a critical output pathway for that memory, then strial lesions should have likewise caused
severely impaired memory—the fact that moderate memory was retained in all lesion
groups suggests that the stria terminalis is not a critical output pathway. Lesions of the
ventral amygdalofugal pathway (connecting the amygdala with the posterior half of the
brain) were similarly without deleterious effect on retention (or modulation for that
59
matter), which shows that neither of these major amygdala pathways is necessary for the
basic performance of this particular avoidance task. Although this study too is not
without alternative explanations, it is supported by a few others demonstrating no
disruptive effect of B nucleus lesion on retention of a single-trial avoidance training
memory (Tomaz, et al., 1992; Roozendaal & McGaugh, 1996). Together, these findings
suggest that one role of the amygdala may be to detect ascending stress hormones and to
communicate a neuromodulatory signal to the forebrain.
Role of the amygdala: US representation? How then to reconcile this with
evidence (discussed above) showing that some lesions of the BLA do impair retention of
the inhibitory avoidance memory? The answer may lie in the particular training protocol
employed, specifically the degree of aversive footshock the rats receive. Studies have
determined that aversive experience induces norepinephrine secretion and utilization in
the amygdala in a manner proportional to shock intensity. An excessive series of intense
footshocks led to increased turnover of norepinephrine in the amygdala, as assessed by
measurement of its major metabolite, MHPG-SO
4
(Iimori, et al., 1982); plasma
corticosterone levels were also elevated in these rats. In another study, a single footshock
triggered the release of norepinephrine in the amygdala, with increased shock amplitude
eliciting significantly greater norepinephrine release (Quirarte, et al., 1998). Thus,
significantly stressful experiences may recruit the BLA to a greater extent, while less
stressful ones would do so to a lesser degree or not at all. The minor footshock used in
the one-trial inhibitory avoidance experiments may not sufficiently engage the amygdala,
so lesions of the B nucleus would not matter. More intensive or prolonged training
60
protocols (Lorenzini, et al., 1991; Killcross, et al., 1997; Vazdarjanova & McGaugh,
1998; Berlau & McGaugh, 2003), on the other hand, would recruit the amygdala, and so
loss of the B nucleus, which mediates the noradrenergic and corticosteroid influences on
memory formation, would be more apparent (as it was in these studies). In fact, this
recruitment may integrate the amygdala into the inhibitory avoidance neural circuit in a
more crucial and permanent way, since lesions made long after training (5-30 d) and
blockade of protein synthesis shortly after training both disrupt avoidance memory.
If there is a critical and lasting contribution that the amygdala makes to inhibitory
avoidance learning, what might that be? Recent reviews of the amygdala literature
postulate that the amygdala, particularly the BLA, may evaluate and store the hedonic
value of the US (Cardinal, Parkinson, Hall & Everitt, 2002; Fanselow & Gale, 2003). In
classical fear conditioning, the value of the US can be altered independently of CS-US
pairings with a technique known as the INFLATION PROCEDURE. This entails first
classically conditioning a conditioned emotional response using a weak footshock US.
Later, some of the animals are given several strong footshocks in the absence of the CS,
that is, exposure to the US only. When subsequently testing the ability of the CS to elicit
fear, those animals given the extra shocks show significantly more fear than the animals
shocked only during training. This is believed to be due to the CS invoking the stored
memory of the US, which was greater or more aversive in the case of the animals who
received additional shocks. A study conducted in Fanselow’s laboratory revealed that
this representation is stored in (or accessed via) the B nucleus of the amygdala (Fanselow
& Gale, 2003). Rats were implanted with cannulae targeting the B-L nuclei, and after
61
recovery they were trained with several CS-mild US pairs. The following day, half of the
rats were infused with a GABA agonist to inactivate the B-L nuclei prior to the inflation
procedure, while the other half were infused just after the inflation procedure. It was
found that B-L nuclei inactivation during the procedure blocked US inflation, whereas
inactivation just after the inflation procedure was without effect. These results were
interpreted as support for the notion that the BLA contains a representation of the US
hedonic or motivational value. While no inhibitory avoidance studies have been
undertaken to evaluate further this specific hypothesis, some of the evidence presented
here could be seen as consonant with the idea that the B nucleus contains a representation
of the hedonic or motivational value of the US, which could in turn motivate avoidance
behavior. In particular, the functional lesion studies that revealed the necessity of protein
synthesis (memory storage) in the B nucleus, as well as the need for functional integrity
of the B nucleus during the recall test, are consistent with a hedonic/motivational
representation hypothesis. Furthermore, the actions of neuromodulatory drugs upon the
B nucleus could also be construed as consolidating or facilitating formation of the US
hedonic representation.
If this is the case, then during passive avoidance conditioning, the CS probably
becomes associated with the hedonic US representation in the B nucleus in a Pavlovian
manner. Later exposure to this CS, be it the context of the shock chamber or some other
more discrete cue, invokes the US hedonic representation, which in turn motivates
instrumental avoidance behavior. The C nucleus is likely not a critical component of this
pathway, since several studies have demonstrated a lack of involvement of the C nucleus
62
in instrumental avoidance behavior (Parent & McGaugh, 1994; Roozendaal & McGaugh,
1997; Killcross, et al., 1997). This in turn seems to indicate that activation of the C
nucleus-mediated components of the fear state (tachycardia, adrenal activation, pressor
responses, freezing, potentiated startle, etc.) are neither necessary nor sufficient to
motivate avoidance responses.
Critical issues remain to be addressed. If the avoidance memory is not being
stored in the amygdala, as McGaugh contends, then where? Answering this question
would allow definitive testing of the hypothesis that the amygdala modulates (and only
modulates) avoidance memory formation. Or alternatively: What neurotransmitters or
neuromodulators are used by the amygdala to modulate these extra-amygdalar avoidance
memories? Intracerebral administration of these agents would modulate memory
formation while at the same time being impervious to amygdala or stria terminalis
lesions. If instead (or additionally) the amygdala houses the US hedonic representation,
thereby allowing CSs to access that representation and motivate avoidance behavior, then
this hypothesis should be tested in an inhibitory avoidance experiment. Specifically, if
this hypothesis is valid, then the inflation procedure should be able to increase avoidance
behavior, i.e., lengthen entrance latencies and reduce time spent in the shock chamber.
Also, one might expect a deflation procedure or effect would weaken well-learned
avoidance behaviors. In conclusion, the amygdala does play a role in inhibitory
avoidance learning and memory, but the nature of that role is still in question. While the
modulatory and US representation hypotheses are plausible, with some evidence
consistent with each, neither can be asserted with confidence, at least based on the current
63
slate of inhibitory avoidance studies. However, as will be demonstrated in the remaining
sections, information gleaned from other instrumental learning paradigms does provide
more support for a role for the amygdala in hedonic valuation.
Active avoidance. The active avoidance task (described in Section 1.2.1) is a
form of negative reinforcement in which an animal performs some behavior in order to
relieve itself from an aversive stimulus. As with passive avoidance, the amygdala seems
to be involved in the learning and memory of active avoidance. Gross amygdala lesions
have been reported to impair the acquisition and retention of active avoidance memory
(reviewed in Sarter & Markowitsch, 1985). In a more recent study, excitotoxic lesion of
the B nucleus retarded acquisition of an active avoidance response (Lorenzini, et al.,
1991), yet the lesions did not seem to alter high-heat pain thresholds, reducing the
likelihood of an explanation based on nociceptive deficits. In an exemplary study by
Amorapanth, LeDoux, and Nader (2000), groups of rats were given bilateral electrolytic
lesions of either the L, B, or C nucleus. Upon recovery, they received 10 tone-shock
pairings in a unique context. Several days following this, the rats were placed into a
different (testing) chamber, which had a doorway opening into an adjoining chamber; the
fear-educing CS was then played for 20 s, and if the rat fled to the adjoining chamber, the
CS was terminated early, thus negatively reinforcing the avoidance response. No matter
the rat’s behavior, after CS termination the rat was briefly removed from the apparatus,
and then replaced into the testing chamber to repeat the trial. No shock was delivered at
any point in the 20 test trials. While sham and C nucleus lesioned animals acquired an
active avoidance response over the course of the 20 trials, animals with lesions of the L
64
or B nuclei failed to increase their avoidance of the CS over trials. Thus the L and B
nuclei seem to be important components of the active avoidance neural circuitry.
A number of other seemingly disparate methods, including shock probe
avoidance, the shuttlebox, and the water maze task, are in fact versions of active
avoidance insofar as they make use of the negative reinforcement paradigm. In SHOCK
PROBE AVOIDANCE, which can also be utilized in a passive avoidance procedure, a rat is
first classically conditioned to fear a Plexiglas rod wrapped in electrified wire. Next, in a
circular testing trough or alley, the probe moves toward the rat, who must flee to avoid
shock. Rats with large electrolytic amygdaloid lesions made far fewer avoidance
responses from an approaching shock probe relative to control rats (Blanchard &
Blanchard, 1972). In a follow-up experiment in the same study, amygdala lesions were
centered on one of the intercalated nuclei, with collateral damage to the C, Co, and M
nuclei, and instead of a shock probe, a cat followed the rat down the corridor (the cat
never attacked any rat, it merely followed). Again, lesioned rats avoided the cat far less
than control rats. Another variant of shock probe avoidance, DEFENSIVE BURYING,
involves a fixed, non-motile probe, which normally would constitute a passive avoidance
task, except that if rats are given the appropriate materials, e.g., bedding chips, they will
bury the shock probe; this represents an active behavior to reduce the presence of an
aversive stimulus. Electrolytic lesion of the C amygdala prior to training seems to
enhance burying during the acquisition session: in a 20 min trial where the probe was
continuously electrified, lesioned rats displayed greater probe burying than controls
(Roozendaal, Koolhaas & Bohus, 1991). However, during a subsequent retention test
65
when the probe was not electrified, the control rats displayed greater burying. This could
suggest a consolidation failure in the lesioned rats, except that the amygdaloid rats in the
Blanchard & Blanchard study (1972) should also have been able to demonstrate shock
probe and cat avoidance, since those trials did not depend upon long-term retention of an
avoidance memory. The C nucleus lesions of Roozendaal, et al. (1991) may have
appeared to enhance burying during acquisition because control rats spent most of their
time freezing, a behavior that is incompatible with burying and is disrupted by C nucleus
lesion; thus during acquisition, the control rats perhaps adopted a passive avoidance
strategy, or perhaps they were simply unable to move due to classically conditioned
freezing. During the subsequent retention test, however, the control rats engaged in
defensive burying.
The SHUTTLEBOX is an apparatus with two independently electrifiable chambers
separated by a small hurdle or a partition with a door. A tone CS sounds shortly before
delivery of shock to the chamber the rat currently occupies, giving the rat the opportunity
to flee to the other chamber to avoid electrocution. This more complex avoidance task
(sometimes called “two-way active avoidance”) normally requires on the order of 50
trials to acquire (Savonenko, et al., 2003). Overtraining (ad nauseam—roughly 150-200
trials) in a shuttlebox apparatus rats seems to preserve active avoidance responses after
posttraining electrolytic BLA damage (Thatcher & Kimble, 1966). Lesion of the BLA in
non-overtrained animals (n=4) appeared to cause retrograde amnesia for active avoidance
memory, but the scant number of animals employed in this study precluded statistical
verification of this effect. More reliable effects were achieved in a recent functional
66
lesion experiment. Blockade of NMDA receptors in the B nucleus during training in a
shuttlebox prevented acquisition of an active avoidance response; this deficit was seen
during training (as opposed to during a subsequent retention test) demonstrating an
acquisition deficit (Savonenko, et al., 2003). Furthermore, freezing responses during
training in the shuttlebox were normal in these animals, meaning that short-term classical
fear conditioning was acquired, but not the instrumental avoidance response. Thus the B
nucleus and its NMDA receptors are required for the acquisition of two-way avoidance
responses.
Another form of active avoidance is the WATER MAZE TASK, wherein an animal is
plunked into a tank of water and must find a small platform hidden just under the surface,
lest it continue to swim. After a few successive trials, the animals quickly learn where to
locate the platform; retention is measured later by their latency to escape to the platform.
Infusion of the opioid peptide β-endorphin into the amygdala just prior to water maze
training impaired memory of the platform location 24 h later, an effect that was
counteracted by simultaneous infusion of a β adrenoreceptor agonist (Introini-Collison, et
al., 1995), just as in inhibitory avoidance. A glucocorticoid antagonist administered in
the B nucleus just prior to a six-trial training session lengthened the latency to locate the
platform 48 h later, while acquisition during training was unimpaired relative to vehicle-
infused controls (Roozendaal & McGaugh, 1997). Thus disruption of neuromodulatory
activity in the B nucleus seemed to prevent consolidation of the avoidance memory.
Intraamygdala naloxone infusion also enhanced retention in an active avoidance Y-maze
task, and, as in inhibitory avoidance, this effect was blocked by simultaneous infusion of
67
the β norepinephrine antagonist propranolol (Introini-Collison, et al., 1989). Hence
memory for two rather distinct active avoidance tasks can be modified by intraamygdala
infusions of neuromodulators.
The fact that amygdala manipulations interfere with all of these different
avoidance types suggests that it does something important for and perhaps common to all
of them. The experimental findings are similar to those of passive avoidance, if not as
fulsome, suggesting similar roles for the amygdala in the two instrumental paradigms.
As one possibility, just as in passive avoidance learning, modulatory hormones and
neurotransmitters act within the amygdala to modify avoidance memory strength, perhaps
indicating a modulatory role, but this data is also consonant with the notion of memory
storage within the B nucleus. As an alternative possibility, the B nucleus of the amygdala
may evaluate and store a representation of US hedonic value. Amorapanth et al. (2000)
suggest that the B nucleus motivates the instrumental avoidance response by acting as a
conditioned reinforcer, which is essentially the same as the idea proposed by Fanselow
and Gale (2003): specifically, the B nucleus is the repository of the US representation of
hedonic or motivational value, and neutral stimuli can be conditioned to access that
representation and motivate behaviors themselves.
1.2.2 Appetitive conditioning. Appetitive conditioning uses rewards to entice
the animal to perform a task. A REWARD is a stimulus that provokes an approach response
from an animal (White, 1989). Food and water are rewards for animals that are hungry or
thirsty. Reward-evoked approach responses are variable and rather arbitrary, and could
for instance take the form of stair climbing, place holding, chain-pulling, or, simply, an
68
approach. In a simple appetitive conditioning setup, a hungry rat is introduced into a
conditioning chamber containing a food hopper and a bar-press-operated switch.
Pressing the bar triggers the hopper to release a food pellet reward. After the rat has
learned this relationship, the number of bar-presses required to release a pellet can be
increased so that the rat will work vigorously to obtain the food reward. The bar-pressing
response in this example is also an approach response, and it is positively reinforced by
the food reward.
Rats are not impaired in simple appetitive conditioning by gross amygdala lesion
(Baxter & Murray, 2002). In the Y-maze inhibitory avoidance study by Cahill and
McGaugh (1990) cited above (Section 1.2.1), prior to avoidance conditioning, dehydrated
rats were first trained to receive their daily allotment of water in one of the arms of the Y-
maze using a positive reinforcement paradigm. One arm of the maze, which was dimly
backlit, bore the water spout, while the other arm, which was dark, was empty. Both
amygdala-lesioned and control groups learned this task handily, showing almost no delay
in approaching the backlit arm after a single training trial. In a follow up experiment in
the same study, amygdala lesioned rats had no difficulty associating an amyl acetate odor
with a sucrose reward, as evidenced by their preference for an amyl acetate-scented room
over an odorless room during the test session. Not surprisingly, more localized lesions
are likewise without effect on simple reward-motivated instrumental learning. In one
study using food-reinforcement, animals with neurotoxic lesions encompassing both the
B and L nuclei acquired appetitive instrumental responses as well as control animals
(Balleine, Killcross & Dickinson, 2003). In a conditioning chamber bearing a bar switch
69
and a chain switch, bar-pressing and chain-pulling were reinforced with food pellets and
maltodextrin solution, respectively, and amygdala lesions were without effect on
instrumental learning. In another study, rats with NMDA lesions of either the B or C
nucleus were perfectly capable of acquiring bar-pressing responses to obtain food pellets
and aqueous sucrose in a two-lever conditioning chamber (Corbit & Balleine, 2005).
Thus neither large nor restricted lesions of the amygdala interfere with the acquisition of
simple appetitive instrumental conditioned responses.
However, the learning of more complex appetitive instrumental tasks is disrupted
by amygdala lesion. For instance, through the process of REWARD DEVALUATION, the
ability of a stimulus to elicit an appetitive response is reduced. One way to devalue a
reward is the technique of SPECIFIC SATIATION, wherein the reward is freely given to the
animal for an hour, letting it consume the reward to its satisfaction; this temporarily
diminishes subsequent instrumental responding to obtain the sated reward, although other
rewards are still pursued. In a follow-up experiment to the bar- and chain-switch task just
described (Balleine, et al., 2003), lesioned and control rats were trained to receive food
pellets and maltodextrin solution, respectively. The next day, half of the rats were given
the maltodextrin to the point of satiation, and then placed into the conditioning chamber
for an extinction test (i.e., no foods were delivered). While control rats made most of
their responses on the pellet-reinforced bar switch, amygdala B-L lesioned rats failed to
differentiate between the two switches, responding equally to both
12
. Through two more
experiments, the authors determined that the lesioned rats could appropriately
12
The other half of the rats received pellet satiation, yielding the same (though complementary) results.
70
differentiate the two switches, but believed that they could not access or encode the
altered reward value of specific foods. Thus the diminished hedonic value of the
maltodextrin could not be activated by cues associated with chain-pulling; since simple
appetitive instrumental conditioning can be mediated by extra-amygdalar circuitry, these
B-L lesioned animals know that chain-pulling gives them maltodextrin, but they do not
seem to remember that they don’t want it. Thus responses on the bar and chain switches
were not distinguished. When then given free access to both reward foods, they ate more
of the food pellets than maltodextrin, demonstrating an ability to differentiate the rewards
in vivo (perhaps through hypothalamic eating and satiety systems?); but during an
extinction test, when all they have to go on is the hedonic representation of the
maltodextrin, these rats fail because they do not have that representation. In a similar
experiment, Corbit and Balleine (2005) trained B-, C-, or sham-lesioned rats to press two
bars in a conditioning chamber for two food rewards. During the devaluation phase, B
nucleus lesions were found to impair discriminated responding, while C nucleus and
sham lesions were without effect on proper responding.
Reward devaluation can also be implemented by simply dispensing less of the
reward (in effect a penalty paradigm). For instance, hungry rats may be trained to
traverse a straight runway in order to obtain 10 food pellets. Over multiple training trials
their run times will decrease substantially. Yet if the reward is then permanently reduced
to 1 food pellet, run times will temporarily increase, and then gradually return to short-
latency asymptotic levels. If a cholinergic agonist is infused into the amygdala
immediately after the first training trial in which the reduced reward was implemented,
71
subsequent run times will be even further augmented (Salinas, Introini-Collison, Dalmaz
& McGaugh, 1997). Also, while a similar single posttraining blockade of amygdalar β-
adrenoreceptors with propranolol does not prevent the devaluation-induced run-time
increase, it does seem to hasten the subsequent return to pre-devaluation run times;
excitotoxic lesion of the BLA has an effect similar to propranolol (Salinas, Parent &
McGaugh, 1996). These findings suggest that blocking norepinephrine in the amygdala
lessens the aversiveness of reward devaluation, while acetylcholine stimulation enhances
it. The modulatory actions of these drugs in the amygdala are in line with the idea that
the amygdala (the BLA in particular) is part of a reward valuation system or memory
representation. Also note the parallel between reward devaluation and the inflation
procedure of inhibitory avoidance, in which the aversiveness of the US is increased, and
in which the BLA is believed to play the same role of an aversiveness valuation system or
memory representation (Cardinal, et al., 2002; Fanselow & Gale, 2003).
A recent experiment is consistent with this notion (Corbit & Balleine, 2005), a
study using a kind of conditioned reward technique known as PAVLOVIAN-INSTRUMENTAL
TRANSFER, in which a neutral stimulus that has been classically associated with a primary
reward can enhance instrumental responding for that reward. In this experiment, rats
with B, C, or sham lesions were classically conditioned to associate one of three unique
auditory stimuli with each of food pellets, sucrose, and polycose (corn starch), i.e., each
reward was paired with a distinct tone. They were then instrumentally conditioned in a
two-lever training box to bar-press in order to receive two of these food rewards, but not
the third: for instance, some rats pressed the left bar for pellets and the right bar for
72
sucrose, but could not obtain polycose. Following this, they were given an extinction test
in which only one of the two levers was available and no rewards were given; bar-
pressing was measured while the three auditory CSs were played in random order every 2
min. In sham rats, bar-pressing responses were greatest in the presence of the auditory
CS associated with the reward obtained with the accessible lever; responses were not
potentiated by the CS associated with the reward earned on the non-accessible lever—this
is termed OUTCOME-SPECIFIC Pavlovian-instrumental transfer, since the CS can only
invigorate responding if it was associated with the accessible outcome (reward).
However, in the presence of the CS associated with the third food reward (which was
never instrumentally conditioned), bar-pressing was also enhanced, a phenomenon named
GENERAL Pavlovian-instrumental transfer. The Pavlovian stimuli invigorated
instrumental responding in the same way that rewards act. Rats with B nucleus lesions
had outcome-specific transfer deficits, while those with C lesions had general transfer
deficits. The authors understood these results to indicate that rats with B or BLA lesions
probably retain a general “motivational arousal” system in the C nucleus that could
influence instrumental responding, as long as specific response-outcome associations are
not required. The B nucleus itself may allow neutral stimuli to access the representations
of specific reward items, which upon activation would stimulate approach responses.
These findings were replicated in an experiment using a much different reward
stimulus (Everitt, Cador & Robbins, 1989). In a conditioning chamber bearing a bar
switch, male rats were allowed access to estrous female rats. A light CS was paired with
entry of the estrous female into the chamber in a Pavlovian manner. Then males were
73
trained to bar-press to gain access to a female, initially with a 1:1 response-reward ratio,
but then with increasing response requirements as follows: bar-pressing 10 times turned
on the light for a short time; turning on the light 10 times finally allowed access to the
female. After training, B nucleus or sham lesions were made. In the test phase, access to
the female was given for the first bar press after a 15 min period, irrespective of the
number of presses (and light toggles) made in that 15 min. In rats with sham lesions, the
light helped to sustain bar-pressing during the long interval, i.e., it acted as a conditioned
reward. Rats with excitotoxic lesions of the L and B nuclei, on the other hand, were
much less vigorous in their bar-pressing during the interval. The sexual performances of
the sham and lesion groups were not significantly different, ruling out a general sexual
deficit in the lesion group. In another test session, deactivating the lightbulb CS did not
significantly alter the rate of bar-pressing by B lesioned animals, but produced a steep
drop in bar pressing by control animals, even though the female was still attained for the
first bar press after 15 min. This further indicates that the light was indeed acting as a
conditioned reward for the sham animals, because in its absence bar-pressing declined,
and that the B nucleus (and perhaps the L) was mediating this conditioned reward effect.
Another more complex conditioning procedure that is affected by BLA lesion is
SECOND-ORDER INSTRUMENTAL CONDITIONING. This involves first using classical
conditioning to associate a neutral CS, a light for example, with delivery of a primary
reinforcer, such as food pellets; acquisition is measured by approach responses to the
light (autoshaping—remember, autoshaped approach responses are reflexive Pavlovian
CRs), which is now a conditioned reward. In the instrumental conditioning phase, the
74
animal is placed into a training box with two bar switches; pressing one bar triggers
exposure to the conditioned reinforcer, while the other bar does nothing. Food pellets are
never delivered in this phase of training, so there is no primary reward stimulus to
associate with bar-pressing. Normal rats will acquire the bar-pressing response to gain
access to the CS, i.e., the light alone can reinforce the novel bar-pressing task. Rats with
BLA lesions, on the other hand, acquire the classical autoshaping task normally, but do
not subsequently acquire a discriminated instrumental response, instead pressing both bar
switches a moderate amount (Burns, Robbins & Everitt, 1993; Everitt, Cardinal, Hall,
Parkinson & Robbins, 2000). Thus these experiments provide further evidence that the
BLA amygdala is involved in the ability of a CS to acquire a conditioned reward
property.
All together, the B nucleus of the amygdala seems to participate in events and
behaviors involving conditioned rewards: the effects seen in reward devaluation,
outcome-specific Pavlovian-instrumental transfer, and second-order instrumental
conditioning all appear to rely upon CS activation of a stored representation of the
hedonic value of the reward. On the other hand, responses to primary rewards
themselves are unaffected. In the larger scheme of brain pathways involved in
hedonically motivated instrumental behavior, it is generally acknowledged that the
nucleus accumbens figures as an important reward center (White, 1989; Ikemoto &
Panksepp, 1999; DeSousa & Vaccarino, 2001; Cardinal, et al., 2002). Not surprisingly,
then, the B nucleus of the amygdala is the major source of amygdalar afferents to the
nucleus accumbens (Pitkänen, 2000). In sum, these facts, together with the results
75
described above, suggest that the B nucleus may be a locus where neutral stimuli (the CS
pathway) can influence appetitive instrumental behaviors, either by activating a hedonic
representation stored in the B nucleus, or by gaining access to important B nucleus inputs
of a central reward nexus in the nucleus accumbens.
1.3 Recapitulation and conclusions
In toto, the amygdala plays a critical role in learning to anticipate events that have
consequences relevant to the survival of the animal. The amygdala associates stimuli that
are concomitant to events of an aversive or appealing nature, and consequently prepares
the animal to respond to those events in the future when re-exposed to those
accompanying conditioned stimuli. This is most patent in the involvement of the
amygdala in emotional reactions related to a fear-inducing stimulus, such as electric
shock. Lesion of the amygdala disrupts fear reactions whether using classical or
instrumental conditioning paradigms, and virtually every constituent part of the fear state
is impaired, including autonomic responses such as heart rate change, increased blood
pressure, stress hormone release, perspiration, and defecation; behavioral responses like
freezing, fleeing, and avoidance; and cognitive responses such as potentiated startle and
hypoalgesia. This panoply of responses is affected by amygdala lesion because the
associative memory mechanism for classically conditioned fear, the engram, resides in
the L and probably the C nucleus of the amygdala. This assertion is supported by a
legion of anatomical and experimental data.
76
Firstly, the L nucleus in particular, but also the C nucleus, is a site where CS and
US sensory pathways converge. Visual and auditory information from the sensory
cortices and thalamus merge with nociceptive information from the thalamus and
parabrachial nucleus. Additionally, electrophysiological recording in the L nucleus has
revealed that contingent activation of the CS and US pathways results in long-term
potentiation and increased synaptic potentials from CS pathway activation. These
potentials correlate with behavioral measures of fear, growing during acquisition training
and abating during extinction. Furthermore, prevention of long-term potentiation by
blocking NMDA receptors in the L nucleus averts fear learning, and blockade of protein
synthesis eliminates long-term, but not short-term fear memory. Although
electrophysiological experiments have not been conducted for the C nucleus, functional
inactivation and blockade of protein synthesis in the C both disrupt classical fear
memory. The C nucleus has widespread efferent projections in the brainstem, contacting
a multitude of brainstem nuclei that control the individual responses seen in the fear state.
Thus the C nucleus forms the control center in the CR pathway, allowing it to activate the
various classically conditioned responses in a coordinated fashion; the L nucleus has
projections to the C nucleus, giving it the ability to activate the CR pathway as well
(Figure 1.4).
Aversive instrumental conditioning also relies upon the amygdala. Lesions of the
amygdala adversely affect aversive instrumental responses such as passive and active
avoidance. While various lesion studies aimed at particular nuclei of the amygdala are
inconsistent in their findings, this variance may be due to the training protocols
77
employed, with multi-trial acquisition methods tending to produce conditioning that is
susceptible to B nucleus lesion. Scads of data indicate that the B nucleus is complicit in
aversive instrumental conditioning. Many of these experiments have focused on the
finding that numerous neuromodulators exert their memory-modulating effects by acting
upon the B nucleus. This has led some to hypothesize that the role of the amygdala is
solely to modulate instrumental memory formation in other brain areas.
However, a more plausible hypothesis is that the B nucleus is an associational and
memory storage site, an engram locus in which CSs can become associated with a
representation of the hedonic value of the US. In line with this, the B nucleus is a
recipient of CS and US pathways. Exteroceptive sensory information transmitted to the
B nucleus arises from the sensory cortices and the hippocampus, which is believed to
contribute complex, polymodal sensory information. Nociceptive information reaches the
B nucleus from the insular cortex, parabrachial nucleus, and cuneiform nucleus. Inputs
from the superjacent L nucleus could contribute either CS or US information. In further
support of this hypothesis, it was found that functional inactivation of the B nucleus
during the inflation procedure blocked the inflation effect; it was believed that
inactivation of the B prevented the inflation procedure from strengthening the hedonic
representation of the US. Thus during fear conditioning, CS inputs could become
associated with the US representation in the B nucleus. Efferent projections of the B
nucleus to forebrain structures such as the nucleus accumbens could then allow the CS to
influence decision-making and instrumental behaviors (Figure 1.4).
78
Figure 1.4 An overall model of the role of the amygdala in fear conditioning. The L to C nucleus
pathway mediates the acquisition and expression of classically conditioned (Pavlovian) responses. The L
to B nucleus pathway mediates the acquisition of a hedonic representation of the US, which may be
stored in the B nucleus, which can be activated by the CS pathway, and which influences decision-
making and instrumental responses. The appetitive classical and instrumental paradigms seem to fall
along exactly the same substrates in the amygdala as the aversive paradigms.
Overall, then, the classical and instrumental response systems are parallel
processes in the amygdala. The L to C nucleus pathway mediates the acquisition of
Pavlovian responses, and expresses them through its connections with brainstem nuclei.
The L to B nucleus pathway probably contains a hedonic representation of the US, and its
efferents target forebrain areas involved in decision-making and instrumental responding.
This division of labor is supported by studies of appetitive classical and instrumental
conditioning, which were discussed tersely in order merely to illustrate this support.
Appetitive classical conditioning paradigms such as conditioned orienting and
autoshaping are impaired by lesions of the C nucleus but are unaffected by lesions of the
B nucleus. Appetitive instrumental conditioning paradigms such as reward devaluation,
79
conditioned reward, and outcome-specific Pavlovian-instrumental transfer are disrupted
by lesions of the B nucleus, but not by lesions of the C nucleus. Although simple
positively reinforced responses are not affected by amygdala lesion, these responses also
do not require CS-activated hedonic representations for performance of the task, since
they are reinforced by the primary rewards themselves, i.e., the rat does not need to
retrieve a memory of the reward in order to perform the response, since it earns the actual
reward. There are a number of different motivational systems for the acquisition of
various primary rewards (food, water, sodium, sex, etc.), and these systems may be
capable of motivating instrumental performance for obtaining their specific reward
without need of the amygdala. But the amygdala (the BLA complex) seems to be part of
a single system that allows conditioned stimuli to access representations of these various
primary rewards, and punishers as well, and consequently can motivate instrumental
responding in the absence of the reinforcers themselves. Together, the C nucleus-
mediated response pathway and the B nucleus-mediated system can allow an animal to
associate CSs with hedonically valent USs and to react appropriately to those CSs in the
future: Pavlovian responses automatically prepare the animal for an encounter with the
US, and instrumental responses affect the animal’s voluntary behaviors by informing
decision-making processes.
80
2
THE AMYGDALA IN CONDITIONED TASTE AVERSION
2.0 Conditioned taste aversion: Aversive memory of a special kind
The amygdala plays a part in one more important learning paradigm: conditioned
taste aversion. Known by many other names
1
, CONDITIONED TASTE AVERSION (CTA) is a
conditioning procedure in which a tastant (the conditioned stimulus, CS) consumed by an
animal is followed by an injection of a toxin (the unconditioned stimulus, US) that causes
illness (the unconditioned response, UR), inducing the animal to avoid the taste in
subsequent exposures (the conditioned response, CR). The phenomenon is traditionally
likened to classical conditioning, though it actually has components of both classical and
instrumental conditioning. While bearing a few unique features, it is rather similar to fear
conditioning of both classical and instrumental forms. Some of these similarities—and
the idiosyncrasies—will be witnessed in the following description of the properties of
CTA.
One property has been termed one-trial learning, whereby a single pairing of CS
and US can result in a strong aversion to the CS. Contextual and cued shock-fear
conditioning are forms of classical conditioning that can be acquired in a single trial,
though asymptotic performance requires around 5 trials to establish (Maren, 1998; see
Section 1.1 for descriptions of these paradigms). The instrumental conditioning
1
Bait-shyness, conditioned taste avoidance, conditioned flavor aversion, conditioned consumption
reduction, the Bérnaise sauce phenomenon, toxiphobic conditioning, or the patronymic “Garcia effect.”
81
paradigm of passive avoidance of shock can also be conditioned in a single trial (see
Section 1.2.1 for description of this paradigm). Similarly, CTA, which has both classical
and instrumental components, is capable of one-trial learning (though only under certain
conditions, see below). This capacity may reflect the vital significance of fear
conditioning, passive avoidance, and CTA for evolutionary fitness.
Another interesting property of CTA concerns the types of stimuli that can serve
as CS and US. For the CS, these types of stimuli comprise a more restricted range of
conditionable stimuli than that seen in fear conditioning. For most species, substances
with some gustatory component serve as the most effective CSs. Standard CS solutions
used in CTA research include saccharin, sucrose, HCl, NaCl, and quinine, all solutions
that activate lingual chemoreceptors. Odorous solutions, like almond-scented water for
instance, are mildly avoided after an acquisition trial, whereas almond-scented saccharin
solution, with both olfactory and gustatory elements, is strongly avoided. Other features
of an ingestible substance such as texture, temperature, and appearance can serve as
conditioning cues to distinguish that substance from another, but only when that cue is in
compound with a gustatory feature. In contrast, recall that in fear conditioning, visual,
auditory, olfactory, and environmental (contextual) stimuli were all effective CSs.
Very little exposure to a gustatory CS is needed to condition a CTA. Yamamoto
found that 500 licks (about 2.5 mL) of a NaCl CS were sufficient for conditioning a
strong CTA in rats (Yamamoto, Shimura, Sako, Yasoshima & Sakai, 1994). Others have
found that even less works. Barker (1976) conditioned a strong CTA by giving about 300
µL of saccharin (consumed in about 8 s). Another experiment in the same study (Barker,
82
1976), revealed that as little as 5 s of exposure to saccharin (approximately 125 µL) was
sufficient to condition a strong CTA. In comparison, cued fear conditioning experiments
typically employ several CS-US pairings to achieve strong conditioning, with individual
CS durations of around 5 or 10 s, and summing to a total of 20-30 s of exposure to the CS
over the several pairings.
Novelty is also an important quality for the CS. While it is easy to establish a
CTA to a novel taste, a familiar CS can resist one-trial aversive conditioning (Berman,
Hazvi, Neduva & Dudai, 2000; Koh & Bernstein, 2005), and may require multiple
acquisition trials to become aversive (St. Andre & Reilly, 2007). Although they can still
be conditioned, highly familiar substances such as water or standard laboratory chow
pellets are more resistant to aversive conditioning (Smith & Schaeffer, 1967; Garcia,
Hankins, Robinson & Vogt, 1972). Similarly, in contextual fear conditioning, a familiar
environment fails to become associated with footshock (Cardinal, Parkinson, Hall &
Everitt, 2002).
Unlike fear conditioning and passive avoidance paradigms, shock is an ineffective
US in taste aversion conditioning. Instead, a multitude of chemical substances can serve
as the US in taste aversion conditioning, from poisons to reinforcing drugs to apparently
innocuous edibles such as sesame oil (Bermudez-Rattoni, 1998). Paradoxically, some
prototypical poisons, such as cyanide, are ineffective as US agents (Ionescu & Buresova,
1977). Thus the actual sensory modalities that are substrates for the US are not well
understood, although visceral malaise is generally believed to be a key modality.
Ionizing radiation was the US in the very first CTA experiment (Garcia, Kimeldorf &
83
Koelling, 1955), and the authors of that work cite evidence that radiation exposure
adversely affects the enteric system (Conard, 1951). An isotonic solution of lithium
chloride (LiCl) is the most frequently employed conditioning agent, and ingestion of this
salt is known to produce visceral malaise within 5-10 min and lasting for approximately 2
hours (Nachman, 1970; Lamprecht & Dudai, 2000). Administration of LiCl through
either intraperitoneal (intra-abdominal) or intraduodenal routes stimulates sensory
portions of the vagus and splanchnic nerves, which innervate the gut; furthermore these
fibers were determined to be pain stimulus-mediating C-fibers (Niijima & Yamamoto,
1994). Consumption of LiCl also produces signs of illness such as listlessness, diarrhea,
pica, immobility, vomiting (though rats are incapable of emesis) and increases of blood
corticosterone levels, and in humans it can cause feelings of nausea, pain, and a negative
hedonic state (Barker, Smith & Suarez, 1977; Braveman, 1977; Karniol, Dalton & Lader,
1978; Mallinckrodt Baker, Inc., Material Safety Data Sheet). Body rotation can also act
as a US, presumably from the feelings of nausea stemming from dizziness (Hutchison,
1973). There are, however, other kinds of stimuli that can function as a US that do not
necessarily produce malaise. Many rewarding drugs have been discovered to attenuate
drinking when their ingestion follows CS exposure (Bermudez-Rattoni, 1998), although
conditioning tends to require more trials with these USs. This exposes an enigma in the
CTA field, namely that a substance that can reinforce its own consumption can
simultaneously condition an animal to avoid a flavored tincture of that drug. One would
expect the pleasurable effects of the drugs to enhance CS consumption, as other learned
behaviors, like bar-pressing, increase when followed by drug reward in an instrumental
84
conditioning paradigm. This raises the question of whether the essential stimulus
comprising the US is actually malaise, or perhaps whether some non-pain sensory system
constitutes or contributes to the US pathway. Clearly, cutaneous pain does not contribute
to this pathway as this type of pain is an ineffective US in CTA (Green, Bouzas &
Rachlin, 1972), in contrast to fear learning where electric shock is the prototypical US.
Another keen feature of CTA, one that was initially quite controversial, is the
long interval between the CS and the US that can sustain learning during the acquisition
phase of CTA. In many kinds of classical conditioning, the CS and US need to be
presented within seconds, at most, of each other (Abrams & Kandel, 1988). In auditory
fear conditioning, the standard CS-US delay period is 10 s, and a 30 s delay period
produces a weaker fear response. Conditioned taste aversion, on the other hand, can
tolerate CS-US trace intervals of hours (Garcia, Ervin & Koelling, 1966; Garcia, Hankins
& Rusiniak, 1974; Riley & Tuck, 1985). Taste aversion has even been reported to bridge
a 24 h interval (Etscorn & Stephens, 1973). In this experiment, separate groups of male
and female rats were used; a single high dose of cyclophosphamide (a nitrogen mustard
compound) given 24 h after a saccharin CS was capable of conditioning a CTA in both
sexes. At the other end of the spectrum, a study by Schafe, Sollars, and Bernstein (1995)
examined the minimum interval a CTA can bear; since most forms of classical
conditioning can take place in seconds, they explored whether CTA is the same. In order
to time delivery of the stimuli accurately, they utilized intraoral cannulae and intravenous
85
catheters for delivery of CS and US, respectively. They found that intervals of 15 or 30
min sustained strong CTA, but that a 10 s interval did not support learning.
2
The behavioral expressions of CTA comprise both classical and instrumental
components. When the CS is sampled during the expression (testing) phase, illness-like
CRs are emitted. When coyotes were taste conditioned against lamb, one behavior they
exhibited upon re-exposure to lamb was eating-of-grass, a behavior that dogs tend to
engage in to alleviate gastric distress. These coyotes were also reported to vomit in
response to US stimulation (LiCl ingestion and injection) and one vomited after exposure
to the CS (Gustafson, Garcia, Hankins & Rusiniak, 1974). Rats show an illness behavior
called lying-on-belly, where they lie prone and immobile; this behavior occurs as a UR to
LiCl and a CR to an aversively conditioned tastant (Meachum & Bernstein, 1990).
Although it is not perfectly clear whether eating-of-grass and lying-on-belly are in fact
Pavlovian URs and CRs, or whether they are instrumental behaviors that the animals emit
as a palliative action, they at least seem to indicate that the animals feel ill. Additionally,
ultrasonic vocalizations that signal misery (22 kHz) are emitted after exposure to
naloxone, which can serve as a US in a CTA paradigm (Stolerman, Pilcher & D’Mello,
1978), and they are evoked by the CS after CTA acquisition (Burgdorf, Knutson,
Panksepp & Shippenberg, 2001). Thus, similar to the fear state in fear learning, illness-
like reactions seem to be classically evoked by the CS after taste aversion learning.
2
Even though they found no significant effect for a 10 s interval, they used only 4 animals per group,
meaning their statistical power to detect a smaller effect was probably low. Thus a 10 s interval may be
sufficient to establish a weak CTA.
86
The orofacial responses to the CS change from positive-ingestive to aversive-
rejective after taste aversion learning, and these changes are considered CRs in CTA
learning (Breslin, Spector & Grill, 1992). When an artificial situation is introduced such
that the sucrose or saccharin CS is infused directly into the mouth through an intraoral
catheter, thus imposing forced exposure to the CS, rats show positive-ingestive responses
(such as such as tongue protrusion, paw lick, and mouth movements or mastication)
before pairing the CS with the US, and after pairing, they display negative-rejective
responses (such as gape, chin rub, passive drip, head shake, and forelimb flail) and a
decrease in positive-ingestive responses. These orofacial responses are viewed as
classical components of learning and are thought to reflect underlying emotional
reactions to the CS.
Avoidance responses constitute the instrumental components of CTA expression.
After tasting the CS at the start of the expression test, a rat will passively avoid the CS,
i.e. it will not approach and consume the CS when allowed direct control over their
exposure to this stimulus. Some rats may even push the sipper bottles out of their cages
(personal observation), an active avoidance response. Coyotes will actively destroy the
aversive food, either by burying it, or by urinating on it (Gustavson, et al., 1974). Hence,
animals will engage in active and passive instrumental avoidance responses to the CS
after taste aversion learning, just as they do in instrumental fear learning when a rat will
not enter an environment in which it was previously exposed to electric shock.
In all, then, CTA has some unique properties, as well as some features akin to
standard fear conditioning paradigms. The CS and US modalities are peculiar to CTA,
87
with gustatory CSs and internally sensed USs. Yet CTA and fear conditioning have a
similarity in their conditionability: both kinds of learning can be acquired fairly robustly
in one trial. They are also similar in that a slate of aversive responses (the fear state or
illness) can be evoked by the CS after conditioning: pairing of a gustatory CS and a LiCl
US allows the CS to evoke a LiCl illness-like response—with a classically evoked
constellation of illness responses. Furthermore, both CTA and fear conditioning can
entail an instrumentally motivated avoidance of the CS. And, as will be seen next, there
is homology in the neural pathways subserving the two learning paradigms.
2.1 Some brain areas involved in CTA
The CS pathway. Much is known about the afferent pathways carrying CS and
US sensory information to higher brain areas. As discussed above, the CS modality is
primarily gustation, while other stimulus properties such as odor, texture, temperature,
viscosity, and astringency act as accessory cues that embellish and combine with the
gustatory stimulus to create the complex perception of a flavor. Tastants activate
ionotropic and metabotropic receptors in the membranes of taste receptor cells. These
cells are grouped into teardrop-shaped structures called taste buds, structures that line the
tongue and throat. Taste receptor cells contact the fibers of several cranial nerves to
communicate taste sensations to the brain.
There are a number of elemental taste qualities: sweet, sour, salty, bitter, and
umami (from the Japanese word “savory”; the taste of high protein foods and amino acids
such as monosodium glutamate). Each of these taste qualities activates a specific
88
chemoreceptor on the taste receptor cells of the oral cavity. Human gastronomic
experience suggests that in addition to the five widely acknowledged taste qualities,
others may exist. For instance, foods high in iron content, like spinach or blood, can have
a metallic component to their flavor. High-fat foods taste differently than their fat-free
counterparts; identification of a TRP (transient receptor potential) channel sensitive to
fatty acids suggests that lipids may constitute another basic taste dimension (Laugerette,
et al., 2005; Gilbertson, et al., 2005).
Taste information is transmitted from taste buds on the tongue, velum, epiglottis,
and pharynx to the central nervous system via the facial (VII
th
), glossopharyngeal (IX
th
),
and vagus (X
th
) cranial nerves (Figure 2.0). Two branches of the facial nerve, the chorda
tympani and the superficial petrosal nerve, innervate the anterior two-thirds of the tongue
and the velum, respectively. The lingual-tonsilar arm of the glossopharyngeal nerve
contacts the taste receptor cells of the posterior third of the tongue, while the superior
laryngeal segment of the vagus nerve innervates the pharyngeal taste buds. These nerves
enter the medulla and terminate most densely in the rostrolateral portion of the nucleus of
the solitary tract (NST), corresponding to its ventrolateral subnucleus and the lateral
division of the rostral subnucleus, with less dense fibers terminating more posteriorly in
the zone flanking the area postrema (Hamilton & Norgren, 1984). Gustatory stimulation
of these nerves can reach the NST in as little as 5 ms (Ogawa & Kaisaku, 1982). As
mentioned above, other stimulus features contribute to the complex perception of flavor,
such as the tactile and thermal properties of a food. These features are detected by the
somatosensory system of the mouth, and are relayed to the central nervous system by the
89
lingual nerve and inferior alveolar branches of the trigeminal nerve (Vth), fibers that
synapse in the same rostral NST areas as the gustatory nerves (Hamilton & Norgren,
1984).
The NST next relays gustatory input up to the parabrachial nucleus (PBN), a
group of neurons that ring the superior cerebellar peduncle in the pons. The rostral NST
projects ipsilaterally to the medial parabrachial nucleus (medial and external medial
subnuclei) and to the medial half of the lateral parabrachial nucleus (ventral lateral
subnucleus and “waist region”; Norgren, 1978; Herbert, Moga & Saper, 1990). These
solitary afferents to the PBN have been shown to respond to electrical stimulation of the
chorda tympani VII
th
and lingual-tonsilar IX
th
(Ogawa & Kaisaku, 1982), providing
further confirmation of the taste pathway.
From the PBN, gustatory information is cast widely about the rest of the brain in a
number of parallel paths. In the diencephalon, PBN waist region efferents reach the
parvocellular ventroposteromedial thalamic nucleus (VPMpc, aka the parvocellular
ventral posterior nucleus), the main thalamic gustatory zone (Krout & Loewy, 2000;
Lundy & Norgren, 2004). Other adjacent thalamic nuclei also receive efferents from the
medial and ventral lateral parabrachial subnuclei, such as the central medial,
parafascicular, paracentral, oval paracentral, and rhomboid nuclei, constituents of a group
(along with the VPMpc) referred to as the midline and intralaminar thalamic complex
[MITC] (Saper & Loewy, 1980; Bester, Bourgeais, Villanueva, Besson & Bernard, 1999;
Krout & Loewy, 2000). Additionally, a portion of the ventral thalamus termed the zona
incerta may be a recipient of PBN gustatory input (Lundy & Norgren, 2004). As for
90
projections to the hypothalamus, the lateral hypothalamic area, a locus of great
importance in eating behavior, is the only target (Saper & Loewy, 1980; Bester, Besson
& Bernard, 1997). The parabrachial gustatory tract also reaches a number of
telencephalic structures. Importantly for this proposal, the PBN projects to the amygdala,
with a higher density of fiber terminal arborization in the central (C) amygdala (medial
subnucleus) than in the VPMpc (Lundy & Norgren, 2004), and with less dense but still
substantial projections to the basal (B), accessory basal (AB), and cortical (Co) nuclei
(Saper & Loewy, 1980; Bernard, Alden & Besson, 1993). Other areas in the
telencephalon are innervated by the PBN gustatory fibers, specifically the bed nucleus of
the stria terminalis, diagonal band of Broca (a part of the septum), substantia innominata,
and globus pallidus (Saper & Loewy, 1980; Bernard, et al., 1993). The PBN also sends a
sparse projection directly up to the gustatory or insular cortex (layers V and VI
specifically), a thin strip of frontal cortex lying just above the rhinal sulcus and the
highest gustatory region of the brain (Saper & Loewy, 1980). However, the role of this
latter projection in gustatory perception is currently unproven (Lundy & Norgren, 2004).
The flow of CS information continues a bit further from some of these areas. The
thalamic gustatory area, the VPMpc, provides the major ascending gustatory input to the
dysgranular/agranular (gustatory) portions of the insular cortex, and while the VPMpc
also projects to the granular insular cortex, electrophysiological evidence indicates that
the granular insular cortex responds primarily to thermal and tactile stimulation of the
oral cavity (Lundy & Norgren, 2004). This projection completes the core gustatory
sensory pathway of the central nervous system: NST-PBN-VPMpc-IC (Figure 2.0). Yet
91
this pathway has important connections with the amygdala. As mentioned, the PBN
innervates the amygdala. The thalamic VPMpc also relays gustatory information to the
amygdala L (lateral), C, and M (medial) nuclei (Turner & Herkenham, 1991; Nakashima,
et al., 2000; Lundy & Norgren, 2004). The thalamic projections to the amygdala and
insular cortex have been shown to arise from distinct subpopulations of neurons in the
VPMpc, rather than one population that inputs to both telencephalic structures
(Nakashima, et al., 2000). The agranular insular cortex also projects (from more to less
dense) to the C, L, B, AB, Co, and M amygdala (Ottersen, 1982; Nakashima, et al.,
2000). Thus, as in the fear conditioning circuit, the amygdala receives both cortical
(agranular insular cortex) and subcortical sensory tracts (PBN, thalamus).
The US Pathway. The US pathway is a little muddier than the CS pathway,
because there are a plethora of agents that can be used to condition an aversion to or
avoidance of a taste, some of which are not obviously poisonous, such as sesame oil
(Deutsch, Molina & Puerto, 1976), intravenous saline (Revusky, Smith & Chalmers,
1971), and high physiological doses of estrogen (Hintiryan, Foster, and Chambers, 2009).
This suggests that multiple routes could exist for the detection of various substances or
body states. However, as mentioned above, it is widely believed that visceral malaise is
an important stimulus for activating the US pathway. Therefore the peripheral visceral
nociceptive fibers and the corresponding ascending pain pathways are the probable roots
of the US pathway.
92
Figure 2.0 The CS Pathway. Beginning with cranial nerves VII, IX,
and X, gustatory information flows upwards, except where indicated
by arrow points. See text for a complete description. The VPMpc is
bracketed because it is a sub-part of the MITC, but is listed separately
because it is the principal taste relay in the thalamus. [Abbreviations:
AMG, amygdala; IC, insular cortex; MITC, midline and intralaminar
thalamic complex; NST, nucleus of the solitary tract; PBN,
parabrachial nucleus; Thal, thalamus; VPMpc, parvocellular
ventroposteromedial nucleus.]
93
The US pathways detailed here follow a general pattern: peripheral, somatic
sensory nerves are funneled up to a critical visceral sensory place in the brainstem, the
PBN (parabrachial nucleus). The PBN projects mainly to an array of nuclei in the
thalamus, which in turn send fibers to the amygdala and the insular cortex. The insular
cortex also projects to the amygdala (Figure 2.1). This framework helps to simplify the
unholy mess of anatomical data presented here.
One important peripheral constituent of the visceral pathway is the X
th
cranial
nerve, or vagus nerve, the esophageal branches of which innervate the stomach,
duodenum, small intestine, pancreas, liver, and gallbladder (Powley, Holst, Boyd &
Kelly, 1994; Gabella, 2004). As described above, the vagus has been demonstrated to
respond to visceral stimulation by LiCl within 5-10 min of infusion (Niijima &
Yamamoto, 1994), a time course that corresponds to the onset of observable signs of
illness. Subdiaphragmatic severance of the vagus has been shown to substantially
weaken CTA induced by intragastric and i.p. infusions of the toxin CuSO
4
, suggesting
that at least some US information is mediated by this nerve (Coil, Rogers, Garcia &
Novin, 1978). Fibers from the vagus enter the rhombencephalon and target several areas
in the medulla. One medullary nucleus, the area postrema (AP), receives a dense
bilateral innervation by sensory fibers of the vagus in the rat (Contreras, Beckstead &
Norgren, 1982; Kalia & Sullivan, 1982). Sensory afferents of the vagus also innervate all
divisions of the caudal NST both fore and aft of the level of the obex, with the densest
projections terminating in the medial NST and commissural NST, more moderate
94
contributions to the ventrolateral portions of the NST, and the lightest projections in the
dorsal and lateral NST.
Another potential channel for US information to ascend to the brain is via the
visceral afferent fibers of the thoracic and lumbar splanchnic nerves, which form part of
the subdiaphragmatic abdominal plexus (Gabella, 2004). The fibers of these nerves
conduct electrical stimuli at latencies characteristic of nociceptive unmyelinated C-fibers
(Akeyson & Schramm, 1994; Niijima & Yamamoto, 1994), suggesting that they are in
fact sensitive to noxious stimulation. They respond to LiCl stimulation with even greater
vigor than the vagus (Niijima & Yamamoto, 1994), and could account for the residual
mild CTA expressed in vagotomized rats poisoned with i.p. CuSO
4
(Coil, et al., 1978).
These sensory neurons enter the spinal gray matter and terminate primarily in the deep
laminae of the dorsal horn, particularly layer X (Akeyson & Schramm, 1994; Willis,
Westlund & Carlton, 2004). The deep layers have ascending fibers through the dorsal
columns of the spinal cord directly up to various brain areas
3
: the parabrachial nucleus;
the thalamus; and a few threadbare fibers in the C amygdala, globus pallidus, basal
nucleus of Meynert, and substantia innominata (Wang, Willis & Westlund, 1999;
Bourgeais, Monconduit, Villanueva & Bernard, 2001). Neurons of these deep laminae of
the spinal gray also project to the medial gracile nucleus (lower abdominal nociception)
and the cuneate/gracile common border area (upper abdominal nociception), which in
3
In the PBN, the specific subnuclear targets are the internal lateral, superior lateral, dorsal lateral, central
lateral, external lateral, Kölliker-Fuse, and medial subnuclei. In the thalamus, weak fiber labeling is seen
throughout the posterior thalamic nuclear group, MITC (paraventricular, intermediodorsal, reuniens,
parafascicular, posterior intralaminar, central medial, and central lateral), peripeduncular, parvocellular
subparafascicular, mediodorsal, ventral lateral, ventral medial, and lateral habenular nuclei. Stronger
labeling was found in the zona incerta.
95
turn project via the medial lemniscus to the ventral posterolateral thalamic nucleus, zona
incerta, reuniens, ventral medial, and posterior thalamic group (Willis & Westlund, 1997;
Villanueva, Desbois, Le Bars & Bernard, 1998; Willis, et al., 2004).
A final starting point for the US pathway is direct sensory transduction of LiCl
and other bloodborne toxins by the AP (the area postrema; Ritter, McGlone & Kelley,
1980). The blood-brain barrier, which functions to prevent entry of most large molecules
and electrically charged or polar chemicals, is “leaky” at the level of the AP (Shapiro &
Miselis, 1985; Oldfield & McKinley, 2004), putting the AP in a position to sample the
blood for toxins whereas the rest of the brain is less capable of doing so and is thus
protected. How the direct detection of LiCl occurs in the AP is not known with certainty.
However, Li is chemically similar to sodium (Na), an element critical for the electrical
properties and activities of neuronal membranes. A significant point of entry for Li is
through voltage-gated Na channels in neuronal membranes. Once in the neuron, Li may
act to unbalance intracellular ion concentrations, particularly those of Na
+
of Ca
2+
, which
would ultimately impair neuronal excitability and neurotransmission (El-Mallakh, 1996).
Therefore, Li could be reducing the activity of neurons in the AP. Another possibility
could be that the AP contains chemoreceptors similar to those found on the lingual
surface and olfactory epithelium, receptors which could specifically detect the presence
of Li (Brizzee & Neal, 1954; Borison, 1974). The AP projects to several zones in the
rhombencephalon, including to parts of the lateral PBN. The core of the AP connects to
the dorsal lateral and central lateral PBN subnuclei, and to the inner sector of the external
lateral subnucleus as well, while the lateral AP projects to the outer sector of the external
96
lateral subnucleus (Shapiro & Miselis, 1985; Herbert, et al., 1990), parts of the PBN that
respond to nociceptive stimuli. (There are a number of other ascending pain pathways,
but to avoid further mind-numbing complexity, they will not be discussed.)
All together, then, the vagal, splanchnic, and postremal nociceptive pathways
either project directly to the PBN, or indirectly via the AP/NST. Both the AP and NST
project to PBN, mainly to terminal fields in the lateral nucleus. Parabrachial efferents
spread all over the brain; areas important for CTA include the thalamus, amygdala, and
insular cortex. The external lateral PBN sends a dense innervation to the ventral medial
thalamus, an important nociceptive relay, and it also contacts various nuclei of the MITC
(the paraventricular, parafascicular, reuniens, central medial, paracentral nuclei, and
VPMpc) and the zona incerta (Saper & Loewy, 1980; Krout & Loewy, 2000). The
central lateral and dorsal lateral PBN also contribute efferents to the MITC (Krout &
Loewy, 2000). Note that many of these midline thalamic nuclei receive gustatory
afferents as well, especially the VPMpc. The internal lateral PBN projects to the
paracentral and to almost all other MITC nuclei (Krout & Loewy, 2000; Gauriau &
Bernard, 2002). The paracentral continues this pathway to the frontal cortices,
importantly the rostral dorsal agranular insular cortex (Berendse & Groenewegen, 1991).
All of the MITC nuclei receiving PBN efferents, including the VPMpc, also project to the
rostral dorsal agranular insular cortex, which is a pain responsive area (Jasmin, Garanto
& Ohara, 2004).
97
Figure 2.1 The US Pathway. Lithium chloride (LiCl) and other toxins affect visceral nociceptors and
neurons within the area postrema. All information flow is upwards, except where indicated by arrow
points. See text for a complete description. Abbreviations: AMG, amygdala; AP, area postrema; IC,
insular cortex; lat., lateral; MITC, midline and intralaminar thalamic complex; n/nn., nucleus/nuclei;
NST, nucleus of the solitary tract; PBN, parabrachial nucleus; Thal, thalamus; VM, ventral medial;
VPL, ventral posterolateral nucleus; VPMpc, parvocellular ventroposteromedial nucleus.
98
The amygdala is innervated by the nociceptive areas of the PBN, thalamus, and
cortex. The lateral PBN has efferent connections with the amygdala and surrounding
structures. A heavy input is provided to the C nucleus of the amygdala, with terminals
found in the capsular and lateral subnuclei of the C (Saper & Loewy, 1980; Bernard, et
al., 1993). The US pathway also has a thalamic route to the amygdala, via the zona
incerta and MITC which relay to the B, L, AB, and C nuclei of the amygdala (Turner &
Herkenham, 1991; Sakai & Yamamoto, 1999). The rostral agranular insular cortex (of
the pain system) innervates the B and AB amygdala, and moderately invests the L
nucleus as well (Jasmin, et al., 2004). The C nucleus is noticeably vacant of this cortical
input. Still, the thalamic and cortical routes to the amygdala have been found to be
functionally mutually redundant (Sakai & Yamamoto, 1999), as in fear conditioning. In
conclusion, the US pathway begins with peripheral sensory structures that feed into the
PBN, the PBN projects to the thalamic MITC and zona incerta, the thalamus continues
the pathway to the amygdala and insular cortex, and the insular cortex projects down to
the amygdala (Figure 2.1).
CS-US Association. Presumably, any brain area that makes an association
between the CS and the US would require a convergence of the CS and US pathways in
that area. In the case of CTA, many of the central nervous system loci subserving the US
pathway are the same as those comprising the CS pathway. Potentially any of the areas
receiving both CS and US information can associate the two. The usual suspects are the
parabrachial nucleus, amygdala, and insular cortex, but most of the brain areas the two
pathways have in common contain at least one subnucleus where the two pathways are in
99
close proximity (Figure 2.2). At the lowest level of the brain, the two pathways synapse
in the NST; while the CS pathway tends to terminate more rostrally in the NST, and the
US pathway terminates more caudally, they both seem to innervate the ventrolateral
subnucleus (Kalia & Sullivan, 1982; Hamilton & Norgren, 1984). The CS and US
pathways next course into the PBN, and while they again tend to assort into mostly
separate termini, they both occupy parts of the medial subnucleus of the medial nucleus
(Wang, et al., 1999; Herbert, et al., 1990). These pathways continue on to the thalamus
wherein they jointly innervate a number of areas, including the VPMpc, zona incerta,
parafascicular nucleus, central medial nucleus, and paracentral nucleus (Saper & Loewy,
1980; Bester, et al., 1999; Krout & Loewy, 2000; Lundy & Norgren, 2004). Ascending
to the highest level of the nervous system, the pathways both synapse in the agranular
insular cortex (Berendse & Groenewegen, 1991; Jasmin, et al., 2004; Lundy & Norgren,
2004). Finally, the amygdala is a recipient of CS and US pathways; in fact, parallel
branches from both pathways arise from the PBN, thalamus, and insular cortex, and
converge in the amygdala (Saper & Loewy, 1980; Ottersen, 1982; Turner & Herkenham,
1991; Bernard, et al., 1993; Sakai & Yamamoto, 1999; Nakashima, et al., 2000; Jasmin,
et al., 2004; Lundy & Norgren, 2004). Within the amygdala, the CS and US pathways
both terminate in the L, B, AB, and C nuclei (Figure 2.3). In the C nucleus, the CS
pathway targets the medial subnucleus, while the US pathway targets the capsular and
lateral subnuclei (Bernard, et al., 1993).
However, these separate subnuclei have their own complex connectivity pattern,
which has been described elsewhere (Sah, et al., 2003) and is depicted graphically in
100
Figure 2.3; thus even though the CS and US pathways innervate separate parts of the C
nucleus, the internal connectivity could allow CS-US associations to occur in the medial,
lateral, or capsular subnuclei. The amygdala, then, has the potential to make separate CS-
US associations in the L, B, AB, and C nuclei. Indeed, the potential exists for
associations to occur throughout almost the entire extent of the CS and US sensory
pathways. This fact is tempered by the recognition that even though the pathways may
comingle in the same part of the brain, they are not necessarily converging onto the same
postsynaptic target. This is significant because the CS and US pathways must synapse
onto the same population of neurons in order for those postsynaptic neurons to make an
association. Thus it is possible that very few of the subnuclei listed in Figure 2.2 are
actually capable of making an association.
The CR Pathway. The CR pathway is a little explored pathway, owing at least
in part to lack of consensus as to where the CTA engram is stored. Still, speculation of a
probable CR pathway (presuming that the amygdala contains a CTA engram) leads one
to deduce that logically such a path includes brain areas that control the behaviors seen in
CTA expression: cessation of ingestion, aversive orofacial and somatic movements,
avoidance of the CS, and lying-on-belly. At its very heart, a CR pathway would need to
modulate the brain areas controlling licking (lateral reticular formation), mastication (the
paragigantocellularis and gigantocellularis nuclei), and swallowing (solitary nucleus) in
order to halt ingestion of a toxic substance. The muscles responsible for effecting these
behaviors are controlled directly by the motor neurons of the trigeminal (V
th
), facial
(VII
th
), and hypoglossal (XII
th
) cranial nerves, i.e., the supratrigeminal nucleus, motor
101
nucleus of the V
th
, VII
th
nucleus, XII
th
nucleus, NST, and nucleus ambiguus (Herbert, et
al., 1990; Lund, 1991).
Figure 2.2 Possible sites of CS-US association. This simplified joint
diagram of the CS and US pathways lists points of overlap between the two
pathways. Abbreviations: AMG, amygdala; IC, insular cortex; NST, nucleus
of the solitary tract; PBN, parabrachial nucleus; Thal, thalamus; VPMpc,
parvocellular ventroposteromedial nucleus.
102
Figure 2.3 Loci of CS-US pathway convergence within
the amygdala. The intranuclear connectivity of the C
nucleus could complete the pathway convergence in this
nucleus. Abbreviations: AB, accessory basal; B, basal; C,
central; cap, capsular; int, intermediate; L, lateral; lat,
lateral subdivision; med, medial.
These motor nuclei are in turn controlled in a coordinated fashion by a central
pattern generator, a nucleus which orchestrates the proper timing and sequence of muscle
movements. In the case of licking and mastication, the pattern generator likely resides in
the mid-medullary reticular formation, constituted of parts of the paragigantocellularis
and gigantocellularis nuclei (Lund, 1991) and the parvocellular part of the lateral reticular
formation (Chen, Travers & Travers, 2001). The C amygdala has direct projections to
these reticular nuclei (Pitkänen, 2000). The central pattern generator for swallowing is
located in the subpostremal zone of the NST (Jean, 2001), and the C amygdala has
103
projections to this area as well. Thus the amygdala is well positioned to inhibit ingestion.
In fact, stimulation of the C amygdala can elicit fictive mastication in rabbits (Applegate,
Kapp, Underwood & McNall, 1983), so its influence over oromotor movements has been
verified.
The oromotor CR can form with a latency of 10-15 min following acquisition, as
evidenced by decreases in ingestive orofacial responses and increases in aversive
orofacial responses in this time frame (Spector, Breslin & Grill, 1988), although in the
laboratory, expression of the CR typically is not tested until at least 24 h after acquisition,
at a time when the effects of LiCl are no longer present. After recovery from the
unconditioned effects of LiCl has taken place, a sucrose CS still elicits aversive orofacial
responses but paradoxically also may elicit (depending upon the acquisition procedure) as
many ingestive orofacial responses as a non-conditioned animal (ibid.), perhaps
indicating that the highly palatable sucrose does not lose its positive hedonic properties,
but rather has negative hedonic features layered upon its neural representation. Reflexive
responses can include vomiting (in species capable of this response, like coyotes;
Gustavson, et al., 1974) and gape (a gag reflex or “dry heave” in species not capable of
vomiting, such as rats; Grill & Norgren, 1978). Gape, like licking, may be controlled by
the lateral reticular formation: muscimol infusion here suppressed both responses,
without impairing appetitive approach responses (Chen, et al., 2001). As stated above,
the C nucleus projects to this area, suggesting an amygdalar influence.
The CR pathway should also activate the expression of lying-on-belly. This
behavior may be organized by the periaqueductal grey, the ventrolateral column of which
104
is involved in “passive emotional coping strategies” such as immobility (“freezing”, see
Section 1.1.1) and listlessness (Gauriau & Bernard, 2002). This response sounds
comparable to lying-on-belly, in which the rat assumes a prone position and does not
move. Lying-on-belly is a natural response to LiCl illness, and seems to occur as a
conditioned reaction to the CS after a latency of about 10 min (Meachum and Bernstein,
1990). The C amygdala has dense connections with the ventrolateral periaqueductal grey
(Rizvi, Ennis, Behbehani & Shipley, 1991), suggesting that this may be a possible
pathway for the conditioned expression of lying-on-belly.
Finally, a CR pathway should be able to induce avoidance of the CS. When an
animal re-encounters a food source that previously poisoned it, it can best avoid ingesting
the tainted food by keeping its distance. Avoidance responses may be mediated by the
nucleus accumbens. The nucleus accumbens has traditionally been associated with
reward. For example, virtually all drugs of abuse have some influence over activity in
the nucleus accumbens, and this nucleus may motivate approach responses to reward-
and drug-related stimuli
4
(Ikemoto & Panksepp, 1999). However, of late the nucleus
accumbens has also been recognized through several lines of evidence as possibly
motivating avoidance responses as well. In particular, permanent and temporary lesions
of the nucleus accumbens have been shown to disrupt conditioned avoidance (ibid.).
Moreover, studies have shown that the nucleus accumbens participates in CTA (Ramirez-
Lugo, Núñez-Jaramillo & Bermudez-Rattoni, 2007). Specifically, c-Fos expression in
4
Specific functional lesion of the nucleus accumbens (dopamine receptor blockade) in rats can dramatically
increase the latency to approach a sipper bottle of highly palatable sucrose solution, without diminishing
the amount of sucrose consumed upon arrival (Ikemoto & Panksepp, 1999). Thus approach behaviors are
impaired but consummatory behaviors are not.
105
the nucleus accumbens has been correlated with CTA expression (Yasoshima, Scott &
Yamamoto, 2006), as has acetylcholine release (Mark, Weinberg, Rada & Hoebel, 1995),
and D
1
dopamine receptor blockade in the nucleus accumbens (shell division) blocks
acquisition of a CTA (Fenu, Bassareo & Di Chiara, 2001). The BLA (all constituent
nuclei) provides the major amygdalar input to the nucleus accumbens, and this projection
has been reported to be a glutamatergic tract (McDonald, 1991). Thus, amygdalar
connections with the nucleus accumbens could mediate the avoidance CR in CTA.
Overall, there are two gross output systems proposed here, systems that would
work together to prevent ingestion of a dangerous food. When an animal encounters a
food source (the CS in this case), it samples the food, and can recognize the food as a CS
that once made it ill—one lick, spanning roughly 60 ms, is enough for this recognition to
occur (Bures, 1998). The C nucleus pathway would trigger reflexive responses to halt
consummatory responses like licking and swallowing, and to cause the animal to spit out
the food. Then the BLA pathway would motivate the animal to avoid the CS entirely.
All together, the amygdala contacts appropriate neural areas for eliciting some of the
primary behaviors seen in the CR.
2.2 Lesion studies demonstrating participation of the amygdala in CTA
The first studies demonstrating a role for the amygdala in CTA created
electrolytic lesions, in which an electrode was used to destroy all neurons, glia, and axons
of passage in a discrete brain region. Nachman and Ashe (1974) made electrolytic
lesions of the basolateral amygdaloid complex (BLA) both before and after CTA training,
106
manipulations which impaired CTA expression in both cases, particularly in the latter
case. A number of subsequent studies have replicated these findings (Aggleton, Petrides
& Iverson, 1981; Borsini & Rolls, 1984; Simbayi, Boakes, & Burton, 1986; Dunn &
Everitt, 1988; Kesner, Berman, & Tardif, 1992; Yamamoto, Fujimoto, Shimura & Sakai,
1995; Morris, Frey, Kasambira & Petrides, 1999; Sakai & Yamamoto, 1999; Reilly & St.
Andre, 2007). The results of one of these studies (Dunn & Everitt, 1988) sparked a
controversy over the involvement of the amygdala in CTA lasting over a decade. This
group made large electrolytic lesions of the amygdala and found deficits in CTA in
accord with the previous studies. In a second part of the experiment, however, they made
lesions using the excitotoxin ibotenic acid, which is a chemical similar to the
neurotransmitter glutamate and which kills neurons but leaves axons of passage
unaffected. Groups receiving these excitotoxic lesions did not evince CTA deficits,
leading the authors to conclude that the amygdala per se is not involved in CTA, but that
axons comprising a part of the CTA neural circuit merely pass through the amygdala.
Thus kill-all electrolytic lesions of the amygdala would disrupt this circuit, but neuron-
selective, axon-sparing excitotoxic lesions would not. Subsequent studies likewise
utilizing excitotoxic amygdala lesions seemed to confirm this conception (Chambers,
1990; Bermudez-Rattoni & McGaugh, 1991; Hatfield, Graham & Gallagher, 1992).
107
Figure 2.4 The CR pathway. The central (C) nucleus of the amygdala projects to areas that could
control lying-on-belly, swallowing, licking, mastication, and gape. The nuclei of the BLA (B, L,
and AB) project to the nucleus accumbens (NAc) which could mediate avoidance behavior. Other
abbreviations: PAG, periaqueductal grey; NST, nucleus of the solitary tract; RF, reticular formation.
Then a 1999 study by Morris and colleagues presented evidence that challenged
this notion, reinstating a role for the amygdala in CTA. This group contended that the
aforementioned excitotoxin studies created lesions that only encompassed part of the
BLA. Morris et al. made ibotenic acid lesions that were histologically verified to have
eliminated at least 90% of the BLA. These animals exhibited CTA deficits, while a
second group receiving lesions of the adjacent C nucleus expressed normal CTAs. These
108
results are even more compelling in light of the fact that Morris et al. demonstrated
through a tract-tracing technique that axons running from the PBN to the IC via the
amygdala were intact in their animals, effectively ruling out a fibers-of-passage
explanation for their results. They concluded that the amygdala, particularly the BLA, is
in fact involved in CTA, a conclusion supported by similar lesion studies (Yamamoto, et
al., 1995; Reilly & St. Andre, 2007).
Other studies attempted to ablate subsections of the BLA in order to further define
the critical circuitry. Fitzgerald and Burton (1981) placed electrolytic lesions in either the
anterior or posterior BLA. Neither lesion group exhibited significantly different behavior
from sham-lesioned controls. Later, Fitzgerald and Burton (1983) advanced their
exploration of the role of the BLA in taste aversion, this time either by electrolytic
lesioning of the BLA entire, or by making knife-cut lesions in an attempt to sever
amygdalar connections with the overlying temporal cortex. Both lesion types impaired
CTA acquisition. Examination of the histology figures presented in both studies reveals
that both the electrolytic and knife-cut lesions made in the later study appeared to include
damage to the lateral nucleus (L), while lesions in the earlier study did not. The histology
photomicrograph presented in Nachman and Ashe (1974) likewise appeared to depict
lesions encroaching upon the L nucleus. Other studies finding basolateral lesion-induced
CTA deficits also created lesions that included damage to the L, at least in part
(Aggleton, et al., 1981; Schafe, Thiele & Bernstein, 1998; Schafe & Bernstein, 1996).
More convincing is a study by Lasiter and Glanzman (1985), in which they
electrolytically lesioned individual nuclei of the amygdala. They observed taste aversion
109
deficits in rats with lesions of the L nucleus. This suggests that the L nucleus may be an
important part of the amygdalar contribution to CTA.
The C nucleus is another candidate for harboring an important part of the CTA
neural circuit. As described above in the Brain Areas section (Section 2.1), the C, like
the BLA, has connections with other brain areas involved in CTA. Furthermore, as
described below (Section 2.3), activity in the C correlates with various phases of CTA,
and gene transcription is triggered in the C in response to LiCl administered peripherally.
Thus it would seem likely that the C plays a role in CTA. Yet numerous lesion studies
have failed to detect deficits in CTA expression in animals with C lesions (Kemble,
Studelska & Schmidt, 1979; Schoenfeld & Hamilton, 1981; Galaverna, et al., 1993;
Yamamoto, et al., 1994; Yamamoto, et al. 1995; Touzani, Taghzouti & Velley, 1997;
Morris, et al., 1999; Sakai & Yamamoto, 1999; St. Andre & Reilly, 2007). This may boil
down to a detection issue, stemming from the possibility that the C and BLA may play
different, complementary roles in CTA. In fear conditioning, the C nucleus mediates the
classically conditioned components and the B nucleus modulates the instrumental
components (see Figure 1.4). Thus, while lesions of the BLA obviously affect the
amplitude of CTA, which are commonly measured with a bottle test, the C may mediate
the Pavlovian reflexes in CTA (i.e., gape/emetic reflexes and perhaps the response lying-
on-belly). Thus the apparent lack of effect of C lesions upon CTA may be due to failure
to measure the correct behaviors or variables, specifically responses tied to an illness
reaction. Lesions of the C do in fact alter orofacial responses to tastants, increasing
rejection responses, and cause ephemeral hypodipsia (Kemble, Levine, Gregoire, Koepp
110
& Thomas, 1972; Touzani, et al., 1997). Some correlative data also support a
dissociation in roles: Wilkins and Bernstein (2006) found that the method of conditioning
a taste aversion affects the pattern of c-fos expression (a gene whose expression
correlates with recent neural activity) in the amygdala. Intra-oral conditioning induces c-
fos expression in the C nucleus, while bottle conditioning induces c-fos in the C and
BLA. Thus the more Pavlovian conditioning method (intra-oral) engages the C nucleus
only, while the instrumental conditioning method (bottle method, which would still have
Pavlovian components to it) engages the C and BLA amygdala. The first general aim of
this proposal entails assessment of the contributions of individual nuclei of the amygdala
to the Pavlovian and instrumental components of CTA.
Other amygdaloid nuclei do not seem to be involved in CTA. The medial (M)
nucleus has been incidentally damaged and specifically targeted with electrolytic
(Aggleton, et al., 1981; Schoenfeld & Hamilton, 1981; Rollins, Stines, McGuire & King,
2001) and excitotoxic lesions (Yamamoto et al., 1995) and deemed uninvolved in CTA
(Reilly & Bornovalova, 2005). The cortical (Co) nucleus is adjacent to the M, and in
some of the above studies incurred damage along with M; additionally, many studies
finding CTA deficits created amygdala lesions that avoided the Co nucleus because of its
extreme position relative to the other major nuclei. Thus, the Co and M nuclei are not
likely involved in CTA.
111
2.3 The role of the amygdala in CTA
While lesion studies are capable of indicating whether or not the amygdala
participates in CTA, they cannot, however, due to their irreversible nature, elucidate the
role of the amygdala in any particular phase of CTA acquisition or expression (see Table
2.0). When a permanent lesion is found to disrupt conditioning or expression, there are a
range of possible reasons why this is so. Findings from the study of long-term
potentiation (LTP) informed the postulation of these roles, as did notions of neural
substrates of classical conditioning. Below are descriptions of some of the possible
explanations for the presence of a deficit in expression of a CTA after lesioning.
A) CS Pathway. Disruption of this pathway produces a sensory deficit. The lesion may
have destroyed neurons critical for taste processing. The animal may have pure ageusia,
and taste nothing (Reilly & Bornovalova, 2005). On the other hand, it may only be
unable to relay gustatory perceptions to the CS-US association area, leaving other taste
perceptual processes intact.
B) US Pathway. Disruption of this pathway also produces a sensory deficit. The lesion
may have destroyed neurons critical for processing the unconditioned stimulus, resulting
in an inability to sense visceral pain, or an inability to send US information to the CS-US
association area.
112
C) CR Pathway. The lesion may have disrupted the motor pathway responsible for
controlling the conditioned response. Thus the animal may realize that the CS it is
consuming once made it sick, but it is unable to avoid or desist consumption. If the
lesion affected only part of this pathway, then some CR components may be expressed
but not others (e.g., avoidance responses but not gapes or lying-on-belly). Multiple
sensitive measures will need to be employed to assess for this possibility.
D) CS Novelty. This is a special case of CS sensory disruption. If perception of the
novelty of a tastant is disrupted, the animal will have difficulty acquiring a CTA to it,
because familiar substances are resistant to aversion conditioning (Reilly and
Bornovalova, 2005).
E) Modulation. This would likely represent a special component of the US pathway.
The lesioned area may modulate the formation of memory elsewhere, that is, while not
required for the actual association of CS and US, this area may be required for
consolidation (i.e., transference to long-term memory) of that associative memory. In
this proposal, such transference is not a physical relocation, but rather a physiological
crystallization within the engram area. The amygdala is thought by some to play a
modulatory role in passive avoidance learning (see Section 1.2.1).
F) Engram. This brain area may actually store the associative memory. For the purposes
of this proposal, our model assumes that the engram involves a convergence of CS and
113
US pathways, and its efferents comprise the CR pathway. During acquisition this area
would make the CS-US association, changing in a way reflective of learning (i.e., LTP),
and the memory would be stored here and retrieved here upon expression.
G) Diaschisis. This effect would manifest as a temporary suppression of CTA
acquisition or expression, due to the fact that traumatic brain injury causes shock or stuns
the rest of the brain. Such an effect could be due to destruction of some of the blood
vessels feeding a nearby area critical to CTA; annihilation of neurons that provide tonic
input to a CTA-relevant area; or the release of chemicals from the dead and dying cells
which interfere with functioning of adjacent normal cells (like proteases). As the shock
subsides, the apparently lost function will reemerge. Diaschisis may account for some,
but not all, of the deficit seen after amygdala lesion, since Nachman and Ashe (1974)
found some recovery of function after 10 days, but no further convalescence after 25, 50,
or 100 days of post-operative recovery.
Table 2.0 Periods of CTA when a given role must
participate in conditioning
Role in CTA CS US Post US* EXP
CS pathway
X X
US pathway X
CR pathway X
Novelty X
Modulation X X
Engram X X X X
*For a period of hours after the US injection or illness
Abbreviation: EXP, expression phase
114
This laundry list of possible explanations for deficits in CTA after discrete neural
lesions illustrates the difficulty of assessing the role of the amygdala in CTA. Further
complicating matters is the fact that several of these deficits could be true all at once, and
some roles even necessitate that others be true simultaneously, e.g., if it is true that the
amygdala holds the engram (H), then it would be sensitive to disruption during CS
presentation (A), US presentation (B), and expression of the CTA (C). Additionally,
several of the nuclei of the amygdala could participate in CTA, with disparate roles for
each nucleus (see Section 2.2, above; Yamamoto, 1998, 2007). In any case, each of the
deficit types discussed as part of the laundry list above, except diaschisis, will be treated
in more detail below. The above description of diaschisis is sufficient and will not be
expanded below.
But before moving on, it should be noted that this list is not exhaustive. Notably
absent from the above list is the UR (unconditioned response) pathway. This is because
if such a path exists, disruption of it would not necessarily affect any of the other things
above and thus would not affect conditioning. For instance, in classical eyeblink
conditioning, the unconditioned blink pathway (mediating the reflexive blink due to an
air puff to the eye) is by and large distinct from the conditioned blink pathway, such that
lesion of one does not affect the other. On the other hand, the UR pathway may be
coextensive with the CR pathway. For example, in classical fear conditioning, C
amygdala lesions can impair both the conditioned and unconditioned aspects of some (but
not all) behaviors, such as freezing. If this were the case for CTA, a CR motor deficit
would be a sufficient explanation for the impairment in CTA expression, since the CR
115
and UR pathways would in fact be just one pathway. Additional explanations receiving
short shrift include US novelty, the ability to bridge the temporary interval between CS
offset and US onset, attention or the ability to bring a stimulus into conscious awareness,
and an acquisition-association area.
This last explanation, acquisition-association, is essentially a variant form or
competitor of the modulation/engram explanations, and is alternately known as SYSTEM
CONSOLIDATION. The function of an acquisition-association area would be to form an
initial associative memory during CTA acquisition; however, after a period of time, that
associative memory is transferred to another part of the brain for permanent storage, and
the acquisition-association area becomes completely dispensable, insofar as expression of
that CTA memory is concerned. However, system consolidation is chiefly associated
with declarative memory (Dudai, 2004), and has not been successfully demonstrated to
occur in the establishment of non-declarative memory (a category that includes CTA).
Moreover, the hippocampal complex has been found to be critical for system
consolidation (ibid.), yet this brain structure is not involved in CTA (Bermudez-Rattoni
& Yamamoto, 1998). Still, it is possible that non-declarative memories utilize non-
hippocampal substrates for system consolidation, and for the amygdala and CTA this
possibility would be easy enough to test: condition a taste aversion in some young rats,
then let enough time pass for system consolidation to reasonably occur (e.g., 1.5 years:
Fanselow & Gale, 2003; Gale, et al., 2004). Then lesion the amygdala in half of these
rats, while retaining the other half as controls. If the amygdala is in fact an acquisition-
association area, then both groups should demonstrate comparable CTAs, because the
116
CTA memory would have been transferred out of the amygdala to another part of the
brain. Such an experiment has not yet been conducted. Acquisition-association and the
other orphaned explanations have received little to no attention in the CTA field of study,
and so cannot be greatly expanded here.
2.3.1 What the evidence shows
A number of techniques other than ablation have been employed in the
exploration of the relationship between the amygdala and CTA: correlative techniques
such as single-unit recording, microdialysis, and c-Fos immunostaining; stimulation
techniques such as direct electrical stimulation and glutamate infusion; and functional
lesion techniques such as novocaine, tetrodotoxin, or muscimol infusion, transient gene
suppression, and protein synthesis inhibition. Coordinated use of these techniques can
tease apart the contribution of the amygdala to CTA learning and memory.
One of the most informative of these methods would be to use a drug that
suppresses neural activity to inhibit the amygdala during the discrete phases of CTA
acquisition and expression. Since CTA can tolerate long intervals between CS and US
presentation, it is possible to inactivate specifically during the CS or US phases. Exactly
this sort of study was attempted by Gallo, Roldan and Bures (1992). To inactivate the
amygdala they infused tetrodotoxin (TTX), a drug that inhibits voltage-gated sodium
channels and hence suppresses action potentials in its area of effect. Targeting the C
nucleus bilaterally, they injected 3 ng of TTX either before saccharin presentation,
immediately after saccharin drinking (30 min before lithium injection), or before the
expression test. The meaning of the results of these procedures will be discussed by
117
applying the possible explanations described above (Section 2.3, The role of the
amygdala in CTA).
When infused before the initial saccharin presentation, the resultant lesion caused
some rats to refuse to drink despite being severely water-deprived, and so were granted
extra time to consume the saccharin. Subsequent expression testing revealed no apparent
deficit in conditioning in this group. There are two important points that should be taken
into consideration when explaining these results. First, TTX induces a three hour period
of inactivation, meaning that both the CS and US phases of acquisition were nearly
completely disrupted. That CTA was still acquired apparently normally and expressed
through an avoidance response fairly eliminates any role for the C nucleus in CS or US
sensory pathways (although acquisition of Pavlovian CRs, which were not assessed,
could have been disrupted). Second, it is likely that the CS was identified as novel during
conditioning since the animals did not show a weakened CTA. Taken together, this
suggests that the C nucleus does not process CS or US properties essential for acquisition
of a CTA, including perception of CS novelty. This also renders the engram role for the
C nucleus unlikely as well.
Rats in the second group were functionally lesioned during the interval between
drinking and poisoning, and they exhibited attenuated CTA. The timing of lesion
induction allowed complete blockade of the illness US and an hour of post-acquisition
processes. Lithium chloride illness, which is measured by its behavioral signs and which
corresponds to amygdala neural activity (Yasoshima, et al., 1995), lasts approximately 2
hours (Lamprecht & Dudai, 2000), while the duration of action of TTX is approximately
118
3 hours (Zhuravin & Bures, 1991). Thus, it is likely that the action of TTX completely
occluded the Li-induced illness and extended into the post-acquisition period. That the
animals expressed a CTA however mild suggests that the animals recognized that they
had become ill after consuming the CS. That they expressed a weakened CTA could be
explained as follows. It may be that the C nucleus was not inactivated for a long enough
period of time. If this were the case, then this could have effectively reduced the strength
of the US and as such resulted in a weaker CTA. Although the C nucleus was blocked
when the US was being processed during acquisition, explanations involving the US
pathway are difficult to entertain since the first group was lesioned during both CS and
US phases and still developed a CTA. Alternatively, it could be that the amount of time
needed for modulation of the long-term CTA memory was reduced. Modulation of the
CTA memory would probably begin shortly after formation of the short-term CTA
memory (10-15 min after US onset) and would continue for at least several hours (see
section Modulation, below). Three hours of inactivation provided by TTX may not have
been enough to block modulatory processes completely, but the partial blockade could
have been enough to weaken long-term memory formation.
The third group of rats received normal, unimpeded acquisition, and then received
TTX before the expression test. Like the group infused before drinking, these rats also
exhibited oligodipsia, and thus were allowed additional time to drink. The authors found
that this group displayed an attenuated CTA compared to control rats, who underwent
unadulterated acquisition and expression tests. Because the C nucleus was blocked after
acquisition of the CTA, then only retrieval of the memory and expression of the CTA
119
could have been affected. That animals expressed a CTA suggests that they are capable
of retrieving the memory and of expressing an avoidance response. That the CTA was
weaker could indicate that the TTX may have affected part of the CR pathway, perhaps
impairing the Pavlovian rejection responses. Or perhaps the lesion reduced their
motivation to avoid the CS. However, the oligodipsia exhibited by this group and the
first group (prior to aversive conditioning) seems to suggest that the lesions enhanced
hedonic motivation or rejection responses, since both groups needed extra time to
consume the CS while lesioned. In fact, some animals in the first group received an extra
30 min to drink the CS on acquisition day, while some animals in the third group
received an extra 85 min (on average) to consume the CS during the expression test; it
appears that the oligodipsia induced by the lesion was able to combine with the taste
aversion in the third group to enhance avoidance of the CS—if these rats were given the
standard half hour test period, they may have even drank less than the control rats. In any
case, the CR pathway or hedonic motivation roles are plausible explanations for the
changes in behavior seen in these animals.
Similar studies by the same group extended this line of work by infusing TTX
into the amygdala at various time points after lithium injection (Roldan & Bures, 1994).
Inactivation induced immediately or 1.5 h after lithium administration impaired
conditioning, but at 6 or 24 h after lithium, TTX seemed to produce no interference.
These findings implicate a role for the C amygdala in modulation of the CTA memory
because the CS and US are administered to an unaltered (unlesioned) animal, particularly
in the 1.5 h group, allowing CS-US association to take place unimpeded, so only post-
120
association processes would be affected, i.e., consolidation. Yet the authors asserted an
alternative explanation: that the functional lesion was impairing CTA strength by
reducing the amount of time the CS and US could associate, presumably through US
disruption. While this is plausible, it seems somewhat incompatible with their data,
because if true, the CTA formed in the 1.5 h delay group should have been fairly strong,
since the malaise-inducing effects of lithium would have already peaked; yet the CTA in
this group was not statistically different from the group infused immediately after lithium
injection. Furthermore, the lesser effects observed in the 6 and 24 h groups cannot be
construed as zero effect, because while they developed significantly stronger aversions
than the 0 and 1.5 h groups, there was no positive control group for statistical
comparison. Thus these latter two groups may have in fact been slightly impaired. In
sum, the findings of the Bures laboratory indicate that the C nucleus may be involved in
the CR, motivation, and modulation pathways. However, as will be covered next, a
multitude of experiments have implicated a role for the amygdala in these and other roles.
CS Pathway. As mentioned in the abbreviated list above, damage to the
conditioned stimulus (CS) pathway would lead to a deficit in the processing of gustatory
information. If damage to this pathway occurred before acquisition or at the time of CS
exposure during acquisition, this would prevent an association between a tastant and
malaise, since the perception of the tastant would be missing. A lesion made after
conditioning would likewise disrupt CTA, given that activation of the CS pathway would
be necessary for retrieval of a previously formed engram.
121
Although not apparent in the TTX studies just described, several experiments
have identified a role for the amygdala in initial CS processing. Single-unit recordings of
the innate (unconditioned) responses of the amygdala to taste stimuli have generally
yielded similar findings: of large numbers of units sampled, few (typically less than 10%)
exhibit salient responses to gustatory stimulation (Azuma, Yamamoto & Kawamura,
1984; Yamamoto, Azuma & Kawamura, 1984; Yasoshima, et al., 1995; Nishijo, Uwano,
Tamura & Ono, 1998). The largest numbers of responses (but not a majority) were
usually detected in the C nucleus, with the next most responses measured in the BLA,
followed by various other amygdaloid nuclei. The responsive units could be categorized
as having best responses to one of four taste modalities tested, but on the whole they were
quite broadly tuned and responded to multiple tastants. In one study, only 7% of units
responded to a variety of tastants, including a saccharin solution, but after conditioning
the saccharin solution with LiCl in these rats, 23% of units responded to this taste
solution (Yasoshima, et al., 1995). The unit responses were characterized as either
increased (exhibiting an increase in action potentials) or decreased (exhibiting a decrease
in activity) in response to saccharin, and there was a tendency for these two kinds of units
to assort into different loci. Units with increased responses were predominantly found in
the BLA while units with decreased responses were typically located in the C nucleus.
Analysis of the responses of these units showed further distinctions. BLA units had
enhanced responses only to the CS (saccharin), or tastants of a similar kind (e.g.,
sucrose), while C units responded differentially to the CS and other naturally aversive
tastants (e.g., HCl and quinine). Yamamoto interpreted this to mean that BLA neurons
122
responded to taste quality, while C neurons responded to taste hedonic properties
(Yasoshima, et al., 1995; Yamamoto, 1998).
Upregulation and translation of the immediate early gene c-fos can indicate recent
neural activity. Several laboratories have examined the expression of c-Fos (or Fos-like)
proteins after ingestion of various CS solutions. c-Fos protein has been detected in the
amygdala after free ingestion of sucrose and saccharin (Yamamoto, Sako, Sakai &
Iwafune, 1997; Koh, Wilkins & Bernstein, 2003; Bernstein & Koh, 2007). These studies
detected a great increase in c-Fos in the C nucleus (particularly its medial subnucleus)
and a minor increase in the B nucleus. Thus both electrical activity and c-Fos protein
expression in the amygdala correlate with the sensation of sapid stimuli.
Evidence of a non-correlational nature also points to a possible contribution of the
amygdala to the CS pathway. Several authors have utilized electrode stimulation of the
amygdala and assessed its effects upon CTA. Stimulation of the B nucleus throughout a
10-min period of exposure to a saccharin CS during conditioning led to a greatly
attenuated CTA when tested two days later with saccharin alone (Phillips & LePiane,
1980). However, when electrical stimulation of the B occurred both during the
acquisition phase encounter with the saccharin CS and during saccharin drinking in the
expression test, the CTA was expressed normally. These results suggest that the
electrical stimulation combined with the gustatory CS to act as compound CS. This
electric CS effect even occurred when tap water was used as the CS, i.e., tap water and
electrical stimulation were co-administered during both the acquisition and expression
phases (Phillips & LePiane, 1978). Although this further indicates that the amygdala is
123
involved in the CS pathway for CTA, state-dependent learning could possibly explain
this intriguing effect. State-dependent learning occurs when an animal is trained while in
an altered cognitive state (for instance, a drugged state). When this animal is tested in a
different cognitive state (in this example, non-drugged), it performs poorer than if tested
in the training state, the drugged state. Similar to a drug, electrical brain stimulation
could be expected to alter an animal’s cognitive state. However, the electric CS effect
discovered by Phillips and LePiane is not likely due to state-dependent learning, because
stimulation of the caudate-putamen nucleus while using a water CS yielded no CTA,
indicating that merely reestablishing the electrical state of the brain that was present
during acquisition was insufficient to allow expression of a CTA to water. Rather,
stimulation of the B nucleus in particular was necessary to condition a taste aversion to
water, a CS that is normally resistant to toxiphobic conditioning due to its unimpeachable
familiarity. The water and the stimulation may have formed a combined stimulus
representation during conditioning, and the water/stimulation compound stimulus was
required to fully retrieve the aversive memory during the expression test.
In other tests of amygdalar involvement in the CS phase of conditioning,
stimulation shortly after CS presentation appeared to have no effect upon expression
(electrode placements were all throughout the amygdala; Kesner, Berman, Burton &
Hankins, 1975). Nor did stimulation of the B nucleus during only the expression test in
rats with normally acquired CTAs have any effect deleterious upon expression of a CTA
(Phillips & LePiane, 1980). Despite these results, the bulk of the correlational and
stimulation data presented here implicates the amygdala in the CS pathway of CTA.
124
Amygdaloid rats are not ageusic. Remember that the core gustatory pathway
(NST-PBN-VPMpc-IC) does not include the amygdala (see Section 2.1), so basic taste
perceptual processes should be unaffected. Aggleton and colleagues (1981) found that
BLA-lesioned rats who developed a weakened CTA to sucrose, responded normally to
.01% and .001% quinine solutions, rejecting both concentrations of the bitter tonic as
strongly as control rats did. Furthermore, rats receiving electrical stimulation of the B
nucleus of the amygdala concurrent with presentation of a novel sucrose solution
demonstrated normal consumption of the solution compared to nonstimulated implanted
controls, and these rats showed attenuation of neophobia with or without stimulation on a
second sucrose exposure (Phillips & LePiane, 1980). These suggest that amygdala
lesions do not eliminate taste perception, since lesioned rats respond appropriately to
various tastants over a range of concentrations. Taken together, the findings laid out in
this section demonstrate that the amygdala (particularly the C and B nuclei) does respond
to gustatory stimulation both before and after toxiphobic conditioning, and it may oversee
the flow of CS information to a CS-US integration area while leaving other gustatory
perceptual processes unimpaired.
US Pathway. The unconditioned stimulus (US) pathway also constitutes a
sensory pathway, one responsible for conducting information about visceral malaise to
higher brain areas. Disruption of this pathway before acquisition or at the time of US
exposure during acquisition would prevent acquisition of a CTA due to failure of US
sensory information to reach the CS-US integration area. Temporary lesion of this
pathway would only be effective during the US phase of conditioning, since this path is
125
typically not invoked during the CS phase of conditioning. Additionally, blockade of the
US pathway during expression testing would have no effect upon CTA since this
pathway is not normally activated in the typical laboratory paradigm for CTA testing.
A variety of evidence implicates the amygdala in US processing. The amygdala
shows altered unit activity in response to intraperitoneal (i.p.) LiCl injection (Reddy &
Bures, 1981; Yamamoto & Fujimoto, 1991; Yasoshima, et al., 1995), with LiCl
responsive units residing in the C and B nuclei (and possibly the L). The unit responses
were varied, with increased activity in some, decreased activity in others, and still others
that cycled between increased and decreased activity at a slow frequency. The numbers
of units responding varied across studies, ranging from a few percent to almost 50%. The
temporal properties of unit responses across studies tended to correspond to the
observable effects of LiCl illness (Lamprecht & Dudai, 2000), with mean response
latencies ranging from about 6 min to almost 30 min, and lasting less than 2 h.
Expression of the immediate early gene c-fos also correlates with LiCl injection
(Lamprecht & Dudai, 1995; Yamamoto, et al., 1997). c-Fos (or Fos-like) proteins have
been immunohistochemically detected in the C and B nuclei (ibid.; St. Andre, Albanos &
Reilly, 2007). The C shows a dramatic increase, while the B produces a small but
significant increase in c-Fos proteins.
Electrical stimulation of the amygdala (4 min: 30 s on, 30 s off) 1 or 30 min after
administration of an apomorphine US disrupted acquisition of a CTA (Kesner, et al.,
1975). On the other hand, a 10 s stimulation delivered 15 min after LiCl injection did not
disrupt CTA acquisition, an effect the author attributed to the rapid establishment of CTA
126
(Arthur, 1975). However, given the former results, this finding could have been due to
insufficient stimulation. Amygdala stimulation does not seem to disrupt the
unconditioned effects of apomorphine illness in general; when severely water-deprived
rats are injected with apomorphine and then given a bottle of water, they will greatly
delay their initiation of drinking (by up to 90 min!). Using this method, Kesner and
colleagues (1975) observed that amygdala stimulation did not hasten or delay onset to
drink water relative to non-stimulated control animals. The disruptive effects of
stimulation on acquisition also cannot be attributed to a brain-state dependent learning
problem since stimulation of the hippocampus or hypothalamus after US injection failed
to interfere with conditioning. Thus it remains possible that stimulation of the amygdala
interfered with transmission of US information to the CS-US association area.
Glutamate release in the amygdala has also been shown to correlate with LiCl
injection. Microdialysis of the amygdala during LiCl injection revealed a significant
increase in glutamate release (Miranda, Ferreira, Ramirez-Lugo & Bermudez-Rattoni,
2002). Additionally, the amount of glutamate released was LiCl-dose dependent, and the
glutamate flux lasted around 2 hours, corresponding with LiCl illness. Intra-amygdala
infusion of glutamate without concurrent US injection can establish a mild flavor
5
aversion (Tucci, Rada & Hernandez, 1998), and glutamate infusion in conjunction with a
subthreshold dose of LiCl (i.p.) can establish a saccharin CTA when neither alone was
5
Recall that flavor is a compound stimulus, with both gustatory and olfactory components. Hence
saccharin (the CS in Miranda, et al., 2002) is a simple tastant, while saccharin-sweetened strawberry Kool-
Aid (the CS in Tucci et al., 1998) is a flavored solution. Amygdalar glutamate infusion alone seems
sufficient to condition a flavor aversion, but apparently cannot condition a taste aversion even though the
dose of glutamate used by Miranda et al. (2002) was 27 times greater than that used by Tucci et al. (1998),
suggesting perhaps that the neural circuitry underlying CTA differs from that subserving conditioned flavor
aversion.
127
sufficient (Miranda, et al., 2002; Ferreira, Miranda, De La Cruz, Rodriguez-Ortiz &
Bermudez-Rattoni, 2005). Furthermore, antagonists of the AMPA, NMDA, and
metabotropic glutamate receptor subtypes administered shortly before US presentation
each disrupted acquisition of a saccharin CTA (Yasoshima, Morimoto & Yamamoto,
2000).
Manipulation of other neurotransmitter systems in the amygdala during the US
phase is likewise seen to affect acquisition. The nonspecific cholinergic agonist
physostigmine interfered with acquisition of a CTA when its action overlapped with an
apomorphine US (Ellis & Kesner, 1981), although preliminary evidence indicates that
specific blockade of either muscarinic or nicotinic acetylcholine receptors in the
amygdala during the US phase is ineffective in preventing acquisition (Castellanos, Salas,
Gonzalez, Roldan & Garcia, 2000). Additionally, while norepinephrine produced no
enhancement or impairment of CTA when infused before the US (Ellis & Kesner, 1981),
similar application of the norepinephrine antagonist propranolol before LiCl injection
impeded acquisition and weakened the resultant CTA (Miranda, LaLumiere, Buen,
Bermudez-Rattoni & McGaugh, 2003). All told, the array of evidence provides strong
support for the notion that the amygdala comprises part of the US pathway.
CR Pathway. The conditioned response (CR) consists of Pavlovian (classical)
and instrumental responses that are elicited by the CS after pairing the CS with the US.
The classical responses typically bear resemblance to the unconditioned responses
triggered by the US, and the instrumental responses allow avoidance of the aversive
substance. In the case of CTA, the preeminent CRs are associated with the CS. These
128
include inhibiting the swallowing of a food recognized as toxic, removing it from the
mouth, and ceasing continued ingestion. Another response, one that is caused both by the
US and by a CS, is LYING-ON-BELLY, a sign of illness in which the rat lays prone,
attempting to have its entire ventral surface touching the floor of its cage (Meachum &
Bernstein, 1990). Still another CR, one peculiar to the CS only, is aversive orofacial and
somatic responses. The CS pathway would affect these various responses through its
control of motor areas responsible for making the CRs, i.e., rejection of CS solution,
presumably through control of swallowing and mastication; illness behaviors, such as
lying-on-belly, perhaps through control of the periacqueductal grey; and aversive
orofacial expressions, through control of facial motor nuclei (motor nuclei of the V
th
,
VII
th
, and XII
th
cranial nerves) and their superordinate pattern generators. The
instrumental responses in CTA would consist of passive avoidance of the sipper bottle, or
active destruction of the CS through physical ejection of the bottle from the cage. Lesion
of the CR pathway would block expression of some or all of these responses. If the CR
pathway emanated from a single brain area, and this area were damaged, one would
expect all CR behaviors to be eliminated. If on the other hand specific components of the
CR (classical versus instrumental responses) are relegated to separate areas, as is thought
to be the case, and as occurs in fear conditioning, then separate lesion of these areas
should eliminate only a subset of the CRs.
The C amygdala showed c-Fos expression in response to the CS after acquisition,
possibly due to activation of the CR pathway (Navarro, Spray, Cubero, Thiele &
Bernstein, 2000), although it could reflect CS pathway activation as well. The
129
electrophysiological findings presented in the CS section above—those that detected
minimal responses to gustatory stimulation prior to conditioning, but enhanced
responsiveness after conditioning—also could reflect CR pathway activation after
conditioning. In any case, both CS and CR pathway activation should correspond with
increased neural activity in some brain areas during CTA expression, so correlative
methods only narrow the range of possible substrates. Disentangling which pathway is
related to which area requires more direct experimental techniques, and even then
dissociation of the two pathways would be difficult. Aside from correlative techniques,
which as stated are difficult to interpret, two other types of experiments would provide
more convincing evidence for a CR pathway. First, stimulation of this pathway should
activate the CR, so that its activation during the consumption of a familiar and nontoxic
solution should still elicit aversive orofacial responses or avoidance. Unfortunately, it
appears that no such study yet exists. Second, temporary lesion of this pathway should
impair CTA expression, and fortunately this kind of experiment has been conducted.
A detrimental effect wrought upon the CR pathway by a transient lesion would
only occur after conditioning in the typical CTA paradigm, because the initial expression
test would be the first time the CR pathway would be invoked (the CR pathway is
presumably not necessary for conditioning to take place). This was purportedly found by
Gallo, Roldan, and Bures (1992): recall that C amygdala inactivation during CS
consumption in the acquisition phase reportedly had no impact upon conditioning, while
inactivation during the expression phase weakened CTA expression. Together, these two
findings suggest that an explanation based on a sensory deficit is unlikely and that the CR
130
pathway was inactivated. In another study, GABA receptors were stimulated in the BLA
amygdala 20 min prior to the expression test through infusion of the benzodiazepine
midazolam. This treatment, effectively a functional lesion, significantly lengthened the
latency to express aversive orofacial and somatic responses to a sucrose CS (Yasoshima
& Yamamoto, 2005). The impaired expression they observed is probably not due to CS
pathway disruption, because systemic administration of midazolam prior to CS ingestion
during acquisition did not inhibit formation of a CTA. Thus this evidence points to a role
for the amygdala in the CR pathway.
Yasoshima, Morimoto, and Yamamoto (2000) found that blockade of AMPA
glutamate receptors (but not NMDA or metabotropic glutamate receptors) in the BLA
complex prevented expression of a CTA when administered just before the expression
test. Antagonism of each of these receptor subtypes disrupted CTA acquisition when it
was interposed between CS and US presentation, which demonstrated that all drugs were
effective in disrupting CTA acquisition, but only AMPA receptors were critical for CTA
expression. Furthermore, a second expression test conducted in the same animals without
blocking AMPA receptors revealed normal CTA expression, demonstrating a transient
effect of AMPA antagonism upon expression, not a permanent memory impairment.
Additionally, a weak quinine solution was rejected by these animals during AMPA
blockade, showing that conditioned aversions were suppressed while innate aversions
were not, i.e., this effect is probably not due to a motivational deficit or pure ageusia.
Still, it remains unclear whether the CS pathway (the branch leading to the CS-US
association area) was disrupted by AMPA receptor blockade. In the microdialysis study
131
of Miranda et al. (2002), glutamate release was assayed in the amygdala during novel
saccharin drinking, yet no significant change in glutamate release was measured. These
results were replicated by Tucci et al. (1998), who detected no change in glutamate
release in response to Kool-Aid drinking by non-conditioned rats. Hence mere gustatory
stimulation does not seem to correspond with glutamate release. However, this latter
study did find that when rats with a CTA were re-exposed to the CS (Kool-Aid), there
was a substantial efflux of glutamate within the amygdala. Thus after conditioning,
consumption of a CS correlates with glutamate release in the amygdala. Ergo it is
possible that a CS pathway disruption could account for the effects of AMPA receptor
blockade obtained by Yamamoto’s laboratory. To engage in a bit of theorization, assume
for a moment that the amygdala is an associative area in CTA, with converging CS and
US pathways and an efferent CR pathway. The CS pathway normally releases little
glutamate and cannot activate the CR pathway it synapses upon, hence the lack of effect
of CS stimulation on glutamate release in the amygdala prior to conditioning. But during
conditioning LTP (long-term potentiation) occurs, causing the CS pathway’s
glutamatergic boutons in this area to strengthen and multiply. Thereafter, stimulation of
the CS pathway will trigger a greater glutamate release—hence the findings of CS-
evoked glutamate release in conditioned rats (Tucci et al., 1998), and the block of
conditioned responding by amygdalar AMPA blockade (Yasoshima, et al., 2000). And
because repeated exposure to the CS alone (without aversive conditioning) does not cause
a change in glutamate release (Tucci et al., 1998), the CTA enhancement of glutamate
132
release cannot be attributed to non-associative, repeated-use strengthening of the CS
pathway. In sum, the data are at least suggestive of a CR pathway in the amygdala.
CS Novelty. Novelty refers to the perception of the familiarity of a substance.
Foods that have never before been encountered are discerned as novel and trigger
neophobia, a concept related to but distinct from that of novelty. NEOPHOBIA is a fear
expressed toward novel foods, and a resulting hesitancy to consume large quantities of
that food (Barnett, 1975; Figure 2.5). If presented with a choice between a familiar food
and novel foods (in a MULTIPLE FOOD PREFERENCE TEST), rats will spend more time eating
the familiar food (Borsini & Rolls, 1984). However, over the course of repeated
sampling of a novel food (3-5 samples), the novelty dissipates and the animal recognizes
the food as familiar as evidenced by increases in the amount of the food consumed, that
is, an attenuation of neophobia (Bermudez-Rattoni, 2004). If no adverse consequences
follow the samplings, then the food becomes familiar-nontoxic, and if the animal is
motivated to ingest the food (because of its nutritive value, caloric value, or palatability),
it will consume greater amounts compared to its first encounter. If illness follows the
samplings, then it becomes familiar-toxic, and the animal will decrease subsequent
consumption in proportion to the magnitude of the illness, i.e., it will develop a CTA.
Familiar substances often fail to be seen as aversive after a single acquisition trial,
perhaps because previous experience with a particular CS has established that taste as
“safe” (or failed to establish it as toxic). Therefore, if a lesion disrupts the perceived
novelty of tastes, by default making them seem familiar, then that lesion is going to
disrupt the ease of conditioning (remember, one-trial conditioning requires a novel CS); it
133
will also disrupt the expression of neophobia, which, of course, depends upon the novelty
of a food.
Note that although lesion-induced destruction of a novelty perception center
would affect both CTA and neophobia, it is possible a lesion could disrupt only the
expression of neophobia without affecting the perception of novelty. We postulate that
such an impairment in neophobia could be detected by measuring the rate of attenuation
of neophobia over the course of several CS exposure trials: lesioned rats should be
drinking at or near asymptotic levels (~25 mL/hour) on every trial, whereas control rats
would drink little on the first trial (5-10 mL/hour), and gradually increase consumption to
asymptotic levels over the subsequent trials. Complementarily, if these lesioned rats have
retained an intact novelty perception, this fact could be assessed by conditioning a CTA,
since CTA strength is inversely proportional to CS novelty. Thus these non-neophobic
rats should develop a CTA equal in strength to that of control rats. This sort of
dissociation of novelty and neophobia has been described by Yamamoto et al. (1995),
who found that lesions of the entorhinal cortex significantly increased acquisition day
consumption of a novel CS, while the aversion conditioned to that CS was just as strong
as control rats. Thus lesion of the entorhinal cortex seemed to disrupt neophobia but not
novelty.
The novelty-familiarity dimension of a CS is independent of its positive
(palatable) and negative (aversive) hedonic qualities and its perceived toxicity. Thus a
new CS can be novel-aversive (e.g., novel quinine), yet once familiarized will still be
aversive.
134
Figure 2.5 A theoretical model of the CS novelty pathway. In the left-hand
model, a novel CS triggers activity in the novelty perception area, which activates
the neophobia and CTA engram areas. The heavy arrows indicate activation of
that connection. The neophobia area can inhibit drinking, and if activation of the
CTA area is followed by US pathway activation, it too becomes capable of
inhibiting drinking. In the right-hand model, a familiar CS fails to induce
activation of the novelty perception area, which in turn does not activate the
neophobia or CTA engram areas (indicated by the light arrows). This blocks CTA
acquisition, and prevents expression of neophobia.
Yet a familiar CS can be toxic and palatable at the same time (e.g., a sucrose
solution that has been toxiphobically conditioned: familiar and toxic-aversive and
palatable). Two interesting questions arise when considering the aversive-palatable
dimension. First, is the toxic dimension truly different from the aversive? Yes, because a
dehydrated animal will drink bitter quinine to rehydrate, but will avoid a poisonous
solution, whether palatable or aversive, even though parched (Yasoshima, et al., 1995).
Second, are the positive and negative hedonic qualities really two separate, independently
adjustable dimensions, and not just one love-hate continuum? Yes, a solution like the
135
sucrose described above is innately highly palatable, and CTA merely adds aversive and
toxic valences to it, without necessarily reducing its palatability as measured by orofacial
expressions. When CS consumption and LiCl illness occur simultaneously, the animal
expresses a high number of both ingestive and rejective orofacial responses when the CS
is resampled during the expression test (Spector, et al., 1988). However, most acquisition
trials utilize the trace conditioning technique, in which CS consumption is concluded
prior to the onset of LiCl illness, and in this case the number of ingestive responses
during the expression test is reduced (Breslin, et al., 1992).
The evidence indicating a role for the amygdala in novelty and/or neophobia is
mixed. If a functional lesion were to affect only the perception of CS novelty, then its
disruptive effects should only be apparent during the CS presentation on the first
acquisition trial, and not during expression testing or subsequent acquisition trials,
because at those subsequent exposures the CS is no longer novel (rather, it is familiar-
toxic). So disrupting novelty perception alone should affect nothing at post-acquisition
time points. Recall that TTX lesion of the C nucleus shortly before acquisition did not
impair CTA formation, and the lesion was effective during and long after novel saccharin
CS presentation (Gallo, et al., 1992). Furthermore, these rats actually drank less during
the saccharin presentation than did control rats, and had to be granted extra time to
drink—if novelty perception were impaired, they would be expected to drink more
saccharin in the allotted time, due to the consequent disruption of neophobia. Thus the C
nucleus, at least, does not seem to be involved in novelty perception.
136
A number of permanent lesion studies have found increased consumption of a
novel CS by amygdala lesioned rats relative to their baseline water consumption or to
sham-operated controls. Specifically, these studies found lessened neophobia after
lesions of the C, M, L, B, or BLA complex (Rolls & Rolls, 1973a,b; Nachman & Ashe,
1974; Aggleton, et al., 1981; Fitzgerald & Burton, 1983; Yamamoto et al., 1995; Sakai &
Yamamoto, 1999; Morris, et al., 1999; Yamamoto & Fujimoto, 1991). Other studies
however have found no effect of amygdala lesion upon neophobic responses when
lesions targeted the C, L, B, BLA complex, or the cortical-medial nuclei (Aggleton, et al.,
1981; Lasiter & Glanzman, 1985; Yamamoto, et al., 1995; Sakai & Yamamoto, 1999;
Morris, et al., 1999). Further blurring the picture are studies showing neophobia
differences between liquid and solid foods. Rollins and colleagues (2001) made lesions
of B or M, or sham lesions in a control group, and detected no differences in amount
consumed or latency to consume a sweetened milk CS upon first exposure between
groups. The rats receiving B lesions still displayed CTA deficits. Intriguingly, this group
did subsequently display impaired neophobia to a novel solid food (Froot Loops).
Similar results in BLA-lesioned rats were obtained by Borsini & Rolls (1984), i.e., they
evinced reduced neophobia to novel solid foods but not to novel sucrose solution, with
moderate disruption of CTA learning. To recapitulate, after lesions of essentially the
same areas of the amygdala, sometimes neophobia deficits are found, sometimes they are
not, and the state of matter of the CS may make a difference.
Other evidence suggests that the amygdala may be important primarily for
aversive learning when the CS is novel, but not when it is familiar. Dunn and Everitt
137
(1988) made large lesions within the amygdala, then gave two CS exposures over 2 days
to assess lesion effects on neophobia. Shortly after the second exposure, rats were
injected with LiCl to create CTA. As will be recalled, they found none. When Morris
and colleagues (1999) carried out their study with careful histology, they included a
replication of the method used by Dunn and Everitt, giving two CS exposures and
reinforcing only the second one with LiCl. They found that even when just one
preexposure to CS was given, normal CTA was acquired by BLA-lesioned rats, i.e., CS
preexposure effaced the ability of amygdala lesions to attenuate CTA. However,
Fitzgerald & Burton (1983) gave a single CS preexposure before CTA acquisition, and
still found CTA learning deficits in BLA-lesioned rats. Again, then, there is ambivalent
evidence regarding the effect of amygdala lesions on CS familiarity and conditioning.
The transcription of c-fos has been correlated with sampling of a novel CS
solution. The BLA expresses more c-Fos protein after subjects received a novel CS-US
pairing than when a familiar CS-US pairing or LiCl alone were given (Koh & Bernstein,
2005). Moreover, the C amygdala (but not the BLA) generates more c-Fos in response to
novel saccharin ingestion than familiar saccharin (Koh, et al., 2003; Koh & Bernstein,
2005).
Examination of several neurotransmitter systems within the amygdala has also
yielded muddled support for a role for the amygdala in novelty perception or neophobia.
Neither neurotoxic catecholaminergic depletion nor depletion of 5-HT within the
amygdala affected the expression of neophobia (Borsini & Rolls, 1984), even though the
former did disrupt expression of a sucrose CTA (though see Fernandez-Ruiz, Miranda,
138
Bermudez-Rattoni & Drucker-Colin, 1993
6
). Because catecholaminergic depletion
affects norepinephrine, epinephrine, and dopamine terminals, this disruption could be due
to one of several factors. Infusion of norepinephrine into the amygdala actually
increased neophobia in a multiple food preference test (Borsini & Rolls, 1984). Post-CS
infusion of the norepinephrine antagonist propranolol prevented familiarization with the
CS, i.e., it sustained novelty or neophobia across several CS exposures (Miranda, et al.,
2003); a similar infusion of norepinephrine itself had the same effect (Ellis & Kesner,
1981). Hence amygdalar norepinephrine may play a role in novelty or neophobia. On
the other hand, serotonin and cholinergic agonist infusion were without apparent effect
on food choice (Ellis & Kesner, 1981; Borsini & Rolls, 1984). Dopamine also appears to
have no effect upon novelty and neophobia. Intra-BLA infusion of either saline, a D
1
, or
D
2
/D
3
dopamine receptor antagonist just prior to CS administration during three
preexposures and one acquisition trial had no effect upon conditioning (Stevenson &
Gratton, 2004), apart from the usual effects of CS preexposure (latent inhibition, i.e.,
attenuation of the subsequent CTA). Although Stevenson and Gratton did not report data
regarding CS consumption during the three preexposure trials (i.e., neophobia),
precluding direct assessment of dopamine antagonism on novelty/neophobia, a lack of
effect of dopamine receptor antagonists on novelty/neophobia can be inferred: since
latent inhibition occurred normally in these animals, dopamine antagonists seemed
6
This study (Fernandez-Ruiz, et al., 1993) employed the taste-potentiated odor aversion paradigm, wherein
a flavor stimulus (almond-scented saccharin) is paired with illness. During the expression tests, the odorant
and tastant are presented separately. This study found that catecholaminergic depletion of the anterior BLA
disrupted odor aversion, but not taste aversion. However, electrolytic lesion of either the anterior or
posterior BLA has been demonstrated not to impair CTA to HCl (Fitzgerald & Burton, 1981), so the lack of
effect of neurotoxic lesion may perhaps be explained by anterior placement.
139
neither to enhance nor disrupt novelty or attenuation of neophobia. In grand total, the
balance of data comes out about evenly split on whether or not the amygdala participates
in novelty perception or expression of neophobia; perhaps it has a minor role that is not
always invoked.
Modulation. This term refers to the regulation of the strength or persistence of a
memory, rather than to the direct mediation of the existence or absence of a memory. As
such, it is concerned with the components and events surrounding the process of
consolidation. Consolidation is “the progressive postacquisition stabilization of long-
term memory” (Dudai, 2004), and in the present work refers to the processes involved in
changes of synaptic efficacy and number that occur during and after conditioning, also
called synaptic consolidation. This model purports that consolidation occurs at the site of
memory storage, i.e., the engram (see next section, Engram). It is thought to work in
conjunction with the heart of the long-term potentiation (LTP) model of synaptic
plasticity, glutamate. Glutamate is an excitatory neurotransmitter, which evidence
suggests is the mediator of memory in LTP, and its receptors come in several varieties.
Two types that are important for LTP, the AMPA ( α-amino-3-hydroxy-5-methyl-4-
isoxazolepropionic acid) and NMDA (N-methyl-D-aspartate) receptors, are ionotropic
membrane channels, directly depolarizing the neuronal membrane by allowing influxes
of Na
+
and sometimes Ca
2+
. The entry of Ca
2+
during acquisition triggers a cascade of
intracellular events that implement a near-immediate increase in synaptic efficacy
through the insertion of additional AMPA receptors into the postsynaptic membrane
(Abraham & Williams, 2003). Yet this increase is short-lived, on the order of 3-4 h
140
(aptly named EARLY LTP), unless the coincident intervention of a modulatory
neurotransmitter takes place. Modulatory neurotransmitters act through metabotropic
receptors
7
to initiate transcription of nuclear genes and thus generate new proteins which
will compensate for the decaying early LTP and greatly extend the life of the LTP (then
termed LATE LTP). Hence the modulation spoken of here concerns the source of the
modulatory neurotransmitter that initiates consolidation; while consolidation, and the
various genetic, molecular synthetic, and other processes it entails, is more properly a
component of the engram (next section). To give a concrete hypothetical example,
suppose that when the noradrenergic neurons of the locus coeruleus are stimulated by US
pathway activation, they secrete the modulatory neurotransmitter norepinephrine into the
amygdala, the site of the CTA engram. The locus coeruleus would be the source of the
modulatory influence, and the amygdala, the place where consolidation would need to
occur, would be the target of that modulatory influence (see Figure 2.7, left-hand model).
There is specific evidence of a need for neuromodulators in consolidation.
Exploring early and late LTP in the hippocampus, O’Carroll and Morris (2004) found
that NMDA receptor-dependent LTP needs to be initiated in conjunction with activation
of receptors for dopamine, a neuromodulator. Induction of LTP in the presence of a
D1/D5 dopamine receptor antagonist resulted in early LTP that decayed back to baseline
within 6 h. Thus unfettered glutamate activity without accompanying dopaminergic
modulatory activity results only in early LTP. In a separate group in the O’Carroll and
Morris study, blockade of NMDA receptors prevented any potentiation whatsoever. So
7
Metabotropic receptor activation can affect neuronal excitability as well.
141
dopamine activity alone is incapable of triggering even early LTP. This further
demonstrates that glutamate mediates early LTP induction, while dopamine modulates
the durability of the already-established potentiation (converting it to late LTP), and that
their simultaneous occurrence is necessary to initiate and sustain LTP, respectively.
In CTA, modulation is a probably a special branch of the US pathway, becoming
active during the US phase of acquisition in order to consolidate or to facilitate
consolidation of the short-term associative memory formed by CS-US integration. In the
inhibitory avoidance paradigm (see Section 1.2.1), the amygdala is believed by some to
be a source of modulation for other brain areas (Barros, et al., 1999; McGaugh, McIntyre
& Power, 2002; McGaugh, 2002). Inhibition of excitatory activity in the amygdala
through blockade of AMPA-type glutamate receptors at 0, 1.5, and 3 h (but not 6 h) after
avoidance training disrupted retention of avoidance memory (Jerusalinsky, et al., 1992),
and total functional inactivation of the amygdala by lidocaine caused amnesia for
avoidance training when administered at 0 and 6 h after training (Parent & McGaugh,
1994). These data seem to suggest that functional inactivation of the source of
modulatory input (i.e., the amygdala, in this case) would disrupt long-term memory
formation when occurring any time from 0 to approximately 6 h after training
8
(Ambrogi
Lorenzini, Baldi, Bucherelli, Sacchetti & Tassoni, 1999). Similarly, one would expect
disruption of CTA to follow an analogous time course should the role of the amygdala in
8
Why does continuous activity seem to be required? Bures, Buresova & Ivanova (1991) speculate that the
reason post-acquisition inhibition of neural activity can interfere with memory is that consolidation requires
a prolonged secretion of “trophic factors” (i.e., modulatory neurochemicals) and a high “synaptic drive” to
sustain plastic processes. It is also possible that late LTP entails enlarging existing synapses or growing
new synapses (Abraham & Williams, 2003), and that activity is required at these new outgrowths for
proper synaptic development to occur, as in motor endplate development (Sanes & Jessell, 2000).
142
fact be to modulate aversive taste memory consolidation elsewhere. Importantly,
however, one should bear in mind that functional inactivation by TTX of either the
source of modulation (the neuromodulatory neurons) or the target of modulation (the
engram) would likely be equally effective in disrupting long-term CTA memory, because
the net result of reduced neuromodulatory input at the engram would be accomplished by
inactivation at either area
9
(see Figure 2.7, right-hand model).
Actual data from CTA experiments regarding this role for the amygdala are sparse
and difficult to interpret. As described in the US Pathway section, strong, brief electrical
stimulation of the amygdala 15 min after an acquisition trial produced no impairment or
enhancement of CTA compared to non-stimulated controls (Arthur, 1975), while weaker,
more prolonged stimulation of the amygdala attenuated CTA when delivered 30 min or 3
h (but not 12 h) after the apomorphine US (Kesner, et al., 1975). The 3 h group in the
latter study is particularly relevant, since its stimulation would have been given well after
CS-US associational processes had begun, that is to say, during consolidation. However,
one might have expected that stimulation of a modulatory source area would have
facilitated CTA rather than diminished it. In another study already discussed, TTX
inactivation of the amygdala partially impaired CTA acquisition when administered at 0
or 1.5 h after LiCl, while at 6 h post-acquisition it had no effect (Roldan & Bures, 1994),
which is roughly congruent with the timeline observed for inhibitory avoidance, for
9
The rationale is this: TTX blocks voltage-gated sodium channels, rendering any neuronal membranes in
its area of effect incapable of active signal propagation, and hence more or less unresponsive. A TTX
lesion in the engram area (Fig 2.7) would not only inactivate the engram neurons, it would inactivate the
axon terminals of efferents from the locus coeruleus. Even though the coerulear neuron somata are firing,
those action potentials are not propagated down to the boutons in the engram area, and thus no
norepinephrine will be released. This possibility is limited to TTX lesions. A muscimol lesion would work
differently, and would not necessarily prevent NE release in its area of effect.
143
which the amygdala is believed to be a source of modulation. Yet this technique as
implemented fails to discriminate between source and target of modulatory influences.
We predict that TTX lesion of the source of modulation would leave short-term CTA
expression intact but prevent long-term memory formation; lesion of the target of
modulation would eliminate short-term CTA expression and long-term memory
formation. Since Roldan and Bures only tested CTA expression at long-term intervals,
they could not differentiate between these two possibilities. All together, this smattering
of data is at least inconclusive, and at most argues against a purely modulatory role for
the amygdala in CTA.
Figure 2.6 Illustration of a hypothetical model of the modulatory pathway in the CTA
neural circuit. The left-hand model shows the modulatory pathway as a branch of the
US pathway, with the locus coeruleus as the source of modulatory norepinephrine
(NE) being released in the engram area, the target of modulation. The engram area can
form a short-term memory, but requires modulatory input to convert the short-term
memory to long-term. In the homologous right-hand model, the encircled area
represents the area of a functional lesion by tetrodotoxin. All axons projecting into and
out of the area are affected by the lesion, meaning that even if the modulatory pathway
is stimulated, it cannot release norepinephrine into the engram area. Should the
amygdala have a modulatory role in CTA, it, not the locus coeruleus, would be
releasing modulatory neurotransmitters into the engram area.
144
Engram. The amygdala may be a site of actual memory storage, the locus of the
engram. If this is the case, then this area would process both CS and US information, and
would connect to motor areas responsible for making the CR. The fact that data support
a role for the amygdala in CS, US, and CR pathways makes a circumstantial case for a
memory storage function. Paired presentation of the CS and US in a conditioning
paradigm would result in lasting changes in this area that allow the CS, on subsequent
encounters, to activate the efferent projections of this area, the CR pathway (or at least
one of the CR pathways). These lasting changes would reflect, we assume, the
establishment of long term potentiation (LTP). Short-term memory of the CTA would be
mediated by early LTP in this engram area, and its transition into long-term memory (via
late LTP [>3-6 h]) would require the intervention of modulatory neurotransmitters. Late
LTP, like long term memory (Davis & Squire, 1984), requires the synthesis of new
proteins in the neurons doing the learning (Kandel, 2000). Hence blockade of protein
synthesis in this brain area would allow conditioning and acquisition to take place
normally, but would prevent the establishment of late LTP and long term memory (Figure
2.8).
Protein synthesis inhibitors have been found to disrupt CTA when applied
intracisternally
10
(Tucker & Oei, 1982). Studies of long term memory in general (Davis
& Squire, 1984), and CTA in specific (Rosenblum, Meiri & Dudai, 1993), indicate that at
least 90% inhibition of protein synthesis is required in order to block the establishment of
10
Intracisternal: an intracerebral route of administration that affects the hindbrain only.
145
memory. Protein synthesis inhibitors are poisonous, and so systemic injection of them
actually enhances CTA, presumably due to their US effects.
Figure 2.7 A hypothesized model of the engram. The large circle represents a neuron in the engram area, which
receives synaptic terminals from the CS pathway (filled triangles), a strong US pathway (hence the two inputs), and
modulatory pathway (open triangles for both), and the efferent projections of this neuron form the start of the CR
pathway (arrows). The CS and US pathways release glutamate onto postsynaptic AMPA receptors (small open circles).
Paired CS-US pathway activation constitutes an acquisition procedure, which induces a rapid increase in synaptic
efficacy at the CS synapse caused by additional AMPA receptors added to the postsynaptic membrane (center figure).
Simultaneously, the modulatory pathway releases norepinephrine (NE), which in turn stimulates gene transcription in
the nucleus (small black circle) and subsequent protein synthesis. These new proteins will act to consolidate this new
memory by causing growth of a new synaptic terminal in the CS pathway (right-hand figure). However, if protein
synthesis is blocked, the gains in synaptic efficacy seen just after acquisition will decay until the nexus returns to its
nascent state (left-hand figure).
Why don’t they still block CTA despite their toxic effects? Knowledge gained
from LTP experiments indicates that protein synthesis inhibition is only effective if
applied within 15-30 min of induction of potentiation (Adams & Dudek, 2005). Thus
protein synthesis inhibitors administered peripherally would have to inhibit 90% of
protein synthesis throughout the body and brain, and do so within roughly 15 min of its
malaise-inducing effects. The most common protein synthesis inhibitor used in recent
years is the antibiotic anisomycin. This drug has various effects, but has repeatedly been
shown to arrest protein synthesis in various preparations in vivo and in vitro. Rosenblum,
146
Meiri, and Dudai (1993) assessed the temporal parameters of protein synthesis inhibition
induced by anisomycin in the insular cortex, finding that anisomycin achieves 90%
efficacy by 20 min, and stays above the critical 90% threshold for 90-100 min. This was
achieved by direct infusion of 100 µg of anisomycin directly into the insular cortex. In
another group, intracerebroventricular injection of 300 µg of anisomycin failed to achieve
the critical threshold—the closest it came was 80% inhibition at roughly 45 min after
infusion. Thus peripherally administered anisomycin probably has trouble penetrating
the blood-brain barrier and amassing in sufficient concentration in the brain to achieve
the threshold of inhibition.
Few studies have interfered with protein synthesis in the amygdala during taste
aversion conditioning. The first study to utilize this technique (Lamprecht & Dudai,
1996) allowed water deprived animals 10 min access to saccharin, followed 40 min later
by an i.p. injection of 0.15 M LiCl (2% bw). Anisomycin was infused into the C nucleus
of the amygdala either once, 30 min before saccharin presentation in one group, or twice,
5 min after saccharin and again 30 min after LiCl in a second group. When tested a few
days later in a two-choice presentation of both saccharin and water bottles, the group
receiving the peri-lithium protein synthesis inhibition (two infusions) exhibited weaker
CTAs than a vehicle-infused control group, which acquired strong CTA, while the group
receiving the pre-saccharin infusion of anisomycin (one infusion) expressed no CTA (i.e.,
virtually equal preferences for saccharin and water). In a similarly designed study
(Bahar, Samuel, Hazvi & Dudai, 2003), anisomycin was infused into either the C or B
nucleus 20 min prior to saccharin access. As in the previous study, protein synthesis
147
inhibition of the C nucleus during acquisition completely blocked CTA, while protein
synthesis inhibition of the B nucleus had essentially no influence on CTA. These results
suggest that anisomycin disrupted the storage of some part of the memory of the taste
aversion in the C nucleus. This hypothesis is supported by the following additional
findings. Anisomycin did not affect expression of the CTA when infused 20 or 30 min
prior to testing, demonstrating no impairment in neural activity, and rats receiving C
nucleus infusions were able to learn subsequent aversions to different flavors,
demonstrating no lasting impairment. Anisomycin delivered into the amygdala did not
produce US effects, as evidenced by the failure of rats receiving anisomycin and
saccharin in the absence of LiCl to develop a CTA. The infusion did not cause any
obvious loss of cells in the infusion area, nor were cells moribund as assessed by means
of cytochrome oxidase staining (an indicator of metabolic activity). These unusual and,
apparently, reproducible results stand in contrast to much of the vast body of amygdalar
CTA research, which have generally found no role for the C and a central role for the
BLA complex in CTA acquisition. However, these results could be incorporated into the
model advanced earlier (Section 2.1), regarding the complementary roles of the C
(Pavlovian) and BLA (instrumental). As described above (see CS Pathway), single unit
recordings of the C nucleus have found inhibitory responses when the animal ingests a
saccharin solution that has been paired with LiCl, but not before pairing (Yasoshima, et
al., 1995). Neurons in this area show similar responses to other unpalatable stimuli such
as quinine and HCl, perhaps indicating that conditioned and unconditioned orofacial
rejection responses are mediated by the C nucleus. Lesion studies support this in that
148
permanent or transient lesions of the C increase aversive orofacial responses but they also
decrease ingestion of a palatable substance (Touzani, et al., 1997; see Section 2.2).
Together, these findings imply that decreasing activity in the C nucleus may trigger
expectoration. After CTA acquisition, the CS would normally be conditioned to inhibit
activity in the C, consequently leading to cessation of ingestion and initiation of rejection
responses. Permanent lesions of this nucleus would essentially be equivalent to sustained
inactivation of the C, and CTA would thus appear normal in these animals. However,
protein synthesis inhibition in the C would leave the neurons and their connections intact,
but block any long-term memory of conditioning, so on subsequent exposure to a
palatable CS like saccharin, the C would not inhibit its activity, and the animal would not
be compelled to cease drinking. What about the apparent lack of participation of the
BLA? One possibility is that the L nucleus is the principal actor in the BLA’s
contribution to CTA acquisition, so protein synthesis inhibition of the B would not be
expected to influence later saccharin rejection. Another possibility is that the B nucleus
contributes to later phases of CTA acquisition, such as the US phase, a phase which was
not completely interrupted in the Bahar et al. (2003) study; recall that pre-CS infusion of
anisomycin into the B nucleus was without effect on CTA. In any case, the results of
Dudai’s studies demonstrate that protein synthesis in the amygdala is important for CTA
acquisition. Based on the LTP model of learning and memory, this suggests that some
memory is being stored there.
However, if the role of the amygdala in fear learning is any guide, then there
could be multiple CTA engrams within the amygdala, to mediate separately the
149
classically conditioned and instrumental responses. In fear learning, a C nucleus pathway
mediates the Pavlovian aspects of fear conditioning, including such CRs as freezing,
tachycardia, pressor responses, and defecation; while a B nucleus pathway seems to
mediate the instrumental responses, such as passive and active avoidance (see Figure
1.4). These pathways and responses have been dissociated in a single lesion study
(Killcross, et al., 1997). In this study, hungry rats were trained in a two-lever
conditioning box to bar-press for food pellets. Then rats received C, B-L, or sham
lesions. After recovery, the rats were classically conditioned to associate an auditory CS
with a brief footshock (CS+). The rats were returned to the conditioning box, only now
there were two reinforcement schedules on the levers: first, as before, food was obtained
from pressing either of the bars, which motivated the bar-pressing response in the rats;
second, using a less frequent reinforcement schedule, pressing one bar periodically
elicited the feared CS+, while pressing the other bar triggered a neutral control tone
(CS-). Control rats began to show the instrumental passive avoidance response by
biasing their bar-pressing away from the lever that elicited the feared CS+ and towards
the lever that triggered the neutral CS-. Additionally, whenever the CS+ sounded, these
rats would show the Pavlovian freeze response and cease all bar-pressing. Rats with
lesions of the C nucleus would passively avoid the lever that activated the aversive CS+,
but failed to show the conditioned freeze response when the CS+ was sounded.
Importantly, rats with lesions of the B-L nuclei appropriately showed the freeze response
when the CS+ was sounded, but failed to bias their responses away from the bar
triggering the CS+. Thus these rats were capable of a normal Pavlovian conditioned fear
150
response, but were incapable of instrumental avoidance responses. In line with this
behavioral deficit, it is thought that the B-L nuclei either contain or allow a CS pathway
access to a hedonic representation of the US (Cardinal, et al., 2002; Fanselow & Gale,
2003). If a CS-US association is occurring in the B (or B-L) nucleus, then this would
constitute an engram. (This is not to say that the classically conditioned responses are not
hedonically motivated [they may or may not be], just that in the amygdala classical CRs
are dissociable from instrumental responses). Thus the Pavlovian responses are mediated
by the C nucleus and the instrumental responses are mediated by the B nucleus. With
efferents targeting different parts of the brain, these separate engrams can elicit or
influence the Pavlovian and instrumental responses (as described in Chapter 1). If this is
true for fear learning, perhaps it is also true for CTA.
Summary of the role of the amygdala. The information presented in the
preceding eight sections pertains to the amygdala as a whole; some experiments targeted
specific nuclei, while others aimed to explore the amygdala in general. Although this
notion was partly, implicitly conveyed in the preceding sections, its explicit statement
here is of value. Manipulations were made to various parts of the amygdala, some of
which may be involved, possibly in separate phases or facets of CTA, and some
uninvolved. Thus a coherent picture of how the amygdala contributes to CTA does not
easily emerge from the mélange of puzzle pieces. From the data presented above, it is
possible that nuclei of the amygdala could be involved in the CS, US, or CR phases, CS
novelty, or memory storage (engram). A systematic application of a combination of
these techniques to one part of the amygdala during various phases of acquisition and
151
expression of CTA would yield a coherent picture of the role of the amygdala (or at least
one part of it) in CTA.
152
3
RECONSOLIDATION
3.0 Consolidation and reconsolidation
CONSOLIDATION is a feature of memory formation whereby a newly learned
memory is transformed from an impermanent and ephemeral form into a durable long-
term storage form. A recently acquired memory is unstable and can be forgotten after
several hours unless changes occur in the neurons engaged in learning the new piece of
information. Such changes include alterations in gene expression and the manufacture of
new proteins. The proteins then help to consolidate the memory, and are postulated to do
so by helping to stabilize synaptic changes concomitant to learning, allowing the memory
to persist past the initial few hours. Prior to crystallization of the memory, new memories
can be eliminated by interventions that disrupt consolidation like drugs that prevent
protein synthesis, electroconvulsive shock, and traumatic brain injury. But as more time
passes, such insults fail to affect those memories, because consolidation has converted the
memories into a stable form that is henceforth impervious to dissolution. Or so it was
generally believed, until the late 1990s (cf. Spear & Mueller, 1984).
At that time, it was found that even fully consolidated memories could be
disrupted if they were recalled prior to amnestic insult: memories, if reactivated, return to
a plastic and revocable state similar to the period of several hours after initial learning,
and must be consolidated again in order to be retained. The first modern appreciation of
153
this phenomenon was by Przybyslawski and Sara (1997). They trained rats to find
chocolate breakfast cereal in an eight arm radial maze until near errorless performance
was achieved, and then injected them systemically with either saline or the NMDA
receptor antagonist MK-801 five minutes after the criterion trial. At the test trial 24 h
later, the drug-treated rats exhibited more errors than saline-treated rats, i.e., they seemed
to have forgotten their training. In a follow up study, rats were injected at various time
points after the criterion trial, from 5 to 180 min, and then retested 24 h later. MK-801
impaired performance when injected up to 90 min after the criterion trial. They
interpreted their results as indicating that a reactivated memory reenters a plastic state,
and that this memory needs to “reconsolidate” in order to be retained further.
Although the article by Przybyslawski and Sara apparently flagged in exciting the
scientific community, Nader, Schafe, and LeDoux (2000) conducted a similar experiment
and reported their results in the prestigious journal Nature, which reached a wide
audience and seeded a deluge of experiments exploring the phenomenon of
reconsolidation. In the decade that has followed, some of the basic features of
reconsolidation have been fleshed out through experiment.
3.1 Some of features and parameters of reconsolidation
Reconsolidation can be induced in a number of vertebrate and invertebrate
species using a variety of learning tasks. The reconsolidation of reactivated memory
has been observed in organisms representing a wide array of phylogenetically diverse
taxa: the snail Lymnaea, a species of Chasmagnathus crab, honeybees, medaka fish
154
(killifish), mice, rats, and humans (Eisenberg, Kobilo, Berman & Dudai, 2003; Pedreira
& Maldonado, 2003; Suzuki, et al., 2004; Stollhoff, Menzel & Eisenhardt, 2005;
Frankland, et al., 2006; Hupbach, Gomez, Hardt & Nadel, 2007). In these species,
reconsolidation can be observed (through its disruption) across an equally diverse range
of memory tasks: cued fear conditioning, contextual fear conditioning, second-order fear
conditioning, eyeblink conditioning, inhibitory avoidance, conditioned taste aversion,
positively reinforced conditioned place preference, appetitive olfactory conditioning,
object recognition memory, radial and water mazes, and semantic and episodic memories
(Przybyslawski and Sara, 1997; Nader, et al., 2000; Debiec, LeDoux & Nader, 2002;
Milekic & Alberini, 2002; Gruest, Richer & Hars, 2004; Inda, Delgado-Garcia &
Carrion, 2005; Torras-Garcia, Lelong, Tronel & Sara, 2005; Akirav & Maroun, 2006;
Debiec, Doyere, Nader & LeDoux, 2006; Morris, et al., 2006; Hupbach, et al., 2007;
Wang, et al., 2008). Interestingly, the reconsolidation of pure contextual memory could
not be disrupted by an amnestic treatment (Biedenkapp & Rudy, 2004), although, as
explained below, memories can resist the induction of reconsolidation under certain
conditions. Notwithstanding the latter study, the near ubiquity of the reconsolidation
effect suggests that it is probably an important ancestral component of memory.
Older memories appear to be more stable and harder to replasticize than
recent memories. Increasing the span of time between the acquisition of a memory and
its reactivation seems to render a memory that is increasingly more difficult to initiate
reconsolidation. In a study using contextual fear conditioning, mice received a mild
single-shock training episode and were then returned to their home cages for 1, 3, or 8
155
weeks. After the waiting period, the mice were given systemic injections of anisomycin
30 min before they were placed in the conditioning box for a 3 min re-exposure period.
The following day they were tested for fear memory expression in the conditioning box.
It was found that the 1 and 3 week-old memories were disrupted by the reactivation-
anisomycin treatment, but the 8 week-old memory was intact as these mice spent as much
time freezing as vehicle-infused controls (Suzuki, et al., 2004). In another study by the
same research group (Frankland, et al., 2006), mice received a more intensive 3-shock
training, were re-exposed to the conditioning box at either 1 or 36 d, and then were
immediately injected with anisomycin in the dorsal hippocampal area. At the subsequent
test session, memory for the 1 d-old memory was significantly diminished while memory
for the 36 d-old fear memory was unaffected by the treatment. Demonstrating the same
effect, Milekic and Alberini (2002) gave rats a single training session in an inhibitory
avoidance training apparatus and then set them aside for 2 d, 1 week, 2 weeks, or 4
weeks. Subsequent re-exposure to the apparatus was followed immediately by systemic
anisomycin injections. During the test session 48 h later, it was learned that the
reactivation-anisomycin procedure disrupted reconsolidation of the 2 d- and 1 week-old
memories, but that 2 and 4 week-old memories were not affected by the treatments.
These experiments suggest that as memories age, they become less susceptible to the
replasticizing effects of recall (Winters, Tucci & DaCosta-Furtado, 2009).
However, two studies have found that older classical fear memories could be
replasticized. Debiec and LeDoux (2004) were able to disrupt reconsolidation of a
classically conditioned fear memory when re-exposure to the CS occurred 2 months after
156
training. Systemic injection of the norepinephrine antagonist propranolol immediately
after the reactivation trial significantly reduced the expression of fear when tested 48 h
later; a retest 1 month later showed no resurgence of the fear, indicating prolonged
memory impairment. Additionally, in another series of experiments (Debiec, et al.,
2002), acquisition-to-reactivation intervals of 3, 15, and 45 d were tested in separate
contextual fear conditioning experiments. Immediately after re-exposure, rats received
infusion of anisomycin into the dorsal hippocampus. Comparison of the data from these
groups revealed no significant diminution of plasticization with increased age of the
memory, i.e., all memories were equally disrupted by anisomycin infusion after re-
exposure. Thus memories aged for two long retention intervals, 45 d and 2 months,
remained susceptible to disruption through the induction of reconsolidation. What
accounts for this discrepancy? One potential factor could be the potency of the amnestic
treatment used in these experiments, a factor that could overcome age-induced stability
and that will be described below.
Well-trained memories are more stable. Increasing the strength of the original
memory through training also increases the stability of the memory, making it more
difficult to initiate reconsolidation (Morris, et al., 2006; Winters, et al., 2009). Suzuki
and colleagues (2004) found that three-trial training of mice in the contextual fear
conditioning paradigm rendered memories that resisted the disruptive effects of re-
exposure followed by systemic anisomycin injection, a treatment protocol that was
effective in disrupting one-trial fear memories. Similarly, while intracerebral anisomycin
infusions can disrupt one-trial CTA memory reconsolidation, anisomycin is ineffective in
157
destabilizing the reconsolidation of a maximum strength (asymptotic) CTA that had been
conditioned with four training trials (Garcia-de la Torre, et al., 2009).
Increasing CS re-exposure time can replasticize more stable memories. As
discussed above, older and stronger memories are more stable and resist the induction of
reconsolidation upon recall. However, with increased duration of the re-exposure
periods, those more stable memories can be replasticized. A study of the reconsolidation
of single-trial contextual fear conditioning in mice found a re-exposure duration of 1 min
insufficient to initiate reconsolidation (as assessed by anisomycin-induced interruption of
memory recrystallization), while a 3-min re-exposure period induced replasticization of
the fear memory (Suzuki, et al., 2004). In further support of this, they used three-trial
training to condition a stronger fear memory, and then 24 h later subjected the mice to 3,
5, or 10 min re-exposures and systemic anisomycin injection. The more rigorous training
rendered the fear memory more durable, such that 3- and 5-min re-exposures were
insufficient to replasticize the memory, but the 10-min session was capable of inducing
reconsolidation. Extending these findings further, they also conditioned a single-trial
contextual fear memory and then let the memory age for 8 weeks, a procedure that results
in a durable memory. After this aging period, the mice were then given systemic
anisomycin and a 10-min re-exposure session. This longer re-exposure period was
successful in replasticizing the aged, stolid memory whereas in a previous experiment
utilizing a 3-min re-exposure, the same durable memory withstood replasticization.
An important factor complementary to re-exposure time may be the potency of
the treatment to disrupt reconsolidation. Indications are that when stable memories are
158
recalled with a short duration re-exposure to the CS (which normally fails to induce
significant replasticization), the memory can still be disrupted if the reconsolidation-
blocking treatment is intensified. Investigating the reconsolidation of contextual fear
memory, Debiec and colleagues (2002) trained rats with a highly memorable eight-trial
conditioning protocol, and then boarded the animals for 45 d. After the aging period, rats
were given a mere 90 s reactivation session in the conditioning cage, and then were
infused with bilateral intra-dorsal hippocampal doses of 250 µg of anisomycin, a massive
localized dose that gave a greater concentration of anisomycin than any other study cited
in this chapter. This powerful treatment blocked reconsolidation. Yet in another group
conditioned with the same protocol, the 250 µg doses were infused intraventricularly,
which gives a more diffuse application of the drug over many more brain areas (250 µg
per hemisphere, still a more concentrated dose than any other study cited in this chapter);
these animals displayed fear normally, showing that the concentrated localized dose in
the hippocampus was key to disrupting reconsolidation. In an additional experiment,
they trained rats and then aged the memories for 45 d, but instead of a drug treatment
after reactivation of the memory, rats received electrical lesion of the dorsal hippocampus
(obviously a powerfully disruptive intervention). Dorsal hippocampal lesions likewise
disrupted reconsolidation of the fear memory, as these rats subsequently displayed little
fear upon retest, while rats that were lesioned without re-exposure to the CS displayed
fear memory normally at the retest session. Interpreting their findings, the authors
posited that reactivation and replasticization of the fear memory triggered the need for a
small number of proteins (fewer than that needed for the initial consolidation of the new
159
memory) in order to restabilize the synapses of the memory trace; thus a very high dose
of anisomycin (or hippocampal ablation) is needed to prevent even a little new protein
synthesis from occurring. Complementary with this interpretation, longer re-exposure
durations could replasticize older, more stable memories because more of the original
trace and its constituent synapses are reengaged and reactivated, and hence are in need of
a greater number of new proteins to restabilize; the increased demand for protein
synthesis would then be easier to disrupt with a lower dose of anisomycin.
The CS re-exposure duration affects whether reconsolidation or extinction is
initiated. Short re-exposure durations reactivate the original memory and trigger
reconsolidation, whereas long exposure durations initiate the formation of an extinction
memory. Examining different re-exposure durations, Suzuki and others (2004) injected
groups of mice systemically with anisomycin or saline vehicle and then re-exposed them
to a fear-conditioned context for 3 or 30 min. Upon retest, anisomycin-injected mice in
the 3-min group exhibited significantly less freezing, while their saline-injected
counterparts showed little extinction after the short re-exposure and spent a majority of
their retest freezing. Therefore the brief re-exposure was adequate to reactivate and
replasticize the fear memory, rendering it sensitive to anisomycin, but insufficient to
induce extinction memory formation. During the 30-min re-exposure, mice in both
anisomycin and saline groups extinguished greatly and at a comparable rate over the
course of the session, showing that anisomycin itself did not interfere with recall or
contemporaneous extinction processes. Yet at the retest, vehicle-injected mice
maintained their extinction training and froze little, while the anisomycin-treated mice
160
froze at a high level comparable to the beginning of their re-exposure session. Thus the
30-min re-exposure seemed to initiate extinction memory formation whose consolidation
was blocked by anisomycin, and was hence forgotten, while preserving and protecting the
original fear memory from erasure.
Similar findings have also been made by Pedreira and Maldonado (2003) with
fear learning in Chasmagnathus, and by Eisenberg and colleagues (2003) in medaka fish.
The Eisenberg group’s experiments also demonstrated the reconsolidation-extinction
boundary in a different way, by holding the re-exposure treatment constant and varying
the strength of the original memory. In two groups of rats they conditioned either a weak
CTA using single-trial training, which extinguishes easily using their protocol, or a strong
CTA with two-trial training, which does not extinguish easily. Their rationale was that
the re-exposure session would be capable of initiating extinction when the CTA memory
was weak, but that same re-exposure session could only induce recall and reconsolidation
when the CTA memory was strong. They tested this notion by infusing anisomycin into
the insular cortex of both weak and strong memory groups after a re-exposure trial, and
several retest sessions took place over the following days. In the weak CTA group, the
treatment blocked extinction learning, resulting in prolonged avoidance of the saccharin
CS. In the strong CTA group, the treatment disrupted the reconsolidation of the memory,
and so the CTA was rapidly diminished. Their conclusion was that whichever process
was dominant during the re-exposure period was the memory trace that became
susceptible to disruption by anisomycin. When the memory is weak, extinction is easy to
initiate and so extinction learning will be the prevailing process of that trial. When the
161
memory is strong, extinction is difficult to initiate, and so reactivation and expression of
the original memory will be the paramount process. In summary, the reconsolidation-
extinction boundary appears to depend upon the interaction of two relative factors: the
strength of the original memory and the duration of re-exposure to the reminder stimulus.
The learned behavior is still exhibited up to several hours after the amnestic
treatment is given and reexposure to CS occurs. One notable feature of reconsolidated
memories is that, like memories undergoing consolidation, they are independent of
protein synthesis for a period of several hours after recall. Using cued fear conditioning
with an auditory CS, Nader, Schafe, and LeDoux (2000) gave rats a recall trial that was
immediately followed by intra-amygdala anisomycin infusion. Four hours later, these
rats displayed freezing behavior comparable to vehicle-infused controls over 3 brief test
trials; yet 20 hours later, these same anisomycin-infused rats exhibited significantly less
freezing than controls. Similar replications of this effect have demonstrated that
memories are present from 2 to 4 hours after reactivation but not 24 hours or more later
when reconsolidation is disrupted (Debiec, LeDoux & Nader, 2002; Kida, et al., 2002;
Lee, Everitt & Thomas, 2004; Suzuki, et al., 2004; Duvarci, Nader & LeDoux, 2005;
Akirav & Maroun, 2006; Debiec, et al., 2006; Debiec & LeDoux, 2006; Doyere, Debiec,
Monfils, Schafe & LeDoux, 2007). Thus a recalled and replasticized memory appears to
reenter a short-term memory phase that can dissipate unless reconsolidated into a long-
term memory by means of new protein synthesis.
162
Reconsolidation is similar, but not identical to consolidation, requiring in
some instances different neural areas and different biochemical underpinnings.
Consolidation and reconsolidation are quite similar in the molecular requirements they
entail. Both processes involve glutamate receptors, kinases, the transcription factor
CREB, and protein synthesis (Kida, et al., 2002; Dudai & Eisenberg, 2004; Lee et al.,
2004; Suzuki, et al., 2004). However, a full accounting of the similarities and differences
between consolidation and reconsolidation would require a detailed description of
consolidation to be laid out here, a task which is beyond the scope of this chapter.
Instead, some of the points of divergence between the two will be described in order to
illustrate that the two memory retention processes are not identical. The discussion will
focus on three types of learning: inhibitory avoidance, contextual fear conditioning, and
cued fear conditioning.
Several experiments have examined the reconsolidation of inhibitory avoidance
memories, using bilateral infusions of anisomycin in the dorsal hippocampus as the
disrupting treatment (Taubenfeld, et al., 2001; Vianna, et al., 2001; Cammarota, et al.,
2004; Tronel, Milekic & Alberini, 2005). These studies used doses of 80 or 125 µg of
anisomycin in each hippocampus, treatments which are effective in preventing
consolidation (ibid.), and yet none were able to reliably disrupt reconsolidation of the fear
memory. In the present author’s survey of the literature, no studies implementing intra-
hippocampal infusions of anisomycin were successful in disrupting reconsolidation of
inhibitory avoidance memory, although disruption was possible when systemic injections
were used (Taubenfeld, et al., 2001; Milekic & Alberini, 2002; Tronel, et al., 2005).
163
Thus the dorsal hippocampus, which plays a role in the consolidation of inhibitory
avoidance memory, does not require new protein synthesis in order to reconsolidate that
memory.
Consistent with this, Taubenfeld and colleagues (2001) gave rats single-trial
inhibitory avoidance training. By suppressing the expression of the transcription factor
C/EBP β (CCAAT enhancer binding protein β) in the dorsal hippocampus at 5 h after
training, they found that consolidation was disrupted and the rats subsequently failed to
avoid the shock chamber. However, when normal unadulterated training occurred, and
then dorsal hippocampal C/EBP β was suppressed at 5 h after re-exposure to the training
box, rats avoided the shock chamber as well as controls. Thus dorsal hippocampal
C/EBP β is required for consolidation but not reconsolidation of inhibitory avoidance
memory.
Complementarily, when a similarly designed study suppressed C/EBP β
expression in the B nucleus of the amygdala at 5 h after either acquisition or reactivation
of inhibitory avoidance memory, it was seen that C/EBP β downregulation impaired
reconsolidation of the memory, but consolidation was unaffected by the treatment
(Tronel, et al., 2005). Hence there exists a double dissociation in the need for C/EBP β in
the hippocampus and amygdala for consolidation and reconsolidation, respectively, in
inhibitory avoidance. All together then, the acquisition of inhibitory avoidance memory
is dependent upon protein synthesis in the hippocampus, whereas reconsolidation
requires protein synthesis in the B amygdala; these anatomical differences perhaps
suggest that inhibitory avoidance memories undergo system consolidation, being initially
164
acquired in the hippocampus, but ultimately being stored in, retrieved from, and
reconsolidated in the amygdala.
Testing several drugs in the B amygdala of rats after contextual fear conditioning,
Bucherelli and others (2006) determined that antagonism of CB1 cannabinoid receptors,
H3 histamine receptors, muscarinic acetylcholine receptors, and total functional
inactivation by TTX all disrupted consolidation. When the same battery of drugs were
applied in separate groups of rats after a re-exposure session, only TTX and CB1
blockade impaired reconsolidation of the memory, while the histamine and muscarinic
antagonists did not affect retention of the fear memory. All drugs were without effect
when their infusion occurred at a point 4 d after training and 3 d before testing,
demonstrating that their effects were specific to learning and memory, as opposed to
neurotoxicity.
Exploring the role of the hippocampus in contextual fear conditioning in rats, Lee,
Everitt, and Thomas (2004) suppressed expression of either the neurotrophic signaling
molecule BDNF or the immediate early gene zif268 in the dorsal hippocampus during the
consolidation and reconsolidation phases of contextual fear memory stabilization.
Knockdown of hippocampal BDNF impaired consolidation, but not reconsolidation,
whereas silencing of hippocampal zif268 expression thwarted reconsolidation, but not
consolidation. This double dissociation provides clear evidence of qualitative differences
in the cellular processes attendant to consolidation and reconsolidation of contextual fear
memory.
165
In a study investigating reconsolidation of cued fear conditioning, Debiec and
LeDoux (2004) found norepinephrine in the L amygdala to be important for
reconsolidation, but interestingly not for consolidation. Similarly, neither systemic
injection nor intra-L infusion of the norepinephrine β-receptor antagonist propranolol
immediately after acquisition had any adverse effect upon consolidation and retention of
the tone-shock fear memory, with drug-treated rats exhibiting as much fear as saline-
treated controls. Yet when infused immediately after a brief re-exposure to the CS, both
drug treatments significantly diminished the fear expressed in the test session 48 h later.
Furthermore, when the intracranial infusion was made 2 mm dorsal to the L amygdala, no
disruption of reconsolidation occurred, demonstrating the topographical specificity of the
effect they detected. Collectively, these biochemical differences may reflect the
underlying mechanics of the two memory stabilization processes, i.e., the growth and
formation of new synapses (consolidation, requiring trophic factors like BDNF) versus
their maintenance (reconsolidation). In sum, these findings indicate that although
reconsolidation in general seems to utilize a subset of the neurochemical pharmacopeia of
consolidation, there are some components specific to each process.
Reconsolidation can be used to enhance memory too. If reconsolidation is
similar to consolidation, then it should be possible to use it to enhance memory, as can be
done with consolidation. In fact it can, as Pavlovian fear memory was enhanced in rats
by stimulating protein kinase A activity in the BLA amygdala during reconsolidation
(Tronson, Wiseman, Olausson & Taylor, 2006). A single-trial acquisition conditioned a
weak fear memory, followed over the next four days by daily reactivation-infusion trials
166
in which the tone CS was played and then rats were immediately infused in the BLA with
a protein kinase A agonist or saline. Agonist-infused rats showed significant increases in
freezing responses to the CS across the daily reactivation trials compared to saline-
infused rats who slightly (non-significantly) decreased responding across trials. The drug
infusion was not merely acting as a US, since a control group of conditioned rats received
a novel tone CS during their “reactivation” trial, which was followed immediately by
kinase agonist infusion; this group did not increase its freezing in response to the novel
CS after this treatment indicating that the agonist was not capable of acting as a
reinforcer per se. Similarly, Lee, Milton, and Everitt (2006) were able to enhance
freezing to a tone CS in a classical fear conditioning experiment by systemically
administering the NMDA agonist D-cycloserine shortly before re-exposure to the tone.
Two subsequent retests given one week apart revealed a persistent augmented freeze
response compared to vehicle-injected controls who showed a steady diminution in
freezing across the re-exposure and retest trials. Repeating the experiment but utilizing a
weaker shock energy during training duplicated the memory enhancement effect, as did
yet a third replication that entailed infusing D-cycloserine directly into the BLA. Hence
treatments that promote or fortify reconsolidation can be utilized to strengthen memory.
Summary of the main features. Overall, the findings appear to coalesce around
several basic features of recalled memory. Reactivated memories seem to go through a
consolidation-like period during which they are vulnerable to elimination or
intensification through pharmacological intervention. Strength of the memory (as
measured by its ability to be disrupted) is dependent upon two main factors: training
167
intensity and age of the memory. The susceptibility of a recalled memory to
replasticization and disruption is dependent on several factors: the strength of the
memory, the duration of the reactivation session, the potency of the amnestic agent, and
the location of the amnestic agent in the brain. And prolonged re-exposure to a CS can
actually prevent the initiation of replasticization and instead induce the formation of an
extinction memory.
3.2 Ruling out alternative explanations
Being an insurgent new model for memory retention, the concept of
reconsolidation is not without critics who question the verisimilitude of reconsolidation
as a probable explanation for the data that has been generated from the exploration of
memory reactivation. The three major alternative hypotheses are that (1) disrupted
memories are not deleted but are instead unretrievable, (2) the memories are simply being
rapidly extinguished, and (3) that amnestic drug interventions are mortifying or stunning
crucial brain tissues. However, a great number of clever and careful control experiments
have been conducted to address alternative explanations for the reconsolidation effect,
experiments which suggest that reconsolidation is not an artifactual misinterpretation of
the data but is instead a genuine feature of memory.
The amnesia observed after ‘disruption of reconsolidation’ is due to failure
to retrieve an intact memory. This is actually an old criticism of consolidation studies,
which is now being levied against reconsolidation studies. Although not conclusive,
there is evidence to suggest that this alternative hypothesis is incorrect. First of all, drug
168
treatments that disrupt reconsolidation do not impair retrieval during the re-exposure
session. Kida and coworkers (2002) gave systemic injections of anisomycin or vehicle to
mice 30 min before reactivation of a contextual fear memory. The drug treatment had no
effect upon ability to express conditioned freezing relative to control mice. In a similar
experiment in the same study, they temporarily inhibited CREB activity in a specially-
bred strain of mutant mice, and timed the inhibition to coincide with the memory
reactivation session. Compared to control groups of wild type mice and mutant mice
without CREB inhibition, CREB knockdown had no effect on performance during the re-
exposure trial. These data and similar findings in other laboratories (Lee, et al., 2004;
Suzuki, et al., 2004) at least demonstrate that several treatments that disrupt
reconsolidation do not adversely affect contemporaneous retrieval of the target memory.
However, some have argued that giving the re-exposure trial in temporal proximity to a
drug state essentially encodes that drug state as part of the original memory (Riccio,
Moody & Millin, 2002). Thus it is argued that subsequent testing in a non-drug state
would be expected to detect impaired memory expression, since not all of the requisite
retrieval cues would be present to fully activate the memory. This is a form of the STATE
DEPENDENT LEARNING HYPOTHESIS, wherein the internal state of the subject at acquisition
(such as drug intoxication or hypothermia) is encoded as part of the CS, and therefore the
internal state of the subject at testing must match that present during acquisition in order
to elicit optimal memory retrieval. Notwithstanding the fact that integrating an altered
state into a memory upon reactivation seems to demand something like the
reconsolidation process, this alternative hypothesis too is improbable. Lee and comrades
169
(2004) infused BDNF antisense oligodeoxynucleotide into the dorsal hippocampus before
acquisition and again before testing 24 h later, finding that consolidation was averted in
these animals compared to control rats infused with missense oligodeoxynucleotides.
Thus it was demonstrated that at least for consolidation, alignment between training and
testing states could not evoke a latent fear memory, suggesting that in fact the memory
was eliminated. These results can only be extrapolated to the reconsolidation effect,
since it seems that no study exists that infuses drugs during both reactivation and retest.
Still, these findings demonstrate that incorporation of internal state into a reactivated
memory is unlikely, since it should be at least as easy to integrate internal state as part of
the CS during initial memory acquisition, and this did not occur. In sum these findings,
together with the fact that reconsolidation can be used to enhance memory as well as
impair it, suggest that reconsolidation-related memory deficits are not due to failure to
retrieve an intact, reticent memory.
The amnesia is due to extinction, perhaps even a drug-induced accelerated
extinction. In a normal extinction paradigm, the CS is repeatedly administered
(unreinforced by the US) until the conditioned behavior is extinguished, i.e., no longer
evident. Conventional theory holds that extinction training initiates the formation of a
new memory that identifies the CS as not predictive of the US; this memory competes
with and inhibits expression of the original CS-US memory, which remains latent in the
brain (Pavlov, 1927; Rescorla, 2001; Bouton, 2004). Various methods can coax this
latent memory to resurface, particularly renewal, spontaneous recovery, and
reinstatement. In RENEWAL the animal is extinguished in one context, and is later
170
introduced into a new context such as a distinctive training box. When the CS is
presented again in the new context, the previously extinguished behavior reappears,
demonstrating that the original CS-US memory had remained more or less intact and that
the extinction was context-dependent. Apropos of reconsolidation, if anisomycin given
after CS re-exposure serves only to enhance or accelerate extinction, rather than to block
reconsolidation, then anisomycin-treated rats should show renewal when placed into a
novel context and exposed again to the CS. In fact they do not exhibit renewal under
those circumstances (Duvarci & Nader, 2004).
Another extinction-based alternative hypothesis is that the amnesia observed after
blocking reconsolidation is transitory. This alternative is in effect a restatement of
SPONTANEOUS RECOVERY, in which extinction training eliminates behavioral expression of
the conditioned response, then a period of time passes, and then during a retest the
original behavior reappears. When this methodology has been applied to memories
eliminated by obstruction of reconsolidation, the lost memories fail to spontaneously
recover. A 45-d-old contextual fear memory in rats was reactivated and then its
reconsolidation was disrupted by dorsal hippocampotomy (Debiec, et al., 2002). After
recovery, five retests occurred at 1 week intervals, during which sham-lesioned controls
exhibited high levels of fear that extinguished over the test trials while lesioned rats
exhibited consistently low levels of fear, and at no point did they evince spontaneous
recovery of the fear memory. Nor was spontaneous recovery detected in other similarly
designed experiments (Debiec & LeDoux, 2004; Duvarci & Nader, 2004; Lee, et al.,
2004). In one of these experiments (Debiec & LeDoux, 2004), a tone-shock fear memory
171
did not spontaneously recover after systemic propranolol prevented reconsolidation, even
with multiple retest episodes at 1 week and 2 weeks following the initial test session; in
fact both propranolol and control groups extinguished over the course of these retests.
Furthermore, a reinstatement procedure was implemented subsequent to the retest
sessions in an attempt to provoke the expression of any latent memory. In
REINSTATEMENT, extinction training occurs and then an unpaired, unexpected delivery of
the US is given, after which the apparently extinguished behavior reemerges. This
procedure failed to reverse the amnesia in propranolol-treated rats, although control rats
did evince a significant increase in freezing behavior. Reinstatement also could not undo
amnesia for conditioned place preference (an appetitive instrumental task) when
reconsolidation for that memory was blocked (Wang, et al., 2008).
Additionally, blockade of reconsolidation can cause amnesia during recall
conditions that do not promote extinction learning. Both extinction training and the re-
exposure stimulus in reconsolidation experiments entail an unreinforced presentation of
the CS. Yet the reconsolidation hypothesis predicts that a complete second acquisition
trial, where CS and US are again administered, should serve to reactivate the memory and
render it in need of reconsolidation; in this situation, no extinction would occur since the
CS faithfully predicts the US for a second time. In a test of this prediction, an auditory
fear memory in rats was reactivated with a second training trial, followed immediately by
anisomycin infusion into the L amygdala. Although these rats exhibited normal fear
reactions during the CS presentation of the second training trial, they showed
significantly less fear in the test session after the second training trial. Hence anisomycin
172
disrupted reconsolidation of the fear memory that was replasticized by the second
training trial; if enhanced extinction was at work, the second training trial would have
kept the auditory fear memory strong, if not stronger (Duvarci & Nader, 2004; reported
anecdotally in Debiec, et al., 2002).
Finally, as described above, extinction entails new learning and new memory
formation. Data accumulated over decades has all but proven long-term memory
formation of all sorts to rely crucially upon the synthesis of new proteins (Davis &
Squire, 1984). If the memory impairments caused by disruption of reconsolidation were
in fact due to extinction enhanced by the blockade of protein synthesis, that would truly
be a paradigm shifting discovery. As such, this alternative hypothesis flies in the face of
current understanding of the neurobiological substrates of memory. Moreover, extinction
memories themselves seem to be subject to reconsolidation and the amnestic effects of
anisomycin after extinction memory reactivation (Garcia-de la Torre, Rodriguez-Ortiz,
Balderas & Bermudez-Rattoni, 2010). In all, extinction is not likely the ultimate cause of
reconsolidation-related memory deficits.
The effect is due to neurotoxicity or some other non-specific effect of the
reconsolidation-disrupting drug. In a study of object recognition memory, Akirav and
Maroun (2006) made infusions of anisomycin or the NMDA receptor antagonist APV in
the ventromedial prefrontal cortex of rats at 0 or 3 h after the re-exposure session. Should
the disruption of reconsolidation actually be due to the poisonous effects of these drugs
on the object recognition neural circuit, then infusions at both time points should be
equally effective in disrupting performance. Instead, at the test session they observed
173
that recognition by the 0 h group was impaired, while recognition by the 3 h group was
unaffected, demonstrating no lingering neurotoxic effects of the drugs in the latter group
(as well as demonstrating the limited window of opportunity for disrupting
reconsolidation after reactivation). Other studies have similarly found that drug
treatments given in the absence of the re-exposure session are without effect (Nader, et
al., 2000; Debiec, et al., 2002; Debiec & LeDoux, 2004; Gruest, Richer & Hars, 2004;
Lee, et al., 2004; Duvarci, et al., 2005). These findings also rule out the possibility that
the treatments are disrupting a very late phase of consolidation, since if that were the
case, the memory would be disrupted whether reactivation occurred or not. Thus brain
damage or dysfunction is also not a likely explanation for reconsolidation-related
amnesia.
Given all of this, the reconsolidation hypothesis remains the most robust model
that accounts for data demonstrating reactivation-induced memory instability.
3.3 What is reconsolidation and why does it exist?
Reconsolidation did not evolve and remain a conserved feature of multitudinous
animal species just so scientists could inject animals with obscure drugs and erase
memories. Two non-mutually exclusive hypotheses of function have been postulated.
One hypothesis explaining the functional role of reconsolidation is that the initial
consolidation after acquisition of a memory and the reconsolidation that can occur
following this are both components of a grand consolidation that takes a great deal of
time to fully realize (Dudai & Eisenberg, 2004). According to this hypothesis, after the
174
initial learning experience, memories go through repeated periods of reconsolidation to
become more and more stable. These rounds of reconsolidation can be triggered by
recall of the memory (either by re-exposure to a cueing stimulus or by self-initiated
recall) or by reactivation during sleep (Walker & Stickgold, 2004). Sleep (both slow
wave and rapid eye movement sleep) is believed to convert recent memories into stable
ones, and reconsolidation may mediate part or all of that transmutation. Thus the
experiments cited in this chapter may present snapshots of memories at various points
along their way to secure and stable storage, and ultimately reconsolidation “may prove
to be good old consolidation in disguise” (Dudai & Eisenberg, 2004).
The other hypothesis of the utility of reconsolidation is that replasticization of a
memory following its reactivation allows for new information to be integrated into the
existing memory. An early forerunner of this idea has been explored in humans in the
work of Elizabeth Loftus (Loftus & Palmer, 1974; Loftus & Yuille, 1984). She found
that when leading questions are used during an interview, they can change the content
remembered by a subject (Loftus, 1975). Her work had serious implications for the
manner that detectives and prosecuting attorneys interviewed witnesses. A more recent
experiment has specifically evaluated the modification of human episodic memories
using a reconsolidation protocol (Hupbach, et al., 2007). College students were taught a
list of mundane items to remember. At the next session 48 h later, as an indirect way of
reactivating the memory of the list, they were asked to recall the previous training session
without being told specific items from the list, and then were taught a second list of items
to remember. A non-reactivated control group was taught the second list without recall
175
of the first list. At the final session 48 h later all participants were asked to report all the
items from list one that they could remember. Students in both groups recalled the same
number of items from list one, but students in the reactivation group included
significantly more items from list two (INTRUSIONS) than students in the non-reactivation
group. Thus reactivation allowed the replasticization of the memory of the first list and
items from the second list were incorporated into it. Moreover, a second experiment
demonstrated that when the final test trial occurred immediately after training of the
second list, no intrusions occurred right away, meaning that the memory had to
reconsolidate before the new hybrid memory was formed.
A series of important experiments involving rats also utilized the plasticity
associated with recall and reconsolidation to transmogrify a cued fear memory (Monfils,
Cowansage, Klann & LeDoux, 2009). The scientists hypothesized that if a cued fear
memory were reactivated and hence replasticized, and extinction training was
implemented before the memory had time to reconsolidate, then the original fear memory
would be rewritten and replaced by the ‘safe’ memory formed during the CS-only
extinction session. To test this reconsolidation-extinction hypothesis, rats were given
three tone-shock acquisition trials, and then the next day half were re-exposed to a single
presentation of the tone CS in the conditioning apparatus (reconsolidation-extinction
rats), while the other half were merely re-exposed to the conditioning apparatus (standard
extinction rats). One hour following the re-exposure phase, all animals were subjected to
extinction training, consisting of 18 (for reconsolidation-extinction) or 19 (for standard
extinction) CS presentations at 3 min intervals; thus all animals had what was effectively
176
a 19-CS extinction session, with only the time between the first and second CS
presentations differing between the reconsolidation-extinction and standard extinction
groups (1 h vs. 3 min, respectively). The extinction curves of the two groups were highly
similar, and freezing responses to the last four CS presentations were not statistically
different between the groups, meaning that both groups had extinguished equivalently
over the one hour extinction session. Next, to assess whether the memories of both
groups had merely extinguished, or if the reconsolidation-extinction group had in fact
overwritten their original fear memory, the rats were tested for renewal, spontaneous
recovery, and reinstatement. In all cases the reconsolidation-extinction rats failed to
exhibit any recovery of memory, while standard extinction rats showed significant
recovery in all cases. These results indicated that in the reconsolidation-extinction
animals, the cued fear memory had not simply been extinguished, since the standard
repertoire of memory evocation tools could not cajole an extinction-repressed memory
trace to resurface. Rather, the single CS preceding the main extinction session in the
reconsolidation-extinction rats reactivated and replasticized the cued fear memory,
allowing the subsequent main extinction session to overwrite or erase the original fear
memory. Furthering this line of investigation, a SAVINGS or reacquisition experiment was
conducted, in which another acquisition trial (CS+US) was administered after extinction;
in a prototypical savings experiment, the latent original memory would reemerge and
boost the efficacy of the secondary acquisition trial such that a stronger fear memory
would be reacquired compared to a group that did not experience the initial acquisition-
extinction, since they had no savings to build upon (cf. Bouton, 2004). When a savings
177
procedure was implemented after fear memory extinction, the reconsolidation-extinction
rats had a marginal reacquisition of the fear memory, while the standard extinction
control rats had a strong reacquisition, significantly greater than the experimental group.
Moreover, the final experiment in that series (Monfils, et al., 2009) was another savings
procedure, this time with 5 reacquisition trials. Again, the reconsolidation-extinction
group showed a significantly weaker savings effect to the extent that freezing during the
last four CS presentations was significantly lower than that of both the standard
extinction group and a group of naïve rats who had not experienced the initial
acquisition-extinction (both of whom had similar reacquisition/acquisition curves). This
latter experiment implies that the reconsolidation-extinction treatment allowed the “safe
CS” memory learned in extinction to overwrite the “unsafe CS” memory learned in
acquisition, and that the “safe CS” memory actually retarded reacquisition of another fear
memory in a kind of latent inhibition effect. Summed up, these results suggest that when
extinction of a fear memory occurs during the plastic period after reactivation of that
memory, the original memory can be significantly modified.
Hence reconsolidation appears to allow the modification of memory. In fact,
older or more well-trained memories can be more easily replasticized (and then
disrupted) when the reactivation session contains new information, providing an
indication that the incorporation of new information into existing memory traces is one of
the adaptive benefits of reconsolidation (Winters, et al., 2009). Reconsolidation may
therefore lend a certain amount of malleability and flexibility to memory, which would
greatly extend the utility of memory as an adaptive trait.
178
3.4 How is the amygdala involved in reconsolidation?
It is clear from the descriptions of experiments already given in this chapter that
the amygdala plays an important role in reconsolidation. For amygdala-dependent
memories it seems that in general the role of the amygdala in learning and consolidation
holds the same for its role in reconsolidation. This is to say that for a particular memory,
if the amygdala is involved in the consolidation and storage of the memory, then it is also
probably involved in the reconsolidation and re-storage of that memory after reactivation.
As with acquisition and consolidation (see Chapter 1), reconsolidation of
instrumental memories, as indicated by limited experimental research into the topic, is
dependent upon the B nucleus, but not the C nucleus. One study examined the effects of
stress on the reconsolidation of a morphine reinforced CONDITIONED PLACE PREFERENCE, in
which a rat that received morphine intoxication in one arm of a maze preferred to spend
more time in that arm compared to another arm in which it received saline injections
(Wang, et al., 2008). Both behaviorally induced stress (by a 5 min forced swim in ice
water) and systemic injections of the stress hormone corticosterone block the
reconsolidation of the place preference memory when administered immediately after
non-reinforced exposure to the preferred context. Importantly, this effect was replicated
by infusion of the glucocorticoid agonist RU28362 directly into the B nucleus, while C
nucleus infusions were without effect. Moreover, the effect of swim stress was avoided
by pre-natatorial infusion of the glucocorticoid antagonist RU486 in the B but not C
nucleus of the amygdala. Thus, the B nucleus but not the C nucleus participates in the
reconsolidation of an instrumental memory.
179
For Pavlovian fear memories most work has evaluated the participation of the L
nucleus in reconsolidation, with findings indicating that, as with consolidation, the L
nucleus participates in reconsolidation of cued fear memories. For instance, Duvarci and
partners (2005) observed that infusion of the protein synthesis inhibitor cycloyheximide
into the L disrupts reconsolidation of tone-shock fear memory, as does blocking the
ERK-MAPK kinase pathway with U0126. Additionally, infusion of the glucocorticoid
antagonist RU486 into the BLA complex (cannula tips were located throughout the L and
B nuclei) just after acquisition or reactivation of a tone-shock fear memory disrupted
consolidation and reconsolidation, respectively, when tested 24 h later (Jin, et al., 2007).
Even more convincing are results gleaned by Doyere and associates (2007), who
conducted a study examining the link between reconsolidation and electrophysiological
activity in the L nucleus. Rats received tone-shock fear conditioning to two distinct CSs
(tone pips and glissandi), one of which was presented again in a reactivation session
during which U0126 inhibited the ERK-MAPK pathway in the L nucleus. It was found
that this treatment disrupted reconsolidation of the activated tone-shock fear memory,
while the non-activated tone-shock fear memory was unaffected. In addition, the team
recorded field potentials evoked by the auditory CSs in the L nucleus during the
experiment. The L nucleus field potentials to the reactivated CS were found to have
attenuated significantly 24 h after reconsolidation was blocked, whereas the non-
reactivated CS still evoked both fear and a strong field potential. Hence CS-specific
tone-evoked field potentials, which are increased in the L amygdala by Pavlovian fear
180
conditioning, can be diminished by blocking reconsolidation in the L amygdala, and this
diminution corresponds to a reduced fear response to the CS.
If in fact preventing reconsolidation acts to corrupt storage of the memory, as
evidence suggests, then any brain area participating in storage of a memory (a locus of
the engram) should also participate in its reconsolidation. Since the amygdala contains
(at least a substantial portion of) the engram for classically conditioned and
instrumentally conditioned memories, particularly aversive memories, it is likely
involved in the reconsolidation of these forms of fear and aversive learning. But is the
amygdala involved in the reconsolidation of CTA?
3.5 What is known about reconsolidation and CTAs?
Several experiments indicate that conditioned taste aversion, like many other
forms of learning, does undergo reconsolidation upon reactivation. A general
demonstration of the effect was accomplished by Gruest, Richer, and Hars (2004), in
which groups of three-day-old rat pups were conditioned an aversion to peach-flavored
milk; subsequent re-exposure to the milk was followed by systemic anisomycin injections
at intervals from 0 to 24 h later. Groups injected at 0, 15, 30 or 60 min after re-exposure
evinced a significantly attenuated CTA at the test trial, while groups injected at 6 or 24 h
exhibited an intense aversion to the milk. Thus reactivated CTA memories are
replasticized and require a protein synthesis-dependent reconsolidation to stabilize once
again. Other experiments attempting to localize the brain substrates of this effect have
implicated the insular cortex and amygdala, two major constituents of the CTA neural
181
circuit, in the reconsolidation of CTA memory. This was already touched upon above, in
the description of the experiment by Eisenberg and partners (2003) in which they found
that reactivation of either weak or strong CTAs in rats followed by anisomycin infusions
in the IC could disrupt either extinction memory or reconsolidation of the original CTA
memory, respectively.
Bahar, Dorfman, and Dudai (2004) failed to replicate these results when the
amygdala was the target of protein synthesis inhibition, finding no role for the amygdala
in the reconsolidation of CTAs. Utilizing a similar methodology to Eisenberg and others
(2003), rats were given either one- or two-trial acquisition training to instill either a
weaker or stronger CTA memory, respectively, using the instrumental bottle method of
taste aversion conditioning. During the re-exposure session, rats were given both water
and saccharin tubules to drink (a preference test), after which subgroups of them were
infused with anisomycin in either the C or B nucleus of the amygdala. Two subsequent
preference retest trials revealed that blocking protein synthesis in the C nucleus after
reactivation of strong CTA memory had no effect: both anisomycin and vehicle-infused
controls maintained a strong aversion across retests (a C nucleus-infusedweak CTA
group was not tested). Anisomycin also had no effect when infused into the B nucleus
after reactivation of a strong CTA memory. However, blocking protein synthesis in the B
after reactivation of a weak CTA memory resulted in disruption of extinction memory
formation, such that vehicle-infused rats demonstrated some extinction across the retest
trials whereas the anisomycin-infused rats maintained their aversion to the saccharin.
182
Thus while the B nucleus is involved with extinction learning in CTA, neither the B nor
the C appear to be involved in reconsolidation of CTA memories.
Yet contrarily, another study found that infusion of an inhibitor of protein kinase
A into the BLA can interfere with reconsolidation of a CTA memory (Koh & Bernstein,
2003). Separate groups of CTA-trained rats received vehicle or inhibitor infusion, either
15 min before or just after intraoral re-presentation of the saccharin solution CS. Three
subsequent retest trials then assessed the effect of these treatments on latency to reject the
saccharin. During the re-exposure session, the pre-CS infusion of inhibitor was not seen
to have any significant effect upon consumption relative to the post-CS and vehicle
groups (although the low n that was utilized precludes any confident judgment on this
matter). However, during the subsequent retest trials, both inhibitor-infused groups were
seen to extinguish their CTAs faster than controls, such that on the second and third retest
trials, the drug-infused groups exhibited significantly longer latencies to reject the CS
than controls. Since the inhibitors were not infused during the three retest sessions, this
was taken to mean that the treatments weakened the original CTA memory during the re-
exposure session, and this then allowed extinction to proceed at a more rapid rate.
Note that these latter results of Koh and Bernstein (2003) are inconsistent with
those of Bahar, Dorfman, and Dudai (2004). In the biochemical chain of events
mediating learning, protein kinase A is an important link that activates the transcription
factor CREB, which in turn induces gene transcription and thus protein synthesis. Hence
an inhibitor of protein kinase A and anisomycin should have similar intracellular effects,
reducing the synthesis of new proteins, and therefore should have similar amnestic
183
effects, interfering with the reconsolidation of the CTA memory. However, a later study
by Garcia-de la Torre and associates (2009) provided evidence that may resolve this
discrepancy. In one of their experiments, the researchers conditioned a taste aversion to
saccharin in a batch of rats, and then used a second conditioning trial as the re-exposure
session. Just prior to the re-exposure trial, they had infused anisomycin (or vehicle) into
either the IC, C amygdala, or B amygdala. Rats were given just saccharin at the test
session 24 h later, where it was seen that IC and C nucleus infusions partially blocked
reconsolidation such that they maintained the level of saccharin consumption that was
expressed during the second conditioning trial, while the B nucleus-infused and control
rats significantly decreased their drinking, reflecting an increased aversion due to the
second conditioning trial. Given the partial effects detected in the C and IC groups, a
second experiment was run in which the same training/testing protocol was used but rats
received simultaneous infusions in either the IC and C nucleus or the IC and B nucleus.
The test session revealed that simultaneous protein synthesis inhibition in the IC and C
nucleus disrupted reconsolidation of the CTA memory as these animals significantly
increased their drinking during the test session. The group infused in the IC and B
nucleus maintained a consistent drinking level across the second acquisition and test
sessions, similar to the group infused only in the IC in the previous experiment, while the
vehicle-infused controls significantly decreased their drinking at the test session,
reflecting their strengthened aversion. Therefore in order to profoundly disrupt
reconsolidation of a CTA memory, protein synthesis needs to be blocked concurrently in
both the amygdala and insula during memory reactivation.
184
3.6 Summary
Reactivation of a memory through recall propels that memory into a plastic and
malleable state. Reconsolidation is the post-reactivation recrystallization of that memory
back into a stolid long term storage form. It is similar though not identical in form to
consolidation, and like consolidation can be disrupted by protein synthesis inhibition and
other pharmacological interventions. Reconsolidation has been demonstrated in a
multitude of animal species and memory paradigms, pointing to the near universality of
the phenomenon. One of the key features of reconsolidation is its training-intensity- and
age-induced resistance to induction, although this resistance may be overcome with a
prolonged reactivation attempt. Another important quality is that an extensive
reactivation attempt will have the dual effects of preventing the replasticization of the
original memory and initiating formation of a countervailing extinction memory.
Numerous experiments have upheld the verity of reconsolidation, ruling out retrieval
failure, enhanced extinction, and non-specific effects of drug infusion as alternative
explanations for the memory deficits obtained after disruption of reconsolidation.
Ecologically, reconsolidation may be an effector of stable memory inscription in the
brain over the long term, and it may also allow the intercalation of new information into
an existing memory. Collectively, these properties reveal much about reconsolidation, an
important aspect of the most important feature of the nervous system.
185
4
METHODOLOGY
4.0 Overview
This experimental series sought to accomplish several aims. Firstly, given the
findings of Monfils and colleagues (2009) that a purely behavioral technique could
disrupt reconsolidation of an auditory fear memory, we wished to determine if this effect
could be replicated in another learning paradigm, conditioned taste aversion. To this end,
two behavioral experiments were conducted, one using the intraoral conditioning method,
the other using the bottle conditioning method. The reconsolidation-extinction technique
was adapted to CTA and utilized in these two experiments, a procedure that entailed a
brief (5 s) intraoral infusion of the sucrose CS to act as a reminder stimulus that would
recall the CTA memory and hence replasticize it; then an extinction trial was
administered during the reconsolidation window. Control groups in these two
experiments received only the standard extinction procedure without prior memory
reactivation. It was expected that extinction would proceed faster in groups receiving
CTA memory reactivation prior to extinction.
A second aim endeavored to detect nuclei of the amygdala involved in the
reconsolidation of CTA. Previous work has generated conflicting findings regarding the
role of amygdala nuclei in CTA, with one group (Koh & Bernstein, 2003) finding that
inhibition of protein kinase A in the B nucleus can interfere with reconsolidation, while
186
another group (Garcia-de la Torre, et al., 2009) found that protein synthesis inhibition in
the C nucleus (but not the B nucleus) can partially disrupt reconsolidation. Therefore two
histochemical experiments examining c-Fos expression were undertaken in order to
assess the participation of the C, B, and L nuclei of the amygdala in reconsolidation of
intraoral- and bottle-conditioned CTA memory.
Finally, a third aim strove to evaluate the contribution of the nuclei of the
amygdala to the performance of the instrumental and classical components of CTA.
Chapter 1 of this opus laid out a neuropsychological model of the amygdala’s role in
learning related to hedonic stimuli, a model in which the C nucleus mediates classical
conditioning and the B nucleus mediates instrumental learning. Thus the two
histochemical experiments described in the preceding paragraph were also used to assess
the applicability of this model to the classical and instrumental components of CTA. The
model predicts that the C nucleus will be activated by the intraoral conditioning method
while the B nucleus will be activated by the bottle conditioning method.
4.1 Materials and methods
Animals and husbandry. Male Sprague-Dawley rats (N=98) were acquired
from Simonsen Laboratories (Gilroy, CA). They were 60 days old upon arrival, and were
left undisturbed for one week (except for occasional handling) to allow rats to
acclimatize. Rats were housed in pairs in opaque plastic cages with wood chip bedding
material. Wire cage tops bore water bottles and food pellets (Harlan Teklad #8604) that
were available ad libitum. Galvanized steel cage dividers were placed in the cages to
187
allow separation of the cagemates when necessary. The vivarium was temperature (21-
22°C) controlled, and the lights were on a reversed 12:12 h schedule (lights off at 0830
h).
Fistulation. After acclimatization, all rats were implanted with a fistula in the
right cheek. The surgery was essentially as described in Hintiryan, Hayes, and Chambers
(2006). A fistula was prepared by fitting a short (3-4 cm) length of low density
polyethylene tubing (0.86 mm i.d. x 1.52 mm o.d., PE/6, Scientific Commodities Inc.,
Lake Havasu City, AZ) onto the blunt end of an 18 gauge hypodermic needle that had
been cut from its luer base (PrecisionGlide needle, Becton Dickinson and Co., Franklin
Lakes, NJ). The free end of the tubing was heat flared using a surgical cautery blade (No.
22-101, McKesson, Richmond, VA). A washer was slid over the needle-tubing assembly
until it rested against the flared end of the tubing, and then this assembly was sterilized
by immersion in 95% ethanol. Washers were fashioned by cutting rounds from a 0.07 cm
thick Teflon sheet (Small Parts Inc., Roanoke, VA) using a standard paper hole punch;
those rounds were then drilled through their centers using a small drill bit powered by a
Dremel tool such that the holes created could easily thread the polyethylene tubing.
Washers destined for the inner surface of an animal’s cheek were trimmed such that they
were given an oval shape that would be smaller and more comfortable in the animal’s
mouth.
Rats were anesthetized with a ketamine/xylazine mixture (100 mg ketamine + 4
mg xylazine per kg body weight, i.p.). The cheek was shorn, washed, and scrubbed with
betadine, and the rat was laid supine. The beveled end of the needle-tubing-washer
188
assembly was pierced through the inner buccal surface at a position several mm superior
to the second mandibular molar, and the assembly was pulled through the cheek until the
washer on the inside was held flush to the epithelium by the flared plastic knob; to
facilitate the penetration of the tubing through the cheek, the puncture created by the
needle was widened slightly with a #12 scalpel blade (Miltex, York, PA) before pulling
the assembly all the way through the cheek. Once through, another washer was threaded
over the needle and tubing and pressed flush against the outside cheek surface. Then the
tubing was cut so that a stub several mm long remained protruding through the center of
the outside washer; this stub was heat flared, thereby securing the outer washer. After the
patency of the new fistula was tested with a cannula, the animal was injected with 0.02
mL of the analgesic Buprenex (subcutaneous, Reckitt Benckiser Pharmaceuticals Inc.,
Richmond, VA), and was returned to his homecage. Cage dividers kept cagemates
separate for 24 h after surgery. Animals were given three days to recover before testing
commenced.
Behavioral procedures. All experiments entailed taste aversion conditioning
comprised of three general phases: preconditioning, acquisition, and extinction. For the
intraoral conditioning method, preconditioning consisted of a 30 second infusion of tap
water (1.5 mL/min flow rate, Sage Instruments syringe pump model 341B) through a 21
gauge cannula inserted through the fistula while the animal was in the orofacial
observation chamber. The rat was then returned to his homecage and the chamber was
cleaned with 70% ethanol. This was done 2 h after the start of the dark phase under red
light illumination (25 W bulbs filtered by Kodak 1A red/orange filters) with the video
189
camera and video monitor turned on, in order to familiarize the animals with the lights
and sounds of the recording equipment. Preconditioning was implemented for 5 days.
On the sixth day, the animals were weighed at the end of the light cycle to
calculate each rat’s lithium dose for the acquisition procedure. Two hours after the start
of the dark phase, the animals were infused with 10% sucrose for 60 seconds while in the
observation chamber and their behaviors were recorded onto a digital video cassette. The
chamber was a homemade clear Plexiglas cylinder with a diameter of 152 mm and a
height-adjustable lid set to 55 mm; this height kept the rat from rearing during recording
Figure 4.0 The orofacial observation chamber, made of clear Plexiglas, allowed recording of the rats from below.
190
which afforded the camera positioned beneath the chamber a macro shot of the rat’s
mouth and forepaws (Figure 4.0). The rat was then removed and immediately injected
with 0.15M LiCl (20 mL/kg, i.p.) and returned to his homecage. The next day, animals
received another 30 second water infusion in the observation chamber in order to keep
the fistulae clean. The following day the extinction procedure began. The animals were
randomly assigned (with constraints, so that cagemates were in the same group) to either
of two groups, reconsolidation-extinction (REX) or standard extinction (SEX). One hour
after the start of the dark phase, the REX group animals received a 5 second exposure
1
to
sucrose in the observation chamber, and then were returned to their cages; the SEX group
animals were briefly held by the experimenter as though a cannula was going to be
inserted into the fistula, and then they were simply placed into the observation chamber
for 5 seconds. One hour later, the REX group animals were returned to the chamber and
infused with sucrose for 60 seconds. The SEX group animals received sucrose exposure
in one 65 second infusion at 2 h after the dark phase began. After each infusion, the
animal was toweled off to remove any sucrose from its fur and the observation chamber
was cleaned with 70% ethanol. The latter 60 seconds of the SEX group were compared
to the 60 second bout of the REX group for statistical analysis. Aversive behaviors were
tallied and compared for the purposes of this analysis, and these behaviors were gape,
passive drip, chin rub, head shake, forelimb flail, and attempted cannula removal (when a
rat swiped or tugged at the cannula to remove it from its fistula). Extinction trials
1
A 5 s infusion was chosen because this amount is sufficient for an animal to recognize the CS; in fact, a single lick is
sufficient for an animal to recognize an aversive CS and to cease ingesting it (Halpern & Tapper, 1971; Bures, 1998).
On the other hand, 5 s at the rate of 1.5 mL/min infuses a paltry 125 µL, and thus should not by itself induce significant
extinction.
191
continued in this fashion every day for each animal until each animal returned to his
acquisition day level of aversive behavior display; this meant that an animal had to show
equal to or less than n+1 aversive behaviors for two days in a row, where n was the
number of aversive behaviors displayed on acquisition day. These criteria ensured that a
rat achieved stable extinction without overextinction (some rats displayed zero aversive
behaviors on acquisition day, and returning them to zero could have entailed
overextinction, hence the n+1 criterion).
For the bottle conditioning method, preconditioning consisted of an intraoral
infusion as described above at 1 h after the start of the dark phase, followed 1 h later by a
bottle of refrigerated tap water given in the homecage for 1 h (cage dividers separated
cagemates during this period). Preconditioning was carried out for 5 days. On the sixth
day, the animals were weighed at the end of the light cycle to calculate each rat’s lithium
dose; an empty cannula was inserted into each rat’s fistula at this time to keep the fistulae
clean. Two hours after the dark phase, the animals were given bottles of refrigerated
10% sucrose for 1 h in their homecages, during which time cagemates were separated.
The bottles were then removed and rats were immediately injected with 0.15M LiCl (20
mL/kg, i.p.) and returned to their homecages. The next day animals received only a 30 s
water infusion in the observation chamber at 2 h after lights off in order to keep the
fistulae clean. The following day the extinction procedure began. The animals were
randomly assigned to either the REX or SEX groups. One hour after the start of the dark
phase, the REX group animals received a 5 s exposure to sucrose in the observation
chamber, and then were returned to their cages; the SEX group animals were briefly held
192
by the experimenter as though a cannula was going to be inserted into the fistula, and
then they were simply placed into the observation chamber for 5 s. One hour after his
exposure to sucrose and/or the observation chamber, each rat was given a bottle of
refrigerated 10% sucrose for 1 h (with cage dividers in place), and the number of mL
consumed was recorded; the animals’ standard food and water bottles were freely
available during this time. Extinction trials continued in this fashion every day for each
animal until each animal returned to his acquisition day level of sucrose consumption or
greater for two consecutive days.
Histological procedures. To collect c-Fos data, rats were rapidly and deeply
anesthetized with halothane vapor (Halocarbon Laboratories, River Edge, NJ). The
thoracic cavity was opened and the descending aorta and esophagus were blocked with a
hemostat clamp. Cardiac perfusion by 225 mL normal saline was followed by perfusion
with 225 mL cold phosphate buffered paraformaldehyde (4%). The brain was extracted
and steeped further in the paraformaldehyde solution for 24 h at 4°C, and then switched
into 30% sucrose in phosphate buffered saline for at least 48 h at 4°C. Fixed brains were
set in a pool of mounting medium on a slicing chuck and frozen in pulverized dry ice.
Coronal sections (35 µm thickness) were alternately collected into two collection wells
filled with phosphate buffered saline. One series of brain sections was placed into vials
with cryoprotectant (30% ethylene glycol, 20% glycerol, 50% 0.05M phosphate buffer)
and stored at -30°C; the other series was processed immunohistochemically, with all
chemical immersions occurring on a shaker. These sections were transferred to a solution
of 3% hydrogen peroxide in absolute methanol for 20 minutes to inactivate the
193
endogenous peroxidase. Then they were rinsed in phosphate buffered saline 3 times (10
min each) and treated with 3% goat serum (Vector Laboratories, Inc., Burlingame, CA),
0.2% Triton (Sigma, St Louis, MO), and 0.7% porcine gelatin (300 bloom) in phosphate
buffered saline for one hour. Without rinsing, the sections were transferred to 1% goat
serum and 0.4% triton phosphate buffered saline with 1:10,000 primary antibody (c-
Fos[4]: sc-52, Santa Cruz Biotechnology, Santa Cruz, CA) for 48 hours at 4 °C. After this
incubation, the slices were rinsed 6 times with phosphate buffered saline (10 min each)
and were transferred to a secondary antibody solution, which was a biotinylated anti-
rabbit IgG (1:200 dilution in phosphate buffered saline; Vector Laboratories, Inc.,
Burlingame, CA) for one hour at room temperature. After they were rinsed 3 times (10
min each), the sections were placed in avidin-biotin complex reagent solution (ABC Elite
kit, Vector Laboratories, Inc., Burlingame, CA) for one hour at room temperature. Then
the sections were rinsed 3 times (10 min each). A peroxidase substrate kit was used to
visualize the antigen-antibody reaction sites in the brain. The sections were placed in a
solution made by 3,3'-diaminobenzidine and nickel enhancer (DAB kit; Vector
Laboratories, Burlingame, CA) for 5-7 minutes and then rinsed 5 times with phosphate
buffered saline (1-2 minutes each) to stop the reaction. The sections were mounted onto
glass slides, air dried, and covered with Cytoseal and cover glass.
The brain sections were examined with a compound light microscope (Olympus
BH-2) and for each rat a right and left hemisphere section of the amygdala was selected
from the level of approximately -3.00 mm posterior to Bregma (Figure 1.0 depicts the
amygdala at this level). Right and left could be distinguished because prior to slicing, the
194
left dorsal cortex of each brain was scored with a razor blade. Each of the selected
amygdalae was photographed through the 10x objective lens with a Q Capture 12 bit
color video camera; 9-12 marginally overlapping photographs of each amygdala were
shot, collectively encompassing all of the C, L, and B nuclei. These photograph series
were imported into Adobe Photoshop CS5 which automatically collated them into single
large high resolution images. The experimenter then examined each image, compared
them with the brain atlas of Paxinos and Watson (1998), and used the program tools to
manually trace out the C, L, and B nuclei and their constituent subnuclei (dorsal lateral,
ventral lateral, and medial subnuclei of the L; capsular and lateral subnuclei of the C;
magnocellular and parvocellular subnuclei of the B). For each subnucleus, the number of
opaque black puncta were counted and recorded.
Statistical analyses. All data were subjected to the percentile bootstrap test for
20% trimmed means, which is a modern robust statistical test equivalent to the t-test. It
has been found to perform well among a range of methods when working with real data
and with small sample sizes (Wu, 2002; Wilcox, 2003a,b). The bootstrap has
independent and dependent variants for assessing between- and within-group differences,
both of which were used. For all tests α = 0.05. The percentile bootstrap test has
previously been used to analyze CTA data (Hintiryan, et al., 2005, 2006, 2009). All
figures present data points depicting 20% trimmed means plus 20% Windsorized
standard errors, descriptive statistics that reduce the effect of outliers and provide a
cleaner view of the central tendency of the data.
195
4.2 Experimental designs
Experiment 1. This experiment tested the ability of the reconsolidation-
extinction treatment to disrupt reconsolidation and thereby facilitate extinction of a CTA
using the intraoral conditioning method. Rats received preconditioning, acquisition, and
extinction as described in Behavioral procedures (REX, n=8; SEX, n=9; see Figure 4.1a).
After each rat had reached his extinction criterion, he began a second phase of
conditioning that was carried out exactly as the first, with preconditioning, acquisition,
and extinction, except that in the second phase all rats received a standard extinction
treatment. This allowed assessment of spontaneous recovery and savings.
Experiment 2. This experiment was similar to Experiment 1, with the primary
difference being the use of the bottle method of CTA. Rats received preconditioning,
acquisition, and extinction as described (REX, n=10; SEX, n=9; see Figure 4.1a). Note
that the reconsolidation-extinction treatment still entailed an intraoral infusion of sucrose
to reactivate the CTA memory. After each rat had reached his extinction criterion, he
began a second phase of conditioning that was carried out similarly to the first, with
preconditioning, acquisition, and extinction, except that in the second phase no intraoral
infusions of water were given (bottles only), and all rats received a standard extinction
treatment. This allowed for the assessment of spontaneous recovery and savings.
Experiment 3. This experiment examined c-Fos expression patterns induced in the
amygdala by the reconsolidation-extinction treatment used in Experiment 1. Rats
received preconditioning, acquisition, and extinction as conducted in Experiment 1.
Aside from the REX and SEX groups, a third group was included that received sucrose
196
only (SUC), i.e., their behavioral treatment was exactly like that of the REX group,
except that on acquisition day SUC rats were injected with equimolar NaCl instead of
LiCl (REX, n=11; SEX, n=10; SUC, n=10). On extinction day 1, one hour after his
extinction treatment, each rat was terminated as described in Histological procedures (see
Figure 4.1b).
Figure 4.1 Schematics of the behavioral procedures used in the experiments; see text for details. (A) Experiments 1
and 2 used this design. During preconditioning, rats received a daily 30 s intraoral infusion of water. Acquisition
entailed delivery of sucrose (CS) intraorally for 60 s (Exp. 1) or by bottle for 1 h (Exp. 2) followed by LiCl injection
(US). Daily extinction trials began 48 h later. For Experiment 1, standard extinction (SEX) rats received CS in a 65 s
intraoral infusion while reconsolidation-extinction (REX) rats received a 5 s infusion and then 1 h later a 60 s infusion.
For Experiment 2, SEX rats were given a bottle of CS for 1 h while REX rats received a 5 s intraoral infusion of CS
and then 1 h later a bottle of CS for 1 h. Extinction trials continued until each rat returned to his baseline, from whence
the rat would begin a second phase of conditioning. Preconditioning and acquisition were similar to the first phase, but
the extinction trial for both groups used the SEX method, which allowed assessment of spontaneous recovery and
savings. (B) Experiments 3 and 4 used this design. The REX, SEX, and SUC groups were treated exactly as in the
previous experiments, with Experiment 3 utilizing the intraoral method and Experiment 4 using the bottle method.
Animals were terminated after the first extinction trial (arrow) for c-Fos analysis.
A
B
H
2
O H
2
O H
2
O
CS+US
H
2
O H
2
O
CS
CS
CS
CS
CS
CS
SEX
H
2
O H
2
O H
2
O H
2
O H
2
O CSCS
REX CS+US
CSCS CSCS
1
6 2 3 4 5 7 8 9
days
H
2
O H
2
O H
2
O
CS+US
H
2
O H
2
O
CS
CS
H
2
O H
2
O H
2
O
CS+US
H
2
O H
2
O
CS
CS
H
2
O
H
2
O
10
H
2
O
H
2
O
Preconditioning Acquisition Extinction Preconditioning Acquisition Extinction
H
2
O H
2
O H
2
O H
2
O H
2
O CSCS
REX CS+US
H
2
O H
2
O H
2
O
CS+US
H
2
O H
2
O
CS
CS
SEX
H
2
O H
2
O H
2
O H
2
O H
2
O CSCS
SUC
CS
1 6 2 3 4 5 7 8
days
H
2
O
H
2
O
H
2
O
Preconditioning Acquisition Extinction
CTA 1 CTA 2
197
Experiment 4. This experiment examined c-Fos expression patterns induced in
the amygdala by the reconsolidation-extinction treatment used in Experiment 2. Rats
received preconditioning, acquisition, and extinction as conducted in Experiment 2.
Aside from the REX and SEX groups, a third group was included that received sucrose
only (SUC), i.e., their behavioral treatment was exactly like that of the REX group,
except that on acquisition day SUC rats were injected with equimolar NaCl instead of
LiCl (REX, n=10; SEX, n=11; SUC, n=10). On extinction day 1, immediately after his
extinction treatment, each rat was terminated as described (see Figure 4.1b).
198
5
RESULTS
5.0 Experiment 1
Experiment 1 entailed conditioning a CTA using the intraoral method, followed
by daily extinction trials wherein the standard extinction (SEX) group received a single
infusion of sucrose over 65 s, while the reconsolidation-extinction (REX) group received
a 5 s reminder infusion followed one hour later by a 60 s infusion, with aversive orofacial
behaviors serving as the dependent variable. The REX treatment was seen to accelerate
extinction (figure 5.0a), such that by extinction day 4 the REX group was expressing
significantly fewer aversive behaviors during sucrose infusion (P = 0.027). Only four
days of extinction could be compared because at that point, rats began to reach their
extinction criteria, and thus were started (individually) on the second CTA regimen.
Comparison of the aversive displays made by each rat on his last extinction day
with those made five days later on acquisition day of the second CTA allowed assessment
of spontaneous recovery (figure 5.0b). Unfortunately some of the subjects’ fistulae
detached, resulting in attrition as the experiment progressed into the second CTA (REX,
n = 6; SEX, n = 5). Within-group tests showed that both REX and SEX groups exhibited
spontaneous recovery (P = 0.011 and P = 0.001, respectively), increasing their display of
aversive behaviors across tests. However, between-group tests revealed that the SEX
group demonstrated significantly more recovery than the REX group (P = 0.014). The
199
second acquisition procedure also allowed assessment for a savings effect, since the
strength of reacquisition could be assayed. As seen in figure 5.0c, both groups reacquired
aversions of similar strength (P = 0.647). Additionally, within-group comparisons across
extinction day 1 of the first CTA and extinction day 1 of the second CTA also revealed
no differences in aversion strength (REX, P = 0.484, SEX, P = 0.974). These findings
indicate no savings effect occurred for either group.
5.1 Experiment 2
Experiment 2 comprised conditioning a CTA using the bottle method, followed
by daily extinction trials such that the REX group received a 5 s intraoral infusion of
sucrose followed one hour later by presentation of a bottle of sucrose for one hour, while
the SEX group simply received the bottle of sucrose. All subjects participated in at least
six days of extinction testing (figure 5.1a), and statistical tests revealed no differences
between groups on any day (P > 0.1 for acquisition and all six extinction days). The
REX treatment did not seem to significantly hasten extinction.
Within-group statistical comparison of the volume drunk on the last extinction
day with that consumed on acquisition day of the second CTA revealed no significant
demonstration of spontaneous recovery for either group (REX, P = 0.297; SEX, P =
0.281; figure 5.1b), and no difference was detected between the groups on acquisition day
of the second CTA (P = 0.270). Moreover, no savings effects were detected for either
group (figure 5.1c). The strength of the second CTA on extinction day 1 was not
different between groups (P = 0.487). Additionally, within-groups tests detected no
200
difference in aversion strength between the first and second CTAs for either group (REX,
P = 0.768; SEX, P = 0.496).
5.2 Experiment 3
Experiment 3 was carried out using the same intraoral method as in Experiment 1,
except that a third group was included that received sucrose only (SUC, i.e., they had no
LiCl during acquisition and thus no CTA) and all animals were terminated after the first
extinction test in order to collect c-Fos data from the amygdala. The behavioral data
(figure 5.2a) mirrored that seen in Experiment 1 (figure 5.0). All groups expressed
similarly low displays of aversive behavior during the acquisition test (P > 0.1 for all
pairwise comparisons). During the extinction trial, the REX and SEX groups performed
significantly more aversive behaviors than the SUC group (P < 0.001 for both
comparisons), indicating that the former two groups acquired an aversion whilst the latter
group did not. The REX and SEX groups were not significantly different from each other
(P = 0.052).
The state of c-Fos expression an hour after the extinction trial was measured in
the subnuclei of the L, C, and B nuclei of the amygdala (figures 5.2b and 5.3). The
groups had differential responses in two subnuclei of the L nucleus. For the dorsal lateral
subnucleus, the REX group possessed significantly more Fos-positive neurons than the
SEX (P = 0.029) and SUC (P = 0.015) groups, while the latter two groups did not differ
from one another (P = 0.383). Additionally, the REX group possessed significantly more
c-Fos-positive neurons than the SEX group in the medial subnucleus (P = 0.046). The
201
ventral lateral subnucleus did not evince any significant differences in c-Fos expression
between groups.
One of the subnuclei of the C nucleus likewise reacted differently amongst the
three groups. The capsular subnucleus showed a marked variance between groups, with
the SUC group bearing significantly more Fos-positive neurons than both the REX (P =
0.011) and SEX (P = 0.007) groups; the REX and SEX groups did not prove to differ (P
= 0.268). The lateral subnucleus on the other hand did not appear to exhibit any
differences between groups (P > 0.1 for all).
Finally, one subnucleus of the B nucleus was also found to respond with partiality
among the three groups. For the large magnocellular subnucleus, there were no
significant differences between groups (P > 0.1 for all). However, for the parvocellular
subnucleus, the REX group had significantly more c-Fos activation than either the SEX
(P = 0.013) or SUC (P = 0.032) groups; the latter two groups did not differ significantly
(P = 0.386).
5.3 Experiment 4
Experiment 4 was executed similarly to Experiment 2, using the bottle
conditioning method of CTA, except a third group was included that was not given LiCl
on acquisition day and thus was a sucrose only (SUC) group; additionally, all subjects
were terminated one hour after the first extinction treatment in order to collect data on c-
Fos expression patterns induced in the amygdala by the behavioral treatments. The
behavioral data (figure 5.4a) indicated that the rats consumed the sucrose equivalently on
202
acquisition day, and that the REX and SEX groups developed a CTA after the
conditioning procedure: the groups were not significantly different on acquisition day (P
> 0.05 for all comparisons), while during the extinction test the REX and SEX groups
drank significantly less than the SUC group (P < 0.001 for both) whereas the REX and
SEX groups did not differ from one another (P = 0.297).
The behavioral treatments were also seen to have an effect upon c-Fos activation
in some of the subnuclei of the amygdala (figures 5.4b and 5.5). The L nucleus, though,
did not respond to the various treatments. The dorsal lateral, ventral lateral, and medial
subnuclei were not seen to have significantly different c-Fos expression patterns among
groups (P > 0.3 for all comparisons). The C nucleus, however, did exhibit varying
expression patterns among the groups. In the capsular subnucleus, the SUC group
contained significantly more c-Fos-positive neurons than either the REX (P = 0.001) or
SEX (P < 0.001) groups, while the latter two groups did not differ (P = 0.354). A similar
pattern of activation was observed in the lateral subnucleus, with the SUC group
expressing significantly more c-Fos than the REX (P = 0.001) and SEX (P < 0.001)
groups, while the REX and SEX groups did not significantly differ (P = 0.354).
The B nucleus was also seen to have disparate c-Fos expression patterns amongst
the groups (figure 5.4b and 5.5). Analysis of the magnocellular subnucleus revealed that
the REX group exhibited the least number of c-Fos-positive neurons, possessing
significantly fewer than the SEX group (P = 0.012), but not significantly fewer than the
SUC group (P = 0.134). The SEX and SUC groups did not differ (P = 0.266). A
disparity was also seen in c-Fos expression in the parvocellular subnucleus. Again, the
203
REX group bore the least number of c-Fos-positive neurons, containing significantly
fewer neurons than the SUC group (P = 0.046), but not significantly fewer than the SEX
group (P = 0.126). Neither did the SEX and SUC groups prove significantly different (P
= 0.368) in this subnucleus. Given that the REX group had the lowest means for c-Fos
expression in both the magnocellular and parvocellular subnuclei, and yet a checkered
pattern of significance was obtained, a whole nucleus analysis was conducted. For each
animal, the magnocellular and parvocellular values were summed to generate a B nucleus
c-Fos expression value (figure 5.4b inset). Statistical analysis of this data revealed that
the REX group possessed significantly fewer c-Fos-positive neurons than the SEX group
(P = 0.009) in the B nucleus, although the REX group did not have significantly fewer
neurons than the SUC group (P = 0.088). The SEX and SUC groups did not differ either
(P = 0.324).
s
50
REX
SEX
A
Aversive behaviors
10
20
30
40
SEX
*
ACQ E1 E2 E3 E4
0
behaviors
8
12
REX
SEX
*
†
†
B
Aversive b
Last EXT test Acquisition 2
0
4
†
C
50
REX C
versive behaviors
10
20
30
40
REX
SEX
Figure 5.0 The results of Experiment 1. (A) The acquisition (ACQ) and extinction (E) curves of
the reconsolidation-extinction (REX) and standard extinction (SEX) groups. The REX treatment
accelerated extinction such that by extinction day 4 the REX group displayed significantly fewer
aversive behaviors than the SEX group (B) Spontaneous recovery occurred in both groups as both
Av
CTA 1, E1 CTA 2, E1
0
10
E1 for CTA1 E1 for CTA 2
aversive behaviors than the SEX group. (B) Spontaneous recovery occurred in both groups, as both
significantly increased their aversive displays from the last extinction test of the first conditioned
taste aversion (CTA 1) to the acquisition of CTA 2 five days later, but the SEX group recovered
significantly more than the REX group. (C) To assess savings effects, CTA strength was evaluated
by comparing aversive displays on the first extinction trials of CTA 1 and CTA 2. Within-group
and between-group tests revealed that all CTAs were of equivalent strength. Thus no savings
effects were detected. All data points represent 20% trimmed means with 20% Windsorized
standarderror. *and†,P<0.05.
204
15
SEX
REX
A
mL
5
10
REX
ACQ E1 E2 E3 E4 E5 E6
0
mL
15
20
25
REX
SEX
B
m
Last EXT test Acquisition 2
0
5
10
C
5
REX
mL
1
2
3
4
SEX
Figure 5.1 The results of Experiment 2. (A) The acquisition (ACQ) and extinction (E) curves
of the reconsolidation-extinction (REX) and standard extinction (SEX) groups. The REX
treatment did not significantly affect extinction over six days of extinction tests. (B)
Spontaneous recovery did not occur in either group, as both essentially maintained their level of
sucrose consumption from the last extinction test of the first conditioned taste aversion (CTA 1)
CTA 1, E1 CTA 2, E1
0
E1 for CTA1 E1 for CTA 2
sucrose consumption from the last extinction test of the first conditioned taste aversion (CTA 1)
to the acquisition of CTA 2 five days later. (C) To assess savings effects, CTA strength was
evaluated by comparing sucrose consumption on the first extinction trials of CTA 1 and CTA 2.
Within-group and between-group tests revealed that all CTAs were of equivalent strength. Thus
no savings effects were detected. All data points represent 20% trimmed means with 20%
Windsorizedstandarderror.
205
60
REX
SEX
*
A
ersive behaviors
20
40
SEX
SUC
Ave
Acquisition Extinction
0
B
lat dl
I
st
c ‐p
C
pc
mc
cap m
vl
urons
30
40
REX
SEX
SUC
Fos-positive neu
10
20
30
SUC
*
*
*
*
Figure 5.2 Behavioral and histological results of Experiment 3. (A) The reconsolidation-extinction
(REX), standard extinction (SEX), and sucrose-only (SUC) groups exhibited similar numbers of
aversive orofacial behaviors on acquisition day. During the extinction test, the REX and SEX groups
exhibited a clear aversion to the sucrose, whereas the SUC group did not since it did not receive
lithiumchloride on acquisition day (B)Photomicrograph depicts the boundaries of the subnuclei of the
c-F
dl vl m cap lat mc pc
0
lithiumchloride on acquisition day. (B)Photomicrograph depicts the boundaries of the subnuclei of the
lateral (dorsal lateral [dl], ventral lateral [vl], medial [m]), central (capsular [cap], lateral [lat]), and
basal (magnocellular [mc], parvocellular [pc]) nuclei of the amygdala. I, optic nerve; st, stria
terminalis;c-p, caudate-putamen. (C) Expressionofc-Fosinsubnucleiof the amygdala varied between
groups. The REX treatment significantly increased Fos in several subnuclei of the lateral (dl and m)
and basal (pc) nuclei, and the SUC group possessed more Fos in the capsular subnucleus of the central
nucleus.*,P<0.05.Datapointsrepresent20%trimmed means with20%Windsorizedstandarderrors.
206
REX SEX SUC
L
c ‐pc ‐p c ‐p
L
st
C
II
I
st
st
st
B
Figure 5.3 Histological data from Experiment 3, which entailed the intraoral
conditioning method.Photomicrographs depictthelateral(L), central(C), and basal(B)
nuclei of the amygdala. In each picture the nucleus of interest is traced while the other
nuclei are shaded. Dark puncta are c-Fos-positive, resulting in part from the treatments
administered to the reconsolidation-extinction (REX), standard extinction (SEX), and
sucrose-only(SUC)groups.SeeFigure5.2bforotherabbreviations.
207
20
REX
*
A
mL
10
15
SEX
SUC
Acquisition Extinction
0
5
20
*
B
ositive neurons
10
15
20
REX
SEX
SUC
B nucleus
Neurons
0
5
10
15
20
*
*
*
*
c-Fos-po
dl vl m cap lat mc pc
0
5
*
Figure 5.4 The behavioral and histological results of Experiment 4. (A) On acquisition day all
groups drank equivalent volumes of sucrose. During the extinction test, the reconsolidation-
extinction (REX) and standard extinction (SEX) groups displayed a strong aversion to the sucrose,
whereas the sucrose-only (SUC) group did not because it received no lithium chloride on
acquisition day. (B) The histological results revealed that the SUC group expressed significantly
more c-Fos than the other groups in both subnuclei of the C nucleus (cap and lat). In the
ll l bl f hbl (B) l h REX idiifi l f magnocellular subnucleus of the basal (B) nucleus, the REX group contained significantly fewer c-
Fos positive neurons than the SEX group, while in the parvocellular subnucleus the REX group had
fewer neurons than the SUC group. The inset presents the c-Fos activity of the whole B nucleus,
and reveals that the REX group had significantly less c-Fos activation than the SEX group, but not
the SUC group. All data points represent 20% trimmed means with 20% Windsorized standard
error.*,P<0.05.
208
REX SEX SUC
L
c ‐p c ‐pc ‐p
I
I
st
st
st
C
I
I
B
Figure 5.5 Histological data from Experiment 4, which entailed the bottle conditioning
method. Photomicrographs depict the lateral (L), central (C), and basal (B) nuclei of the
amygdala. In each picture the nucleus of interest is traced while the other nuclei are
shaded. Dark puncta are c-Fos-positive, resulting in part from the treatments administered
to the reconsolidation-extinction (REX), standard extinction (SEX), and sucrose-only
(SUC)groups.SeeFigure5.2bforotherabbreviations.
209
210
6
DISCUSSION
6.0 Summary of results
The results of Experiment 1 demonstrate that a behavioral procedure designed to
interfere with reconsolidation can elicit a fleet extinction of intraorally-conditioned taste
aversion. Rats given a brief reminder taste of sucrose one hour before each extinction
session demonstrated a significantly reduced aversion by extinction day 4 compared to
rats administered only the standard extinction session. Moreover, spontaneous recovery
of the aversion after extinction, which occurred in both groups, was significantly
attenuated in the group receiving the reconsolidation-extinction treatment. Yet a second
acquisition trial using the same conditioning parameters as the first yielded aversions of
equal strength in the two groups, and their second aversions were of equal strength to the
first, which indicates that savings did not occur in either group. Thus the reconsolidation-
extinction treatment exerted a rather weak effect on the CTA memory compared to the
findings of Monfils and colleagues (2009; see Chapter 3) of cued fear conditioning, but
their results were replicated here and extended to the CTA paradigm.
The immunohistochemical findings of Experiment 3 revealed several associations
between subnuclei of the amygdala and the intraoral treatments the groups received.
Relative to the SEX and SUC groups, the REX group had a significantly increased
number of neurons expressing c-Fos in the dorsal lateral subnucleus of the L nucleus and
211
the parvocellular subnucleus of the B nucleus, as well as a significant increase over just
the SEX group in the medial subnucleus of the L. In the capsular subnucleus of the C
nucleus, the mere exposure to highly palatable sucrose that the non-conditioned SUC
group received caused a significant increase in c-Fos expression compared to the
aversively conditioned REX and SEX groups.
The outcome of Experiment 2 was incongruent with that of Experiment 1.
Animals in the REX group received an intraoral reminder infusion of sucrose one hour
prior to bottle presentation in each daily extinction trial, a treatment that exerted no
significant influence upon extinction compared to the SEX control group. Neither group
demonstrated spontaneous recovery or savings. Hence the reconsolidation-extinction
treatment was ineffective in Experiment 2.
The denouement of Experiment 4 revealed that the REX protocol was again found
to influence c-Fos expression, in this case exerting a suppressive effect upon c-Fos
regulation in the B nucleus. In the magnocellular subnucleus, the REX group had
significantly fewer c-Fos-positive neurons than the SEX group, but not significantly
fewer than the SUC group. In the parvocellular subnucleus, the REX group held fewer c-
Fos-positive neurons than the SUC group, but not significantly fewer than the SEX
group. In the C nucleus, the SUC group bore more c-Fos than either the REX or SEX
groups in both the capsular and lateral subnuclei. The L nucleus expressed no differences
in c-Fos across groups.
212
6.1 Interpretation of the behavioral results
The primary aim of this experimental series was to investigate the
reconsolidation-extinction effect in the CTA paradigm. The reconsolidation-induced
memory modification achieved by Monfils and others (2009) was duplicated here, but the
present results demonstrated a weaker effect than that obtained previously. Those potent
prior effects upon classical fear conditioning included a complete blockade of
spontaneous recovery and a powerful inhibition of reacquisition, as well as a prevention
of renewal and reinstatement. Our application of the reconsolidation-extinction
technique to intraorally-conditioned taste aversion yielded a more rapid attenuation of
aversive responses, as well as a partial disruption of spontaneous recovery, but no
inhibition of reacquisition (renewal and reinstatement were not tested). The reason for
this discrepancy may lie in the fact that in the previous study extinction could be carried
out to completion in a single session that wholly fit within the reconsolidation window,
whereas in CTA, extinction is necessarily a long affair that cannot be conducted in a short
period of time since it presumably deals with evaluation of post-ingestive consequences
(Berman, et al., 2003). Ergo a rapid and complete overwriting of the original CTA
memory could not be accomplished, and instead the memory was reactivated and
amortized a little bit each day, resulting in a less powerful effect. Such a moderate effect
is also similar to the pharmacological disruptions of CTA reconsolidation achieved by
Koh and Bernstein (2003), wherein a more rapid extinction was seen in rats infused with
a protein kinase A inhibitor in the B nucleus, and by Garcia-de la Torre and colleagues
213
(2009), who observed a partial disruption of CTA reconsolidation due to protein
synthesis inhibition in the C nucleus.
One particularly important implication of this work and others like it is that
memories can be erased. The conventional view of extinction is that extinguished
memories are not unlearned or eliminated but are rather suppressed by a newly learned
inhibitory extinction memory (Pavlov, 1927; Bouton, 2004), and that the “extinguished”
memories are actually preserved in nigh undiminished capacity (Rescorla, 2001).
However, evidence contrary to this dogma is beginning to accumulate. For instance pre-
weanling rats, conditioned to fear an auditory CS and then extinguished, failed to exhibit
reinstatement after a reminder shock or renewal after a context change; these young rats
showed no differences in initial or final fear levels or different rates of extinction
compared to post-weanling rats that were merely a week older, yet the older rats
expressed reinstatement and renewal (Kim & Richardson, 2007a, 2007b, 2010). Further
evidence comes from studies demonstrating that adult rats that were extinguished within
an hour of acquisition exhibited little to no spontaneous recovery, renewal, or
reinstatement compared to rats extinguished 24 or 72 h after training (Myers, Ressler &
Davis, 2006; Quirk, et al., 2010). Adding to this growing body of work, the present
results, along with those of Monfils and colleagues (2009) and a similar study in humans
by Schiller and colleagues (2010), indicate that the reconsolidation-extinction treatment
too actually eliminates the original aversive memory due to the fact that the repertoire of
techniques normally implemented to exhume an extinguished memory fail to recover any
latent trace (or only a weakened trace, in the case of the present results) after
214
reconsolidation-extinction. The fact that the reconsolidation-extinction technique is a
simple procedure that could conceivably occur in a naturalistic setting (rather than a
wholly artificial scientific machination such as intracranial drug infusion) suggests that
perhaps some occasions of extinction do in fact entail true forgetting.
The reconsolidation-extinction paradigm could have clinical value as a technique
used in a behavioral therapeutic approach to treat psychopathological disorders of
aversive conditioning (Schiller, et al., 2010). Imagine an American infantryman
returning from a combat tour in Iraq and suffering from posttraumatic stress disorder as a
result of witnessing an improvised explosive device attack. Stateside, he exhibits a
conditioned fear response every time he sees a trash heap alongside the road, a stimulus
that could conceal a bomb in Iraq and one that is a non-threatening sight in America. His
debilitating fear conditioning could potentially be cured with a therapy akin to the
reconsolidation-extinction treatment demonstrated by Monfils and associates (2009) or
Schiller and coworkers (2010), suffering naught from spontaneous recovery, renewal, or
reinstatement. Yet such a treatment bears the possible side effect that the relearning of a
similar fear response could be inhibited. If this soldier were then returned to Iraq for a
second tour of duty, his failure to relearn what in a combat zone is an appropriate and
adaptive fear response could prove extraordinarily dangerous. Our results suggest a more
moderate approach may avoid such a problem, using a reconsolidation-extinction therapy
that utilizes incomplete extinction so that fear conditioning is abated and spontaneous
recovery in a civilian setting is attenuated, and yet facile relearning of a similar fear
response is permitted if necessary.
215
The argument could be raised that the reconsolidation-extinction effect detected in
Experiment 1 was entirely due to enhancement of normal extinction without any
influence on the reconsolidation of the original memory. Indeed, pharmacological
treatment with the NMDA agonist D-cycloserine concurrent with extinction training can
enhance the efficacy of fear extinction without preventing renewal, showing that such an
effect is possible (Walker, Ressler, Lu & Davis, 2002; Ledgerwood, Richardson &
Cranney, 2005; Woods & Bouton, 2006). Yet rebutting this alternative argument is the
fact that both the REX and SEX groups received the same amount of sucrose during each
of the extinction sessions, and thus unreinforced exposure to the CS was the same, which
means there was no volumetric difference to account for greater extinction in the REX
group. Moreover, massed (temporally dense) extinction has been shown to be more
effective than spaced (temporally rarefied) extinction in rapidly extinguishing aversive
conditioning, including human phobias and CTA in rodents (Pavlov, 1927; Foa, Jameson,
Turner & Payne, 1980; Berman, et al., 2003; Cain, Blouin & Barad, 2003; Bouton, 2004).
Ergo if only extinction were at work, the SEX group should have recovered more quickly
since their exposure to the CS (a 65 s infusion) occurred in a more temporally condensed
fashion than the REX group (a 5 s infusion followed 1 h later by a 60 s infusion). To
further test the hypothesis that the results of the REX group in Experiment 1 were due to
disruption of reconsolidation of the original CTA memory, then an additional experiment
should be conducted. In this experiment, a group similar to the REX group would receive
the 5 s reactivation stimulus followed 6 h later by its main extinction infusion. This
would allow time for reconsolidation to complete after induction by the 5 s reminder, and
216
the main extinction infusion would thus occur outside of the reconsolidation window and
have only the standard effect. Such groups have been included in other studies of
reconsolidation and were found to perform equivalently to standard extinction control
groups (Monfils, et al., 2009; Schiller, et al., 2010). Such a group was not included in the
present series because it was not practically feasible.
In contrast to Experiment 1, Experiment 2 failed to replicate the reconsolidation-
extinction effect, as did another recently published fear conditioning study (Chan, Leung,
Westbrook & McNally, 2010). It is possible that no significant effect was generated in
the REX group due to stimulus mismatch. As described in Chapter 3 (section 3.2), in
order to induce replasticization and reconsolidation in a memory, that memory must be
reactivated. Since Experiment 2 entailed bottle conditioning and bottle extinction, it is
possible that the brief intraoral reactivation procedure was insufficient in recalling the
CTA memory because the constellation of sensory stimuli associated with an involuntary
intraoral infusion are different than the stimuli associated with a voluntary instrumental
interaction with a sipper bottle, despite the overlap in the most important sensory
modality, taste. Along these same lines, it is also possible that the reactivation infusion
was too brief, since a certain minimum amount of time is needed to reactivate a memory
(see section 3.1). Given the stimulus mismatch, a longer infusion may have been
necessary so that the qualities of the reminder infusion that overlapped with the stored
memory had sufficient time to activate that engram. In support of this notion, the graph
of the Experiment 2 extinction data (figure 5.4a) appears to depict a trend towards
separation of the REX and SEX groups that may suggest a trifling effect was taking
217
place, and that could be either validated and amplified or ultimately disproven in a future
study utilizing a longer reactivation cue.
Alternatively, the results of Experiment 2 could reflect the success of both groups
to self-administer the reconsolidation-extinction treatment equally: the instrumental bottle
method of CTA affords the rat control over its exposure to the sucrose, meaning that all
rats could have sampled the sucrose, waited 10 or more minutes, and then consumed
more sucrose in an extinction bout of drinking, in effect creating their own
reconsolidation-extinction treatments. The 5 s sucrose exposure for the REX group
would not have been sufficient to differentiate the groups and thus the resultant extinction
curves would have appeared the same and no spontaneous recovery would have occurred;
the methodology employed here allowed no way to assess or control for this problem.
However, as mentioned above, there was a trend in the extinction data suggesting a
difference between groups may have been taking place. Furthermore, during extinction
of the second CTA in which both groups received only standard bottle extinction, the
extinction curves of the two groups were more closely approximated over the same
period (data not shown), suggesting that the reconsolidation-extinction treatment could
have been exerting a weak effect in the first CTA. A future replication of this experiment
could utilize food or fluid deprivation in conjunction with the bottle method; this would
motivate the rats to drink quickly, allowing the bottles to be delivered for a shorter period
of time (e.g., 10 min total) which would circumvent this problem altogether.
218
6.2 Interpretation of the histological results
C nucleus. Aversive conditioning to sucrose seems to depress c-Fos expression
in the C nucleus. In Experiment 3 the SUC group displayed greater c-Fos production in
the capsular subnucleus of the C nucleus relative to the REX and SEX groups.
Additionally, in Experiment 4 both the lateral and capsular subnuclei contained
significantly elevated c-Fos production in the SUC group over the REX and SEX groups,
findings which have precedence in the CTA literature (Yamamoto, et al., 1997; Wilkins
& Bernstein, 2006; Yasoshima, Scott & Yamamoto, 2006). In a study by Yasoshima,
Scott, and Yamamoto (2006), c-Fos expression was marginally elevated by exposure to
sucrose in non-conditioned rats in approximately the same posterior region of the C
nucleus measured presently: one sub-area of the C nucleus they measured (equivalent to a
combination of the capsular and lateral subnuclei) was nonsignificantly elevated in non-
conditioned rats compared to rats with a strong CTA (P = 0.104). Moreover,
electrophysiological recording of saccharin-responsive units in the C nucleus has
revealed that CTA causes a decrease in activity in these units during saccharin
stimulation (Yasoshima, et al., 1995). Together, these findings indicate that exposure to
an aversively conditioned tastant reduces neural activity in the C nucleus. It is possible
that depression of C nucleus output is involved in cessation of ingestion, because in
rabbits, electrical stimulation of the C nucleus resulted in fictive mastication (Applegate,
et al., 1983), while in rats, excitotoxic lesion of the C nucleus decreased ingestion of a
palatable substance (Touzani, et al., 1997). Thus a conditioned inhibition of C nucleus
219
output could be one of the means by which the CTA neural circuit prevents consumption
of a toxic food.
L nucleus. After intraoral conditioning in Experiment 3, the REX group’s
treatment on extinction day caused a significant increase in c-Fos production over the
SEX and SUC groups in the dorsal lateral subnucleus of the L nucleus. Additionally, in
the medial subnucleus the REX group expressed significantly more c-Fos than the SEX
group, although the lack of a difference between the REX and SUC groups challenges the
meaningfulness of this finding. The similar effects obtained in the SEX and SUC groups
in all subnuclei of the L are consistent with a previous c-Fos study which failed to find
any significant changes in c-Fos activity in the L nucleus induced by CTA (Yasoshima, et
al., 2006). Yet interestingly the present results suggest a potential role for the L nucleus
in reconsolidation of a CTA memory, since the reconsolidation-extinction treatment
clearly stimulated c-Fos translation in the dorsal lateral subnucleus. This is consistent
with experiments identifying the L nucleus as playing an important role in the
reconsolidation of classical fear memory (Debiec & LeDoux, 2004; Duvarci, et al., 2005;
Doyere, et al., 2007). As discussed already (see Chapter 3, sections 3.3 and 3.4),
reconsolidation is the means by which a reactivated memory re-stores itself in long term
memory, from which it can be inferred that any brain area involved in reconsolidation is
likewise involved in long term storage of a memory. Direct application of this logic to
the present results suggests that if the L nucleus is involved in reconsolidation of CTA
then it is also involved in storage of the CTA engram. Researchers have had a devilish
time attempting to unveil the role of the amygdala in CTA (Lamprecht & Dudai, 2000),
220
and the present results add further complexity to the inconsistent pool of research that
sometimes implicates the C nucleus and other times the B nucleus in CTA
reconsolidation (Koh & Bernstein, 2003; Bahar, Dorfman & Dudai, 2004; Garcia-de la
Torre, et al., 2009). However, there is reason to entertain the notion that the L amygdala
contains part of the CTA engram. As detailed in Chapter 2, the L amygdala receives
convergent sensory inputs from both the CS and US pathways (see Figures 2.0 and 2.1).
Moreover, permanent lesion studies that incidentally damaged or specifically targeted the
L nucleus observed CTA deficits (Nachman & Ashe, 1974; Aggleton, et al., 1981; Lasiter
& Glanzman, 1985; Schafe & Bernstein, 1996; Schafe, et al., 1998). Furthermore,
whereas inhibition of protein synthesis in the B nucleus during acquisition had no effect
upon CTA (Bahar, et al., 2003), functional lesion of the BLA complex by inhibition of
AMPA receptors during CTA acquisition or expression caused both block of CTA
acquisition and transitory suppression of CTA expression (Yasoshima, et al., 2000);
deductive reasoning suggests that inclusion of the L nucleus in the latter study may have
allowed those lesions to affect CTA. Thus, it is reasonable to suspect that the L nucleus
bears some of the CTA engram, and the implication from the present results that the L
nucleus participates in CTA reconsolidation is congruent with this.
If in fact the reconsolidation-extinction treatment in Experiment 3 induced
reconsolidation in CTA neurons of the L nucleus, then the increased activation in the
dorsal lateral subnucleus is likely reflecting an updating of the CTA memory because this
is one of the hypotheses of the utility of reconsolidation. Recall that reactivation of a
memory allows the integration of new content into that memory which then
221
reconsolidates into a new hybrid whole (Hupbach, et al., 2007; Winters, et al., 2009).
Applied to the present experiment, the brief reminder infusion of the reconsolidation-
extinction treatment replasticized the CTA memory and then the unreinforced main
infusion acted to modify or undo the memory, which then reconsolidated into a weaker
form. This supposition could be tested by infusing a protein synthesis inhibitor into the L
nucleus prior to reactivation of CTA memory with a reconsolidation-extinction treatment,
with the expected effect that reconsolidation would be disrupted and the CTA memory
abolished.
B nucleus. Also of note is the c-Fos activation in the parvocellular portion of the
B nucleus in REX animals in Experiment 3, a finding that could relate to extinction
memory circuitry. Extinction memory formation and expression is believed to depend
upon the infralimbic portion of the medial prefrontal cortex in CTA and other aversive
learning paradigms (Akirav, et al., 2006; Quirk & Mueller, 2008). For classical fear
conditioning in particular, the infralimbic cortex is thought to harbor a considerable
fraction of the extinction memory, and to express that memory in part by controlling C
nucleus output (Quirk & Mueller, 2008; Laurent & Westbrook, 2009). Neural activity in
the infralimbic cortex during the peri-extinction time period is critical for the retention of
a newly learned extinction memory (ibid.). It has been demonstrated that electrical
activation of the B nucleus suppresses neural activity in the infralimbic cortex (Pèrez-
Jaranay & Vives, 1991). Similarly, a CS that elicits a strong fear response likewise
inhibits infralimbic cortical activity in mice, and unilateral lesion of the BLA blocks the
CS-induced suppression of ipsilateral infralimbic activity (Garcia, Vouimba, Baudry &
222
Thompson, 1999). These facts are relevant to the present results because the densest
known amygdalocortical pathway originates in the parvocellular subnucleus of the B
nucleus and terminates in the infralimbic cortex (Pitkänen, 2000). Hence the increased c-
Fos activity in the parvocellular subnucleus of REX rats may indicate that the
reconsolidation portion of the reconsolidation-extinction treatment activated the
inhibitory amygdalocortical pathway, causing an inhibition of the infralimbic cortex, and
thereby temporarily suppressing the ability of the rat to form an extinction memory; the
main sucrose bolus constituting the extinction portion of the reconsolidation-extinction
procedure then only modified the original memory instead of producing a new memory.
This hypothesis could be tested by analysis of c-Fos expression in the infralimbic cortex
of these animals: animals in the REX group should express less c-Fos in the infralimbic
cortex than animals in the SEX group, who should exhibit increased c-Fos expression in
the infralimbic cortex as a result of forming an extinction memory.
The REX group in Experiment 1 exhibited both a rapid attenuation of aversive
responding and diminished spontaneous recovery, which has two important implications:
(1) both of those significant effects suggest that there was a reduction in the original CTA
memory, and (2) the bare fact that spontaneous recovery occurred, however diminished,
indicates that the formation of an extinction memory also occurred in these rats since
spontaneous recovery only occurs in the context of extinction (otherwise there is nothing
from which to spontaneously recover). Available data demonstrates that when a memory
is reactivated, either reconsolidation takes place in the original memory or an extinction
memory forms—they do not both occur. The selection between the two processes is a
223
function of the amount of time in exposure to the CS (see Section 3.1). For example,
contextually fear conditioned mice were injected with anisomycin just before re-exposure
to the aversive context for either 3 or 30 min (Suzuki, et al., 2004). The short 3 min
exposure induced negligible extinction but it triggered reconsolidation, which was
disrupted by the anisomycin, causing the mice to show little fear to the context
subsequently. The lengthy 30 min exposure triggered extinction memory formation
which greatly suppressed fear over the course of the session, but the anisomycin
prevented consolidation of this extinction memory so that at retest the next day the mice
exhibited a resurgence of fear; this resurgence makes manifest that the original fear
memory was not eliminated, but was somehow protected from erasure by the induction of
extinction. Hence it seems that re-exposure to a CS can either replasticize the original
associative memory or form an extinction memory, depending on the duration of
stimulation.
Such has been demonstrated in a complementary manner with CTA as well,
wherein the re-exposure treatment was held constant across groups, and either
reconsolidation or extinction was induced by this re-exposure depending on whether the
animals had a strongly or a weakly conditioned aversion, respectively (Eisenberg, et al.,
2003; see Chapter 3). These findings could explain the dual effects of the
reconsolidation-extinction treatment in Experiment 1: the REX group initially had their
CTA memory replasticized and modified by the reconsolidation-extinction treatment, but
once the CTA memory was weakened, extinction memory formation became the
dominant process to be induced by the treatment. The REX group daily received both
224
short and long re-exposures to the CS. During the first extinction session, the short
component of the reconsolidation-extinction treatment reactivated and replasticized the
strong CTA memory; the long component then could not also induce an extinction
memory, so its effect was to modify the original memory. After several days of this
treatment, the original CTA memory became weakened to the point that a threshold was
crossed, like a flip-flop circuit, and the reconsolidation-extinction treatment elicited
extinction memory formation rather than reconsolidation. This psychological model is
consonant with the neural models described in the preceding paragraphs (Figure 6.0).
During the initial reconsolidation-extinction treatments, the short CS re-exposure
activates the strong CTA memory, which causes activation of the parvocellular
subnucleus, which in turn inhibits the infralimbic cortex and staves off extinction
memory. As the CTA memory weakens over the course of the reconsolidation-
extinction treatments, it becomes less capable of activating the parvocellular subnucleus
and thus cannot inhibit the infralimbic cortex. An uninhibited infralimbic cortex permits
extinction memory formation to occur, and this formation can occur in a more rapid and
unfettered fashion since the original CTA memory is diminished; standard extinction is a
balkier process since the original CTA memory remains intact and constantly works
against extinction memory formation. Thus the parvocellular subnucleus may be an
important node in the neural network regulating the boundary between reconsolidation
and extinction in certain kinds of learning.
Fi
nu
su
tu
D
in
th
re
pa
ex
de
co
d
o
te
av
su
an
p
in
igure 6.0 Neu
ucleus or pathw
ucrose activates
urn activates the
uring the main s
nhibition. Instead
hen reconsolidate
eminder infusion
athway. The ma
xtinction memor
egradation) as w
In Exp
onditioned a
istinguishing
f Experimen
endered by
verages, wi
ubnucleus, s
nd significan
ossible that t
ntraoral meth
uropsychological
way. During the
the strong CTA
e parvocellular (
sucrose infusion
d, the main infus
es into a weake
n acts upon a we
ain infusion then
ry prevents furth
well as influencin
periment 4,
aversion. Gi
g the REX a
nt 4 relate to
the rats in
ith significa
significantly
ntly less c-F
the bottle m
hod, one in w
l model of the
initial reconsoli
A memory stored
(pc) portion of t
n one hour later,
sion acts to elim
er form. During
ak CTA memory
n has encounter
her replasticizati
ng central (C) nu
rats were as
iven that Ex
and SEX gro
o reconsolid
the B nucle
antly less c
less c-Fos t
os activity th
ethod engag
which recon
reconsolidation-
idation-extinctio
d in the lateral (
the basal (B) nu
extinction mem
minate part of the
the following r
y, which genera
rs no resistence
ion of the CTA
ucleus activity to
sessed for c
xperiment 2
oups, it is un
dation. Of p
eus. The R
c-Fos than
than the SUC
han the SEX
ges a fundam
nsolidation in
-extinction effec
n treatment (lef
(L) nucleus (repr
ucleus, thereby
mory formation i
e original CTA m
reconsolidation-
ates insufficient a
to extinction m
memory in the
allow ingestion
-Fos activity
lacked any s
nclear if any
particular sa
REX group
the SEX g
C group in t
X group in th
mentally diffe
nduces a sup
ct. Bolded lines
ft panel), the 5
resented by the
inhibiting the in
n the IL is not p
memory stored in
extinction treatm
activity to activa
memory formati
L (and hence p
n of the sucrose.
y after expre
significant b
y of the histo
alience are th
expressed t
group in th
the parvocel
he B nucleus
erent neural
ppression of
s indicate an ac
s reminder infu
dark shading). T
nfralimbic corte
possible because
n the L, a memo
ment (right pane
ate the pc-IL inh
on in the IL, an
protects it from
ession of a b
behavioral ef
ochemical re
he c-Fos pat
the lowest g
he magnoce
llular subnuc
s as a whole.
circuitry tha
c-Fos. Yet
225
ctivated
sion of
This in
ex (IL).
e of the
ory that
el), the
hibitory
nd this
further
ottle-
ffects
esults
tterns
group
llular
cleus,
It is
an the
there
226
appears to be no other instance of reconsolidation-induced suppression of immediate
early genes or transcription factors in the published literature, only activation (Tronson &
Taylor, 2007). However, there is evidence describing an extinction-induced suppression
of the transcription factor cyclic AMP response element binding protein (CREB) in the B
nucleus. Using the cued fear paradigm, Lin and coworkers (2003) observed that fear
memory recall increases and subsequent extinction training decreases phosphorylated
(i.e., activated) CREB in the B and L nuclei. Since CREB is a transcription factor that
ultimately controls the expression of c-Fos protein, extinction training would be expected
to inhibit c-Fos expression in the B nucleus. Additionally, the B nucleus is known to
participate in the extinction of a bottle-conditioned taste aversion (Bahar, et al., 2003).
Moreover, electrical activation of the medial prefrontal cortex (as would occur in
extinction memory formation) causes an inhibition of activity in the BLA (Sotres-Bayon,
Bush & LeDoux, 2004). Thus it is possible that the abatement in c-Fos production
measured in the REX rats is a reflection of extinction learning inhibiting the
phosphorylation of CREB and hence downregulating c-Fos production. However, it is
unclear why such a decrease was not also seen in SEX rats. Since there were no
significant differences in behavior between REX and SEX rats in either Experiment 2 or
Experiment 4, there is nothing to which these immunohistochemical findings can be tied.
Assessment of the amygdala model. A secondary aim of the histological
experiments was to evaluate the applicability of the model of the amygdala’s role in
learning and memory explicated in Chapter 1 to the CTA paradigm. The model
postulates that CS and US information converges in the L, C, and B nuclei of the
227
amygdala, where associations form between the stimuli. Upon reactivation by a CS,
these associative neurons activate adaptive responses appropriate for the US that is
expected to come. The L-C-brainstem pathway actuates Pavlovian CRs such as freezing,
attention, or hormone release. The L-B-forebrain pathway motivates instrumental
approach or avoidance responses. Thus the intraoral conditioning method would be
expected to stimulate the L-C-brainstem pathway, and the bottle conditioning method
would be expected to activate the L-B-forebrain pathway. In agreement with part of this
neuropsychological model, rats expressing a CTA conditioned with the intraoral method
presented altered c-Fos activity in the C nucleus: both REX and SEX groups had
significantly less c-Fos induction than the SUC group. And while L nucleus c-Fos
expression was not different between the SEX and SUC groups, the upregulated
expression levels in the REX group indicate that reconsolidation, and hence memory
activation, occurred in the L nucleus in the REX rats; this finding too is consistent with
the model. However, the B nucleus c-Fos activity seen in the REX rats is inconsistent
with the model, but as postulated above that activity may be due to a suppression of
extinction; the model does not currently account for extinction, although the infralimbic
cortical circuitry mentioned earlier could provide the basis for an expansion of the model
that would incorporate extinction learning. Still, by and large the intraoral data are
consistent with the model.
On the other hand, the bottle conditioning method used in Experiment 4 generated
data that seems somewhat dissonant with the model. As with the intraoral method in
Experiment 3, expression of a bottle-conditioned aversion led to reduced c-Fos activity in
228
the C nucleus in both the REX and SEX groups. This finding is unexpected and on its
face is not consistent with the model. An explanation for this result may rest with the
notion that instrumental conditioning essentially entails classical conditioning (i.e.,
pairing CS with US after an animal has voluntarily approached the CS) followed
subsequently by voluntary approach/avoidance responses in the presence of that
Pavlovian CS. In CTA, the CS is a tastant, a stimulus verging on the interoceptive and
one that the rat must sample before engaging in approach/avoidance responses (i.e., the
instrumental pathway cannot be activated without CS stimulation). Therefore even bottle
conditioned CTAs could be expected to activate a Pavlovian CR (aversive orofacial
responses) during the initial CS sampling, which would necessitate C nucleus activity.
This may warrant a minor revision to the model reflecting the possibility that both
classical and instrumental conditioning create CS-US associations in both the forebrain
and hindbrain pathways. Continuing on, the histological results of Experiment 4 depict
no difference between the SEX and SUC groups in c-Fos activity in the B and L nuclei,
and since bottle conditioning was used, this finding is inconsistent with the model.
Moreover, this finding is inconsistent with data obtained by Wilkins and Bernstein
(2006), who observed that bottle conditioning (but not intraoral conditioning) led to an
uptick in c-Fos expression in the B and L nuclei. These discrepancies may possibly be
explained by the fact that the c-Fos experiments conducted presently were not
specifically designed to assess for differences between the CTA conditioning methods;
rather they were designed to assess for a reconsolidation-extinction effect. These
discrepancies could also be the fault of CTA itself, which has been quite a pickle in
229
giving up its relationship with the amygdala to investigators. Overall, the data obtained
here are mixed with regard to the applicability of the model of amygdala function to the
CTA paradigm.
6.3 Conclusions
This body of experiments has yielded several interesting findings. Firstly, a
reconsolidation-extinction effect was effectuated in CTA akin to that obtained in classical
fear conditioning (Monfils et al., 2009). In rats receiving a brief reactivation of their
intraoral CTA memory one hour before each extinction session, extinction was
accelerated and spontaneous recovery was attenuated. This effect is believed to have
resulted from disruption of reconsolidation of the CTA memory, and not solely from
enhanced extinction memory formation. An application of this procedure to human
psychotherapy may help to relieve stubborn pathological memories. On the other hand,
no reconsolidation-extinction effect was obtained in rats with bottle-conditioned taste
aversions, perhaps due to the mismatch between the intraoral reactivation stimulus and
the original conditioning stimulus, a sipper bottle, which would have led to reactivation
failure. Histological analysis of c-Fos expression in the amygdala shortly after
implementing the reconsolidation-extinction procedure in intraorally-conditioned rats
suggests that reconsolidation of CTA memory may occur in the L nucleus. Further, the
parvocellular subnucleus of the B nucleus may participate in controlling whether
reconsolidation or extinction is initiated upon memory reactivation. The histological data
gathered from bottle-conditioned rats showed an anomalous pattern of reduced c-Fos
230
activity in the B nucleus of rats given the reconsolidation-extinction procedure, which
could be consistent with a role for the B nucleus in extinction of CTA. However, given
that the standard extinction treatment did not elicit a similar pattern of c-Fos expression,
and also that in general there were no significant behavioral findings onto which the c-
Fos data could be mapped, this conclusion is tenuous. Jointly, the histological results
from the intraoral and bottle conditioning studies indicate that the C nucleus can
participate in the expression of both types of CTA. Conjointly they also lend patchy
support to the neuropsychological model of the amygdala proposed in Chapter 1, and at
the least suggest that some addenda to the model are justified.
231
REFERENCES
Abraham, WC & Williams, JM (2003) “Properties and mechanisms of LTP
maintenance.” The Neuroscientist, 9(6):463-474
Abrams, TW & Kandel, ER (1988) “Is contiguity detection in classical conditioning a
system or a cellular property? Learning in Aplysia suggests a possible molecular site.”
Trends in Neurosciences, 11(4):128-135
Adams, JP & Dudek, SM (2005) “Late-phase long-term potentiation: getting to the
nucleus.” Nature Reviews Neuroscience, 6:737-743
Aggleton, JP & Passingham, RE (1981) “Syndrome produced by lesions of the amygdala
in monkeys (Macaca mulatta).” Journal of Comparative and Physiological
Psychology, 95(6):961-977
Aggleton, JP, Petrides, M & Iversen, SD (1981) “Differential effects of amygdaloid
lesions on conditioned taste aversion learning by rats.” Physiology & Behavior,
27:397-400
Akeyson, EW & Passingham, LP (1994) “Processing of splanchnic and somatic input in
thoracic spinal cord of the rat.” American Journal of Physiology: Regulatory,
Integrative and Comparative Physiology, 266:R257-R267
Akirav, I, Khatsrinov, V, Vouimba, R-M, Merhav, M, Ferreira, G, Rosenblum &
Maroun, M (2006) “Extinction on conditioned taste aversion depends on functional
protein synthesis but not on NMDA receptor activation in the ventromedial prefrontal
cortex.” Learning & Memory, 13:254-258
Akirav, I & Maroun, M (2006) “Ventromedial prefrontal cortex is obligatory for
consolidation and reconsolidation of object recognition memory.” Cerebral Cortex,
16:1759-1765
Ambrogi Lorenzini, CG, Baldi, E, Bucherelli, C, Sacchetti, B & Tassoni, G (1999)
“Neural topography and chronology of memory consolidation: a review of functional
inactivation findings.” Neurobiology of Learning and Memory, 71:1-18
Amorapanth, P, LeDoux, JE & Nader, K (2000) “Different lateral amygdala outputs
mediate reactions and actions elicited by a fear-arousing stimulus.” Nature
Neuroscience, 3(1):74-79
Antoniadis, EA & McDonald, RJ (2000) “Amygdala, hippocampus and discriminative
fear conditioning to context.” Behavioural Brain Research, 108:1-19
232
Applegate, CD, Kapp, BS, Underwood, MD & McNall, CL (1983) “Autonomic and
somatomotor effects of amygdala central n. stimulation in awake rabbits.” Physiology
& Behavior, 31:353-360
Ardenghi, P, Barros, D, Izquierdo, LA, Bevilaqua, L, Schröder, N, Quevedo, J,
Rodriques, C, Madruga, M, Medina, JH, Izquierdo, I (1997) “Late and prolonged
post-training memory modulation in entorhinal and parietal cortex by drugs acting on
the cAMP/protein kinase A signalling pathway.” Behavioural Pharmacology, 8:745-
751
Arthur, JB (1975) “Taste aversion learning is impaired by interpolated amygdaloid
stimulation but not by posttraining amygdaloid stimulation.” Behavioral Biology,
13:369-376
Azuma, S, Yamamoto, T & Kawamura, Y (1984) “Studies on gustatory responses of
amygdaloid neurons in rats.” Experimental Brain Research, 56:12-22
Bahar, A, Dorfman, N & Dudai, Y (2004) “Amygdalar circuits required for either
condolidation or extinction of taste aversion memory are not required for
reconsolidation.” European Journal of Neuroscience, 19:1115-1118
Bahar, A, Dudai, Y & Ahissar, E (2004) “Neural signature of taste familiarity in the
gustatory cortex of the freely behaving rat.” Journal of Neurophysiology, 92:3298-
3308
Bahar, A, Samuel, A, Hazvi, S & Dudai, Y (2003) “The amygdalar circuit that acquires
taste aversion memory differs from the circuit that extinguishes it.” European Journal
of Neuroscience, 17(7):1527-1530
Balleine, BW (2001) “Incentive processes in instrumental conditioning.” In RR Mowrer
& SB Klein (Eds.) Handbook of Contemporary Learning Theories. Lawrence
Earlbaum Associates: Mahwah, NJ
Barker, LM (1976) “CS duration, amount, and concentration effects in conditioning taste
aversions.” Learning and Motivation, 7:265-273
Barker, LM, Smith, JC & Suarez, EM (1977) “‘Sickness’ and the backward conditioning
of taste aversions.” In LM Barker, MR Best & M Domjan (Eds.) Learning
Mechanisms in Food Selection, USA: Baylor University Press
Barnett, SA (1975) The rat: A study in behavior. University of Chicago Press: Chicago,
IL
233
Barros, DM, Izquierdo, LA, Sant’Anna, MK, Quevedo, J, Medina, JH, McGaugh, JL &
Izquierdo, I (1999) “Stimulators of the cAMP cascade reverse amnesia induced by
intra-amygdala but not intrahippocampal KN-62 administration.” Neurobiology of
Learning and Memory, 71, 94-103
Basbaum, AI & Jessell, TM (2000) “The perception of pain.” In ER Kandel, JH Schwartz
& TM Jessell (Eds.) Principles of Neural Science, 4
th
Ed. McGraw-Hill: San
Francisco
Baxter, MG & Murray, EA (2002) “The amygdala and reward.” Nature Reviews
Neuroscience, 3:563-573
Bechara, A, Tranel, D, Damasio, H, Adolphs, R, Rockland, C & Damasio, AR (1995)
“Double dissociation of conditioning and declarative knowledge relative to the
amygdala and hippocampus in humans.” Science, 269(5227):1115-1118
Berendse, HW & Groenewegen, HJ (1991) “Restricted cortical termination fields of the
midline and intralaminar thalamic nuclei in the rat.” Neuroscience, 42(1):73-102
Berlau, DJ & McGaugh, JL (2003) “Basolateral amygdala lesions do not prevent memory
of context-footshock training.” Learning & Memory, 10:495-502
Berman, DE, Hazvi, S, Neduva, V & Dudai, Y (2000) “The role of identified
neurotransmitter systems in the response of insular cortex to unfamiliar taste:
activation of ERK1-2 and formation of a memory trace.” Journal of Neuroscience,
20(18):7017-7023
Berman, DE, Hazvi, S, Stehberg, J, Bahar, A & Dudai, Y (2003) “Conflicting processes
in the extinction of conditioned taste aversion: behavioral and molecular aspects of
latency, apparent stagnation, and spontaneous recovery.” Learning & Memory, 10:16-
25
Berman, RF, Kesner, RP & Partlow, LM (1978) “Passive avoidance impairment in rats
following cycloheximide injection into the amygdala.” Brain Research, 158:171-188
Bermudez-Rattoni, F (1998) “Pharmacology of CTA.” In J Bures, F Bermudez-Rattoni &
T Yamamoto (Eds.) Conditioned Taste Aversion: Memory of a Special Kind. Oxford
University Press: New York
Bermudez-Rattoni, F (2004) “Molecular mechanisms of taste-recognition memory.”
Nature Reviews Neuroscience, 5:209-217
234
Bermudez-Rattoni, F, Introini-Collison, I, Coleman-Mesches, K & McGaugh, JL (1997)
“Insular cortex and amygdala lesions induced after aversive training impair retention:
effects of degree of training.” Neurobiology of Learning and Memory, 67:57-63
Bermudez-Rattoni, F & McGaugh, JL (1991) “Insular cortex and amygdala lesions
differentially affect acquisition on inhibitory avoidance and conditioned taste
aversion.” Brain Research, 549(1):165-170
Bermudez-Rattoni, F & Yamamoto, T (1998) “Neuroanatomy of CTA: lesion studies.” In
J Bures, F Bermudez-Rattoni & T Yamamoto (Eds.) Conditioned Taste Aversion:
Memory of a Special Kind. Oxford University Press: New York
Bernard, JF, Alden, M & Besson, JM (1993) “The organization of the efferent projections
from the pontine parabrachial area to the amygdaloid complex: a Phaseolus vulgaris
leucoagglutinin (PHA-L) study in the rat.” Journal of Comparative Neurology,
329(2):210-229
Bernstein, IL & Koh, MT (2007) “Molecular signaling during taste aversion learning.”
Chemical Senses, 32:99-103
Bester, H, Besson, J-M & Bernard, J-F (1997) “Organization of efferent projections from
the parabrachial area to the hypothalamus: a Phaseolus vulgaris-leucoagglutinin
study in the rat.” Journal of Comparative Neurology, 383(3):245-281
Bester, H, Bourgeais, L, Villanueva, L, Besson, J-M & Bernard, J-F (1999) “Differential
projections to the intralaminar and gustatory thalamus from the parabrachial area: a
PHA-L study in the rat.” Journal of Comparative Neurology, 405(4):421-449
Biedenkapp, JC & Rudy, JW (2004) “Context memories and reactivation: constraints on
the reconsolidation hypothesis.” Behavioral Neuroscience, 118(5):956-964
Bolles, RC & Fanselow, MS (1980) “A perceptual-defensive-recuperative model of fear
and pain.” The Behavioral and Brain Sciences, 3:291-323
Borison, HL (1974) “Area postrema: chemoreceptor trigger zone for vomiting—is that
all?” Life Sciences, 14(10):1807-1817
Borsini, F & Rolls, ET (1984) “Role of noradrenaline and serotonin in the basolateral
region of the amygdala in food preferences and learned taste aversions in the rat.”
Physiology & Behavior, 33:37-43
Bourgeais, L, Monconduit, L, Villanueva, L & Bernard, J-F (2001) “Parabrachial internal
lateral neurons convey nociceptive messages from the deep laminas of the dorsal horn
to the intralaminar thalamus.” Journal of Neuroscience, 21(6):2159-2165
235
Bouton, ME (2004) “Context and behavioral processes in extinction.” Learning &
Memory, 11:485-494
Braveman, NS (1977) “What studies on pre-exposure to pharmacological agents tell us
about the nature of the aversion-inducing treatment.” In LM Barker, MR Best & M
Domjan (Eds.) Learning Mechanisms in Food Selection, Baylor University Press:
USA
Breslin, PAS, Spector, AC & Grill, HJ (1992) “A quantitative comparison of taste
reactivity behaviors to sucrose before and after lithium chloride pairings: a
unidimensional account of palatability.” Behavioral Neuroscience, 106(5):820-836
Brizzee, KR & Neal, LM (1954) “A re-evaluation of the cellular morphology of the area
postrema in view of recent evidence for a chemoreceptor function.” Journal of
Comparative Neurology, 100(1):41-61
Bucherelli, C, Baldi, E, Mariottini, C, Passani, MB & Blandina, P (2006) “Aversive
memory reactivation engages in the amygdala only some neurotransmitters involved
in consolidation.” Learning & Memory, 13:426-430
Bucherelli, C, Tassoni, G & Bures, J (1992) “Time-dependent disruption of passive
avoidance acquisition by post-training intra-amygdala injection of tetrodotoxin in
rats.” Neuroscience Letters, 140:231-234
Bures, J (1998) “The CTA paradigm: terminology, methods, and conventions.” In J
Bures, F Bermudez-Rattoni & T Yamamoto (Eds.) Conditioned Taste Aversion:
Memory of a Special Kind. Oxford University Press: New York
Bures, J, Buresova, O & Ivanova, SF (1991) “Brain stem mechanisms of conditioned
taste aversion learning in rats.” Archives Internationales de Physiologie, de Biochimie
et de Biophysique, 99(5):A131-134
Burgdorf, J, Knutson, B, Panksepp, J & Shippenberg, TS (2001) “Evaluation of rat
ultrasonic vocalizations as predictors of the conditioned aversive effects of drugs.”
Psychopharmacology, 155:35-42
Burns, LH, Robbins, TW & Everitt, BJ (1993) “Differential effects of excitotoxic lesions
of the basolateral amygdala, ventral subiculum and medial prefrontal cortex on
responding with conditioned reinforcement and locomotor activity potentiated by
intra-accumbens infusions of D-amphetamine.” Behavioural Brain Research, 55:167-
183
236
Bussey, TJ, Everitt, BJ & Robbins, TW (1997) “Dissociable effects of cingulate and
medial frontal cortex lesions on stimulus-reward learning using a novel Pavlovian
autoshaping procedure for the rat: implications for the neurobiology of emotion.”
Behavioral Neuroscience, 111(5):908-919
Cahill, L & McGaugh, JL (1990) “Amygdaloid complex lesions differentially affect
retention of tasks using appetitive and aversive reinforcement.” Behavioral
Neuroscience, 104(4):532-543
Cahill, L, Vazdarjanova, A & Setlow, B (2000) “The basolateral amygdala complex is
involved with, but is not necessary for, rapid acquisition of Pavlovian ‘fear
conditioning’.” European Journal of Neuroscience, 12:3044-3050
Cahill, L, Weinberger, NM, Roozendaal, B & McGaugh, JL (1999) “Is the amygdala a
locus of “conditioned fear”? Some questions and caveats.” Neuron, 23:227-228
Cain, CK, Blouin, AM & Barad (2003) “Temporally massed CS presentation generate
more fear extinction than spaced presentations.” Journal of Experimental Psychology:
Animal Behavior Processes, 29(4):323-333
Calandreau, L, Desmedt, A, Decorte, L & Jaffard, R (2005) “A different recruitment of
the lateral and basolateral amygdala promotes contextual or elemental conditioned
association in Pavlovian fear conditioning.” Learning & Memory, 12(4):383-388
Cammarota, M, Bevilaqua, LRM, Medina, JH & Izquierdo, I (2004) “Retrieval does not
induce reconsolidation of inhibitory avoidance memory.” Learning & Memory,
11:572-578
Cardinal, RN, Parkinson, JA, Hall, J & Everitt, BJ (2002) “Emotion and motivation: the
role of the amygdala, ventral striatum, and prefrontal cortex.” Neuroscience and
Biobehavioral Reviews, 26:321-352
Castellanos, C, Salas, JA, Gonzalez, H, Roldan, G & Garcia, D (2000) “Muscarinic
blockade but not acetylcholine depletion impairs taste aversion learning in rats.”
Society for Neuroscience Abstracts, 26:75.10
Chambers, KC (1990) “A neural model for conditioned taste aversions.” Annual Review
of Neuroscience, 13:373-385
Chan, WYM, Leung, HT, Westbrook, RF & McNally, GP (2010) “Effects of recent
exposure to a conditioned stimulus on extinction of Pavlovian fear conditioning.”
Learning & Memory, 17:512-521
237
Chapman, WP, Schroeder, HR, Geyer, G, Brazier, MAB, Fager, C, Poppen, JL, Solomon,
HC & Yakovlev, PI (1954) “Physiological evidence concerning importance of the
amygdaloid nuclear region in the integration of circulatory function and emotion in
man.” Science, 120(3127):949-950
Chen, Z, Travers, SP & Travers, JP (2001) “Muscimol infusions in the brain stem
reticular formation reversibly block ingestion in the awake rat.” American Journal of
Physiology-Regulatory, Integrative and Comparative Physiology, 280:R1085-R1094
Cliffer, KD, Burstein, R & Giesler, Jr, GJ (1991) “Distributions of spinothalamic,
spinohypothalamic, and spinotelencephalic fibers revealed by anterograde transport
of PHA-L in rats.” Journal of Neuroscience, 11(3):852-868
Cohen, DH (1975) “Involvement of the avian amygdalar homologue (archistriatum
posterior and mediale) in defensively conditioned heart rate change.” Journal of
Comparative Neurology, 160:13-36
Coil, JD, Rogers, RC, Garcia, J & Novin, D (1978) “Conditioned taste aversions: vagal
and circulatory mediation of the toxic unconditioned stimulus.” Behavioral Biology,
24(4):509-519
Conard, RA (1951) “Effect of x-irradiation on intestinal motility of the rat.” American
Journal of Physiology, 165(2):375-385
Contreras, RJ, Beckstead, RM & Norgren, R (1982) “The central projections of the
trigeminal, facial, glossopharyngeal and vagus nerves: an autoradiographic study in
the rat.” Journal of the Autonomic Nervous System, 6(3):303-322
Cook, M & Mineka, S (1990) “Selective associations in the observational conditioning of
fear in rhesus monkeys.” Journal of Experimental Psychology: Animal Behavior
Processes, 16(4):372-389
Corbit, LH & Balleine, BW (2005) “Double dissociations of basolateral and central
amygdala lesions on the general and outcome-specific forms of Pavlovian-
instrumental transfer.” Journal of Neuroscience, 25(4):962-970
Danielson, EH, Magnuson, DJ & Gray, TS (1989) “The central amygdaloid nucleus
innervation of the dorsal vagal complex in rat: a Phaseolus vulgaris leucoagglutinin
lectin anterograde tracing study.” Brain Research Bulletin, 22(4):705-715
Davis, M (1989) “Sensitization of the acoustic startle reflex by footshock.” Behavioral
Neuroscience, 103(3):495-503
238
Davis, M (2000) “The role of the amygdala in conditioned and unconditioned fear and
anxiety.” In JP Aggleton (Ed.) The Amygdala: A Functional Analysis, 2
nd
Ed. Oxford
University Press: New York
Davis, HP & Squire, LR (1984) “Protein synthesis and memory: a review.”
Psychological Bulletin, 96(3):518-559
Debiec, J, Doyere, V, Nader, K & LeDoux, JE (2006) “Directly reactivated, but not
indirectly reactivated, memories undergo reconsolidation in the amygdala.”
Proceedings of the National Academy of Sciences of the United States of America,
13(9):3428-3433
Debiec, J & LeDoux, JE (2004) “Disruption of reconsolidation but not consolidation of
auditory fear conditioning by noradrenergic blockade in the amygdala.”
Neuroscience, 129:267-272
Debiec, J & LeDoux, JE (2006) “Noradrenergic signaling in the amygdala contributes to
the reconsolidation of fear memory: treatment implications for PTSD.” Annals of the
New York Academy of Sciences, 1071:521-524
Debiec, J, LeDoux, JE & Nader, K (2002) “Cellular and systems reconsolidation in the
hippocampus.” Neuron, 36:527-538
DeSousa, NJ & Vaccarino, FJ (2001) “Neurobiology of reinforcement: interaction
between dopamine and cholecystokinin systems.” In RR Mowrer & SB Klein (Eds.)
Handbook of Contemporary Learning Theories. Lawrence Earlbaum Associates:
Mahwah, NJ
Deutsch, JA, Molina, F & Puerto, A (1976) “Conditioned taste aversion caused by
palatable nontoxic nutrients.” Behavioral Biology, 16(2):161-174
Dickinson, A & Balleine, B (1994) “Motivational control of goal-directed action.”
Animal Learning and Behavior, 22(1):1-18
Doron, NN & LeDoux, JE (1999) “Organization of projections to the lateral amygdala
from auditory and visual areas of the thalamus in the rat.” Journal of Comparative
Neurology, 412:383-409
Doyere, V, Debiec, J, Monfils, M-H, Schafe, GE & LeDoux, JE (2007) “Synapse-specific
reconsolidation of distinct fear memories in the lateral amygdala.” Nature
Neuroscience, 10(4):414-416
Dudai, Y (2004) “The neurobiology of consolidations, or, how stable is the engram?”
Annual Review of Psychology, 55:51-86
239
Dudai, Y & Eisenberg, M (2004) “Rites of passage of the engram.” Neuron, 44(1):93-100
Dunn, LT & Everitt, BJ (1988) “Double dissociations of the effects of amygdala and
insular cortex lesions on conditioned taste aversion, passive avoidance, and
neophobia in the rat using the excitotoxin ibotenic acid.” Behavioral Neuroscience,
102(1):3-23
Duvarci, S & Nader, K (2004) “Characterization of fear memory reconsolidation.”
Journal of Neuroscience, 24(42):9269-9275
Duvarci, S, Nader, K & LeDoux, JE (2005) “Activation of extracellular signal-regulated
kinase mitogen-activated protein kinase cascade in the amygdala is required for
memory reconsolidation of auditory fear conditioning.” European Journal of
Neuroscience, 21:283-289
Eisenberg, M, Kobilo, T, Berman, DE & Dudai, Y (2003) “Stability of retrieved memory:
inverse correlation with trace dominance.” Science, 301:1102-1104
Ellis, ME & Kesner, RP (1981) “Physostigmine and norepinephrine: effects of injection
into the amygdala on taste associations.” Physiology & Behavior, 27:203-209
El-Mallakh, RS (1996) “Ionic mechanisms of lithium action.” In D Spiegel (Ed.)
Lithium: Actions and Mechanisms. American Psychiatric Press, Inc.: Washington, DC
Etscorn, F & Stephens, R (1973) “Establishment of conditioned taste aversions with a 24-
hour CS-US interval.” Physiological Psychology, 1(3):251-253
Everitt, BJ, Cardinal, RN, Hall, J, Parkinson, JA & Robbins, TW (2000) “Differential
involvement of amygdala subsystems in appetitive conditioning and drug addiction.”
In JP Aggleton (Ed.) The Amygdala: A Functional Analysis, 2
nd
Ed. Oxford
University Press: New York
Fanselow, MS & Gale, GD (2003) “The amygdala, fear, and memory.” Annals of the
New York Academy of Sciences, 985:125-134
Fanselow, MS & Kim, JJ (1994) “Acquisition of contextual Pavlovian fear conditioning
is blocked by application of an NMDA receptor antagonist D,L-2-amino-5-
phosphonovaleric acid to the basolateral amygdala.” Behavioral Neuroscience,
108(1):210-212
Fenu, S, Bassareo, V & Di Chiara, G (2001) “A role for dopamine D1 receptors of the
nucleus accumbens shell in conditioned taste aversion learning.” Journal of
Neuroscience, 21(17):6897-6904
240
Fernandez-Ruiz, J, Miranda, MI, Bermudez-Rattoni, F & Drucker-Colin, R (1993)
“Effects of catecholaminergic depletion of the amygdala and insular cortex on the
potentiation of odor by taste aversions.” Behavioral and Neural Biology, 60:189-191
Ferreira, G, Miranda, MI, De la Cruz, V, Rodriguez-Ortiz, CJ & Bermudez-Rattoni, F
(2005) “Basolateral amygdala glutamatergic activation enhances taste aversion
through NMDA receptor activation in the insular cortex.” European Journal of
Neuroscience, 22:2596-2604
Ferry, B & McGaugh, JL (1999) “Clenbuterol administration into the basolateral
amygdala post-training enhances retention in an inhibitory avoidance task.”
Neurobiology of Learning and Memory, 72:8-12
Ferry, B, Roozendaal, B & McGaugh, JL (1999) “Basolateral amygdala noradrenergic
influences on memory storage are mediated by an interaction between β- and α
1
-
adrenoceptors.” Journal of Neuroscience, 19(12):5119-5123
Fitzgerald, RE & Burton, MJ (1981) “Effects of small basolateral amygdala lesions on
ingestion in the rat.” Physiology & Behavior, 27:431-437
Fitzgerald, RE & Burton, MJ (1983) “Neophobia and conditioned taste aversion deficits
in the rat produced by undercutting temporal cortex.” Physiology & Behavior,
30:203-206
Foa, EB, Jameson, JS, Turner, RM & Payne, LL (1980) “Massed vs. spaced exposure
sessions in the treatment of agoraphobia.” Behaviour Research & Therapy, 18:333-
338
Frankland, PW, Ding, H-K, Takahashi, E, Suzuki, A, Kida, S & Silva, AJ (2006)
“Stability of recent and remote contextual fear memory.” Learning & Memory,
13:451-457
Gabella, G (2004) “Autonomic nervous system.” In G Paxinos (Ed.) The rat nervous
system, 3
rd
Ed. Elsevier Academic Press: San Diego
Galaverna, OG, Seeley, RJ, Berridge, KC, Grill, HJ, Epstein, AN & Schulkin, J (1993)
“Lesions of the central nucleus of the amygdala I: effects on taste reactivity, taste
aversion learning and sodium appetite.” Behavioural Brain Research, 59:11-17
Gale, GD, Anagnostaras, SG, Godsil, BP, Mitchell, S, Nozawa, T, Sage, JR, Wiltgen, B
& Fanselow, MS (2004) “Role of the basolateral amygdala in the storage of fear
memories across the adult lifetime of rats.” Journal of Neuroscience, 24(15):3810-
3815
241
Gallagher, M, Graham, PW & Holland, PC (1990) “The amygdala central nucleus and
appetitive Pavlovian conditioning: lesions impair one class of conditioned behavior.”
Journal of Neuroscience, 10(6):1906-1911
Gallagher, M, Kapp, BS, Musty, RE & Driscoll, PA (1977) “Memory formation:
evidence for a specific neurochemical system in the amygdala.” Science,
198(4315):423-425
Gallo, M, Roldan, G & Bures, J (1992) “Differential involvement of gustatory insular
cortex and amygdala in the acquisition and retrieval of conditioned taste aversion in
rats.” Behavioral Brain Research, 52:91-97
Garcia, J, Ervin, FR & Koelling, RA (1966) “Learning with prolonged delay of
reinforcement.” Psychonomic Science, 5(3):121-122
Garcia, J, Hankins, WG, Robinson, JH & Vogt, JL (1972) “Bait-shyness: tests of CS-US
mediation.” Physiology & Behavior, 8(5):807-810
Garcia, J, Hankins, WG & Rusiniak, KW (1974) “Behavioral regulation of the milieu
interne in man and rat.” Science, 185(4154):824-831
Garcia, J, Kimeldorf, DJ & Koelling, RA (1955) “Conditioned aversion to saccharin
resulting from exposure to gamma radiation.” Science, 122(3160):157-158
Garcia, R, Vouimba, R-M, Baudry, M & Thompson, RF (1999) “The amygdala
modulates prefrontal cortex activity relative to conditioned fear.” Nature, 402:294-
296
Garcia-de la Torre, P, Rodriguez-Ortiz, CJ, Arreguin-Martinez, JL, Cruz-Castaneda, P &
Bermudez-Rattoni, F (2009) “Simultaneous but not independent anisomycin infusions
in insular cortex and anygdala hinder stabilization of taste memory when updated.”
Learning & Memory, 16:514-519
Garcia-de la Torre, P, Rodriguez-Ortiz, CJ, Balderas, I & Bermudez-Rattoni, F (2010)
“Differential participation of temporal structures in the consolidation and
reconsolidation of taste aversion extinction.” European Journal of Neuroscience,
32:1018-1023
Gauriau, C & Bernard, J-F (2002) “Pain pathways and parabrachial circuits in the rat.”
Experimental Physiology, 87.2:251-258
Gilbertson, TA, Liu, L, Kim, I, Burks, CA & Hansen, DR (2005) “Fatty acid responses in
taste cells from obesity-prone and –resistant rats.” Physiology & Behavior, 86:681-
690
242
Goldstein, LE, Rasmusson, AM, Bunney, BS & Roth, RH (1996) “Role of the amygdala
in the coordination of behavioral, neuroendocrine, and prefrontal cortical monoamine
responses to psychological stress in the rat.” Journal of Neuroscience, 16(15):4787-
4796
Goosens, KA & Maren, S (2001) “Contextual and auditory fear conditioning are
mediated by the lateral, basal, and central amygdaloid nuclei in rats.” Learning &
Memory, 8:148-155
Grill, HJ & Norgren, R (1978) “Chronically decerebrate rats demonstrate satiation but not
bait shyness.” Science, 201:267-269
Green, L, Bouzas, A & Rachlin, H (1972) “Test of an electric-shock analog to illness-
induced aversion.” Behavioral Biology, 7:513-518
Groenewegen, HJ & Witter, MP (2004) “Thalamus.” In G Paxinos (Ed.) The rat nervous
system, 3
rd
Ed. Elsevier Academic Press: San Diego
Grove, EA (1988) “Neural associations of the substantia innominata in the rat: afferent
connections.” Journal of Comparative Neurology, 277:315-346
Gruest, N, Richer, P & Hars, B (2004) “Memory consolidation and reconsolidation in the
rat pup require protein synthesis.” Journal of Neuroscience, 24(46):10488-10492
Gustavson, CR, Garcia, J, Hankins, WG & Rusiniak, KW (1974) “Coyote predation
control by aversive conditioning.” Science, 184(4136):581-583
Gutiérrez, R, Téllez, LA & Bermudez-Rattoni, F (2003) “Blockade of cortical muscarinic
but not NMDA receptors prevents a novel taste from becoming familiar.” European
Journal of Neuroscience, 17:1556-1562
Halpern, BP & Tapper, DN (1971) “Taste stimuli: quality coding time.” Science,
171:1256-1258
Hamilton, RB & Norgren, R (1984) “Central projections of gustatory nerves in the rat.”
Journal of Comparative Neurology, 222(4):560-577
Han, J-S, McMahan, RW, Holland, P & Gallagher, M (1997) “The role of an amygdalo-
nigrostriatal pathway in associative learning.” Journal of Neuroscience, 17(10):3913-
3919
Hatfield, T, Graham, PW & Gallagher, M (1992) “Taste-potentiated odor aversion
learning: role of the amygdaloid basolateral complex and central nucleus.”
Behavioral Neuroscience, 106(2):286-293
243
Helmstetter, FJ (1992a) “Contribution of the amygdala to learning and performance of
conditional fear.” Physiology & Behavior, 51:1271-1276
Helmstetter, FJ (1992b) “The amygdala is essential for the expression of conditional
hypoalgesia.” Behavioral Neuroscience, 106(3):518-528
Herbert, H, Moga, MM & Saper, CB (1990) “Connections of the parabrachial nucleus
with the nucleus of the solitary tract and the medullary reticular formation in the rat.”
Journal of Comparative Neurology, 293(4):540-580
Hintiryan, H, Foster, NN & Chambers, KC (2009) “Dissociating the conditioning and the
anorectic effects of estradiol in female rats.” Behavioral Neuroscience, 123(6):1226-
1237
Hintiryan, H, Hayes, U & Chambers, KC (2005) “The role of histamine in estradiol-
induced conditioned consumption reductions.” Physiology & Behavior, 84:117-128
Hintiryan, H, Hayes, U & Chambers, KC (2006) “Intraoral cheek fistulae: a refined
technique.” Laboratory Animals, 40:456-464
Hitchcock, JM & Davis, M (1986) “Lesions of the amygdala, but not of the cerebellum or
red nucleus, block conditioned fear as measured with the potentiated startle reflex.”
Behavioral Neuroscience, 100(1):11-22
Hitchcock, JM & Davis, M (1987) “Fear-potentiated startle using an auditory conditioned
stimulus: effect of lesions of the amygdala.” Physiology & Behavior, 39:403-408
Hitchcock, JM & Davis, M (1991) “Efferent pathway of the amygdala involved in
conditioned fear as measured with the fear-potentiated startle paradigm.” Behavioral
Neuroscience, 105(6):826-842
Hitchcock, JM, Sananes, CB & Davis, M (1989) “Sensitization of the startle reflex by
footshock: blockade by lesions of the central nucleus of the amygdala or its efferent
pathway to the brainstem.” Behavioral Neuroscience, 103(3):509-518
Hitchcott, PK & Phillips, GD (1998) “Double dissociation of the behavioural effects of
R(+) 7-OH-DPAT infusions in the central and basolateral amygdala nuclei upon
Pavlovian and instrumental conditioned appetitive behaviours.”
Psychopharmacology, 140:458-469
Holahan, MR & White, NM (2004) “Intra-amygdala muscimol injections impair freezing
and place avoidance in aversive contextual conditioning.” Learning & Memory,
11:436-446
244
Huff, NC & Rudy, JW (2004) “The amygdala modulates hippocampus-dependent context
memory formation and stores cue-shock associations.” Behavioral Neuroscience,
118(1):53-62
Hupbach, A, Gomez, R, Hardt, O & Nadel, L (2007) “Reconsolidation of episodic
memories: a subtle reminder triggers integration of new information.” Learning &
Memory, 14:47-53
Hutchison, Jr., SJ (1973) “Taste aversion in albino rats using centrifugal spin as an
unconditioned stimulus.” Psychological Reports, 33(2):467-470
Iimori, K, Tanaka, M, Kohno, Y, Ida, Y, Nagakawa, R, Hoaki, Y, Tsuda, A & Nagasaki,
N (1982) “Psychological stress enhances noradrenaline turnover in specific brain
regions in rats.” Pharmacology Biochemistry and Behavior, 16:637-640
Ikemoto, S & Panksepp, J (1999) “The role of nucleus accumbens dopamine in motivated
behavior: a unifying interpretation with special reference to reward-seeking.” Brain
Research Reviews, 31:6-41
Inda, MC, Delgado-Garcia, JM & Carrion, AM (2005) “Acquisition, consolidation,
reconsolidation, and extinction of eyelid conditioning responses require de novo
protein synthesis.” Journal of Neuroscience, 25(8):2070-2080
Introini-Collison, IB, Dalmaz, C & McGaugh, JL (1996) “Amygdala β-noradrenergic
influences on memory storage involve cholinergic activation.” Neurobiology of
Learning and Memory, 65:57-64
Introini-Collison, IB, Ford, L & McGaugh, JL (1995) “Memory impairment induced by
intraamygdala β-endorphin is mediated by noradrenergic influences.” Neurobiology
of Learning and Memory, 63:200-205
Introini-Collison, IB, Miyazaki, B & McGaugh, JL (1991) “Involvement of the amygdala
in the memory-enhancing effects of clenbuterol.” Psychopharmacology, 104:541-544
Introini-Collison, IB, Nagahara, AH & McGaugh, JL (1989) “Memory enhancement with
intra-amygdala post-training naloxone is blocked by concurrent administration of
propranolol.” Brain Research, 476:94-101
Ionescu, E & Buresova, O (1977) “Failure to elicit conditioned taste aversion by severe
poisoning.” Pharmacology, Biochemistry, and Behavior, 6(3):251-254
245
Iwata, J, Chida, K & LeDoux, JE (1987) “Cardiovascular responses elicited by
stimulation of neurons in the central amygdaloid nucleus in awake but not
anesthetized rats resemble conditioned emotional responses.” Brain Research,
418:183-188
Jasmin, L, Granato, A & Ohara, PT (2004) “Rostral agranular insular cortex and pain
areas of the central nervous system: a tract-tracing study in the rat.” Journal of
Comparative Neurology, 468:425-440
Jean, A (2001) “Brain stem control of swallowing: neuronal network and cellular
mechanisms.” Physiological Reviews, 81(2):929-969
Jellestad, FK & Bakke, HK (1985) “Passive avoidance after ibotenic acid and radio
frequency lesions in the rat amygdala.” Physiology & Behavior, 34:299-305
Jerusalinsky, D, Ferreira, MBC, Walz, R, Da Silva, RC, Bianchin, M, Ruschel, AC,
Zanatta, MS, Medina, JH & Izquierdo, I (1992) “Amnesia by post-training infusion of
glutamate receptor antagonists into the amygdala, hippocampus, and entorhinal
cortex.” Behavioral and Neural Biology, 58(1):76-80
Jin, X-C, Lu, Y-F, Yang, X-F, Ma, L & Li, B-M (2007) “Glucocorticoid receptors in the
basolateral nucleus of amygdala are required for postreactivation reconsolidation of
auditory fear memory.” European Journal of Neuroscience, 25:3702-3712
Jolkkonen, E & Pitkänen, A (1998) “Intrinsic connections of the rat amygdaloid complex:
projections originating in the central nucleus.” Journal of Comparative Neurology,
395:53-72
Kalia, M & Sullivan, JM (1982) “Brainstem projections of sensory and motor
components of the vagus nerve in the rat.” Journal of Comparative Neurology,
211(3):248-264
Kalin, NH, Shelton, SE, Davidson, RJ & Kelley, AE (2001) “The primate amygdala
mediates acute fear but not the behavioral and physiological components of anxious
temperament.” Journal of Neuroscience, 21(6):2067-2074
Kandel, ER (2000) “Cellular mechanisms of learning and the biological basis of
individuality.” In ER Kandel, JH Schwartz & TM Jessell (Eds.) Principles of Neural
Science, 4
th
Ed. McGraw-Hill: San Francisco
Kapp, BS, Frysinger, RC, Gallagher, M & Haselton, JR (1979) “Amygdala central
nucleus lesions: effect on heart rate conditioning in the rabbit.” Physiology &
Behavior, 23:1109-1117
246
Kapp, BS, Pascoe, JP & Bixler, MA (1984) “The Amygdala: A neuroanatomical systems
approach to its contribution to aversive conditioning. In LR Squire and N Butters
(Eds.) The Neuropsychology of Memory. The Guilford Press: New York
Kapp, BS, Supple, Jr, WF, & Whalen, PJ (1994) “Effects of electrical stimulation of the
amygdaloid central nucleus on neocortical arousal in the rabbit.” Behavioral
Neuroscience, 108(1):81-93
Karniol, IG, Dalton, J & Laber, MH (1978) “Acute and chronic effects of lithium
chloride on physiological and psychological measures in normals.”
Psychopharmacology, 57:289-294
Kemble, ED, Levine, MS, Gregoire, K, Koepp, K & Thomas, TT (1972) “Reactivity to
saccharin and quinine solutions following amygdaloid or septal lesions in rats.”
Behavioral Biology, 7:503-512
Kemble, ED, Studelska, DR & Schmidt, MK (1979) “Effects of central amygdaloid
nucleus lesions on ingestion, taste reactivity, exploration and taste aversion.”
Physiology & Behavior, 22:789-793
Kesner, RP, Berman, RF, Burton, B & Hankins, WG (1975) “Effects of electrical
stimulation of amygdala upon neophobia and taste aversion.” Behavioral Biology,
13:349-358
Kesner, RP, Berman, RF & Tardif, R (1992) “Place and taste aversion learning: role of
basal forebrain, parietal cortex, and amygdala.” Brain Research Bulletin, 29:345-353
Kida, S, Josselyn, SA, Pena de Ortiz, S, Kogan, JH, Chevere, I, Masushige, S & Silva, AJ
(2002) “CREB required for the stability of new and reactivated fear memories.”
Nature Neuroscience, 5(4):348-355
Killcross, S, Robbins, TW & Everitt, BJ (1997) “Different types of fear-conditioned
behaviour mediated by separate nuclei within the amygdala.” Nature, 388:377-380
Kim, JH & Richardson, R (2007a) “A developmental dissociation in reinstatement of an
extinguished fear response in rats.” Neurobiology of learning and memory, 88:48-57
Kim, JH & Richardson, R (2007b) “A developmental dissociation of context and GABA
efects on extinguished fear in rats.” Behavioral Neuroscience, 121(1):131-139
Kim, JH & Richardson, R (2010) “New findings on extinction of conditioned fear early
in development: theoretical and clinical implications.” Biological Psychiatry, 67:497-
303
247
Kim, JJ, Rison, RA & Fanselow, MS (1993) “Effects of amygdala, hippocampus, and
periaqueductal gray lesions on short- and long-term contextual fear.” Behavioral
Neuroscience, 107(6):1093-1098
Klüver, H & Bucy, P (1939) “Preliminary analysis of functions of the temporal lobes in
monkeys.” Archives of Neurology and Psychiatry, 42(6):979-1000
Koekkoek, SKE, Hulscher, HC, Dortland, BR, Hensbroek, RA, Elgersma, Y, Ruigrok,
TJH & De Zeeuw, CI (2003) “Cerebellar LTD and learning-dependent timing of
conditioned eyelid responses.” Science, 301:1736-1739
Koh, MT & Bernstein, IL (2003) “Inhibition of protein kinase A activity during
conditioned taste aversion retrieval: interference with extinction or reconsolidation of
a memory?” NeuroReport, 14(3):405-407
Koh, MT & Bernstein, IL (2005) “Mapping conditioned taste aversion associations using
c-Fos reveals a dynamic role for insular cortex.” Behavioral Neuroscience,
119(2):388-398
Koh, MT, Wilkins, EE & Bernstein, IL (2003) “Novel tastes elevate c-fos expression in
the central amygdala and insular cortex: implication for taste aversion learning.”
Behavioral Neuroscience, 117(6):1416-1422
Krout, KE & Loewy, AD (2000) “Parabrachial nucleus projections to midline and
intralaminar thalamic nuclei of the rat.” Journal of Comparative Neurology, 428:475-
494
LaBar, KS, LeDoux, JS, Spencer, DD & Phelps, EA (1995) “Impaired fear conditioning
following unilateral temporal lobectomy in humans.” Journal of Neuroscience,
15(10):6846-6855
Lamprecht, R & Dudai, Y (1995) “Differential modulation of brain immediate early
genes by intraperitoneal LiCl.” NeuroReport, 7:289-293
Lamprecht, R & Dudai, Y (1996) “Transient expression of c-Fos in rat amygdala during
training is required for encoding conditioned taste aversion memory.” Learning &
Memory, 3:31-41
Lamprecht, R & Dudai, Y (2000) “The amygdala in conditioned taste aversion: it’s there,
but where?” In JP Aggleton (Ed.) The Amygdala: A Functional Analysis, 2
nd
Ed.
Oxford University Press: New York
248
Lasiter, PS & Glanzman, DL (1985) “Cortical substrates of taste aversion learning:
involvement of dorsolateral amygdaloid nuclei and temporal neocortex in taste
aversion learning.” Behavioral Neuroscience, 99(2):257-276
Laugerette, F, Passilly-Degrace, P, Patris, B, Niot, I, Febbraio, M, Montmayeur, J-P &
Besnard, P (2005) “CD36 involvement in orosensory detection of dietary lipids,
spontaneous fat preference, and digestive secretions.” Journal of Clinical
Investigation, 115(11):3177-3184
Laurent, V & Westbrook, RF (2009) “Inactivation of the infralimbic but not the prelimbic
cortex impairs consolidation and retrieval of fear extinction.” Learning & Memory,
16:520-529
Lavond, DG, Kim, JJ & Thompson, RF (1993) “Mammalian brain substrates of aversive
classical conditioning.” Annual Review of Neuroscience, 44:317-342
Ledgerwood, L, Richardson, R & Cranney, J (2005) “D-Cycloserine facilitates extinction
of learned fear: effects on reacquitision and generalized extinction.” Biological
Psychiatry, 57:841-847
LeDoux, JE (1987) “Emotion.” In Handbook of Physiology, Section 1: The Nervous
System, Volume V: Higher Functions of the Brain, Part 1 (pp.419-459). American
Physiological Society: Bethesda, MD
LeDoux, JE, Cicchetti, P, Xagoraris, A & Romanski, LM (1990) “The lateral amygdaloid
nucleus: sensory interface of the amygdala in fear conditioning.” Journal of
Neuroscience, 10(4):1062-1069
LeDoux, JE, Farb, C & Ruggiero, DA (1990) “Topographic organization of neurons in
the acoustic thalamus that project to the amygdala.” Journal of Neuroscience,
10(4):1043-1054
LeDoux, JE, Iwata, J, Cicchetti, P & Reis, DJ (1988) “Different projections of the central
amygdaloid nucleus mediate autonomic and behavioral correlates of conditioned
fear.” Journal of Neuroscience, 8:2517-2529
LeDoux, JE, Sakaguchi, A, Iwata, J & Reis, DJ (1986) “Interruption of projections from
the medial geniculate body to an archi-neostriatal field disrupts the classical
conditioning of emotional responses to acoustic stimuli.” Neuroscience, 17(3):615-
627
LeDoux, JE, Sakaguchi, A & Reis, DJ (1984) “Subcortical efferent projections of the
medial geniculate nucleus mediate emotional responses conditioned to acoustic
stimuli.” Journal of Neuroscience, 4(3):683-698
249
Lee, JLC, Everitt, BJ & Thomas, KL (2004) “Independent cellular processes for
hippocampal memory consolidation and reconsolidation.” Science, 304:839-843
Lee, JLC, Milton, AL & Everitt, BJ (2006) “Reconsolidation and extinction of
conditioned fear: inhibition and potentiation.” Journal of Neuroscience,
26(39):10051-10056
Lee, H & Kim, J (1998) “Amygdalar NMDA receptors are critical for new fear learning
in previously fear-conditioned rats.” Journal of Neuroscience, 18:8444-8454
Liang, KC, Juler, RG & McGaugh, JL (1986) “Modulating effects of posttraining
epinephrine on memory: involvement of the amygdala noradrenergic system.” Brain
Research, 368:125-133
Liang, KC, McGaugh, JL, Martinez, Jr., JL, Jensen, RA, Vasquez, BJ & Messing, RB
(1982) “Post-training amygdaloid lesions impair retention of an inhibitory avoidance
response.” Behavioural Brain Research, 4:237-249
Liang, KC, McGaugh, JL & Yao, H-Y (1990) “Involvement of amygdala pathways in the
influence of post-training intra-amygdala norepinephrine and peripheral epinephrine
on memory storage.” Brain Research, 508:225-233
Lin, C-H, Yeh, S-H, Lu, H-Y & Gean, P-W (2003) “The similarities and diversities of
signal pathways leading to consolidation of conditioning and consolidation of
extinction of fear memory.” Journal of Neuroscience, 23(23):8310-8317
Loftus, EF (1975) “Leading questions and the eyewitness report.” Cognitive Psychology,
7:560-572
Loftus, EF & Palmer, JC (1974) “Reconstruction of automobile destruction: an example
of the interaction between language and memory.” Journal of Verbal Learning and
Verbal Behavior, 13:585-589
Loftus, EF & Yuille, JC (1984) “Departures from reality in human perception and
memory.” In Weingartner, H & Parker, ES (Eds.) Memory consolidation:
psychobiology of cognition. Lawrence Earlbaum Associates: Hillsdale, NJ
Lorenzini, CA, Bucherelli, C, Giachetti, A, Mugnai, L & Tassoni, G (1991) “Effects of
nucleus basolateralis amygdalae neurotoxic lesions on aversive conditioning in the
rat.” Physiology & Behavior, 49:765-770
Lund, JP (1991) “Mastication and its control by the brain stem.” Critical Reviews in Oral
Biology and Medicine, 2(1):33-64
250
Lundy, Jr, RF & Norgren, R (2004) “Gustatory system.” In G. Paxinos (Ed.) The rat
nervous system, 3
rd
Ed. Elsevier Academic Press: San Diego
Malenka, RC & Nicoll, RA (1997) “Never fear, LTP is hear [sic].” Nature, 390:552-553
Maren, S (1998) “Overtraining does not mitigate contextual fear conditioning deficits
produced by neurotoxic lesions of the basolateral amygdala.” Journal of
Neuroscience, 18(8):3088-3097
Maren, S, Aharonov, G & Fanselow, MS (1996) “Retrograde abolition of conditional fear
after excitotoxic lesions in the basolateral amygdala of rats: absence of a temporal
gradient.” Behavioral Neuroscience, 110(4):718-726
Maren, S & Fanselow, MS (1995) “Synaptic plasticity in the basolateral amygdala
induced by hippocampal formation stimulation in vivo.” Journal of Neuroscience,
15(11):7548-7564
Mark, GP, Weinberg, JB, Rada, PV & Hoebel, BG (1995) “Extracellular acetylcholine is
increased in the nucleus accumbens following the presentation of an aversively
conditioned taste stimulus.” Brain Research, 688:184-188
McDonald, AJ (1991) “Organization of amygdaloid projections to the prefrontal cortex
and associated striatum in the rat.” Neuroscience, 44(1):1-14
McDonald, AJ (1998) “Cortical pathways to the amygdala mammalian amygdala.”
Progress in Neurobiology, 55:257-332
McGaugh, JL (2000) “Amygdala: role in modulation of memory storage.” In JP Aggleton
(Ed.) The Amygdala: A Functional Analysis, 2
nd
Ed. Oxford University Press: New
York
McGaugh, JL (2002) “Memory consolidation and the amygdala: a systems perspective.”
Trends in Neurosciences, 25(9):456-461
McGaugh, JL (2004) “The amygdala modulates the consolidation of memories of
emotionally arousing experiences.” Annual Review of Neuroscience, 27:1-28
McGaugh, JL, Introini-Collison, IB & Nagahara, AH (1988) “Memory-enhancing effects
of posttraining naloxone; involvement of β-noradrenergic influences in the
amygdaloid complex.” Brain Research, 446:37-49
McGaugh, JL, McIntyre, CK & Power, AE (2002) “Amygdala modulation of memory
consolidation: interaction with other brain systems.” Neurobiology of Learning and
Memory, 78:539-552
251
McKernan, MG & Shinnick-Gallagher, P (1997) “Fear conditioning induces a lasting
potentiation of synaptic currents in vitro.” Nature, 390:607-611
Meachum, CL & Bernstein, IL (1990) “Conditioned responses to a taste conditioned
stimulus paired with lithium chloride administration.” Behavioral Neuroscience,
104(5):711-715
Meunier, M, Bachevalier, J, Murray, EA, Malkova, L & Mishkin, M (1999) “Effects of
aspiration versus neurotoxic lesions of the amygdala on emotional responses in
monkeys.” European Journal of Neuroscience, 11(12):4403-4418
Milekic, MH & Alberini, CM (2002) “Temporally graded requirement for protein
synthesis following memory reactivation.” Neuron, 36:531-525
Miranda, IM, Ferreira, G, Ramirez-Lugo, L & Bermudez-Rattoni, F (2002)
“Glutamatergic activity in the amygdala signals visceral input during taste memory
formation.” Proceedings of the National Academy of Sciences, 99(17):11417-11422
Miranda, MI, Guillaume F, Ramirez-Lugo, L & Bermudez-Rattoni, F (2002)
“Glutamatergic activity in the amygdala signals visceral input during taste memory
formation.” Proceedings of the National Academy of Sciences of the United States of
America, 99(17):11417-11422
Miranda, MI, LaLumiere, RT, Buen, TV, Bermudez-Rattoni, F & McGaugh, JL (2003)
“Blockade of noradrenergic receptors in the basolateral amygdala impairs taste
memory.” European Journal of Neuroscience, 18:2605-2610
Miserendino, MJD, Sananes, CB, Melia, KR & Davis, M (1990) “Blocking of acquisition
but not expression of conditioned fear-potentiated startle by NMDA antagonists in the
amygdala.” Nature, 345:716-718
Monfils, M-H, Cowansage, KK, Klann, E & LeDoux, JE (2009) “Extinction-
reconsolidation boundaries: key to persistent attenuation of fear memories.” Science,
324:951-955
Morris, R, Frey, S, Kasambira, T & Petrides, M (1999) “Ibotenic acid lesions of the
basolateral, but not the central, amygdala interfere with conditioned taste aversion:
evidence from a combined behavioral and anatomical tract-tracing investigation.”
Behavioral Neuroscience, 113(2):291-302
Morris, RGM, Inglis, J, Ainge, JA, Olverman, HJ, Tulloch, J, Dudai, Y & Kelly, PAT
(2006) “Memory reconsolidation: sensitivity of spatial memory to inhibition of
protein synthesis in dorsal hippocampus during encoding and retrieval.” Neuron,
50:479-489
252
Muramoto, K, Ono, T, Nishijo, H & Fukuda, M (1993) “Rat amygdaloid neuron
responses during auditory discrimination.” Neuroscience, 52(3):621-636
Myers, KM, Ressler, KJ & Davis, M (2006) “Different mechanisms of fear extinction
dependent on length of time since fear acquisition.” Learning & Memory, 13:216-223
Nachman, M (1970) “Learned taste and temperature aversions due to lithium chloride
sickness after temporal delays.” Journal of Comparative and Physiological
Psychology, 73(1):22-30
Nachman, M & Ashe, JH (1974) “Effects of basolateral amygdala lesions on neophobia,
learned taste aversions, and sodium appetite in rats.” Journal of Comparative and
Physiological Psychology, 87(4):622-643
Nader, K, Schafe, GE & LeDoux, JE (2000) “Fear memories require protein synthesis n
the amygdala for reconsolidation after retrieval.” Nature, 406:722-726
Nagel, JA & Kemble, ED (1976) “Effects of amygdaloid lesions on the performance of
rats in four passive avoidance tasks.” Physiology & Behavior, 17:245-250
Nakashima, M, Uemura, M, Yasui, K, Ozaki, HS, Tabata, S & Taen, A (2000) “An
anterograde and retrograde tract-tracing study on the projections from the thalamic
gustatory area in the rat: distribution of neurons projecting to the insular cortex and
amygdaloid complex.” Neuroscience Research, 36:297-309
Navarro, M, Spray, KJ, Cubero, I, Thiele, TE & Bernstein, IL (2000) “cFos induction
during conditioned taste aversion expression varies with aversion strength.” Brain
Research, 887:450-453
Neylan, TC (1997) “Temporal lobe and behavior: Klüver and Bucy’s classic.” Journal of
Neuropsychiatry, 9(4):606
Niijima, A & Yamamoto, T (1994) “The effects of lithium chloride on the activity of the
afferent nerve fibers from the abdominal visceral organs in the rat.” Brain Research
Bulletin, 35(2):141-145
Nishijo, H, Uwano, T, Tamura, R & Ono, T (1998) “Gustatory and multimodal neuronal
responses in the amygdala during licking and discrimination of sensory stimuli in
awake rats.” Journal of Neurophysiology, 79:21-36
Norgren, R (1978) “Projections from the nucleus of the solitary tract in the rat.”
Neuroscience, 3(2):207-218
253
O’Carroll, CM & Morris, RGM (2004) “Heterosynaptic co-activation of glutamatergic
and dopaminergic afferents is required to induce persistent long-term potentiation.”
Neuropharmacology, 47:324-332
Ogawa, H & Kiasaku, J (2004) “Physiological characteristics of the solitario-parabrachial
relay neurons with tongue afferent inputs in rats.” Experimental Brain Research,
48(3):362-368
Oldfield, BJ & McKinley, MJ (2004) “Circumventricular organs.” In G. Paxinos (Ed.)
The rat nervous system, 3
rd
Ed. Elsevier Academic Press: San Diego
Ottersen, OP (1982) “Connections of the amygdala of the rat. IV: Corticoamygdaloid and
intraamygdaloid connections as studied with axonal transport of horseradish
peroxidase.” Journal of Comparative Neurology, 205:30-48
Parent, MB & McGaugh, JL (1994) “Posttraining infusion of lidocaine into the amygdala
basolateral complex impairs retention of inhibitory avoidance training.” Brain
Research, 661:97-103
Parent, MB, Quirarte, GL, Cahill, L & McGaugh, JL (1995) “Spared retention of
inhibitory avoidance learning after posttraining amygdala lesions.” Behavioral
Neuroscience, 109(4):803-807
Parent, MB, Tomaz, C & McGaugh, JL (1992) “Increased training in an aversively
motivated task attenuates the memory-impairing effects of posttraining N-methyl-D-
aspartate-induced amygdala lesions.” Behavioral Neuroscience, 106(5):789-797
Parent, MB, West, M & McGaugh, JL (1994) “Memory of rats with amygdala lesions
induced 30 days after footshock-motivated escape training reflects degree of original
training.” Behavioral Neuroscience, 108(6):1080-1087
Parkinson, JA, Robbins, TW & Everitt, BJ (2000) “Dissociable roles of the central and
basolateral amygdala in appetitive emotional learning.” European Journal of
Neuroscience, 12:405-413
Pascoe, JP & Kapp, BS (1985) “Electrophysiological characteristics of amygdaloid
central nucleus neurons during Pavlovian fear conditioning in the rabbit.” Behavioral
Brain Research, 16:117-133
Pavlov, IP (1927) Conditioned reflexes: An investigation of the physiological activity of
the cerebral cortex (Translated by GV Anrep), Oxford University Press: London
Paxinos, G & Watson, C (1998) The rat brain in stereotaxic coordinates, Fourth Ed.
Academic Press, Inc: San Diego, CA
254
Pedreira, ME & Maldonado, H (2003) “Protein synthesis subserves reconsolidation or
extinction depending in reminder duration.” Neuron, 38:863-869
Pellegrino, L (1968) “Amygdaloid lesions and behavioral inhibition in the rat.” Journal
of Comparative and Physiological Psychology, 65:483-491
Perez-Jaranay, JM & Vives, F (1991) “Electrophysiological study of the response of
medial prefrontal cortex neurons to stimulation of the basolateral nucleus of the
amygdala in the rat.” Brain Research, 564:97-101
Phillips, AG & LePiane, FG (1978) “Electrical stimulation of the amygdala as a
conditioned stimulus in a bait-shyness paradigm.” Science, 201:536-538
Phillips, AG & LePiane, FG (1980) Disruption of conditioned taste aversion in the rat by
stimulation of amygdala: a conditioning effect, not amnesia.” Journal of Comparative
and Physiological Psychology, 94(4):664-674
Phillips, RG, & LeDoux, JE (1992) “Differential contribution of amygdala and
hippocampus to cued and contextual fear conditioning.” Behavioral Neuroscience,
106(2):274-285
Pitkänen, A (2000) “Connectivity of the rat amygdaloid complex.” In JP Aggleton (Ed.)
The Amygdala: A Functional Analysis, 2
nd
Ed. Oxford University Press: New York
Pitkänen, A, Stefanacci, L, Farb, CR, Go, G-G, LeDoux, JE & Amaral, DG (1995)
“Intrinsic connections of the rat amygdaloid complex: projections originating in the
lateral nucleus.” Journal of Comparative Neurology, 356:288-310
Powley, TL, Holst, M-C, Boyd, DB & Kelly, JB (1994) “Three-dimensional
reconstructions of autonomic projections to the gastrointestinal tract.” Microscopy
Research and Technique, 29:297-309
Przybyslawski, J & Sara, SJ (1997) “Reconsolidation of memory after its reactivation.”
Behavioural Brain Research, 84:241-246
Quirarte, GL, Galvez, R, Roozendaal, B & McGaugh, JL (1998) “Norepinephrine release
in the amygdala in response to footshock and opioid peptidergic drugs.” Brain
Research, 808:134-140
Quirarte, GL, Roozendaal, B & McGaugh, JL (1997) “Glucocorticoid enhancement of
memory storage involves noradrenergic activation in the basolateral amygdala.”
Proceedings of the National Academy of Sciences of the United States of America,
94:14048-14053
255
Quirk, GJ & Mueller, D (2008) “Neural mechanisms of extinction learning and retrieval.”
Neuropsychopharmacology, 33:56-72
Quirk, GJ, Pare, D, Richardson, R, Herry, C, Monfils, M-H, Schiller, D & Vicentic, A
(2010) “Erasing fear memories with extinction training.” Journal of Neuroscience,
30(45):14993-14997
Quirk, GJ, Repa, JC & LeDoux, JE (1995) “Fear conditioning enhances short-latency
auditory responses of lateral amygdala neurons: parallel recordings in the freely
behaving rat.” Neuron, 15:1029-1039
Ramirez-Lugo, L, Núñez-Jaramillo, L & Bermudez-Rattoni, F (2007) “Taste memory
formation: role of nucleus accumbens.” Chemical Senses, 32:93-97
Reddy, MM & Bures, J (1981) “Unit activity changes elicited in amygdala and neocortex
of anaesthetized rats by intraperitoneal injection of lithium chloride.” Neuroscience
Letters, 22:169-172
Reilly, S & Bornovalova, MA (2005) “Conditioned taste aversion and amygdala lesions
in the rat: a critical review.” Neuroscience and Biobehavioral Reviews, 29:1067-1088
Rescorla, RA (1974) “Effect of inflation of the unconditioned stimulus value following
conditioning.” Journal of Comparative and Physiological Psychology, 86(1):101-106
Rescorla, RA (2001) “Experimental extinction.” In Mowrer, RR & Klein, SB (Eds.)
Handbook of contemporary learning theory. Lawrence Earlbaum Associates: Mawah,
NJ
Revusky, SH, Smith, Jr, MH & Chalmers, DV (1971) “Flavor preference: effects of
ingestion-contingent intravenous saline or glucose.” Physiology & Behavior,
6(4):341-343
Riley, AL & Tuck, DL (1985) “Conditioned taste aversions: a behavioral index of
toxicity.” Annals of the New York Academy of Sciences, 443:272-292
Ritter, S, McGlone, JJ & Kelley, KW (1980) “Absence of lithium-induced taste aversion
after area postrema lesion.” Brain Research, 201(2):501-506
Rizvi, TA, Ennis, M, Behbehani, MM & Shipley, MT (1991) “Connections between the
central nucleus of the amygdala and the midbrain periaqueductal gray: topography
and reciprocity.” Journal of Comparative Neurology, 303:121-131
Rogan, MT, Stäubli, UV & LeDoux, JE (1997) “Fear conditioning induces associative
long-term potentiation in the amygdala.” Nature, 390:604-607
256
Roldan, G & Bures, J (1994) “Tetrodotoxin blockade of amygdala overlapping with
poisoning impairs acquisition of conditioned taste aversion in rats.” Behavioural
Brain Research, 65:213-219
Rollins, BL, Stines, SG, McGuire, HB & King, BM (2001) “Effects of amygdala lesions
on body weight, conditioned taste aversion, and neophobia.” Physiology & Behavior,
72:735-742
Rolls, BJ & Rolls, ET (1973a) “Effects of lesions in the basolateral amygdala on fluid
intake in the rat.” Journal of Comparative and Physiological Psychology, 83(2):240-
247
Rolls, ET & Rolls, BJ (1973b) “Altered food preferences after lesions in the basolateral
region of the amygdala in the rat.” Journal of Comparative and Physiological
Psychology, 83(2):248-259
Roman, C & Reilly, S (2007) “Effects of insular cortex lesions on conditioned taste
aversion and latent inhibition in the rat.” European Journal of Neuroscience,
26(9):2627-2632
Romanski, LM, Clugnet, M-C, Bordi, F & LeDoux, JE (1993) “Somatosensory and
auditory convergence in the lateral nucleus of the amygdala.” Behavioral
Neuroscience, 107(3):444-450
Romanski, LM & LeDoux, JE (1992) “Equipotentiality of thalamo-amygdala and
thalamo-cortico-amygdala circuits in auditory fear conditioning.” Journal of
Neuroscience, 12(11):4501-4509
Roozendaal, B, Koolhaas, JM & Bohus, B (1991) “Central amygdala lesions affect
behavioral and autonomic balance during stress in rats.” Physiology & Behavior,
50:777-781
Roozendaal, B & McGaugh, JL (1996) “Amygdaloid nuclei lesions differentially affect
glucocorticoid-induced memory enhancement in an inhibitory avoidance task.”
Neurobiology of Learning and Memory, 65:1-8
Roozendaal, B & McGaugh, JL (1997) “Glucocorticoid receptor agonist and antagonist
administration into the basolateral but not central amygdala modulates memory
storage.” Neurobiology of Learning and Memory, 67:176-179
Roozendaal, B, Quirarte, GL & McGaugh, JL (2002) “Glucocorticoids interact with the
basolateral amygdala β-adrenoceptor-cAMP/PKA system in influencing memory
consolidation.” European Journal of Neuroscience, 15:553-560
257
Rosen, JB & Davis, M (1988) “Enhancement of acoustic startle by electrical stimulation
of the amygdala.” Behavioral Neuroscience, 102(2):195-202
Rosen, JB Hitchcock, JM, Sananes, CB, Miserendino, MJD & Davis, M (1991) “A direct
projection from the central nucleus of the amygdala to the acoustic startle pathway:
anterograde and retrograde tracing studies.” Behavioral Neuroscience, 105(6):817-
825
Rosenblum, K, Meiri, N & Dudai, Y (1993) “Taste memory: the role of protein synthesis
in gustatory cortex.” Behavioral and Neural Biology, 59(1):49-56
Rudy, JW, Barrientos, RM, & O’Reilly, RC (2002) “Hippocampal formation supports
conditioning to memory of a context.” Behavioral Neuroscience, 116(4):530-538
Rudy, JW, Huff, NC & Matus-Amat, P (2004) “Understanding contextual fear
conditioning: insights from a two-process model.” Neuroscience and Biobehavioral
Reviews, 28:675-685
Sah, P, Faber, SL, de Armentia, ML & Power, J (2003) “The amygdaloid complex:
anatomy and physiology.” Physiological Reviews, 83:803-834
Sakai, N & Yamamoto, T (1999) “Possible routes of visceral information in the rat brain
in formation of conditioned taste aversion.” Neuroscience Research, 35(1):53-61
Salinas, JA, Introini-Collison, IB, Dalmaz, C & McGaugh, JL (1997) “Posttraining
intraamygdala infusions of oxotremorine and propranolol modulate storage of
memory for reductions in reward magnitude.” Neurobiology of Learning and
Memory, 68:51-59
Salinas, JA, Parent, MB & McGaugh, JL (1996) “Ibotenic acid lesions of the amygdala
basolateral complex or central nucleus differentially effect [sic] the response to
reductions in reward.” Brain Research, 742:283-293
Sananes, CB & Davis, M (1992) “N-methyl-D-aspartate lesions of the lateral and
basolateral nuclei of the amygdala block fear-potentiated startle and shock
sensitization of startle.” Behavioral Neuroscience, 106(1):72-80
Sanes, JR & Jessell, TM (2000) “The formation and regeneration of synapses.” In ER
Kandel, JH Schwartz & TM Jessell (Eds.) Principles of Neural Science, 4
th
Ed.
McGraw-Hill: San Francisco
Saper, CB & Loewy, AD (1980) “Efferent connections of the parabrachial nucleus in the
rat.” Brain Research, 197(2):291-317
258
Savonenko, A, Werka, T, Nikolaev, E, Zieliñski, K & Kaczmarek, L (2003) “Complex
effects of NMDA receptor antagonist APV in the basolateral amygdala on acquisition
of two-way avoidance reaction and long-term fear memory.” Learning & Memory,
10:293-303
Schafe, GE & Bernstein, IL (1996) “Forebrain contribution to the induction of a
brainstem correlate of conditioned taste aversion: I. The amygdala.” Brain Research,
741:109-116
Schafe, GE, Doyere, V & LeDoux, JE (2005) “Tracking the fear engram: the lateral
amygdala is an essential locus of fear memory storage.” Journal of Neuroscience,
25(43):10010-10015
Schafe, GE LeDoux, JE (2000) “Memory consolidation of auditory Pavlovian fear
conditioning requires protein synthesis and protein kinase A in the amygdala.”
Journal of Neuroscience, 20(RC96):1-5
Schafe, GE, Sollars, SI & Bernstein, IL (1995) “The CS-US interval and taste aversion
learning: a brief look.” Behavioral Neuroscience, 109(4):799-802
Schafe, GE, Thiele, TE & Bernstein, IL (1998) “Conditioning method dramatically alters
the role of amygdala in taste aversion learning.” Learning & Memory, 5:481-492
Schiller, D, Monfils, M-H, Raio, CM, Johnson, DC, LeDoux, JE & Phelps, EA (2010)
“Preventing the return of fear in humans using reconsolidation update mechanisms.”
Nature, 463(7):49-54
Schoenfeld, TA & Hamilton, LW (1981) “Disruption of appetite but not hunger of satiety
following small lesions in the amygdala of rats.” Journal of Comparative and
Physiological Psychology, 95(4):565-587
Selden, NRW, Everitt, BJ, Jarrard, LE & Robbins, TW (1991) “Complimentary roles for
the amygdala and hippocampus in aversive conditioning to explicit and contextual
cues.” Neuroscience, 42(2):335-350
Senba, E & Ueyama, T (1997) “Stress-induced expression of immediate early genes in
the brain and peripheral organs of the rat.” Neuroscience Research, 29(3):183-207
Shapiro, RE & Miselis, RR (1985) “The central neural connections of the area postrema
of the rat.” Journal of Comparative Neurology, 234(3):344-364
Simbayi, LC, Boakes, RA & Burton, MJ (1986) “Effects of basolateral amygdala lesions
on taste aversions produced by lactose and lithium chloride in the rat.” Behavioral
Neuroscience, 100(4):455-465
259
Slotnick, BM (1973) “Fear behavior and passive avoidance deficits in mice with
amygdala lesions.” Physiology & Behavior, 11:717-720
Smith, JC & Schaeffer, RW (1967) “Development of water and saccharin preferences
after simultaneous exposures to saccharin solution and gamma rays.” Journal of
Comparative and Physiological Psychology, 63(3):434-438
Smith, OA & DeVito, JL (1984) “Central neural integration for the control of autonomic
responses associated with emotion.” Annual Review of Neuroscience, 7:43-65
Sotres-Bayon, F, Bush, DEA & LeDoux, JE (2004) “Emotional preservation: an update
in prefrontal-amygdala interactions in fear extinction.” Learning & Memory, 11:525-
535
Spear, NE & Mueller, CW (1984) “Consolidation as a function of retrieval.” In
Weingartner, H & Parker, ES (Eds.) Memory consolidation: psychobiology of
cognition. Lawrence Earlbaum Associates: Hillsdale, NJ
Spector, AC, Breslin, P & Grill, HJ (1988) “Taste reactivity as a dependent measure of
the rapid formation of conditioned taste aversion: a tool for the neural analysis of
taste-visceral associations.” Behavioral Neuroscience, 102(6):942-952
St. Andre, J, Albanos, K & Reilly, S (2007) “C-fos expression in the rat brain following
lithium chloride-induced illness.” Brain Research, 1135:122-128
St. Andre, J & Reilly, S (2007) “Effects of central and basolateral amygdala lesions on
conditioned taste aversion and latent inhibition.” Behavioral Neuroscience,
121(1):90-99
Stevenson, CW & Gratton, A (2004) “Role of basolateral amygdala dopamine in
modulating prepulse inhibition and latent inhibition in the rat.” Psychopharmacology,
176(2):139-145
Stolerman, IP, Pilcher, CWT & D’Mello, GD (1978) “Stereospecific aversive property of
narcotic antagonists in morphine-free rats.” Life Sciences, 22(19):1755-1762
Stollhoff, N, Menzel, R & Eisenhardt, D (2005) “Spontaneous recovery from extinction
depends on the reconsolidation of the acquisition memory in an appetitive learning
paradigm in the honeybee (Apis mellifera).” Journal of Neuroscience, 25(18):4485-
4492
Suzuki, A, Josselyn, SA, Frankland, PW, Masushige, S, Silva, AJ & Kida, S (2004)
“Memory reconsolidation and extinction have distinct temporal and biochemical
signatures.” Journal of Neuroscience, 24(20):4787-4795
260
Swanson, LW & Petrovich, GD (1998) “What is the amygdala?” Trends in
Neurosciences, 21:323-331
Taubenfeld, SM, Milekic, MH, Monti, B & Alberini, CM (2001) “The consolidation of
new but not reactivated memory requires hippocampal C/EBP β.” Nature
Neuroscience, 4(8):813-818
Terman, GW, Shavit, Y, Lewis, JW, Cannon, JT & Liebeskind, JC (1984) “Intrinsic
mechanisms of pain inhibition: activation by stress.” Science, 226(4680):1270-1277
Thatcher, RW & Kimble DP (1966) “Effect of amygdaloid lesions on retention of an
avoidance response in overtrained and non-overtrained rats.” Psychonomic Science,
6(1):9-10
Thomas, GM & Huganir, RL (2004) “MAPK cascade signalling and synaptic plasticity.”
Nature Reviews Neuroscience, 5:173-183
Tomaz, C, Dickinson-Anson, H & McGaugh, JL (1992) “Basolateral amygdala lesions
block diazepam-induced anterograde amnesia in an inhibitory avoidance task.”
Proceedings of the National Academy of Sciences of the United States of America,
89:3615-3619
Touzani, K, Taghzouti, K & Velley, L (1997) “Increase of the aversive value of taste
stimuli following ibotenic acid lesion of the central amygdaloid nucleus in the rat.”
Behavioral Brain Research, 88:133-142
Tronel, S, Milekic, MH & Alberini, CM (2005) “Linking new information to a
reactivated memory requires consolidation and not reconsolidation mechanisms.”
Public Library of Science Biology, 3(9):1630-1638
Tronson, NC & Taylor, JR (2007) “Molecular mechanisms of memory reconsolidation.”
Nature Reviews Neuroscience, 8:262-275
Tronson, NC, Wiseman, SL, Olausson, P & Taylor, JR (2006) “Bidirectional behavioral
plasticity of memory reconsolidation depends on amygdalar protein kinase A.”
Nature Neuroscience, 9(2):167-169
Tucci, S, Rada, P & Hernandez, L (1998) “Role of glutamate in the amygdala and lateral
hypothalamus in conditioned taste aversion.” Brain Research, 813:44-49
Tucker, AR & Oei, TPS (1982) “Protein synthesis inhibition and amnesia for saccharin
aversion memory in rats after intra-cisternal administration of cycloheximide.”
Physiology & Behavior, 28:1025-1028
261
Turner, BH & Herkenham, M (1991) “Thalamoamygdaloid projections in the rat: a test
of the amygdala’s role in sensory processing.” Journal of Comparative Neurology,
313:295-325
Ursin, H (1965) “Effect of amygdaloid lesions on avoidance behavior and visual
discrimination in cats.” Experimental Neurology, 11:298-317
Vazdarjanova, A & McGaugh, JL (1998) “Basolateral amygdala is not critical for
cognitive memory of contextual fear conditioning.” Proceedings of the National
Academy of Sciences of the United States of America, 95:15003-15007
Vianna, MRM, Szapiro, G, McGaugh, JL, Medina, JH & Izquierdo, I (2001) “Retrieval
of memory for fear-motivated training initiates extinction requiring protein synthesis
in the rat hippocampus.” Proceedings of the National Academy of Sciences of the
United States of America, 98(21):12251-12254
Villanueva, L, Desbois, C, Le Bars, D & Bernard, J-F (1998) “Organization of
diencephalic projections from the medullary subnucleus reticularis dorsalis and the
adjacent cuneate nucleus: a retrograde and anterograde tracer study in the rat.”
Journal of Comparative Neurology, 390:133-160
Walker, DL, Ressler, KJ, Lu, K-T & Davis, M (2002) “Facilitation of conditioned fear
extinction by systemic administration or intra-amygdala infusions of D-cycloserine as
assessed with fear-potentiated startle in rats.” Journal of Neuroscience, 22(6):2243-
2351
Walker, MP & Stickgold, R (2004) “Sleep-dependent learning and memory
consolidation.” Neuron, 44:121-133
Wang, C-C, Willis, WD & Westlund, KN (1999) “Ascending projections from the area
around the spinal cord central canal: a Phaseolus vulgaris leucoagglutinin study in
rats.” Journal of Comparative Neurology, 415:341-367
Wang, X-Y, Zhao, M, Ghitza, UE, Li, Y-Q & Lu, L (2008) “Stress impairs
reconsolidation of drug memory via glucocorticoid receptors in the basolateral
amygdala.” Journal of Neuroscience, 28(21):5602-5610
Weiskrantz, L (1956) “Behavioral changes associated with ablation of the amygdaloid
complex in monkeys.” Journal of Comparative and Physiological Psychology,
49(4):381-391
Weisz, DJ, Harden, DG & Xiang, Z (1992) “Effects of amygdala lesions on reflex
facilitation and conditioned response acquisition during nictitating membrane
response conditioning in rabbit.” Behavioral Neuroscience, 106(2):262-273
262
Whalen, PJ, Kapp, BS & Pascoe, JP (1994) “Neuronal activity within the nucleus basalis
and conditioned neocortical electroencephalographic activation.” Journal of
Neuroscience, 14(3):1623-1633
White, NM (1989) “Reward or reinforcement: What’s the difference?” Neuroscience and
Biobehavioral Reviews, 13:181-186
Wilcox, RR (2003a) Applying contemporary statistical techniques. Academic Press: San
Diego
Wilcox, RR (2003b) Fundamentals of modern statistical methods: Substantially
improving power and accuracy. Springer-Verlag: New York
Wilensky, AE, Schafe, GE, Kristensen, MP & LeDoux, JE (2006) “Rethinking the fear
circuit: the central nucleus of the amygdala is required for the acquisition,
consolidation, and expression of Pavlovian fear conditioning.” Journal of
Neuroscience, 26(48):12387-12396
Wilensky, AE, Schafe, GE & LeDoux, JE (2000) “The amygdala modulates memory
consolidation of fear-motivated inhibitory avoidance learning but not classical
conditioning.” Journal of Neuroscience, 20(18):7059-7066
Wilkins, EE & Bernstein, IL (2006) “Conditioning method determines patterns of c-fos
expression following novel taste-illness pairing.” Behavioural Brain Research,
169:93-97
Willis, WD & Westlund, KN (1997) “Neuroanatomy of the pain system and of the
pathways that modulate pain.” Journal of Clinical Neurophysiology, 14(1):2-31
Willis, WD, Westlund, KN & Carlton, SM (2004) “Pain system.” In G Paxinos (Ed.) The
rat nervous system, 3
rd
Ed. Elsevier Academic Press: San Diego
Winters, BD, Tucci, MC & DaCosta-Furtado, M (2009) “Older and stronger object
memories are selectively destabilized by reactivation in the presence of new
information.” Learning & Memory, 16:545-553
Woods, AM & Bouton, ME (2006) “D-Cycloserine facilitates extinction but does not
eliminate renewal of the conditioned emotional response.” Behavioral Neuroscience,
120(5):1159-1162
Wu, PC (2002) Central limit theorem comparing means, trimmed means, one-step M-
estimators and modified one-step M-estimators under non-normality. Unpublished
dissertation, University of Southern California: Los Angeles
263
Yamamoto, T (1998) “Electrophysiology of CTA.” In J Bures, F Bermudez-Rattoni & T
Yamamoto (Eds.) Conditioned taste aversion: Memory of a special kind. Oxford
University Press: New York
Yamamoto, T (2007) “Brain regions responsible for the expression of conditioned taste
aversion in rats.” Chemical Senses, 32:105-109
Yamamoto, T, Azuma, S & Kawamura, Y (1984) “Functional relations between the
cortical gustatory area and the amygdala: electrophysiological and behavioral studies
in rats.” Experimental Brain Research, 56:23-31
Yamamoto, T & Fujimoto, Y (1991) “Brain mechanisms of taste aversion learning in the
rat.” Brain Research Bulletin, 27:403-406
Yamamoto, T, Fujimoto, Y, Shimura, T & Sakai, N (1995) “Conditioned taste aversion in
rats with excitotoxic brain lesions.” Neuroscience Research, 22(1):31-49
Yamamoto, T, Sako, N, Sakai, N & Iwafune, A (1997) “Gustatory and visceral inputs to
the amygdala of the rat: conditioned taste aversion and induction of c-fos-like
immunoreactivity.” Neuroscience Letters, 226:127-130
Yamamoto, T, Shimura, T, Sako, N, Yasoshima, Y & Sakai, N (1994) “Neural substrates
for conditioned taste aversion in the rat.” Behavioural Brain Research, 65(2):123-137
Yasoshima, Y, Morimoto, T & Yamamoto, T (2000) “Different disruptive effects on the
acquisition and expression of conditioned taste aversion by blockades of amygdalar
ionotropic and metabotropic glutamatergic receptor subtypes in rats.” Brain
Research, 869:15-24
Yasoshima, Y, Scott, TR & Yamamoto, T (2006) “Memory-dependent c-Fos expression
in the nucleus accumbens and extended amygdala following the expression of a
conditioned taste aversive [sic] in the rat.” Neuroscience, 141:35-45
Yasoshima, Y, Shimura, T & Yamamoto, T (1995) “Single unit responses of the
amygdala after conditioned taste aversion in conscious rats.” NeuroReport, 6:2424-
2428
Yasoshima, Y & Yamamoto, T (2005) “Effects of midazolam on the expression of
conditioned taste aversion in rats.” Brain Research, 1043:115-123
264
Zhang, X & Giesler, Jr, GJ (2005) “Response characteristics of spinothalamic tract
neurons that project to the posterior thalamus in rats.” Journal of Neurophysiology,
93:2552-2564
Zhuravin, IA & Bures, J (1991) “Extent of the tetrodotoxin induced blockade examined
by pupillary paralysis elicited by intracerebral injection of the drug.” Experimental
Brain Research, 83:687-690
Abstract (if available)
Abstract
Reconsolidation is a temporary plasticity that memories undergo when recalled, and these memories can be disrupted, usually through pharmacological intervention. Recently, a purely behavioral treatment was devised that was shown to disrupt reconsolidation of fear memories (Monfils, Cowansage, Klann & LeDoux, 2009). We applied a modified form of this treatment to the classically conditioned taste aversion paradigm, wherein rats received intraoral infusion of sucrose followed by injection of lithium chloride. During extinction each day, treatment rats received reconsolidation-extinction (REX), consisting of a brief reminder infusion of sucrose one hour before a one minute extinction infusion. Control rats received standard extinction (SEX), wherein they received only the one minute infusion. Aversive orofacial behaviors were tallied and compared by percentile bootstrap tests. Each rat was extinguished until it had recovered, and then 5 days later sucrose was given again to test spontaneous recovery, measured by comparing aversive behaviors on the last extinction day to behaviors displayed on the spontaneous recovery test. The REX group recovered faster, showing significantly fewer aversive behaviors by extinction day 4. Both groups showed spontaneous recovery, but the SEX group exhibited significantly more aversive behaviors during the spontaneous recovery test. These results indicate that the REX treatment disrupted the taste aversion memory. A second experiment, conducted similarly to the first but with an additional unconditioned sucrose-only group, examined expression of the immediate early gene c-Fos in the subnuclei of the lateral, central, and basal nuclei of the amygdala. Neural tissue was collected one hour after the first extinction test, allowing assessment of amygdala participation in reconsolidation.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The neural basis of estradiol conditioned taste aversion learning and estradiol anorexia
PDF
Inhibition of sexual behavior in naive male rats by infusion of vasopressin or V1b vasopressin receptor antagonist into the olfactory bulb
PDF
Age-related changes in cognitive functioning and their implications for the reacquisition of conditioned taste avoidances
PDF
Neuropsychological functioning in older adult long-term cancer survivors
PDF
The development of electrodermal fear conditioning from ages 3 to 8 years and relationships with antisocial behavior at age 8 years
Asset Metadata
Creator
Foster, Nicholas N.
(author)
Core Title
The amygdala and conditioned taste aversion
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Psychology
Publication Date
04/13/2011
Defense Date
03/03/2011
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
amygdala,c-Fos,conditioned taste aversion,extinction,learning and memory,OAI-PMH Harvest,reconsolidation
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Chambers, Kathleen C. (
committee chair
), Bechara, Antoine (
committee member
), Gatz, Margaret (
committee member
), Lavond, David G. (
committee member
), Levin, Janet (
committee member
)
Creator Email
nfoster@usc.edu,pharaoh.six@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m3742
Unique identifier
UC1170911
Identifier
etd-Foster-4465 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-450640 (legacy record id),usctheses-m3742 (legacy record id)
Legacy Identifier
etd-Foster-4465.pdf
Dmrecord
450640
Document Type
Dissertation
Rights
Foster, Nicholas N.
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
amygdala
c-Fos
conditioned taste aversion
extinction
learning and memory
reconsolidation