University of Southern California Dissertations and Theses

Synaptic circuits for information processing along the central auditory pathway

(USC Thesis Other)

Synaptic circuits for information processing along the central auditory pathway

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content I
Synaptic Circuits for Information Processing along the Central
Auditory Pathway
MU ZHOU
A DISSERTATION
PRESENTED TO THE FACULTY
OF THE UNIVERSITY OF SOUTHERN CALIFORNIA
IN CANDIDACY FOR THE DEGREE
OF DOCTOR OF PHILOSOPHY
(PHYSIOLOGY AND BIOPHYSICS) MENTOR: LI I ZHANG
December 2014
© Copyright by Mu Zhou, 2014.
All rights reserved.
II
Acknowledgements
First I want to thank my PhD advisor Dr. Li Zhang, for providing me the privilege of working in
his lab, for sharing with me his time and insights, and for supporting me tremendously during my
PhD studies. Many things I learnt from him will influence my career in the future. Also I want to
thank Dr. Huizhong Tao, who gave me many good advices on my projects and technical details.
Thanks are also due to my committee members, Dr. Alapakkam Sampath, Dr. Sarah Bottjer and
Dr. Robert Farley, for their helpful inputs through the qualifying and dissertation process.
Meanwhile, I feel very fortunate to work with and learn from many excellent young
scientists in the lab. Dr. Guangying Wu taught me about the recording techniques and DCN
surgery. Dr. Baohua Liu taught me general animal surgery procedures when I first joined the lab
and later helped with the neuron modeling work. Dr. Feixue Liang helped to optimize the awake
recording setup and helped substantially with data collections. In fact everyone I worked with in
Zhang and Tao labs helped me in the form of data collection, data analysis, programming,
molecular techniques and discussions. I am very grateful to be in this helpful and encouraging
environment.
I wish to thank my parents and my late grandparents for their unconditional love and
support.
Finally, my thanks to my wife, who encourages me a lot. I will never forget the time
explaining to her about my cochlear nucleus research on our honeymoon cruise trip.
III
Table of Contents
Acknowledgements.........................................................................................................................II
Table of Figure legends ................................................................................................................ VI
Abstract ...................................................................................................................................... VIII
Chapter 1: Introduction............................................................................................................... 1
1.1 Overview of auditory system ........................................................................................................ 1
1.1.1 Ear and auditory nerve.......................................................................................................... 1
1.1.2 Auditory brainstem ...............................................................................................................2
1.1.3 Inferior colliculus...................................................................................................................4
1.1.4 Medial geniculate body......................................................................................................... 5
1.1.5 Auditory cortex .....................................................................................................................5
1.2 Dorsal cochlear nucleus ................................................................................................................6
1.2.1 Anatomy................................................................................................................................7
1.2.2 Physiology .............................................................................................................................7
1.3 Methodology................................................................................................................................. 8
Chapter 2 Generation of Intensity Selectivity ......................................................................... 10
2.1 Introduction ................................................................................................................................10
2.2 Methods......................................................................................................................................12
2.2.1 Animal preparation and mapping of the DCN ....................................................................12
2.2.2 In vivo whole-cell and Loose-patch recordings...................................................................13
2.2.3 Data analysis .......................................................................................................................15
2.3 Results.........................................................................................................................................16
2.3.1 Intensity tuning functions of rat DCN neurons ...................................................................16
2.3.2 Excitatory and inhibitory synaptic inputs to DCN neurons .................................................21
2.3.3 Synaptic tuning profiles and I/E ratio..................................................................................25
IV
2.3.4 Neuronal modeling of the generation of intensity selectivity ............................................27
2.4 Discussions ..................................................................................................................................29
2.4.1 Intensity selectivity in central auditory system ..................................................................29
2.4.2 Synaptic mechanisms underlying intensity selectivity........................................................31
2.4.3 Potential underlying circuits ...............................................................................................32
Chapter 3 Generation of Temporal Diversity .......................................................................... 35
3.1 Introduction ................................................................................................................................35
3.2 Methods......................................................................................................................................36
3.2.1 Animal preparation .............................................................................................................36
3.2.2 In vivo whole-cell and loose-patch recordings....................................................................37
3.2.3 Sound stimulation ...............................................................................................................38
3.2.4 Data analysis .......................................................................................................................39
3.2.5 Fitting synaptic currents .....................................................................................................41
3.2.6 Dynamic clamp....................................................................................................................41
3.2.7 Statistics ..............................................................................................................................41
3.3 Results.........................................................................................................................................42
3.3.1 Temporal response patterns of rat DCN pyramidal neurons..............................................42
3.3.2 Excitatory and inhibitory synaptic inputs to DCN pyramidal neurons................................44
3.3.3 Diversity in temporal dynamics of excitation .....................................................................47
3.3.4 A synaptic mechanism for the generation of response diversity .......................................48
3.3.5 Potential excitatory sources for pyramidal neurons...........................................................51
3.4 Discussion....................................................................................................................................54
Chapter 4 Synaptic Mechanisms for Generating Off Responses ............................................ 57
4.1 Introduction ................................................................................................................................57
4.2 Methods......................................................................................................................................58
4.3 Results.........................................................................................................................................59
4.3.1 Characterization of DCN off responding neurons...............................................................59
4.3.2 Membrane potential responses of off responding neurons...............................................60
V
4.3.3 Synaptic mechanism for generating off responses.............................................................61
4.4 Discussion....................................................................................................................................62
4.4.1 Circuit basis for delayed decay of excitation ......................................................................63
4.4.2 Second component of inhibitory input...............................................................................63
4.4.3 Functional significance........................................................................................................64
Chapter 5 Modulation of Synaptic inputs in Auditory Cortex by Active Behavioral States... 66
5.1 Introduction ................................................................................................................................66
5.2 Methods......................................................................................................................................67
5.2.1 Awake animal preparation..................................................................................................67
5.2.2 In vivo whole-cell and loose-patch recordings in awake animals.......................................68
5.2.3 Optogenetically guided loose-patch recordings from PV neurons.....................................70
5.2.4 Silencing L1 with TTX...........................................................................................................71
5.2.5 Sound stimulation ...............................................................................................................72
5.2.6 Data analysis .......................................................................................................................72
5.2.7 Statistics ..............................................................................................................................75
5.3 Results.........................................................................................................................................76
5.3.1 Laminar-specific down-regulation of auditory responses ..................................................76
5.3.2 Behavioral state-dependent gain modulation....................................................................79
5.3.3 Balanced excitation and inhibition in quiescent state........................................................82
5.3.4 Scaling down of excitation and inhibition in active states..................................................85
5.3.5 Change of membrane properties in active states...............................................................88
5.3.6 Modulation of PV neuron activity.......................................................................................90
5.3.7 Contribution of L1-mediated suppression ..........................................................................92
5.4 Discussion....................................................................................................................................94
5.4.1 Behavioral state-dependent gain modulation ....................................................................94
5.4.2 Balanced excitation and inhibition in awake cortex ...........................................................96
5.4.3 L1 mediated suppression of L2/3 activity ...........................................................................97
Bibliography ............................................................................................................................... 100
VI
Table of Figure legends
Figure 2.1 Receptive field properties of rat DCN neurons .......................................................... 17
Figure 2.2 Summary of intensity tuning properties of several types of DCN neurons................ 19
Figure 2.3 Synaptic inputs to example DCN pyramidal neurons ................................................ 23
Figure 2.4 Summary of properties of excitation and inhibition underlying intensity tuning
functions of DCN pyramidal neurons ........................................................................................... 24
Figure 2.5 Synaptic mechanisms underlying the generation of intensity selectivity................... 26
Figure 2.6 A proposed model for the generation of intensity selectivity in the DCN ................. 33
Figure 3.1 Response diversity in the rat DCN ............................................................................. 43
Figure 3.2 Synaptic inputs underlying discharge patterns of three example pyramidal neurons 46
Figure 3.3 Summary of properties of synaptic inputs to the three types of pyramidal neuron.... 47
Figure 3.4 A synaptic mechanism for the response diversity in the DCN................................... 50
Figure 3.5 Potential circuit mechanisms for generating response diversity in the DCN ............. 53
Figure 4.1 Characterizing off responses in DCN......................................................................... 59
Figure 4.2 Membrane potential responses in off responding neurons ......................................... 60
Figure 4.3 Synaptic mechanism for generating off responses ..................................................... 62
Figure 5.1 Behavioral state-dependent modulation of spike responses in the mouse A1............ 78
Figure 5.2 Gain modulation of auditory responses by behavioral state....................................... 80
VII
Figure 5.3 Activity of thalamic neurons and summary of SNR under different behavioral states
....................................................................................................................................................... 82
Figure 5.4 Properties of synaptic responses in quiescence .......................................................... 83
Figure 5.5 Modulation of synaptic responses by behavioral state ............................................... 87
Figure 5.6 Modulation of resting membrane potential and resting conductance by behavioral state
....................................................................................................................................................... 89
Figure 5.7 Changes of activity of PV neurons between different behavioral states .................... 91
Figure 5.8 Role of L1 in the behavioral state-dependent L2/3-specific gain modulation ........... 93
Figure 5.9 Our proposed model for state-dependent suppression in L2/3 of A1......................... 97
VIII
Abstract
Sound information is transmitted into electrical signals in the inner ear and these signals, i.e. action
potentials, are passed and processed along the central auditory pathway. For perception of sound
to occur, the basic attributes of sound (e.g. location, intensity, duration, et al.) need to be extracted
and represented by neurons in the central auditory pathway. The synaptic circuit mechanisms
underlying the extracting of these sound information attributes, however, remain poorly
understood. During my PhD studies, I applied in vivo whole cell voltage clamp recording technique
in studying how fusiform neurons in the Dorsal Cochlear Nucleus (DCN) of rat integrate excitatory
and inhibitory synaptic inputs and generate novel spike patterns that presumably can encode
intensity and temporal properties of sound information. In an independent but related project, I
used head-fixed awake mice as a model to study how sound information processing could be
modulated by different animal behavioral states.
In my first project, we carried out in vivo whole-cell recordings from pyramidal neurons in
the rat DCN, where intensity selectivity first emerges along the auditory pathway. Our results
revealed that intensity-selective cells received fast-saturating excitation but slow-saturating
inhibition with intensity increments, whereas in intensity-nonselective cells excitation and
inhibition were similarly slow-saturating. The differential intensity tuning profiles of the non-
intensity-tuned excitation and inhibition qualitatively determined the intensity selectivity of output
responses. In addition, the selectivity was further strengthened by significantly lower
excitation/inhibition ratios at high intensity levels compared to intensity-nonselective neurons. Our
results demonstrate that intensity-selectivity in the DCN is generated by extracting the difference
between tuning profiles of intensity-nonselective excitatory and inhibitory inputs, which we
propose can be achieved through a differential circuit mediated by feedforward inhibition.
IX
In my second project, we studied how different temporal firing patterns (primary-like,
pauser and buildup) were generated in pyramidal neurons in the rat DCN. We found that primary -
likeneurons received strong fast-rising excitation, whereas pauser or buildup neurons
received accumulating excitation with a relatively weak fast rising phase followed by a slow rising
phase. Pauser cells had stronger fast-rising excitation than buildup cells. On the other hand,
inhibitory inputs to the three types of cells exhibited similar temporal patterns with a strong fast-
rising phase. Dynamic-clamp recordings demonstrated that the differential temporal patterns of
excitation could primarily account for the different discharge patterns. In addition, discharge
patterns in a single neuron varied in a stimulus-dependent manner, which could be attributed to a
modulation of excitation/inhibition ratio. Further studies of excitatory inputs to vertical and
cartwheel cells suggested that fast-rising and accumulating excitation are separately conveyed by
auditory nerve and parallel fibers respectively. The differential summation of excitatory input from
two sources may thus contribute to the generation of response diversity.
In my third project, we studied how transient offset response was generated in DCN
fusiform cells. We found that in off responding neurons, their excitatory input started to decay
later than inhibitory input did. This sharp increase of net excitation at sound offset caused the
strong offset depolarization. Shortly after this offset depolarization, a second phase of inhibitory
input accounted for the transient feature of the offset depolarization. Our results demonstrate that
a fine temporal interaction of excitatory and inhibitory inputs can generate precisely timed off
response in the rat DCN.
In my fourth project, we developed in vivo whole cell voltage-clamp recording techniques
in the auditory cortex of head-fixed awake mouse. We reported that sensory-evoked spike
responses of layer 2/3 (L2/3) excitatory cells were scaled down with preserved sensory tuning
X
when animals transitioned from quiescence to active behaviors, while L4 and thalamic responses
were unchanged. Whole-cell voltage-clamp recordings further revealed that tone-evoked synaptic
excitation and inhibition exhibited a robust functional balance. Changes of behavioral state caused
scaling down of excitation and inhibition at an approximately equal level in L2/3 cells, but no
synaptic changes in L4 cells. This laminar-specific gain control could be attributed to an
enhancement of L1-mediated inhibitory tone, with L2/3 parvalbumin inhibitory neurons
suppressed as well. Thus, L2/3 circuits can adjust the salience of output in accordance with
momentary behavioral demands while maintaining the sensitivity and quality of sensory
processing.
1
Chapter 1: Introduction
1.1 Overview of auditory system
Sound information greatly enriches our life experience. Our auditory system is evolved to be
sophisticated enough to distinguish even a tiny variation in symphony. The processing of sound
begins in the cochlear, where hair cells translate mechanical vibration into electrical signals.
These signals are relayed by spiral ganglion neurons via the eighth cranial nerve into the central
auditory system, where auditory signals are processed and passed along auditory brainstem,
midbrain, thalamus and cortex nuclei. Here I will provide the basic background information
about the auditory system.
1.1.1 Ear and auditory nerve
The ear is consisted of the external ear, middle ear and inner ear. The external ear collects sound
wave by the auricle and sends them along the ear canal to the eardrum. Since the auricle surface
is asymmetric, it is most efficient in collecting sound wave emanating from certain locations in
space relative to head position. Our ability to locate sound source, especially at the vertical plane,
critically depends on this asymmetric shape of the auricle(Oertel and Young, 2004).
Sound wave causes vibration of the eardrum, which causes vibration of three tiny ossicles
in the middle ear: the malleus, incus and stapes. The stapes inserts into the oval window, which
is an opening of the bony structure covering the cochlear. In this way the three tiny ossicles
transform the airborne vibration into vibration of the cochlear fluid. The lever actions of the three
2
tiny ossicles are regulated by middle ear muscles, which may protect the ear during loud noise
environment(Pickles, 2008).
The inner ear, or cochlear, is a fluid filled, coiled structure. The vibration of cochlear
fluid is transduced into electrical signals by hair cells. Hair cells are named for their hair bundles,
which comprise hundreds of stereocilia with different lengths. There are links emerging from the
shorter tips of stereocilia, called tip links. These tip links are coupled to mechanotransducer
channels. When cochlear fluid flows towards tallest stereocilia, the mechanotransducer channels
will be open, causing influx of cations into hair cells. Therefore, the bidirectional flow of
cochlear fluids will cause depolarization and hyperpolarization of hair cells. Different sound
frequency will cause maximum vibration at different location of the cochlear, so the tonotopic
map is preserved by locations along the cochlear.
The hair cells are targeted by bipolar spiral ganglion neurons, whose central processes
form the auditory nerve. At least 90% of the spiral ganglion neuron axons terminate on inner hair
cells, which form a single row in the cochlear. Each axon contacts with only one inner hair cell,
while each inner hair cell contacts with about 10 nerve fibers. Relatively fewer spiral ganglion
neurons target outer hair cells, which form three rows in the cochlear. Each nerve fiber branches
and contacts numerous outer hair cells. Whether the auditory nerves that carry information from
the outer hair cells contribute to our perception of hearing is still unclear.
1.1.2 Auditory brainstem
The auditory nerve fibers enter the central nervous system via the internal acoustic meatus and
terminate in the cochlear nucleus. Upon entering the cochlear nucleus, each auditory nerve splits
into two branches. One branch innervates the anterior ventral cochlear nucleus (AVCN), while
3
the other branch innervates the posterior ventral cochlear nucleus (PVCN) and the dorsal
cochlear nucleus. Since each auditory nerve innervates four different types of principal neurons,
its assumed that different attributes of auditory information are processed in parallel in the
cochlear nucleus.
Bushy cells in AVCN most likely encode the precise onset timing information. They
form endbulb of held synapses with auditory nerve and calyx of held synapses with
contralateral medial nucleus of trapezoid body (MNTB) neurons. Both of these two specialized
synapses ensure rapid transmission(Kandel et al., 2013). The axons of bushy cells are also larger
than those of other cells. Octopus cells in PVCN may detect the onset transients and synchrony
of sound patterns(Oertel et al., 2000). Their dendrites span a large area of afferent fibers,
receiving inputs from broad range of spectral inputs. Stellate/multipolar cells in PVCN provide a
continuous representation of sound energy over a narrow frequency range. They might be
important in sound localization based on intensity cues. Fusiform cells in DCN are proposed to
utilize the spectral cues to localize sound source in the vertical plane(Oertel and Young, 2004).
Meanwhile, fusiform cells are different from the other principal cell types in that they receive
stronger inhibition. This fact should enable them to generate many new features based on
integration of different patterns of synaptic inputs. More detailed background of DCN structure
and physiology will be given in Chapter 1.2.
Bushy cells project bilaterally to the medial superior olive(MSO) and lateral superior
olive(LSO) in the superior olivary complex(SOC). They also project contralaterally to MNTB.
Neurons in LSO compare the intensities of the stimuli at the two ears and derive the location of
sound source. Neurons in MSO compare the timing of the stimuli at the two ears. In birds and
mammals different circuit mechanisms are proposed for how MSO neurons detect sound location
4
based on interaural time difference(Grothe et al., 2010). Stellate/multipolar cells project to
ipsilateral DCN, LSO, periolivary nuclei and contralateral ventral nucleus of the lateral
lemniscus. Octopus cells project to contralateral periolivary nuclei and the ventral nucleus of the
lateral lemniscus. Fusiform cells directly project to contralateral inferior colliculus (IC). This is
the shortest auditory pathway to IC.
1.1.3 Inferior colliculus
Inferior colliculus is an obligatory stage for all sound information. All auditory pathways
ascending through the auditory brainstem converge there. While in the auditory brainstem many
features of sound information are already encoded by distinct neuronal firing patterns, these
features are further processed in IC. For example, intensity selective neurons first emerge in
fusiform cells in DCN. These cells have the so called type IV receptive field structures. The local
inhibitory circuits in DCN are thought to help generate these response patterns which are quite
different from those in the auditory nerve. In IC, some principal cells have O shape receptive
field structures. These cells are found to receive input from type IV cells in DCN and further
process this input to generate O shape receptive field, which is supposed to be more efficient
in coding sound intensity(Young and Davis, 2001).
Neurons in IC also generate novel response properties that can encode certain aspect of
sound information. One good example is the generation of duration selective neurons. This type
of neurons, which are tuned to sound duration, have not been observed at lower levels of the
auditory pathway. Studies suggest that these duration selective neurons receive two sources of
excitatory inputs: onset excitation and offset excitation(Casseday et al., 2000). If for a specific
duration of sound stimulation, the delayed onset excitation overlaps with the offset excitation,
5
the neuron will exhibit maximum firing rate in response to this sound duration. The fact that IC
receives inputs from all lower levels of auditory pathway makes it an idea place to receive
synaptic inputs of various latencies and generate duration selective responses.
1.1.4 Medial geniculate body
The medial geniculate body (MGB) represents the obligatory thalamic relay between IC and the
auditory cortex. It is divided into ventral, dorsal and medial divisions. Only the ventral part of
the medial geniculate body (MGBv) principally projects to the core areas of the auditory cortex.
MGBv mainly receives ascending inputs from the central nucleus of the inferior colliculus. There
are two relatively homogeneous neuronal groups in MGBv. The principal neurons send
thalamocortical projections to the auditory cortex. The interneurons target principal neurons
within the MGBv. Both of these two neuronal types receive inhibitory input from the reticular
nucleus of the thalamus.
1.1.5 Auditory cortex
The auditory cortex consists of core areas, surrounded by belt and parabelt areas. MGBv
principal neurons directly project to core areas. Auditory information is first analyzed in the core
areas and then in the belt and parabelt areas. In different species, the auditory core areas
generally consist of three regions: the primary auditory cortex (A1), anterior auditory field
(AAF) and posterior auditory field. Each of these three regions has a complete representation of
tonotopic map and their gradients form mirror images. For example, in A1, neurons represent
high frequencies in anterior regions and low frequencies in posterior regions, while in AAF it is
the opposite gradient.
6
The major projection neurons in the auditory cortex are pyramidal neurons. They are
organized into six layers. A simplified hierarchy structure in the auditory cortex is that sound
information flow in the sequences of MGBv, Layer 4 (L4), Layer 2/3 (L2/3) and deep layers.
Neurons in Layer 1(L1) and deep layers receive inputs from other cortical regions. Neurons in
L2/3 project to other cortical areas. Some neurons in Layer 5 (L5) project to subcortical regions,
such as IC. Some neurons in Layer 6 (L6) project to MGB.
The inhibitory neurons in the auditory cortex only consist of about 20% of the total
neuronal population. However, they play very important roles in processing the sound
information. All the inhibitory neurons in the auditory cortex could be divided into three non-
overlapping groups: parvalbumin(PV), somatostatin(SST) and 5-HT positive interneurons
(Rudy et al., 2011). PV cells target the soma or axon initiation segments of pyramidal cells and
provide the major inhibitory input to pyramidal cells. 5-HT interneurons consist of vasoactive
intestinal peptide (VIP) positive cells that specialize in disinhibition and non-VIP cells that
mainly exist in L1. The function of SST neurons are still less understood.
1.2 Dorsal cochlear nucleus
The dorsal cochlear nucleus receives major inputs directly from the auditory nerve. Nevertheless,
plenty of novel firing patterns are generated in fusiform cells, the projection neurons in DCN.
These new features of response properties are good candidates to encode different aspects of
sound information. Therefore, it is very important to understand how local circuits in DCN
achieve this initial sorting and encoding of sound information. Here I provide more details about
the anatomical organization and physiological properties of DCN neurons. If not specified, the
7
information provided in this chapter is consistent among previous studies in different species,
including cat, gerbil, rat and mouse.
1.2.1 Anatomy
DCN is a three layers structure. The middle layer is called fusiform layer, which consists of
fusiform cells. Fusiform cells are the principal neurons in DCN. They directly project to
contralateral IC. The superficial layer consists of local inhibitory neurons, including cartwheel
and stellate cells. The deep layer consists of inhibitory neurons called vertical/tuberculoventral
neurons and another type of projection neurons called giant cells(Young and Davis, 2001).
Fusiform cells receive auditory nerve inputs onto their basal dendrites and receive
parallel fiber inputs onto their apical dendrites. Parallel fibers are axons of granule cells which
receive mossy fiber inputs from both auditory and non-auditory sources. T-type multipolar cells
in PVCN also project into DCN and its still unclear whether they directly innervate fusiform
cells. Fusiform cells receive inhibitory inputs from vertical, cartwheel and stellate cells. D-type
multipolar cells in PVCN also project into DCN and its unknown whether they also provide
inhibitory inputs to fusiform cells.
1.2.2 Physiology
Historically, the response properties of DCN neurons are characterized by the response maps
which are based on responses to tones of different frequencies and sound intensities(Young and
Brownell, 1976)(Figure 2.1C-E). Three distinct cell types are distinguished based on these
response maps, or receptive fields. Type II cells are characterized by almost zero spontaneous
activities. Type III cells are characterized by the clear sideband inhibition. Type IV cells are
generally suppressed at high sound intensities near characteristic frequencies (CF). Intracellular
8
recordings with dye fillings demonstrate that in general type III and type IV correspond to
fusiform cells and type II corresponds to vertical cells(Smith and Oertel, 1983; Hancock and
Voigt, 2002a). In cat majority of fusiform cells show type IV response properties while in
rodents fusiform cells mostly show type III responses.
The response properties of DCN cells could also be characterized based on their temporal
response patterns to CF tone stimuli(Figure 3.1B). Buildup and pauser responses are
characteristic for fusiform cells. Chopper responses were recorded from fusiform cells in gerbils
(Hancock and Voigt, 2002a). Vertical cells exhibit primary-like responses, which presumably
inherit the response properties of the auditory nerves. In cat vertical cells also exhibit onset
responses(Rhode, 1999). Cartwheel cells can be easily distinguished by their complex spiking
responses(Manis et al., 1994). Giant cells are generally considered to have similar response
properties with fusiform cells. The response patterns of stellate cells and granule cells are still
unknown.
1.3 Methodology
The function of our brain is in principle determined by its structure. The structures of neural
circuits, including the central auditory pathway, are very complex. The total number of neurons
is huge and each neuron makes thousands of synapses with other neurons. Therefore, its very
challenging to understand the detailed neuronal connections, not to mention that its equally
important to understand the functional meaning of each connection.
We dissect the functional auditory circuits by recording from principal neurons in
different auditory nuclei. We divide the thousands of synapses received by the recorded neurons
9
into two groups: excitatory and inhibitory. By revealing the underlying excitatory and inhibitory
inputs, we can figure out how these individual neurons integrate their received synaptic inputs
and generate spike outputs.
The major technique we use is in vivo whole cell voltage clamp recording. By achieving
high quality whole cell recordings, we can clamp the cell membrane potential at different levels.
By clamping the cell at 0 mV, which is the reversal potential for glutamate receptor-mediated
excitatory currents, we can record pure inhibitory synaptic current. Similarly, by clamping the
cell at- 70 mV, which is the reversal potential for GABAA or Glycine receptor-mediated Cl

currents, we can record pure excitatory synaptic current. Combined with in vivo single unit
extracellular recordings and current clamp recordings, we can record the synaptic inputs,
subthreshold membrane potential change and suprathreshold spiking respond from the same
group of neurons. In this way, we can figure out the computations performed by individual
neurons.
10
Chapter 2 Generation of Intensity Selectivity
2.1 Introduction
In the sensory system of each modality, stimulus intensity must be represented as a fundamental
aspect of sensory input. Most sensory neurons have monotonic spike rate versus intensity level
functions, i.e. they encode intensity by increasing spike rate as intensity is increased. In the
central auditory system, however, another strategy is used by intensity selective neurons. The
spike rate of these neurons initially increases and peaks as sound intensity is increased, and then
decreases as sound intensity is further increased, resulting in a nonmonotonic response-level
function (Phillips and Kelly, 1989; Phillips et al., 1995). Intensity-selective neurons are observed
at every stage of the ascending central auditory pathway (Greenwood and Maruyama, 1965;
Brugge et al., 1969; Aitkin and Webster, 1972; Young and Brownell, 1976; Rouiller et al., 1983;
Aitkin 1991; Schreiner et al., 1992; Kuwabara and Suga, 1993; Phillips et al., 1995). In the rat,
the number of intensity-selective cortical neurons was found to increase in animals trained to
perform an intensity discrimination task, suggesting that intensity-selective neurons may be
required for the precise coding of sound loudness (Polley et al., 2004, 2006).
The neural basis for intensity selectivity remains not well understood. Previous studies
have mostly focused on late stages of the auditory neuraxis, in particular the cortex. There
several synaptic mechanisms have been proposed for sharpening intensity selectivity or even
generating intensity selectivity de novo (Shamma, 1985; Ojima and Murakami, 2002; Sutter and
Loftus, 2003; Wu et al., 2006; Tan et al., 2007; de la Rocha et al., 2008). Intensity selectivity in
the cortex may be partially inherited from intensity-selective outputs of previous processing
11
stages, as evidenced by the intensity-tuned excitatory inputs to cortical neurons (Wu et al., 2006;
Tan et al., 2007). Along the auditory pathway, intensity selective neurons are first observed in
the dorsal cochlear nucleus (DCN) (Young and Brownell, 1976). Since the ascending input to
cochlear nuclei is provided by the auditory nerve (AN), which responses are non-intensity-tuned
(Kiang et al., 1965; Sachs and Abbas, 1974), it is interesting to examine DCN neurons as to
understand how intensity selectivity is initially generated (Figure 2.1A). Based on spike
response properties of DCN neurons, it has been proposed that the generation of intensity
selectivity in the DCN may be attributed to a lower intensity threshold of excitation than
inhibition (Voigt and Young 1980; Nelken and Young, 1994; Davis et al., 1996; Spirou et al.
1999; Young and Davis 2001). However, since sound-evoked excitatory and inhibitory synaptic
inputs to individual DCN neurons have not been experimentally elucidated, this theory remains
to be tested.
In the current study, we performed in vivo whole-cell recordings from DCN neurons to
directly examine the synaptic mechanisms underlying intensity tuning. By isolating the
excitatory and inhibitory inputs evoked by the same sound stimuli, we found that the difference
in intensity threshold between excitation and inhibition alone is not sufficient for explaining
intensity selectivity in the DCN. On the other hand, the differential intensity tuning profiles of
excitation and inhibition is key to generating intensity selectivity.
12
2.2 Methods
2.2.1 Animal preparation and mapping of the DCN
All experimental procedures used in this study were approved under the Animal Care and Use
Committee at the University of Southern California. Experiments were carried out in a sound-
attenuation booth (Acoustic Systems). Female Sprague-Dawley rats (about 3 months old and
weighing 250 300 g) were anaesthetized with ketamine and xylazine (ketamine: 60 mg/kg;
xylazine: 8 mg/kg; i.p.). The body temperature was maintained at 37.5 by a feedback heating
system (Harvard Apparatus, MA). The animal was positioned with the left ear facing a calibrated
free-field speaker (Vifa, Denmark), and with a sound-attenuating plug placed in the right ear.
After opening the left part of the occipital bone, part of the cerebellum was aspirated to expose
the left DCN. Artificial cerebrospinal fluid (ACSF; in mM: 124 NaCl, 1.2 NaH2PO4, 2.5 KCl, 25
NaHCO3, 20 Glucose, 2 CaCl2, 1 MgCl2) was applied when necessary during the surgery and the
experiment to clean the DCN surface. To map the tonotopy, multiunit spike responses were
recorded with a parylene-coated tungsten microelectrode (2 M©, FHC) at 100- 300 ¼m below
the DCN surface. Signals were amplified, band-pass filtered between 0.3 and 10 kHz (Plexon
Inc.), and then thresholded to extract the timing of each spike using a custom-made software
(LabView, National Instrument). To obtain tonal receptive fields, pure tones (0.5- 64 kHz at 0.1
octave intervals, 50 ms duration, 3 ms ramp) at eight 10 dB spaced sound intensities (0- 70 dB
SPL) were delivered in a pseudo-random sequence. The time interval and intensity difference
between two sequential tones were set at 0.5- 1 s and no more than 30 dB, respectively, as to
minimize cross-interactions between the two stimuli. Characteristic frequency (CF), the
13
frequency at which neurons reliably responded to tones with the minimum intensity, was
determined for each recording site. As illustrated in Figure 2.1B, a clear low-to-high frequency
gradient along the lateral-medial axis can be observed in the rat DCN, similar as described
previously (Yajima and Hayashi, 1989; Kaltenbach and Lazor, 1991).
2.2.2 In vivo whole-cell and Loose-patch recordings
Whole-cell recordings (Wu, et al., 2008; Zhou et al., 2010, 2012; Sun et al., 2010) were made
with an Axopatch 200B amplifier (Molecular devices). Patch pipettes made from borosilicate
glass capillaries (Kimax) had an impedance of 3- 4 M©. Pipettes contained a potassium-based
solution(in mM): 130 K-gluconate, 4 MgATP, 0.3 GTP, 8 phosphocreatine, 10 HEPES, 11
EGTA, 5 KCl, 1 CaCl2, 2.5 fluorescein dextran, pH 7.3. The patch pipette was lowered into the
DCN at an angle of about 85
o
. The brain stem was then covered with 3.5% agar prepared in
warm ACSF. Whole cell capacitance was fully compensated and the initial series resistance (<
30 M©) was compensated for 40- 60%. Signals were filtered at 5 kHz and sampled at 10 kHz.
Only cells with resting membrane potential lower than 50 mV were studied. A- 10 mV
junction potential was corrected. Excitatory and inhibitory currents were recorded by clamping
the cell at- 70 mV and 0 mV respectively. The morphology of some recorded cells was
reconstructed by confocal fluorescence imaging of the cells filled with fluorescein dextran (0.2
mM, Invitrogen) in fixed brainstem sections.
In this study we specifically focused on fusiform cells in the pyramidal cell layer. We
controlled our recording depths according to the travel distance of the pipette tip. Because there
is no pial tissue covering the DCN and pipette penetrations caused little dimpling of the DCN
14
surface, there was a relatively precise correspondence between travel distance and recording
depth. Also for a consistent control of recording depth, we chose to record from the middle
frequency representation region, as in this region the DCN surface is parallel to the horizontal
plane. Our reported pyramidal neurons were recorded at 100- 250 ¼m depths below the surface,
corresponding to the pyramidal cell layer (Mugnaini et al., 1980; Wouterlood and Mugnaini,
1984). Due to the depth control, giant cells, another type of DCN projection neurons, were
presumably rarely encountered. The pyramidal cell type was confirmed by completely
constructed dendritic morphologies in four cells (Figure 2.3K, see Hancock and Voigt, 2002a,
b), and the somata location in the pyramidal cell layer was also confirmed by six partially
recovered cells (data not shown). In addition, the response properties of the recorded pyramidal
cells were consistent with those previously reported (Evans and Nelson, 1973; Young and
Brownell, 1976). For loose-patch recordings, pipettes were filled with ACSF. Loose seal
(100- 200 M©) was made from neurons, allowing spikes only from the patched cell to be
recorded. Recording was under the voltage-clamp mode. A command voltage was applied to
achieve a zero baseline current. Signals were filtered at 0.3- 10 kHz and sampled at 20 kHz. For
recordings from inhibitory neurons, pipettes of smaller tip openings (impedance ~10 M©) were
used as to overcome a bias towards larger cells (Wu et al., 2008). We specifically searched for a
type of inhibitory neurons, type II cells, at 250- 700 ¼m depths ( Mugnaini et al., 1980; Rhode,
1999) using tone and noise stimuli. We identified type II cells by their lack of spontaneous
activity and absence of responses to broadband noise stimuli at 60 dB sound pressure level
(SPL). Cells with complex spikes, potentially cartwheel cells (Portfors and Roberts, 2007), were
excluded from the analysis in this study.
15
2.2.3 Data analysis
Spikes recorded in the loose-patch configuration could be detected without ambiguity because
their amplitudes were normally higher than 100 pA, while the baseline fluctuation was less than
5 pA. Tone-driven spikes were counted within a 4- 60ms time window after the tone onset. The
average spontaneous spike number within the same time duration was subtracted from the
number of evoked spikes. Intensity threshold was defined at the tip of the TRF by the minimal
intensity level at which CF tones reliably evoked responses. The evoked response should be
higher than the average baseline activity by three standard deviation of the baseline fluctuation.
All the synaptic response traces evoked by the same test stimulus were averaged. The peak
amplitude of the average trace was determined by averaging pixels within a ±5 ms time window
centered on the maximum value. The integrated conductance was calculated by deriving the
synaptic conductance and then integrating conductance within 4- 60 ms after the tone onset.
Excitatory and inhibitory synaptic conductance were derived(Borg-Graham et al., 1998;
Anderson et al., 2000; Zhang et al., 2003; Wu et al., 2008; Sun et al., 2010) according to I = Ge
* (V- Ee) + Gi * (V- Ei). I is the amplitude of the synaptic current response at any time point
after subtraction of the baseline current; G e and Gi are the excitatory and inhibitory synaptic
conductance; V is the holding voltage, and E e(0 mV) and Ei(- 70mV) are the reversal potentials.
The clamping voltage V was corrected from the applied holding voltage (V h): V = Vh Rs * I,
where Rs is the effective series resistance. By holding the recorded cell at two different voltages
(the reversal potentials for excitatory and inhibitory currents respectively), G e and Gi were
resolved from the equation.
16
The membrane potential change caused by synaptic conductances could be derived
according to an integrate-and-fire neuron model (Somers et al., 1995; Liu et al., 2007; Zhang M
et al., 2011):
()()() [ ])()()()()()()( t V E t V G E t V t G E t V t G
C
dt
dt t V
m r m r i m i e m e m
+ - + - * + - * - = +
where Vm(t) is the membrane potential at time t, C the whole-cell capacitance, Gr the resting
leaky conductance, E r the resting membrane potential ( 60 mV). To simulate the spike response,
20 mV above the resting membrane potential was set as the spike threshold and a 5 ms refractory
period was used. C (20- 50 pF) was measured during the experiment and G r was calculated based
on the equation Gr = C * Gm / Cm, where Gm, the specific membrane conductance is 2e 5 S/cm
2
,
and Cm, the specific membrane capacitance is 1e 6 F/cm
2
(Hines, 1993; Stuart and Nelson,
1998).
2.3 Results
2.3.1 Intensity tuning functions of rat DCN neurons
It was previously proposed in the cat and gerbil DCN that intensity-selective neurons have lower
intensity thresholds of spike response than inhibitory neurons in DCN, and that this difference in
intensity threshold can explain the generation of the selectivity (Voigt and Young 1980; Nelken
and Young, 1994; Spirou et al. 1999; Young and Davis 2001; Oertel and Young 2004). We
examined this proposed mechanism in the rat by first characterizing intensity tuning functions of
its DCN neurons. After exposing the dorsal surface of the cochlear nucleus in the brain stem, in
vivo cell-attached loose-patch recordings were made from individual neurons in the DCN to
17
obtain their spike responses evoked by tones of different frequencies and intensities (see
Methods). In this study, we specifically focused on fusiform/pyramidal cells, the DCN projection
neurons located in layer 2 (Mugnaini et al., 1980; Wouterlood and Mugnaini, 1984, see
Figure 2.1 Receptive field properties of rat DCN neurons
A, Schematic drawings of intensity tuning function (rate vs. intensity level) of the auditory nerve (AN) as
well as of intensity-selective neurons in the DCN. Intensity selectivity is first generated in the DCN, and
may be inherited by later processing stages. B, Left, schematic diagram of the experimental setup. Sound
is applied to the ear ipsilateral to the recorded DCN. Right, an image of the exposed dorsal surface of the
medullar. The tonotopic map in the DCN is indicated by colored lines representing different CFs. The
labeling of the frequency axis is based on multiunit recordings in this animal, but is highly reproducible
between animals. L, lateral; A, anterior. Scale bar, 1mm. C, Spike tonal receptive field (TRF) of an
intensity-nonselective DCN pyramidal neuron, examined by the cell-attached recording. Top, each small
trace is a record of evoked spikes (one trial) by a 50ms tone of a particular frequency and intensity
combination. SPL, sound pressure level. Bottom left, color map depicts the average spike TRF (five
repetitions). Color represents the average spike rate (Hz). Bottom right, plot of spike rate evoked by a
characteristic frequency (CF) tone (6.1 kHz for this cell) vs. tone intensity. Bar = SEM. Note that the
spike rate monotonically increases with increasing intensity. D, An example intensity-selective DCN
pyramidal neuron. Data are presented in the same way as in C. Note that the cells intensity tuning at CF
(11.8 kHz) is a nonmonotonic function. E, An example type II neuron, identified by the sharp TRF, the
extremely low level of spontaneous activity, as well as the absence of responses to broadband noise
stimuli(not shown).
18
Methods). The frequency-intensity tonal receptive field (TRF) was reconstructed from recorded
spikes, and the characteristic frequency (CF) was determined at the intensity threshold of the
spike TRF (see Methods). The response-level function (i.e. intensity tuning) of the cell was then
tested by applying CF tones at different intensities for a number of repetitions. Two types of
pyramidal neurons were observed. Examples are given in Figure 2.1C and 2.1D. For the first
type, neurons exhibited typical V-shaped spike TRFs, and inhibitory side bands were apparent
when the spontaneous firing rate was relatively high (Figure 2.1C, upper panel and color map).
They exhibited monotonic response-level functions, as indicated by a continuous increase in
response level as the intensity of CF tones increased (Figure 2.1C, lower right). These neurons
can thus be viewed as intensity-nonselective, and are functionally similar to previously reported
type III or type I/III cells (Evans and Nelson, 1973; Young and Voigt, 1982; Davis et al., 1996).
In this work, we did not further separate them into type III (with inhibitory sidebands) and type
I/III (without apparent inhibitory sidebands) neurons. For the second type, the response-level
function for CF tones was monotonic: the spike rate first reached a peak and then declined at
higher intensity levels (Figure 2.1D). These neurons can be viewed as intensity-selective, and
are functionally consistent with previously reported type IV neurons (Evans and Nelson, 1973;
Young and Voigt, 1982; Davis et al., 1996). Their TRFs were not typical V-shaped, since the
responses at or near CF at high intensity levels were reduced.
We used an intensity selectivity index (ISI) to quantify the selectivity level. It was
defined as the ratio of the response at 30 dB above the preferred intensity or at the highest testing
intensity over the maximum response. A perfect monotonic response-level function would
generate an ISI of 1. As shown by the distribution of ISIs, pyramidal cells appeared to cluster
into three groups (Figure 2.2A). The first group (NS) had ISIs of 1 or close to 1, and the
19
average response-level function for this group was clearly monotonic (Figure 2.2B). According
to the ISI distribution and the previous criterion for defining intensity-selective neurons (Ding
and Voigt, 1997; Wu et al. 2006), in this study we defined intensity-selective neurons as having
an ISI of < 0.6. The second and third group (Figure 2.2A, Si and Sii) were therefo re
collectively intensity-selective neurons, with their average response-level function exhibiting a
strong nonmonotonicity (Figure 2.2C). The Si group, however, exhibited much strong
intensity selectivity than the Sii group. In the Si group, CF tones at high intensities could
completely suppress spiking activity, resulting in a negative response relative to the baseline
activity (Figure 2.2E). In the Siigroup, on the other hand, spike rate at the highest testing
Figure 2.2 Summary of intensity tuning properties of several types of DCN neurons
A, Distribution of intensity selectivity indices of pyramidal cells. Note that only a subset of nonselective
cells (ISI = 1) are shown in this graph. B, Average intensity tuning profile for nonselective (NS) pyramidal
cells. Bar = SEM. N = 17. C, Average tuning profile for selective (S) pyramidal cells. N = 21. D, Average
tuning profile for inhibitory type II neurons. Inset, distribution of ISIs for type II neurons. N = 20. E,
Average tuning profile for the Si group of intensity -selective pyramidal cells. N = 9. F, Average tuning
profile for the Sii group of intensity -selective pyramidal cells. N = 8. G, Average spontaneous firing rate
for different types of cells. NS, intensity-nonselective pyramidal; S, intensity-selective pyramidal. Bar =
SEM. *, p < 0.001, ANOVA with post hoc test. H, Average intensity threshold for different types of cells.
Bar = SEM. *, p < 0.05, ANOVA with post hoc test.
20
intensity was only reduced to about 40% of the maximum (Figure 2.2F). These two groups, one
strongly selective and the other moderately selective, may correspond to previously reported
type IV and type IV -t neurons respectively (Evans and Nelson, 1973; Davis et al., 1996).
Within the pyramidal cell population, only 13% (17 out of 129) of cells were intensity-selective.
This relatively small percentage of intensity-selective neurons was also reported in other rodents
(Davis et al., 1996; Navawongse and Voigt, 2009; Ma and Brenowitz, 2011).
Previously type II neurons have been thought to play an important role in creating
intensity selectivity. They have been demonstrated to be vertical cells in morphology and be
inhibitory in function (Davis and Voigt, 1997; Rhode, 1999). We specifically searched type II
cells in the deep layer (see Methods). They could be identified by little or no spontaneous firing
activity, absence of response to broadband noise stimuli but strongly responding to CF tones
(Evans and Nelson, 1973; Young and Voigt, 1982; Young and Davis, 2001). An example type II
neuron is shown in Figure 2.1E. The cell exhibited almost zero spontaneous activity, a narrow
V-shaped TRF, and monotonic intensity-tuning function. A summary of response-level function
for 20 type II neurons is shown in Figure 2.2D. All type II neurons were intensity-nonselective
with ISIs > 0.6 (Figure 2.2D, inset). Compared to pyramidal cells (both intensity nonselective
and selective) which have 20- 30 Hz spontaneous spike rates, the extremely low spontaneous
rate of type II neurons made them easily distinguished (Figure 2.2G). The type II neurons had
significantly higher intensity thresholds than both the intensity-nonselective and selective
pyramidal cells(Figure 2.2H), consistent with previous reports in cats and gerbils. The two
types of pyramidal cells did not differ in intensity threshold (Figure 2.2H). This result suggests
that in both types of pyramidal neurons the intensity threshold of inhibition may be higher than
21
that of excitation. Therefore, the generation of intensity selectivity may not be simply attributed
to a difference in intensity threshold between excitation and inhibition.
2.3.2 Excitatory and inhibitory synaptic inputs to DCN neurons
To reveal the synaptic basis for intensity selectivity, we carried out whole-cell current-clamp and
voltage-clamp recordings from the same DCN neuron (see Methods). Firstly in the current-clamp
recording mode, the spike TRF was obtained to determine the cells CF and whether the cell was
intensity-selective. CF-tone evoked excitatory and inhibitory synaptic responses at different
intensities were then recorded under the voltage-clamp mode. An example intensity-nonselective
cell is shown in Figure 2.3A-2.3C. The mapping of spike TRF revealed that the cells CF was at
15.3 kHz (Figure 2.3A). At this frequency, changes of spike response with increasing intensity
indicated that the cell was monotonic (Figure 2.3B, Sp). The subsequent voltage clamp
recording revealed that at the intensity threshold of spike response (marked as 0 dB), there was
significant tone-evoked excitatory input, but not inhibitory input(Figure 2.3B, Ex and In
respectively). From the plotting of peak amplitude of evoked synaptic currents (Figure 2.3C), it
is clear that both the excitatory and inhibitory inputs were gradually strengthened as intensity
increased, and that the intensity threshold of the inhibitory input was higher than the excitatory
input. An example intensity-selective cell is shown in Figure 2.3D-2.3F. In this cell, the
amplitudes of excitatory and inhibitory inputs also increased with intensity increments,
indicating that synaptic inputs per se are not intensity-selective. However, different from the
intensity-nonselective neuron, its excitatory input appeared to quickly reach a plateau level as
early as the intensity threshold for spike response was reached (Figure 2.3F).
22
More example intensity-nonselective and selective neurons are shown in Figure 2.3G,
2.3H and Figure 2.3I, 2.3J, respectively. For the intensity-nonselective neurons, excitatory and
23
inhibitory inputs increased with similar paces. In contrast, for the intensity-selective neurons
whose spike responses decreased at high intensities, the excitatory response increased faster than
the inhibitory response with increasing intensity.
The whole-cell recording allowed labelling of the recorded cells with dyes. Complete
dendritic morphologies were reconstructed for two intensity-nonselective and two selective
neurons (Figure 2.3K). The cells were all pyramidal cells located in layer 2 (i.e. pyramidal cell
layer), as evidenced by the presence of apical and basal dendrites as well as the spindle-like cell
body shape (Young and Davis, 2001). Six partially recovered cells were also located in the
pyramidal cell layer and were identified as fusiform cells based on the soma shape (data not
shown). There was no apparent difference in morphology between intensity-nonselective and
selective pyramidal neurons (Figure 2.3K). We also estimated the quality of our voltage-clamp
recordings by plotting the current-voltage (I/V) relationship for the recorded synaptic currents.
As shown by an example cell (Figure 2.3L), synaptic currents evoked by CF tones were
recorded under four different membrane potentials. The current amplitude at 1ms after the
Figure 2.3 Synaptic inputs to example DCN pyramidal neurons
A, Spike TRF of an example intensity-nonselective pyramidal neuron, examined by the whole-cell current-clamp
recording. The dotted box outlines the responses to CF (15.3 kHz) tones. Right, color map depicts the average spike
TRF. Color represents the average spike rate. B, Records of CF-tone evoked spikes, average traces (four repetitions) of CF-tone evoked excitatory (Exc) and inhibitory (Inh) responses at different intensities for the same cell as in A.
The intensity threshold of spike response is denoted as 0 dB. The short line on top of the trace indicates the duration
of the tone stimulation (50 ms). Color bar on the right represents the magnitude of responses. Maximum value for
the color scale is 100Hz, 700pA, 365pA respectively. Baseline is indicated by the dotted line. C, Plot of normalized
peak amplitude of excitation and inhibition vs. tone intensity. The maximum amplitude is given. Bar = s.d.. D, E, F,
An example intensity-selective pyramidal neuron. Data are displayed in a similar manner. Maximum value for the
color scale in E is 120Hz, 265pA, 425pA respectively. G, H, Two more example intensity-nonselective cells.
Maximum value for the color scale is (from left to right): 150Hz, 690pA, 580pA in G; 120Hz, 435pA, 370pA in H.
I, J, Two more example intensity-selective cells. Maximum value for the color scale is: 160Hz, 335pA, 430pA in I;
140Hz, 505pA, 500pA in J. K, Reconstructed morphology for two intensity-nonselective and two selective cells.
DCN layers are labeled. Note that the cell bodies were all located in layer 2. Scale, 50¼m. L, I-V curve for recorded
synaptic currents in an example cell. Red, average current amplitude within a 1 ms window 1 ms after the synaptic
response onset. Black, average current amplitude within a 1ms window 11 ms after the onset. R, correlation
coefficient. Note that the y-axis scale is different for the red and black plot. Inset, response traces at different
voltages (mV). Scale, 100pA, 10ms. M, amplitudes of CF-tone evoked synaptic currents at two voltages in all the
recorded cells. Amplitudes were measured within a 1 ms window at 1ms after the synaptic response onset.
24
response onset (Figure 2.3L, red) changed linearly with voltage, with the reversal potential
matching that expected for excitatory currents (0 mV). The current amplitude at 11ms after the
onset (Figure 2.3L, black) also changed linearly with voltage, but the reversal potential was- 46
mV, indicating a combination of excitatory and inhibitory currents at this time point. We next
examined CF tone-evoked synaptic currents at 1ms after the response onset in all the recorded
pyramidal cells (Figure 2.3M). The currents were all close to zero when the membrane was
Figure 2.4 Summary of properties of excitation
and inhibition underlying intensity tuning
functions of DCN pyramidal neurons
A, B, Average normalized intensity tuning
functions of intensity-nonselective (N = 9) and
selective (N = 5) neurons. The peak response
amplitude was measured. Bar = SEM. *, p < 0.05,
paired t-test. C, D, Average normalized intensity
tuning functions based on the integrated
conductance. Bar = SEM. *, p < 0.05, paired t-test.
E, F, Comparison of excitatory and inhibitory
tuning profiles between intensity-nonselective and
selective neurons based on peak amplitudes. Bar =
SEM. *, p < 0.05, t-test. G, Average intensity
threshold of excitation and inhibition for the two
types of neurons. Bar = SEM. *, p < 0.05, paired t-
test.
25
clamped at the excitatory reversal potential. The linear I/V relationship as well as the relatively
precise excitatory reversal potential as revealed suggest that voltage-clamp quality was
reasonably good in our experiments.
2.3.3 Synaptic tuning profiles and I/E ratio
We summarized synaptic tuning profiles from nine intensity-nonselective and five intensity
selective pyramidal neurons. As shown in Figure 2.4A and 2.4B by the normalized peak
response amplitude, intensity-nonselective and selective neurons differed dramatically in the
relationship between excitatory and inhibitory intensity tuning profiles. Excitation and inhibition
were similarly increasing with intensity increments in nonselective neurons, while in selective
neurons excitation saturated much faster than inhibition. The same conclusion can be made when
the integrated conductance was measured (Figure 2.4C, 2.4D, see Methods). Between intensity-
selective and nonselective cells, there was no difference in inhibitory tuning profile, while their
excitatory tuning profiles significantly differed in terms of slope (Figure 2.4E, 2.4F). Consistent
with the loose-patch recording results, excitation had a lower intensity threshold than inhibition
in both intensity-selective and nonselective neurons (Figure 2.4G). This intensity threshold
difference between excitation and inhibition was not different between the two types of
pyramidal cells (p > 0.1, t-test).
Besides intensity tuning profile, the relative and absolute strengths of excitatory or
inhibitory inputs may also contribute to intensity selectivity. We further compared the
amplitudes of excitatory and inhibitory inputs between the two types of pyramidal neurons. As
shown in Figure 2.5A, at the intensity threshold, excitation was stronger in selective than
nonselective neurons. However, at 50dB above the intensity threshold, excitation in selective
26
neurons became weaker than nonselective neurons, due to its early saturation with intensity
increments. On the other hand, the absolute level of inhibition was not significantly different
between the two types of pyramidal neurons at either intensity (Figure 2.5A). Figure 2.5B plots
the change of inhibition/excitation (I/E) ratio with intensity increments. The I/E ratio was more
Figure 2.5 Synaptic mechanisms underlying the
generation of intensity selectivity
A, Average peak amplitudes of excitation and
inhibition in intensity-nonselective (NS, N = 9) and
selective (S, N = 5) pyramidal neurons evoked by CF
tones at 0dB and 50dB (relative). *, p < 0.05, t-test. B,
Average I/E ratio plotted as a function of relative
intensity for nonselective and selective neurons. Bar =
SEM. *, p < 0.05, t-test. C, The difference between
onset latencies of excitatory and inhibitory responses
plotted against tone intensity. Bar = SEM. D, The onset
latency of excitation for all the recorded pyramidal
neurons (N = 14) and that of spiking response of
inhibitory neurons (N = 20). Bar = SEM. E, F, Average
intensity tuning profiles of spike responses recorded
experimentally and those derived by integrating the
experimentally obtained excitatory and inhibitory
responses in the neuron model for nonselective (N = 9) and selective (N = 5) neurons. Bar = SEM. G, H,
Average intensity tuning responses and scaled
excitatory responses (magenta, cyan). Top panel
profiles of spike responses derived by integrating
experimentally recorded excitatory and inhibitory
responses (black), and by integrating the recorded
inhibitory depicts how swapping procedures were
carried out. One swap, the scaling of excitatory
responses was according to the average excitatory
tuning profile of the other neuronal type, with the
response at the highest intensity fixed. Two swaps,
excitatory responses were further scaled (without
changing the intensity tuning profile) to achieve an I/E
ratio at the highest intensity the same as that observed
for the other neuronal type.
27
or less constant above the intensity threshold in nonselective neurons, whereas it kept increasing
in selective neurons. At the highest testing intensity, the selective neurons received relatively
stronger inhibition compared to the nonselective neurons (Figure 2.5B). As for the temporal
properties of synaptic responses, we found that the onset delay of inhibition relative to excitation
reduced with intensity increments for both intensity-selective and nonselective neurons (Figure
2.5C). This shortening of relative delay of inhibition can be accounted for at least partially by the
intensity-dependent changes of the excitatory response onset and the spike response onset of
inhibitory type II neurons (Figure 2.5D).
2.3.4 Neuronal modeling of the generation of intensity selectivity
We have shown that intensity-selective and nonselective neurons differ in excitatory tuning
profile and absolute excitatory strength. Intuitively, the differential intensity tuning profile of
excitation can account well for the functional difference between intensity-selective and
nonselective neurons. In the intensity-selective neuron, excitation is close to saturation at the
intensity threshold (it reaches more than 80% of the peak level), whereas at this intensity level
inhibition is nearly the lowest (Figure 2.4A, 2.4B). This may lead to a strong output response at
the intensity threshold. With intensity further increasing, inhibition catches up, and the level of
spike response may gradually reduce.
To demonstrate how much the observed excitatory and inhibitory tuning profiles can
account for the intensity tuning profile of output responses in a quantitative manner, we
employed a conductance-based single-compartment integrate-and-fire neuron model (Liu et al.,
2007; Zhou et al., 2012). Feeding it with experimentally obtained excitatory and inhibitory
synaptic conductances, we derived the expected membrane potential and spike responses at
28
different tone intensities on a cell by cell basis (see Methods). As shown in Figure 2.5E and
2.5F, the intensity tuning profile of derived spike responses could largely match that of recorded
responses for both intensity-nonselective and selective neurons. This suggests that the interplay
of somatically recorded excitatory and inhibitory responses can qualitatively determine the
tuning property of the cell. Next, to test whether the differential intensity tuning profiles of
excitatory input play a critical role in determining intensity-selective or nonselective properties,
we exchanged the excitatory tuning profiles between the intensity-selective and nonselective
groups. This was achieved in each individual cell by keeping the excitatory response at the
highest intensity fixed while scaling the response amplitudes at other intensities according to the
average excitatory tuning profile for the other neuronal type (Figure 2.5G, 2.5H, top panel). The
scaled excitatory response was then integrated with the original inhibitory response to derive
spike response. Through this one-step swapping procedure, the tuning of intensity-nonselective
neurons became slightly nonmonotonic (Figure 2.5G, magenta), while that of intensity-selective
neurons was changed to monotonic (Figure 2.5H, magenta).
We noticed that the swapping of excitatory tuning profiles only produced modestly
nonmonotonic spike responses. This apparent deviation from recorded intensity-selective
responses may be due to a difference in absolute strength of synaptic inputs between the two
types of pyramidal neurons (Figure 2.5A). For example, in nonselective neurons, the excitatory
response at the highest testing intensity is strong and may result in nearly saturated spike
response, leaving little room for the input at lower intensities to generate even stronger spike
response. In the following procedure, in addition to swapping the excitatory tuning profile we
also scaled the excitatory response amplitudes in general so that the I/E ratio at the highest
intensity was the same as that observed for the other neuronal type. In another word, we also
29
swapped the I/E ratio between the two types of cells. As shown in Figure 2.5G and 2.5H(cyan),
the two-step swapping procedure converted the intensity-nonselective responses to more in vivo-
like intensity-selective responses, while the intensity-selective responses were completely
switched to monotonic responses. These modeling results demonstrate that the differential tuning
relationship between excitation and inhibition plus the differential excitatory input strength can
explain the functional difference between intensity-selective and nonselective neurons. Thus,
these two factors may play key roles in generating intensity selectivity in the DCN.
2.4 Discussions
In this study, we applied in vivo whole-cell recordings to examine synaptic mechanisms
underlying the creation of intensity selectivity in the DCN. To our knowledge, this is the first
time excitatory and inhibitory synaptic inputs underlying functional selectivity at this earliest
stage of central auditory processing are elucidated. Our results confirmed that synaptic inputs to
DCN neurons all exhibit monotonic intensity tuning functions. Two synaptic mechanisms
contribute to the converting of monotonic inputs to nonmonotonic outputs. First, the differential
intensity tuning profiles of excitation and inhibition, i.e. fast saturating excitation but slow
saturating inhibition, are a fundamental basis for the generation of intensity selectivity. Second,
the lower level of excitation or relatively stronger inhibition at high intensities further
strengthens the intensity selectivity of output responses.
2.4.1 Intensity selectivity in central auditory system
Nonmonotonic intensity-tuned neurons have been widely observed along the ascending pathway
of the central auditory system (e.g. in the cochlear nucleus: Greenwood and Maruyama, 1965;
Young and Brownell, 1976, Ramachandran et al.,1999, inferior colliculus: Aitkin 1991;
30
Kuwabara and Suga, 1993, medial geniculate body: Aitkin and Webster, 1972; Rouiller et
al.,1983, and auditory cortex: Davies et al., 1956; Evans and Whitfield, 1964; Brugge et al.,
1969; Schreiner et al., 1992; Phillips et al., 1995; Wu et al., 2006; Tan et al., 2007; Sadagopan
and Wang, 2008). It has been thought that the selectivity can be generated de novo at later stages
of the auditory pathway. For example, in earlier studies of the auditory cortex it has been
postulated that an intensity-dependent increase of inhibition can result in nonmonotonic
response-level functions (Ojima and Murakami, 2002; Sutter and Loftus, 2003). These studies
were based on extracellular recordings of spike responses under two-tone stimuli or intracellular
recordings of tone-evoked membrane potential responses, both of which detect inhibition in an
indirect way. More recent whole-cell recording studies have revealed that excitatory inputs to
intensity-selective auditory cortical neurons are already well tuned, indicating that intensity
selectivity in the cortex can be partially inherited from previous stages, although local cortical
inhibition can further sharpen the selectivity (Wu et al., 2006; Tan et al., 2007; Zhang et al.,
2011). Thus, it is worthwhile to examine the first stage where the selectivity is initially
generated.
The receptive field maps of intensity-selective pyramidal cells in this study look different
from the circumscribed, O -shaped response maps often observed in the cortex (e.g. Phillips
and Kelly, 1989; Wu et al., 2006, Sadagopan and Wang, 2008), which have perfect intensity
selectivity. The DCN receptive field maps are not completely enclosed, and intensity tuning
functions at off-CF can be in fact monotonic. Possibility, along the ascending auditory pathway,
recruitments of inhibition (e.g. lateral inhibition) at each higher processing stage progressively
sculpt the receptive field map, in particular by silencing CF and off-CF responses at high
intensity levels (Sutter and Loftus, 2003, Sadagopan and Wang, 2008; de la Rocha et al., 2008;
31
Zhang LI et al., 2011; Wu et al., 2011). This may eventually lead to the emergence of O -
shaped receptive field maps in the cortex.
2.4.2 Synaptic mechanisms underlying intensity selectivity
To convert monotonic response-level functions to nonmonotonic functions, synaptic inhibition is
generally thought to be required (Shamma, 1985; Phillips, 1988; Calford and Semple, 1995;
Ding and Voigt, 1997). Previously three different scenarios have been postulated with regard to
how the excitatory-inhibitory interplay might generate intensity tuning. First, based on the
observation that inhibitory type II neurons exhibit a higher intensity threshold than their principal
neuron counterparts (Voigt and Young 1980; Spirou et al. 1999; Hancock and Voigt 2002a), it is
suggested that a lower intensity threshold of excitation than inhibition may result in the highest
E/I ratio and the strongest spike response at the excitatory intensity threshold (Nelken and Young
1996; Young and Davis 2001; Hancock and Voigt 2002a). Our data however indicate that in both
intensity-selective and nonselective neurons, the intensity threshold of excitation is about 10 dB
lower than inhibition. Therefore, a threshold difference alone does not necessarily produce
intensity tuning. Second, it has been found for intensity-selective cortical neurons that the onset
latency of inhibition relative to excitation shortens with intensity increments (Wu et al., 2006).
Modeling work further suggests that this intensity-dependent decrease in the temporal delay of
inhibition can sufficiently produce intensity-tuned output responses (Wu et al., 2006). In the
DCN, we found that the relative delay of inhibition was only slightly reduced with intensity
increments and that it was not different between intensity-selective and nonselective neurons
(Figure 2.5C). In addition, because DCN responses are much more sustained than cortical
responses (Wehr and Zador, 2003; Zhang et al., 2003; Tan et al., 2004), any small variations of
relative inhibitory delay would not affect the overall response level significantly. Third, a
32
theoretic study of thalamocortical networks has demonstrated that under certain network
properties unbalanced recruitments of excitation and inhibition can occur such that inhibition
grows slower than excitation with intensity increments (de la Rocha et al., 2008). Under such
unbalanced excitation and inhibition, the integrated output response peaks well before inhibition
reaches maximum. Consistent with this postulate, unbalanced recruitments of excitation and
inhibition occur in the DCN, and play a key role in generating intensity selectivity.
2.4.3 Potential underlying circuits
How can the unbalanced recruitments of excitation and inhibition be achieved? Previously, a
circuit with broad lateral inhibition was proposed for cortical neurons to result in a slower
recruitment of inhibition than excitation as the number of inhibitory neurons activated gradually
increases with increasing intensity (Sutter and Loftus, 2003; de la Rocha et al., 2008). This
mechanism requires that the receptive fields of inhibitory neurons are relatively broad, are
spectrally offset but have overlapping regions. Here in the DCN, we notice that the speed of
recruitment of inhibition does not differ between intensity-selective and nonselective neurons
(Figure 2.4F). It appears that it is the recruitment pattern of excitation that makes the difference.
The two different excitatory tuning patterns observed in intensity-nonselective and
selective neurons were reminiscent of the two types of intensity-dependent responses of auditory
nerve fibers reported previously: fast saturating and slow saturating (Sachs and Abbas, 1974;
Winter et al., 1990). This raises an interesting possibility that the different excitatory tuning
patterns may be conveyed from the two different types of auditory nerve inputs, as it has been
well documented that pyramidal neurons receive direct auditory nerve inputs (Oertel and Young,
2004). In addition, we found in two intracellularly recorded type II neurons that they received
33
slowly saturating excitatory input similar to intensity-nonselective pyramidal neurons (Figure
2.6A). This observation suggests that the slow-saturating inhibition to pyramidal neurons could
be relayed by inhibitory type II neurons innervated by slow-saturating auditory nerve fibers.
Based on the above evidences, we propose a simple model that can explain DCN
intensity selectivity (Figure 2.6B). In this model, the intensity tuning properties of excitation and
inhibition are simply inherited from auditory nerve fibers. The intensity-selective pyramidal
neuron may receive fast-saturating excitatory input from fast saturating auditory nerve fibers
(Input A), and slow-saturating inhibitory input from inhibitory neurons driven by slow-saturating
auditory nerve input (Input B). Through such a feedforward inhibitory circuit likely mediated by
type II/vertical cells (Voigt and Young, 1980; Oertel and Young, 2004; Kuo et al., 2012), the
Figure 2.6 A proposed model for the
generation of intensity selectivity in the DCN
A1, A2, two example type II neurons. The spike and
excitatory responses evoked by CF tones at
different intensities are displayed in a similar
manner as in Figure 2. Middle, maximum value for
the color scale is 180Hz, 285pA in A1, and 150Hz,
510pA in A2. Right, black marks the average
intensity tuning curve for intensity-nonselective
pyramidal neurons. B, In our proposed model, the
intensity-selective pyramidal neuron receives
excitatory input from afferent that shows fast
saturating firing rate (FR) with intensity
increments. Another type of afferent that shows
slow saturating FR innervates inhibitory type II
neurons, the output of which thresholds and
reverses the sign (-) of their afferent input. Due to
the spike thresholding effect on inhibitory neuron
responses, the threshold of inhibition to the
pyramidal neuron becomes elevated compared to
the excitation. Through this differential circuit, the
pyramidal neuron is able to detect the difference
between the intensity tuning patterns of the two
types of afferent, as reflected by the differential
tuning patterns of excitation and inhibition, and to
report this difference as the intensity-selective
output.
34
threshold of inhibition becomes higher than excitation due to the spike thresholding effect on
inhibitory neuron responses, while the intensity tuning function of inhibition remains slow-
saturating. In addition, the low spontaneous activity of type II neurons (Figure 2.2G) may also
contribute to their higher intensity threshold, since relatively stronger auditory nerve input is
required to drive these cells compared to pyramidal neurons. By this analogous differential
circuit, the pyramidal neuron can extract the difference between the intensity tuning patterns of
Input A and Input B to create intensity-selective output responses. The overall structure of our
proposed circuit is in line with the one previously proposed for type IV neurons (Voigt and
Young 1980; Spirou et al. 1999; Hancock and Voigt 2002a), except that differential AN fiber
inputs are incorporated. It is unknown why AN fibers have different saturation slopes. Is it due to
distinct properties of different types of spiral ganglion cells? How specific is the connectivity
between different types of DCN neurons and different types of AN fibers? The proposed
synaptic circuitry mechanism for the generation of intensity selectivity raises many interesting
questions for future investigations.
35
Chapter 3 Generation of Temporal Diversity
3.1 Introduction
While auditory nerve fibers exhibit more or less uniform temporal discharge patterns in response
to a certain sound stimulus (Kiang et al., 1965), a diversity of discharge patterns is observed for
neurons at almost every processing stage along the central auditory pathway. For example, in
response to tone stimulation, although a population of neurons does inherit the primary-like
discharge pattern from the auditory nerve, many other neurons in the dorsal cochlear nucleus
(DCN) (Godfrey et al., 1975; Rhode and Smith, 1986; Rhode and Kettner, 1987; Hancock and
Voigt, 2002a), inferior colliculus (Semple and Kitzes, 1985; Ehret and Merzenich, 1988;
Kuwada et al., 1997), and even auditory cortex (Recanzone, 2000), exhibit pauser and
buildup response patterns distinct from auditory nerve responses. These diversified discharge
patterns may be important for the representation and processing of specific sound information
components. However, how such functional diversity is created at the first central station along
the afferent auditory pathway remains unclear.
One recent model proposes that the variation in the availability of an intrinsic rapidly
inactivating A-type K
+
conductance and its activation status can result in the variation in the
first-spike latency and first inter-spike interval as observed for different discharge patterns
(Manis 1990; Kanold and Manis, 1999; Kanold and Manis, 2005). On the other hand, it has been
suggested that diverse spike patterns may result from synaptic circuitry of specific organizations
(Godfrey et al., 1975; Kuwada et al., 1997; Hancock and Voigt, 2002a). However, what synaptic
mechanisms are underlying the response diversity remains unknown.
36
Synaptic connections in the DCN have mostly been studied in in vitro preparations. It is
generally thought that pyramidal neurons receive direct auditory nerve input on their basal
dendrites in the deep layer (Brown and Ledwith, 1990; Ryugo and May, 1993), as well as
auditory and non-auditory input on their apical dendrites in the superficial layer (Mugnaini et al.,
1980; Oertel and Young, 2004). The pyramidal neurons also receive inhibitory input from
various sources, including cartwheel cells in the superficial layer, vertical/tuberculoventral cells
in the deep layer and possibly the D-multipolar cells in the posterior ventral part of cochlear
nucleus (Nelken and Young, 1994; Golding and Oertel, 1997; Oertel and Young, 2004; Kuo et
al., 2012). It is possible that this rich repertoire of synaptic inputs can result in summed
excitation and inhibition with diverse temporal patterns. In this study, we performed sequential
loose-patch/current-clamp recording and voltage-clamp recording from the same DCN pyramidal
neurons in vivo. This allowed us to directly correlate the spike response pattern of the cell with
the temporal patterns of underlying excitatory and inhibitory synaptic inputs, and to determine
how excitatory and inhibitory interplay shapes the output spike response. Our results revealed
that the response diversity of DCN pyramidal neurons could be largely explained by the
differential temporal patterns of synaptic inputs they received.
3.2 Methods
3.2.1 Animal preparation
All experimental procedures used in this study were approved under the Animal Care and Use
Committee at the University of Southern California. Experiments were performed on 101 Female
Sprague-Dawley rats (about 3 months old and weighing 250 300 g, raised up with 12 hours
light/dark cycle). Rats were anaesthetized with ketamine and xylazine (ketamine: 60 mg/kg;
37
xylazine: 8 mg/kg; i.p.). During the experiments, level of anesthesia will be monitored by
checking animals response to toe pinch. Ketamine one third of initial dosage is given when
needed. The body temperature was maintained at 37.5 by a feedback heating system (Harvard
Apparatus, MA). The animal was positioned with the left ear facing a calibrated free-field
speaker (Vifa, Denmark) in a sound-attenuation booth (Acoustic Systems). A sound-attenuating
plug was inserted in its right ear. After opening the left part of the occipital bone, part of the
cerebellum was aspirated to expose the left DCN. During the experiment, DCN was covered with
artificial cerebrospinal fluid (ACSF; in mM: 124 NaCl, 1.2 NaH 2PO4, 2.5 KCl, 25 NaHCO3, 20
Glucose, 2 CaCl2, 1 MgCl2). We mapped the DCN tonotopy with extracellular recordings, which
showed a low-high frequency gradient along the lateral-medial axis, as described in our previous
study (Zhou et al., 2012).
3.2.2 In vivo whole-cell and loose-patch recordings
Whole-cell recordings were made with an Axopatch 200B amplifier (Molecular Devices). Patch
pipettes made from borosilicate glass capillaries (Kimax) had an impedance of 4- 5 M©. Pipettes
contained a potassium-based solution (in mM): 130 K-gluconate, 4 MgATP, 0.3 GTP, 8
phosphocreatine, 10 HEPES, 11 EGTA, 5 KCl, 1 CaCl 2, 0.25 fluorescein dextran, pH 7.3. The
patch pipette was lowered into the DCN at an angle of about 85
o
. The brainstem surface was
covered with 3.5% agar prepared in warm ACSF. Whole cell capacitance was fully compensated
and the initial series resistance (25- 50 M©) was compensated for 40- 60% to achieve an
effective series resistance of 15- 30 M©. Signals were filtered at 5 kHz and sampled at 10 kHz.
Only cells with resting membrane potential lower than 50 mV were studied. Pyramidal neurons
were recorded at 100- 250 µm depth below the surface, corresponding to the pyramidal cell layer
38
(Mugnaini et al., 1980; Wouterlood and Mugnaini, 1984). The reconstructed morphologies of
some recorded neurons confirmed the pyramidal neuron type (see our previous study, Zhou et
al., 2012). As we showed previously (Zhou et al., 2012), under our whole-cell voltage-clamp
recording condition, the pyramidal neurons in DCN can be reasonably clamped for the detected
evoked synaptic currents.
Two major types of inhibitory neuron were also recorded in this study. Vertical cells
were recorded at 250- 700 µm depth (Mugnaini et al., 1980; Rhode, 1999) and identified by their
lack of spontaneous activity and absence of responses to broadband noise stimulation (Young
and Brownell, 1976; Rhode, 1999; Spirou et al., 1999). Cartwheel cells were recorded at 80- 150
µm depth (Wouterlood and Mugnaini, 1984) and identified by firing of complex spikes (Manis et
al., 1994; Parham and Kim, 1995; Portfors and Roberts, 2007). Loose-patch recordings were
performed with pipettes of smaller tip openings (impedance ~10 M©) so as to overcome the
recording bias towards larger cells (Wu et al., 2008). Pipettes were filled with ACSF. Loose seal
(100- 200 M©) was made from neurons, allowing spikes only from the patched cell to be
recorded. Recording was made under the voltage-clamp mode. Signals were filtered at 0.3- 10
kHz.
3.2.3 Sound stimulation
Softwares for sound stimulation and data acquisition were custom-developed with LabVIEW
(National Instrument). For each successful whole-cell recording, we first mapped the tonal
receptive field for the recorded cell under current-clamp mode. Pure tones (0.5- 64 kHz at 0.1
octave intervals, 50 ms duration, 3 ms ramp) at eight 10 dB spaced sound intensities (0- 70 dB
sound pressure level (SPL)) were delivered pseudo-randomly. The time interval and intensity
39
difference between two sequential tones were set at 0.5- 1 s and no more than 30 dB,
respectively, to ensure minimal cross-interaction between the stimuli. Characteristic frequency
(CF), the frequency at which neurons responded to tones with minimum intensity, was
determined on-line. Then the responses of the same cell to 10 repetitions of CF tone stimulation
at 60 dB SPL were recorded when the cell was voltage-clamped at- 70 mV and 0 mV
respectively. An inter-stimulus interval of 4.5 s was used to minimize cross-trial interaction.
After this, the cells responses to 40 repetitions of 60 dB CF tone stimulation were recorded
under current-clamp mode. Four different durations were applied: 25 ms, 50 ms, 100 ms and 200
ms. The same sound stimulation protocol in current-clamp recordings was applied for loose-
patch recordings.
3.2.4 Data analysis
Spikes from either current-clamp recordings or loose-patch recordings were sorted offline. The
neuronal response type was determined based on the peri-stimulus spike time histogram(PSTH) pattern in response to 60 dB CF tone stimulation. First spike latency was determined by the time
point in PSTH where firing rate exceeded the average spontaneous firing rate by 3 standard
deviations of baseline activity. Evoked firing rate was calculated by subtracting the spontaneous
rate firing from the firing rate within defined time windows. All the synaptic response traces
evoked by the same test stimulus were averaged. Onset latency was determined by the time
point where the current amplitude exceeded the average baseline by 3 standard deviations. In
current clamp recordings, subthreshold V m responses were obtained by removing spikes with a 3
ms median filter (each data point was replaced by the median of data points within a time
window ±3 ms of this data point).
40
Excitatory and inhibitory synaptic conductance were derived (Borg-Graham et al., 1998;
Anderson et al., 2000; Zhang et al., 2003; Wu et al., 2008; Sun et al., 2010) according to I = Ge
* (V- Ee) + Gi * (V- Ei). I is the amplitudeof the synaptic current response at any time point
after subtraction of the baseline current; G e and Gi are the excitatory and inhibitory synaptic
conductance; V is the holding voltage, and E e(0 mV) and Ei(- 70 mV) are the reversal
potentials. The clamping voltage V was corrected from the applied holding voltage (V h): V = Vh
Rs * I, where Rs is the effective series resistance. By holding the recorded cell at two different
voltages (the reversal potentials for excitatory and inhibitory currents respectively), Ge and Gi
were resolved from the equation.
The expected membrane potential change caused by synaptic conductances was derived
with an integrate-and-fire neuron model (Somers et al., 1995; Liu et al., 2007):
where Vm(t) is the membrane potential at time t, C the whole-cell capacitance, Gr the resting
leaky conductance, E r the resting membrane potential ( 60 mV). To simulate the spike response,
20 mV above the resting membrane potential was set as the spike threshold and a 5 ms refractory
period was used. C (20- 50 pF) was measured during the experiment and G r was calculated based
on the equation Gr = C * Gm / Cm, where Gm, the specific membrane conductance is 2e 5 S/cm
2
,
and Cm, the specific membrane capacitance is 1e 6 F/cm
2
(Hines, 1993; Stuart and Nelson,
1998).
()()() [ ])()()()()()()( t V E t V G E t V t G E t V t G
C
dt
dt t V
m r m r i m i e m e m
+ - + - * + - * - = +
41
3.2.5 Fitting synaptic currents
Each phase of the synaptic current was fitted with an exponential function using MATLAB
(Mathworks). The equations to fit the two phases of the accumulating excitation (Figure 3.3A) were: 1- exp (- (t- t1)/Ä1) and k * (1- exp (- (t- t2)/Ä2)) + (1- exp (-(t2- t1)/Ä1)), where
t1, t2 were onsets of each phase and Ä1, Ä2 were time constants. The equations to fit the two
phases of the fast-rising excitation and inhibition (Figure 3.3A) were: 1- exp (- (t- t1)/Ä1) and
exp (-(t- t2)/Ä2) * (1- exp (-(t2- t1)/Ä1)).
3.2.6 Dynamic clamp
Dynamic clamp recordings were carried out according to our previous studies (Liu et al., 2011;
Li et al., 2012). The current injected to the cell was calculated in real time according to:
()() i m i e m e
E t V t G E t V t G t I - * + - * =)()()()()( The time-dependent Ge and Gi were simulated synaptic conductances. E e and Ei(reversal
potentials) were set as 0 mV and- 70 mV respectively. The membrane potential V m(t) was
sampled at 5 kHz.
3.2.7 Statistics
Shapiro-Wilk test were first applied to exam whether samples had a normal distribution. In the
case of a normal distribution, t-test or one way ANOVA with post hoc (Turkey test) test was
applied. Otherwise, a non-parametric test was applied. Data were presented as mean ± SEM if
not otherwise specified.
42
3.3 Results
3.3.1 Temporal response patterns of rat DCN pyramidal neurons
In the rat DCN, we first characterized the discharge patterns of pyramidal neurons located in
layer 2 (see Methods). Cell-attached recordings (Wu et al., 2008, Zhou et al., 2012) were
performed to record spikes from individual pyramidal neurons in the middle frequency region
(11.8 ± 3.7 kHz, mean ± s.d.). When the cells were tested with characteristic frequency (CF) tones, buildup (30%), pauser (35%) and primary-like (35%) response patterns were widely
observed (Figure 3.1A, 3.1B), consistent with previous studies (Rhode and Smith, 1986;
Hancock and Voigt, 2002a). The primary-like discharge pattern started with a strong onset
response, as shown by the PSTH (Figure 3.1A, 3.1B). The firing rate was then reduced to a level
that was sustained in the remaining duration of the tone (Figure 3.1B). The primary-like pattern
in the rat DCN may correspond to the previously reported chopper pattern in other species,
although for the latter inter-spike intervals are much more uniform, resulting in regularly spaced
peaks in the PSTH (Rhode and Smith, 1986; Hancock and Voigt, 2002a). The pauser discharge
pattern started with a sharp transient onset response, which was followed by a ~ 15-ms period of
silence in firing (Figure 3.1B). The firing rate was then slowly increased to a level that was
sustained in the remaining tone duration. The buildup pattern exhibited an initial silent period of
about 15 ms, after which the first spike appeared and the firing rate was then gradually increased
(Figure 3.1B).
The three response patterns appeared to be fully distinguishable by their spikes within the
first 25-ms window after the stimulus onset. The buildup response exhibited a much longer first
spike latency (19.8 ± 5.6 ms, mean ± s.d.) than the primary-like (5.2 ± 1.3 ms) and pauser
43
responses (5.4 ± 0.8 ms; Figure 3.1C). The primary-like response exhibited a significantly
Figure 3.1 Response diversity in the rat DCN
A, Spike patterns of three example pyramidal cells in response to CF tones of different durations (marked by thick
bars). For each cell, sample recorded traces (in one trial) are shown in the upper panel, and PSTHs (bin size = 1
ms) for responses in all trials are shown in the lower panel. B, Average PSTHs (bin size = 3 ms) for all the recorded
cells of three response types. Cell numbers are marked. C, Average spike latency for the three groups of cell. Bar
= SEM. ***, p < 0.001, one-way ANOVA and post hoc test (same for the below). D, Average firing rate within the
0 25 ms window after the tone onset. E, Average inter-spike interval between the first and second spikes. Note
that the first inter-spike interval was relatively long in buildup cells, due to their low firing rates at the onset of
tones. This is different from pauser cells, whose responses were suppressed after the first spike. Cell numbers in
C-E are the same as in B. F, Plot of evoked firing rate (after subtraction of spontaneous firing) within the 4- 9 ms
window after the tone onset versus that within the 9- 14 ms window for all the cells. Note that primary-like (Pr),
pauser (Pa) and buildup (Bu) cell groups are well segregated.
44
higher firing rate during the first 25 ms window (Figure 3.1D), and a shorter inter-spike interval
between the first and second spikes (i.e. first inter-spike interval) compared to the pauser and
buildup responses (Figure 3.1E). As shown by the plot of evoked firing rate within the 4- 9 ms
time window after the tone onset (which contains the first evoked spike in primary-like and
pauser patterns) against that within the 9- 14 ms window after the tone onset (which contains the
silent period in pauser and buildup patterns), the cells were segregated into three distinct clusters
(Figure 3.1F), further demonstrating that the different discharge patterns are clearly
distinguishable. The spontaneous firing rate, on the other hand, was not different among the three
types of cell (primary-like: 21.8 ± 3.2 Hz; pauser: 26.8 ± 4.1 Hz; buildup: 18.5 ± 4.3 Hz; one-
way ANOVA, F = 1.22, P = 0.30).
3.3.2 Excitatory and inhibitory synaptic inputs to DCN pyramidal neurons
We next carried out whole-cell recordings to reveal the synaptic inputs underlying different
discharge patterns. The discharge pattern of the recorded cell was first examined under the
current-clamp mode, by applying repeated CF tones at 60 dB SPL (Figure 3.2A, 3.2B). Primary-
like, pauser and buildup spike patterns were observed, similarly as in cell-attached recordings
(Figure 3.2B). Notably, they were associated with different subthreshold membrane potential
response patterns after filtering out the action potentials (Figure 3.2C). The primary-like
response was marked by a fast and strong depolarization, which largely sustained during tone
stimulation. The pauser response was marked by a sharp transient onset depolarization followed
by a brief hyperpolarization of about 10 ms duration, after which the membrane potential
gradually depolarized. The buildup response lacked a clear onset depolarization and the
membrane potential slowly depolarized from an early brief hyperpolarization which started at
about 5 ms after the tone onset and lasted for about 10 ms.
45
46
We subsequently recorded excitatory and inhibitory currents under the voltage-clamp
mode by clamping the cell at- 70 mV and 0 mV respectively (Figure 3.2D, 3.2E). Excitatory
and inhibitory conductances were derived from the recorded currents (Figure 3.2F; see
Methods). While the inhibitory inputs to the three types of cell exhibited similar temporal
profiles, the excitatory inputs were strikingly different (Figure 3.2F). The primary-like cell
exhibited fast rising excitation, i.e. t he excitation rose rapidly to the peak before slowly
decaying to a sustained level. In contrast, the pauser and buildup cells exhibited accumulating
excitation: the excitation rose gradually without decaying. The pauser and buildup cell
apparently differed in the kinetics of their excitatory rising phase.
We next examined whether the observed synaptic inputs can sufficiently explain the
corresponding spike patterns of the recorded cells. With a single-compartment, integrate-and-fire
neuron model, we derived the expected membrane potential and spike responses from the
experimentally determined excitatory and inhibitory synaptic conductances (see Methods). As
shown for the three example cells, both the derived membrane potential (Figure 3.2G) and spike
response (Figure 3.2H) patterns qualitatively resembled the recorded membrane potential and
spike response patterns respectively, suggesting that the temporal interplay of sound-evoked
synaptic inputs may sufficiently result in distinct spike patterns.
Figure 3.2 Synaptic inputs underlying discharge patterns of three example pyramidal neurons
A, A raw trace of current-clamp recording (in one trial) in response to the CF tone (duration = 200 ms).
Dash vertical lines mark the onset and offset of the tone stimulation. B, PSTH (bin size = 1 ms) for
spikes recorded in all trials. C, Average subthreshold V m responses. Scale bar was presented as a
vertical bar (same for the below). D, Average excitatory currents (10 trials) recorded in voltage-clamp
mode. E, Average inhibitory currents. F, Superimposed derived excitatory (red) and inhibitory (blue) synaptic conductances. G, Derived membrane potential responses by integrating the synaptic
conductances shown in F with the neuron model. H, Derived spike responses after a 20 mV spike
threshold (refractory period = 5 ms) was applied. Each vertical line represents a spike.
47
3.3.3 Diversity in temporal dynamics of excitation
To quantify the temporal properties of synaptic responses, we performed curve fittings to the
conductance traces using exponential functions (see Methods). The accumulating excitation
exhibited two rising phases (I and II) of different time constants with phase I being fast and
phase II being slow, whereas the fast-rising excitation as well as the inhibition exhibited only a
single fast rising phase (Figure 3.3A). We quantified the amplitude of each rising phase.
Figure 3.3 Summary of properties of synaptic inputs to the three types of pyramidal neuron
A, Example traces of accumulating excitation, fast-rising excitation and inhibition. The best-fit curve for
each trace is shown in red, and measurements of phase I and phase II amplitudes are illustrated with
double arrowheads. B, Average phase I amplitudes of excitation (red) and inhibition (blue) to three types
of cell. Bar = SEM. ***, p < 0.001; *, p < 0.05, one-way ANOVA and post hoc test. Cell numbers are
marked. C, Average E/I ratios in phase I. D, Average timing of the peak of phase I excitation relative to
the tone onset. E, Average amplitudes of phase II excitation. F, Average onset latencies of excitation and
inhibition. Cell numbers in C-F are the same as in B
48
Primary-like, pauser and buildup cells were substantially different in the amplitude of phase I
excitation, while they were not different in the amplitude of inhibition (Figure 3.3B).
Consequently, the E/I ratio between phase I excitation and inhibition was different among the
three types of cell (Figure 3.3C). The amplitude of phase I excitation, as well as the E/I ratio,
was the largest in primary-like cells and smallest in buildup cells (Figure 3.3B, 3.3C). On the
other hand, the timing of the peak phase I excitation was not different among the three types of
cell (Figure 3.3D). As for phase II excitation, pauser and buildup cells did not differ in its
amplitude, whereas the excitation in primary-like cells did not have a second rising phase
(Figure 3.3E). Furthermore, we did not find a difference in the onset latency of either excitation
or inhibition among the three types of cell (Figure 3.3F). Excitation always preceded inhibition
by about 1.5 ms (Figure 3.3F), suggesting a universal disynaptic nature of the inhibitory input.
3.3.4 A synaptic mechanism for the generation of response diversity
To further test whether synaptic inputs of differential temporal patterns as we observed can
account for the diversity of spike response patterns, we carried out in vivo dynamic-clamp
recordings (see Methods) to examine the membrane potential responses to simulated excitatory
and inhibitory conductances injected into the cell (Figure 3.4A). Synaptic conductances were
described by the arithmetic functions used for fitting experimental data (Figure 3.4B, 3.4C, left
panel). To be consistent with the experimental observation (Figure 3.3B, 3.3E), the inhibitory
conductance was fixed, while the excitatory conductance was varied. To simulate pauser and
buildup responses (Figure 3.4B), the amplitude of phase II excitation was fixed while that of
phase I excitation was varied. Notably, a fast transient depolarization at the onset of the
membrane potential response was generated, and its amplitude was dependent on the E/I ratio
49
between the phase I excitation and inhibition (Figure 3.4B, middle panel). The larger the E/I
ratio, the greater was the amplitude of the transient onset depolarization (Figure 3.4B, right
panel). After a delay, the membrane voltage gradually increased to a level above the spike
threshold (Figure 3.4B, middle panel). Therefore, the putative spike pattern would be of pauser
or buildup type, depending on whether the transient onset depolarization was large enough to
cross the spike threshold. In the same recorded cell, the simulated excitation and inhibition of
50
primary-like responses generated a large fast transient depolarization followed by a smaller
sustained depolarization (Figure 3.4C, middle panel). The amplitude of the sustained
depolarization depended on the E/I ratio (Figure 3.4C, right panel). Therefore, the putative spike
pattern would be primary-like if the sustained depolarization was large and crossed the spike
threshold, or of onset type if it was below the spike threshold.
It was previously reported that the same DCN neuron could exhibit different discharge
patterns in a stimulus-dependent manner under different tone frequencies and intensities(Rhode
et al., 1983; Hancock and Voigt, 2002a). This could potentially be attributed to a similar synaptic
mechanism, since the strength of synaptic inputs or E/I ratio might be modulated by frequency
and intensity. We thus examined the spike and synaptic responses to all effective tones within the
tonal receptive field (TRF) of individual cells. The example cell in Figure 3.4D was identified as
a pauser cell based on its spike response pattern to the CF tone (13.9 kHz) at a relatively high
intensity (60 dB SPL). At this intensity level, while the cell exhibited pauser discharge patterns
Figure 3.4 A synaptic mechanism for the response diversity in the DCN
A, Schematic illustration of dynamic-clamp recording. V m is recorded from the patched cell. The synaptic current
(I syn) is calculated based on the V m and modeled synaptic conductances, and is injected into the cell in real time. B,
Left, temporal profiles of inhibition (blue) and accumulating excitation (red) simulated for pauser and buildup
response patterns. The amplitude of phase I excitation was varied. Middle, membrane potential responses in the in
vivo dynamic-clamp recording from a DCN pyramidal neuron. The E/I ratio between phase I excitation and
inhibition is marked. Bar represents a 50 ms putative tone stimulus. Dash line marks the putative spike threshold.
Right, peak amplitude of the transient onset depolarization (marked by arrows in the middle panel) plotted against
the E/I ratio between phase I excitation and inhibition. Bar = SEM. N = 4 cells. C, Left and middle, dynamic-
clamp recording in the same cell as shown in B(middle panel). Fast-rising excitation with the peak amplitude
varied was injected to simulate primary-like and onset cells. Right, amplitude of sustained depolarization (marked
by arrows in the middle panel) plotted against the E/I ratio. Bar = SEM. N = 4 cells. D, Left panel, color map of
frequency-intensity TRF of a pyramidal neuron (top), and the type of discharge pattern for tone-evoked responses
within the TRF (bottom). Horizontal arrows points to the intensity level at 60 dB SPL. Vertical arrow points to the
CF. None, no significant tone-evoked response; Other, delayed sustained response pattern which resembles the
primary-like pattern except that the response onset delay was longer and there was no apparent onset peak response.
Right panel (from top to bottom), recorded trace of spike response, PSTH (bin size = 1 ms), excitatory and
inhibitory currents to effective tones at 60 dB SPL in the same cell. Vertical arrow points to the CF. E, Plot of E/I
ratio versus amplitude of phase I excitation for all the identified buildup and pauser responses within the TRF of the
cell shown in D. F, TRF and response patterns for another example cell. G, Plot of E/I ratio versus amplitude of
phase I excitation for all the identified onset and primary-like responses within the TRF of the cell shown in F.
51
to 11.3, 12.1, 13, and 13.9 kHz tones, it displayed buildup discharge patterns to 10.6 and 14.9
kHz tones (Figure 3.4D, right panel). Such variation in discharge pattern was observed at other
tone intensities as well (Figure 3.4D, left panel). We plotted the synaptic parameters for the
corresponding identified discharge patterns within the TRF (Figure 3.4E). It became clear that
buildup patterns were associated with small excitatory input amplitudes and small E/I ratios
during phase I, while pauser patterns were associated with larger amplitudes of phase I excitation
and larger E/I ratios. Another example cell (Figure 3.4F) was identified as primary-like based on
its spike response pattern to the CF tone (9.8 kHz) at 60 dB SPL. However, its discharge patterns
to frequencies above the CF were clearly of onset type, consisting only of a transient onset
response (Figure 3.4F, right panel). For all the tone-evoked responses, onset patterns were
associated with smaller phase I excitation and smaller E/I ratios compared to primary-like
patterns. Together, these observations were consistent with the prediction from the dynamic-
clamp experiment that the strength of excitation and the E/I ratio during phase I are critical for
distinguishing different temporal response types.
3.3.5 Potential excitatory sources for pyramidal neurons
That the excitatory input may contain two distinct components of fast and slow rising kinetics is
reminiscent of previous findings that DCN pyramidal neurons primarily receive excitatory input
from two sources: the auditory nerve input onto their basal dendrites and the parallel fiber input
onto their apical dendrites (Figure 3.5A). Zhang and Oertel (1994) showed that DCN pyramidal
cells can respond to electrical stimulation of the auditory nerve with a slowly rising excitatory
postsynaptic potential (EPSP), which remains after severing the branches of auditory nerve fibers
directly innervating the pyramidal cells. Kuo and Trussell (2011) showed that parallel fibers
exhibit strong short-term facilitation. These results suggest that parallel fiber input is delayed
52
relative to the direct auditory nerve input and that the second excitatory component we observed
in vivo might be contributed by parallel fiber input. While it is difficult to separate the two
sources of excitatory input in vivo, we may be able to estimate their temporal properties by
examining the sound-evoked excitatory inputs to cartwheel and vertical cells respectively.
Previous studies suggest that vertical cells are primarily driven by auditory nerve input (Spirou et
al., 1999), while cartwheel cells predominantly receive excitation from parallel fibers (Mugnaini
et al., 1980; Fujino and Oertel, 2003) (Figure 3.5A). Therefore, the temporal property of
excitation to cartwheel and vertical cells may reflect that of parallel fiber and auditory nerve
input, respectively.
In our recordings, cartwheel cells were identified by the existence of complex spikes
(Figure 3.5B), according to previous studies (Manis et al., 1994; Parham and Kim, 1995;
Portfors and Roberts, 2007). Vertical cells were identified by the extremely low spontaneous
firing rate (less than 0.05 spikes s
-1
) and the lack of responses to noise stimuli (Young and
Brownell, 1976; Rhode, 1999; Spirou et al., 1999) (Figure 3.5C). Subsequent voltage-clamp
recordings from the same cells then revealed their tone-evoked excitation (Figure 3.5B, 3.5C).
Notably, the excitation in vertical cells was similar to that in primary-like pyramidal cells, since
they both exhibited a very short delay to the peak response (i.e. fast rising) (Figure 3.5E) and a
large amplitude of this peak (Figure 3.5D). This result supports the notion that the fast rising
excitation in primary-like neurons could be mainly attributed to the auditory nerve input that
drives fast spiking of these and vertical cells with similarly short latencies (Figure 3.5F). On the
other hand, the slowly accumulating excitation in cartwheel cells resembled the phase II
excitation in pauser and buildup neurons, as demonstrated by the long delay to the peak
amplitude (Figure 3.5E). This result suggests that parallel fiber input may assume a slowly
53
accumulating property and contribute largely to the phase II excitation in pauser and buildup
neurons. We also noticed that the phase I excitation in pauser and buildup neurons resembled the
fast-rising excitation in primary-like neurons in terms of onset latency (Figure 3.3F) and rising
speed (pauser, Ä = 7.08 ± 0.99 ms, mean ± s.d.; buildup, Ä = 7.64 ± 2.57 ms; primary -like, Ä =
5.19 ± 1.26 ms, one-way ANOVA, F = 1.62, P = 0.27). This suggests that the phase I excitation
may be attributed to direct auditory nerve input.
Figure 3.5 Potential circuit mechanisms for generating response diversity in the DCN
A, Schematic drawing of DCN circuits. A.N., auditory nerve; P.F., parallel fiber; Vt, vertical cell; Cw, cartwheel
cell; Pyr, pyramidal cell. Red and blue dots represent excitatory and inhibitory synapses respectively. B, Left panel,
color map of the TRF (top) of an example cartwheel cell and the PSTHs for responses to noise and CF-tone stimuli
(marked by the bar). Vertical arrow points to the CF. Right panel, cell-attached recording of spike responses (top),
superimposed excitatory currents of four trials (middle) and average excitatory current (bottom). Arrow points to
the complex spike. Tone duration is marked by the bar. C, TRF and responses to noise and CF-tone stimuli of a
vertical cell. Note that it exhibited zero spontaneous firing rate and no response to noise stimulation. Right, evoked
spikes and excitatory currents by CF tones. D, Average amplitude of phase I excitation for different types of cell.
Bar = SEM. Cell numbers are marked. E, Average timing of peak excitatory amplitude relative to the tone onset.
Cell numbers are the same as in D. F, Average first spike latency. Cell numbers are marked.
54
3.4 Discussion
In this study, we applied in vivo whole-cell recordings to examine the synaptic mechanisms
underlying different temporal response patterns in the DCN. We found that the differential
temporal profiles of excitatory synaptic inputs received by DCN pyramidal neurons were
sufficient to generate different discharge patterns. DCN neurons receiving dominant fast-rising
excitation exhibit primary-like or onset responses, depending on the E/I ratio in the initial
response rising phase. Those receiving accumulating excitation exhibit pauser or buildup
responses, again depending on the E/I ratio in the initial fast response rising phase.
The difference in excitatory temporal profiles may be due to a large variation of how
DCN pyramidal cells sample from two potential excitatory input sources. For primary-like cells,
their inputs can be primarily from auditory nerve fibers, given their earliest arrival after the onset
of tone stimulus and its temporal profile closely resembling that of auditory nerve firing (Kiang
et al., 1965). Previous studies also show that T-type multipolar cells in PVCN are excitatory and
project to DCN. Its possible that this T -type multipolar cells also innervate fusiform cells.
Considering that this T-type multipolar cells relay input from auditory nerve and exhibit
sustained/chopper responses, this cell type may also contribute to the fast-rising component of
excitatory input received by fusiform cells. For pauser and buildup cells, their input is a mixture,
from both auditory nerve and parallel fibers. The auditory nerve input to pauser cells is likely
strong enough to trigger the fast onset spike response, as the first spike latency of these cells is
similar to that of primary-like and vertical cells. The auditory nerve input to buildup cells is
relatively weak, and their spiking has to rely on the second phase excitation to accumulate to a
certain level to overcome inhibition, resulting in a much longer delay for the first spike to appear.
55
Since all cells exhibit similar inhibitory input patterns, these differential excitatory input patterns
would directly lead to distinct temporal patterns of membrane potential responses, with the
distinction lying on whether a fast onset spike response can be generated and whether there is a
suppression period after that. The functional type of the cells was defined based on their
responses to CF tones at a high intensity. For the same cell, the excitatory input pattern as well as
the E/I ratio can vary depending on the tone frequency and intensity, resulting in a stimulus-
dependent variation in discharge pattern.
Our current data, however, do not exclude the intrinsic channel property hypothesis. In a
series of studies(Manis 1990; Kanold and Manis, 1999; Kanold and Manis, 2005), Paul Manis
and colleagues demonstrated that in DCN pyramidal neurons, three distinct temporal discharge
patterns could be generated in vitro and in vivo, depending on the voltage from which the cell
was depolarized. Buildup and Pauser responses were seen if the cells were held hyperpolarized
before they were depolarized. Our current clamp recordings do not support the existence of such
preceding hyperpolarization. Nevertheless, inhibitory inputs from spontaneous events or non-
auditory innervations might provide this hyperpolarization before sound onset. In these
scenarios, the previously proposed IKIF currents might work synergistically together with the
synaptic inputs to efficiently and specifically generate diversified response patterns.
While the DCN has been shown to use spectral cues to localize sound in the vertical
plane (Oertel and Young, 2004), the functional significance of the response diversity in the DCN
is unclear. It is possible that by generating the response diversity, DCN can play a role in an
initial sorting of distinct fundamental components of acoustic information, which sets a
foundation for the parallel processing in higher-level auditory centers. Different from the other
parts of cochlear nucleus, DCN response properties are more strongly influenced by inhibition,
56
which may endow this processing center with an advantage in generating novel response
properties that allow detections of important features in acoustic stimuli. While our current study
provides a plausible synaptic circuitry mechanism for the generation of diverse temporal
response patterns in DCN pyramidal cells, it should be noted that it remains a great challenge to
resolve the circuitry underlying various specific processing functions. Future studies with
genetic/optogenetic tools will be needed to further dissect the role of each different type of
neuron in the functional circuitry of DCN.
57
Chapter 4 Synaptic Mechanisms for Generating Off
Responses
4.1 Introduction
Neurons in different sensory modalities respond to the offset of the sensory stimulus as well as the
onset (Kandel et al., 2013). This phenomenon is best demonstrated in the retina, where on -center
and off -center retinal ganglion cells were observed. In the auditory system, these off responding
neurons were described at many nuclei along the central auditory pathway, including the dorsal
cochlear nucleus (Rhode et al., 1983; Shofner and Young, 1985), superior periolivary nucleus
(Kadner et al., 2006; Kopp-Scheinpflug et al., 2011), inferior colliculus (Faingold et al., 1986),
medial geniculate body (He, 2001) and auditory cortex (Qin et al., 2007; Scholl et al., 2010). Sound
offsets are important cues for perceptual grouping(Bregman et al., 1994). The mechanisms for
generating offset responses in the auditory pathway, however, is still not well understood.
At least three hypotheses have been proposed to explain the generation of offset responses.
The most popular one is about the post-inhibitory rebound depolarization. In this hypothesis, the
neurons fire action potentials after they are released from sustained suppression during the sound
stimulation. The hyperpolarization-activated cation channel, IH channel, was proposed to
contribute significantly to this rebound depolarization (Kuwada and Batra, 1999). Indeed, this IH
channel, together with a T-type calcium channel, were found to generate offset firings in the
superior periolivary nucleus (SPON) of mice(Kopp-Scheinpflug et al., 2011). However, the SPON
neurons are GABAergic. The offset responses in other nuclei could not simply be inherited from
SPON neurons. In the second hypothesis, Scholl et al. (2010) studied off firing neurons in rat
58
primary auditory cortex (A1) and found that these off firing neurons in A1 directly receive offset
excitatory inputs, presumably from lower stages of the auditory pathway. Again, it is still unclear
how the offset responses are initially generated. In the third hypothesis, it is proposed that a delayed
excitation compared to inhibition can generate offset firings (Grothe 1994; Yang and Pollak,
1997). However, without a method to directly measure the detailed temporal pattern of excitation
and inhibition, this hypothesis was never directly tested.
In this study, we recorded from fusiform cells in DCN, where off responses were first
described in the central auditory pathway. With in vivo whole cell recordings, we directly measured
the membrane potential responses and the synaptic inputs into the same cells. We found that off
responding fusiform cells exhibited depolarized membrane potential right before their offset
responses, suggesting that the off responses were not due to post-inhibitory rebound mechanism.
We also found that the excitatory inputs into fusiform cells did not have an offset component,
indicating that the off responses were generated de novo in DCN. Instead, the generation of offset
responses in DCN could be explained by a synaptic mechanism. Excitatory input outlasts
inhibitory input by about 2 ms, causing an offset depolarization. This offset depolarization is
curtailed by an offset inhibitory input, which helps to generate a single offset spike with temporal
precision.
4.2 Methods
The methods (animal preparation, electrophysiological recording, sound stimulation and data
analysis) for this part of my work are the same as in Chapter 3.2.
59
4.3 Results
4.3.1 Characterization of DCN off responding neurons
We first performed in vivo loose patch recordings to characterize the response properties of DCN
off responding neurons. In the example cell shown in Figure 4.1A-C, it exhibited pauser temporal
response patterns when stimulated with CF tones at 60 dB (Figure 3.1A and 3.1B). After the
sound stopped, it consistently fired a single spike across different trials. This response was phase
locked to the sound offset since it was precisely timed for different durations of sound stimulation
(Figure 4.1A-C). A small percentage of neurons recorded in the middle frequency region of DCN
Figure 4.1 Characterizing
off responses in DCN
A, An example cell that
exhibits off responses. Top,
sample recorded trace (one
trial); Middle, PSTH of spike
responses to 50 repetitions of
60 dB CF tone stimulation;
Bottom, raster plotting of the
spike responses (each dot
represents one spike). Blue
shaded region indicate tone
stimulation duration. B, C,
Responses to 50 ms (B) and
100 ms (C) tone stimulation
of the same cell. Data are
presented in the same way as
in A. D, Summary of
normalized off spiking
possibility in response to
different durations of tone
stimulations. E, Summary of
the jitter of off response
timing in response to
different duration of tone
stimulations.
60
(11.8 ± 3.7 kHz, mean ± s.d.) could be defined as off responding neurons. These neurons had very
high probability of firing at sound offset (Figure 4.1D) and their off spikes had a very small jitter
of spike timing (Figure 4.1E). In addition, a strong suppression could be observed for these off
responding neurons (Figure 4.1A-C).
4.3.2 Membrane potential responses of off responding neurons
According to the post-inhibitory rebound depolarization hypothesis, the off responding neurons
should exhibit hyperpolarized membrane potentials during the total duration of sound stimulation.
We directly tested this hypothesis with current clamp recordings. In the example cell shown in
Figure 4.2A and 4.2B, it clearly exhibited off responses to different duration of tone stimulations.
However, its membrane potential was depolarized right before the offset response, although it was
Figure 4.2 Membrane potential
responses in off responding
neurons
A, Current clamp recordings from an
example cell that exhibits off
responses. Top, sample recorded trace
(one trial); Middle, averaged trace of
recorded membrane potentials;
Bottom, PSTH of spiking responses.
B, Responses to 100 ms tone
stimulation of the same cell. Data are
presented in the same way as in A. C,
Summary of the averaged membrane
potential value in a 5 ms window right
before offset depolarization. D, The
membrane potential response of an
example off responding neuron to a
hyperpolarized current pulse
injection. Top, averaged membrane
potential responses to 60 dB CF tone
stimulation; Bottom, membrane
potential responses to a 30 mV step
hyperpolarization.
61
briefly hyperpolarized at sound onset, which was typically for pauser response neurons (Figure
3.2C). Among all the 16 neurons we successfully recorded with current clamp recordings, most of
them exhibited depolarized membrane potentials right before sound offset (Figure 4.2C).
Therefore, the sustained hyperpolarization during sound stimulation was not observed in rat DCN
off responding neurons. Furthermore, we directly tested the contributions of IH channels. In an
example cell which exhibited off responses (Figure 4.2D), a hyperpolarized current pulse injection
did not cause the voltage sag characteristic of I H conductances(Kopp-Scheinpflug et al., 2011),
indicating that IH channel might not contribute significantly to shape the offset responses in DCN
fusiform neurons.
4.3.3 Synaptic mechanism for generating off responses
To determine whether a synaptic mechanism could explain the generation of offset firing, we
performed sequentially in vivo whole cell current clamp and voltage clamp recordings from DCN
off responding cells. An example cell exhibiting offset responses was shown(Figure 4.3A and
4.3B). The long lasting hyperpolarization after sound stimulation was consistent with the after-
stimulus suppression in spiking responses. Voltage clamp recordings from the same cell revealed
that the temporal patterns of excitation and inhibition during the sound stimulation were similar as
characterized in Figure 3.2F. After sound stopped, inhibition decayed first, followed by the decay
of excitation. Before excitation decreased to basal level, inhibitory conductance increased again
and then decreased. Simulation based on an integrate and f ire model demonstrate d that these
dynamic temporal profiles of excitatory and inhibitory conductances qualitatively determined the
off responding properties, since the derived membrane potential responses were similar as the
recorded membrane potential responses (Figure 4.3D).
62
The offset depolarization appears to be attributed to two factors. First, the decay of
excitation was slightly delayed compared to that of the first inhibitory component, resulting in an
outstanding net excitation at sound offset. This phenomenon was consistent among all the whole
cell recordings we obtained (Figure 4.3E). Second, the shortly arrived second inhibitory
component quickly hyperpolarized the membrane potential, resulting in only one spike generated
at sound offset. The off spike normally appeared very closely to the dip of the two inhibitory
components (Figure 4.3F).
4.4 Discussion
In this study we applied in vivo whole cell voltage clamp recordings to determine the synaptic
mechanisms for generating precisely timed off responses in DCN fusiform cells. We found that
the off responses in rat DCN could not be attributed to post-inhibitory rebound mechanism, since
Figure 4.3 Synaptic mechanism for generating off responses
A, PSTH of spiking responses to different duration of 60 dB CF tone stimulation. B-D,
Averaged membrane potential responses (B), calculated synaptic conductances (C) and
simulated membrane potential responses (D) to different duration of tone stimulation from
the same cell as in A. E, Summary of the timings when excitatory input and first phase of
inhibitory input start to decay relative to the sound ending point. F, plotting of the off
response timing versus the time point when inhibition level reaches minimum between the
two phases. Unity line is shown.
63
the membrane potential right before sound offset was depolarized. Instead, the temporal patterns
of excitatory and inhibitory conductances qualitatively generated the transient offset firing. The
delayed decay of excitation relative to the first phase of inhibition caused the offset depolarization.
A second phase of inhibition helped to make sure that this offset response was transient.
4.4.1 Circuit basis for delayed decay of excitation
Our results showed that in DCN off responding neurons, it was very consistent that the excitation
decayed later than the first phase of inhibition (Figure 4.3E). To understand the neural circuit
basis for this observation, it would be very helpful to consider the neural circuits of DCN (Figure
3.5A). All the off responding neurons were either buildup or pauser patterns. They were supposed
to receive strong parallel fiber inputs. The delayed auditory information received by granule cells
(Oertel and Young, 2004) and the accumulating pattern of parallel inputs (Figure 3.5B) very likely
will contribute to the delayed decay of excitatory input. The inhibitory inputs provided by vertical
and presumably D-type multipolar cells are relayed directly from auditory nerve and should decay
earlier. Inhibitory inputs from cartwheel cells might be relatively weaker. Therefore, the delayed
decay of excitatory input could be explained by the local circuits in DCN. If a cell receives
dominant excitatory input from auditory nerve, it will not generate off responses.
4.4.2 Second component of inhibitory input
While the delayed decay of excitation generated the offset depolarization, it was the second phase
of inhibitory inputs that curtailed the time window of this depolarization and made sure that only
a single spike was generated. The source of this second phase of inhibition is still unknown and
here we provide two possible candidates. First, there might be a small group of local inhibitory
neurons in DCN. They receive strong inhibition from either vertical or D-type multipolar cells and
64
generate offset firing due to the post-inhibitory rebound mechanism. This inhibitory cell type then
provides fusiform cells with offset inhibitory input. Stellate cells in the superficial layer might be
good candidates. They exhibit IH channels characteristics(Apostolides and Trussell, 2014),
although the physiological response patterns of stellate cells are totally unknown. Second, this
offset inhibition might come from a top-down projections from SPON. SPON neurons are
GABAergic and only respond to sound offset (Kadner et al., 2006; Kopp-Scheinpflug et al., 2011).
SPON neurons project to the cochlear nucleus as well as the inferior colliculus (Schofield, 1991;
Ostapoff et al., 1997). Therefore, SPON neurons are also good candidates to provide the offset
inhibitory inputs to DCN fusiform neurons.
The precisely timed second phase of inhibitory input is critical in generating off responses.
In buildup and pauser responding neurons that did not exhibit offset responses, their second
component of inhibitory input started before the fully decay of the first phase of inhibitory input,
preventing the neuron from generating off responses (data not shown). Therefore, the inhibitory
neurons that provide offset inhibitory inputs to fusiform neurons should be consisted of at least
two subgroups with different offset latencies.
4.4.3 Functional significance
In the central auditory pathway, offset responses are observed in different levels from the cochlear
nucleus to the auditory cortex (Rhode et al., 1983; Shofner and Young, 1985; Faingold et al., 1986;
He, 2001; Kadner et al., 2006; Qin et al., 2007; Scholl et al., 2010; Kopp-Scheinpflug et al., 2011).
Recent studies in SPON(Kopp-Scheinpflug et al., 2011) and A1(Scholl et al., 2010) revealed how
an inhibitory nucleus could generate off responses and how A1 neurons could inherit offset firing
from lower stages. However, to the best of my knowledge, how this response pattern is generated
65
at the first place has never been tested. In this study, we demonstrated that a novel synaptic
mechanism could account for the generation of offset responses in DCN fusiform cells. Since
auditory nerve fibers, AVCN and PVCN neurons do not exhibit off responses (Kiang et al., 1965;
Rhode and Smith, 1986), the off responses in the auditory system are very likely to be generated
in DCN and then propagated along the auditory pathway.
While potentially providing excitatory sources for all offset firing neurons along the
auditory pathway, the off responding neurons in DCN might play another important role in
generating duration selective neurons. Duration selectivity is found to be generated in IC
(Casseday et al., 1994). Three factors are proposed to be required to generate duration selective
firings: delayed onset excitation, offset excitation and onset inhibition (Casseday et al., 2000).
When delayed onset excitation temporally overlaps with the offset excitation, the neuron will fire
maximally to this specific duration. While the IC neurons are well known to receive onset
excitatory inputs with various latencies, the sources of this offset excitatory inputs are less clear.
Our findings raise the possibility that DCN off responding neurons provide offset excitatory inputs
to duration selective neurons in IC. Besides contributing to the generation of duration selective
neurons in IC, the transient off spiking neurons themselves might directly contribute to the
encoding of sound duration. When responses to both onset and offset of a sound stimulation are
represented in single neurons (so called on-off neurons) or neuronal populations, the duration of
the sound should theoretically be easy to derive. The transient feature of the offset firing generated
in DCN will further help to encode the fast-changing sound components in the real world.
66
Chapter 5 Modulation of Synaptic inputs in Auditory
Cortex by Active Behavioral States
5.1 Introduction
The processing of sensory information in cortical neurons is achieved through the spatiotemporal
integration of converging synaptic inputs evoked by sensory inputs(Oswald et al., 2006; Wu et
al., 2011; Petersen and Crochet, 2013). Such synaptic integration is largely determined by the
structure of the underlying functional cortical synaptic circuits(Callaway, 1998; Douglas and
Martin, 2004; Wu et al., 2011), but can also be influenced by behavioral and cognitive states of
the animal(Fanselow and Nicolelis, 1999; Reynolds and Chelazzi, 2004; Ferezou et al., 2007;
Lee et al., 2008; Otazu et al., 2009; Niell and Stryker, 2010) which modulate the internally
generated brain activities(Crochet and Petersen, 2006; Poulet and Petersen, 2008; Goard and
Dan, 2009; Constantinople and Bruno, 2011). In visual and somatosensory cortices, it has been
shown that behaviorally active states, such as locomotion and whisking, result in a depolarization
of the membrane potential and a more desynchronized state of cortical neurons(Gentet et al.,
2010; Zagha et al., 2013; Polack et al., 2013), which alters the level or reliability of their spike
responses to sensory stimulation(Ferezou et al., 2007; Niell and Stryker, 2010; Zagha et al.,
2013; Polack et al., 2013; Bennett et al., 2013). Despite the observed changes in membrane
potential dynamics, how behavioral states directly modulate cortical synaptic circuits, as
reflected by potential changes of excitatory and inhibitory synaptic inputs to a cortical neuron,
remains largely unknown.
67
In this study, by achieving high-quality in vivo whole-cell voltage-clamp recordings in
awake head-fixed mice, we were able to reveal excitatory and inhibitory synaptic inputs to the
same cortical neurons under different behavioral states of the animal. In middle layers of the
primary auditory cortex (A1), our results revealed a robust functional balance between sound-
evoked excitatory and inhibitory inputs to a cortical neuron under various behavioral states,
which is a salient synaptic circuit property previously demonstrated in anesthetized animal
models(Wehr and Zador, 2003; Zhang et al., 2003; Tan et al., 2004; Wu et al., 2008; Wu et al.,
2011). The balanced synaptic excitation and inhibition were found scaled down at a similar level
during active states as compared to the quiet resting state in layer 2/3 but not layer 4 excitatory
cells, resulting in well preserved sensory tuning of the former cells. We also provided evidence
that layer 1 interneurons were activated in active states, which contributed to the reduced
response gain of layer 2/3 excitatory cells. Together, our results suggest that balanced excitation
and inhibition is a fundamental synaptic circuit basis for auditory cortical processing in the
awake A1, and that behavioral state-dependent scaling of excitatory and inhibitory inputs may be
a general strategy for cortical circuits to adjust the representation of sensory information
according to momentary behavioral and task demands.
5.2 Methods
5.2.1 Awake animal preparation
All experimental procedures used in this study were approved by the Animal Care and Use
Committee at the University of Southern California. Female C57BL/6J mice aged 5- 7 weeks
were used in this study. Animals for awake recordings were prepared in a similar way as
previously described(Olsen et al., 2012; Xiong et al., 2013). Mice were housed with 12 hours
68
light/dark cycle and with flying saucer pet exercise wheels placed in their home cages. One week
before the recording, the mouse was anesthetized with isoflurane (1.5%) and a screw for head
fixation was mounted on top of the skull with dental cement. An adaptor for connecting to an
enclosed sound delivery system was attached to the left ear. Afterward the mouse was injected
subcutaneously with 0.1 mg/kg buprenorphine and returned to its home cage. During the
recovery period, the mouse was trained to get accustomed to the head fixation on the recording
setup. To fix the head, the screw was tightly fit into a metal post. The animal was allowed to run
freely on a flat plate rotating smoothly around its center. On the day of recording, the mouse was
anesthetized with isoflurane. Surgery was performed in a sound-attenuation booth (Acoustic
Systems). Craniotomy over the A1 region was performed and the dura was removed. The animal
was positioned with the left ear connected to a calibrated closed acoustic delivery system using a
TDT EC1 speaker. The right ear was plugged. Multi-unit recordings were made with a tungsten
electrode (2 M©, FHC) to identify the A1 based on response properties and the tonotopic
gradient, as described in previous studies (Guo et al., 2012). The animal head was tilted so that
the electrode could penetrate the A1 surface at an angle of 80
o
. The animal was allowed to
recover from isoflurane for at least 1 hour. Recording was started after the animal exhibited
normal running. The recording session lasted for about 4 hours. The animal was given drops of
5% sucrose through a pipette every hour.
5.2.2 In vivo whole-cell and loose-patch recordings in awake animals
Whole-cell recordings were made with an Axopatch 200B amplifier (Molecular Devices). Patch
pipettes (impedance of 4 5 M©) contained a cesium-based solution (in mM): 125 Cs-gluconate,
5 TEA-Cl, 4 MgATP, 0.3 GTP, 10 phosphocreatine, 10 HEPES, 10 EGTA, 2 CsCl, 1.5 QX-314,
1% biocytin or 0.25 fluorescent dextrans, pH 7.3. The patch pipette, controlled by a
69
micromanipulator (Siskiyou), was lowered into the A1 at the same angle as in multi-units
recordings. The cortical surface was covered with 3.5% agar prepared in a warm artificial
cerebrospinal fluid (ACSF; in mM: 124 NaCl, 1.2 NaH2PO4, 2.5 KCl, 25 NaHCO3, 20 Glucose,
2 CaCl2, 1 MgCl2). Whole cell capacitance was fully compensated and the initial series
resistance (Rs; 15 50 M©) was compensated for 40 60% to achieve an effective R s of 10 30
M©. For some recordings, we regularly monitored Rs before, during and after epochs of animal
movements. The running epochs were relatively evenly spaced during each of recording
sessions, which usually last for less than 10 minutes, enough for us to collect data for different
states. There was no significant change (< 10%) of R s during our effective recording sessions.
Signals were low pass filtered at 2 kHz and sampled at 10 kHz. Only cells with resting
membrane potential lower than 50 mV were studied. A- 10 mV junction potential was
corrected. Excitatory and inhibitory synaptic currents were recorded under the voltage-clamp
mode with the cell clamped at- 70 mV and 0 mV respectively. Membrane potentials were
recorded under the current-clamp mode with pipettes containing a potassium-based solution (in
mM): 130 K-gluconate, 4 MgATP, 0.3 GTP, 8 phosphocreatine, 10 HEPES, 10 EGTA, 5 KCl, 1
CaCl2, 0.25 fluorescein dextran, pH 7.3. Signals were low pass filtered at 5 kHz and sampled at
10 kHz. As demonstrated before(Wu et al., 2008; Zhou et al., 2010; Li et al., 2013b), the blind
whole-cell recording method with relatively large pipette openings resulted in almost exclusive
sampling from excitatory cortical neurons. Loose-patch recordings were performed as
previously described (Li et al., 2013a), with a pipette filled with ACSF. Signals were recorded in
voltage-clamp, with a command voltage applied to adjust the baseline current to be zero. Loose-
patch recordings from MGBv neurons were made by vertically penetrating the brain (2.8 3.6
mm from Bregma, 1.7 2.2 mm from midline, 2.8 3.2 mm below the pia surface). MGBv was
70
discriminated from other auditory thalamic regions as previously described (Li et al., 2013a).
LFP recordings were made with the same recording pipette as in loose-patch recordings. During
recordings, behaviors of the animal were recorded with a video camera. The rotating speed
(without distinguishing the rotation direction) of the plate was detected with an optical sensor
and recorded simultaneously. The behavioral and rotating speed recordings were precisely timed
with the electrophysiological recording. The behavioral state of the animal was analyzed both
online and offline.
On average, one good whole-cell recording (maintained for 20 40 minutes) or two loose-
patch recordings (maintained for more than 1 hour) was obtained in each well-trained animal.
The recording sites were marked. The laminar locations of the recorded neurons were determined
based on the micromanipulator reading, and in some cases confirmed by histology of the track of
pipette penetration and/or fluorescence or biocytin labeled cell bodies. We found a relatively
good correspondence between the traveling depth of the recording pipette from the pia and the
reconstructed laminar location of the recorded neuron. The depth range of different layers in
mouse A1 was determined based on the results from Nissl staining and fluorescence expression
pattern in a layer-4-specific Cre line (Scnn1a-Tg3-cre, Jackson Laboratory) crossed with the
Ai14 reporter mouse. The L2/3 neurons were sampled at a cortical depth of 250 350 µm from
the pial surface, L4 neurons at a depth of 375 500 µm(Li et al., 2013a), and L1 neurons were
within 100 µm from the pia.
5.2.3 Optogenetically guided loose-patch recordings from PV neurons
Adult PV-Cre (Jackson Laboratory) female mice were anesthetized with 1.5% isoflurane. A
small cut was made on the skin covering the right A1 and the muscles were removed. Two ~0.2
71
mm craniotomies were made in the A1 region (temporal lobe, 2.7 and 3.2 mm caudal to
Bregma). Adeno-associated viruses (AAVs) encoding ChR2 were purchased from the
University of Pennsylvania Viral Vector Core: AAV2/9.EF1±.DIO.hChR2(H134R)-
EYFP.WPRE.hGH (Addgene 20298). The virus was delivered using a beveled glass
micropipette (tip diameter: ~40 µm) attached to a microsyringe pump (World Precision
Instruments). Injections were performed at two locations and two depths (300 and 600 µm), at a
volume of 100 nl per injection and at a rate of 20 nl min
1
. Right after each injection, the pipette
was allowed to rest for 4 minutes before withdrawal. We then sutured the scalp, injected 0.1 mg
kg
1
buprenorphine and returned the mouse to its home cage. Mice were allowed to recover for
3 4 weeks. On the day of recording, loose-patch recordings using pipettes of smaller tip
openings (pipette impedance ~10 M©) were performed. An optic fiber connecting to a blue LED
source (470 nm, Thorlabs) was positioned close to the cortical surface of the recording site. We
actively searched for neurons exhibiting LED evoked spikes, which were identified as PV
neurons. After each experiment that brain was sectioned and imaged to confirm the expression of
ChR2-EYFP.
5.2.4 Silencing L1 with TTX
This method was adapted from a previous study(Shlosberg et al., 2006). Firstly we examined the
time course of TTX effects in each layer. A glass pipette containing 1 M NaCl was used for
recording multi-unit spikes. Multi-unit recording was made at 70 µm (L1), 250 µm (L2/3) or 425
µm (L4) below the pia surface of the A1. Responses to repetitive 50 dB CF tones (ISI = 4 s) were
measured before and after TTX application. TTX solution (5 µM) was applied through a glass
micropipette(~100 µm opening) attached via polyethylene tubing to a syringe. Each time we
loaded ~ 2 µl TTX in the pipette and applied a very small pressure so that the TTX solution
72
could be gently applied onto the A1 surface. To determine the contribution of L1 activity to gain
changes in L2/3, loose-patch recordings were performed in L2/3 or L4, while the behavior and
running speed of the animal were simultaneously recorded. Responses to repetitive 50 dB CF
tones (ISI = 2 s) were measured before and after TTX application. Spike responses during a ~
150 seconds window after TTX application when L2/3 firing rates became stable were analyzed
and compared to responses before TTX application.
5.2.5 Sound stimulation
Software for sound stimulation and data acquisition was custom-developed in LabVIEW
(National Instruments). For loose-patch recordings, pure tones (2 32 kHz spaced at 0.1 octave,
50 ms duration, 3 ms ramp) at eight intensities (0 70 dB SPL spaced at 10 dB) were delivered
pseudo-randomly. Inter-stimulus interval (ISI) was 0.5 s. Spike TRFs were continuously mapped
to obtain TRFs in different states. It took ~2.5 hours to map TRFs for more than 50 repetitions,
from which we were able to reconstruct TRFs of about 10 repetitions for active states. For
whole-cell recordings, either best frequency tones at seven sound intensities (10- 70 dB SPL
spaced at 10 dB) were delivered randomly with ISI = 2 s, or 40 dB tones at twenty-one
frequencies (2- 32 kHz, spaced at 0.2 octave) were delivered pseudo-randomly with ISI = 1 s.
Thus, data collection was randomized.
5.2.6 Data analysis
We performed data analysis with custom-developed software (MATLAB, MathWorks).
Analysis performers were partially blind to the conditions of the experiments, since the data from
all the recorded neurons were first pooled together for a randomized batch processing without
categorizing the neurons according to their laminar locations.
73
The three behavioral states (Figure 5.1B), quiescence (Q), active without locomotion
(A- L), and locomotion(L), were identified based on the body movement and the rotation
speed of the plate. In Q state, there was no obvious body movement and the average rotation
speed (in each 1-s epoch) was lower than 0.5 cm s
1
. Animal normally stayed in Q state for
more than 70% of the recording time. In A- L state, the animal showed whisking and/or
facial/jaw/paw movements, which caused small back and forth movements of the plate with an
average speed below 2 cm s
1
. In L state, the animal was running forward, with the rotating
speed consistently above 2 cm s
1
. During locomotion the mouse also whisked. The two active
states each will take about 10% of the recording time. Well trained animals to spend relatively
more time on running. In a typical experiment, quiescence and active states of the animal were
intermingled.
LFP signals were low-pass filtered at 300 Hz. After fast Fourier transformation of the
signal, normalized power spectrum was obtained and power ratio was calculated as the ratio of
the area under 1 10 Hz band over that under 20 80 Hz band. Loose-patch recording signals
were filtered with a 100 5000 Hz band-pass filter. Spontaneous firing rates were calculated
from spikes within a 200 ms window before tone onsets, or within 6 s segments of records during
which only spontaneous spikes were recorded. Spike TRF was determined as the frequency-
intensity space where firing rates exceeded the average spontaneous level by 2 s.d. of baseline
fluctuations. Evoked firing rates were calculated by subtracting the average spontaneous firing
rate. CF was defined as the frequency at which tones evoked a significant spike response with
minimum intensity. This minimum intensity was the intensity threshold of TRF. PSTHs were
derived from CF-tone evoked responses. Spike response latency was defined as the lag between
the stimulus onset and the time point in the PSTH where evoked firing rate exceeded the average
74
spontaneous firing rate by 2 s.d. of baseline activity. Signal-to-noise ratio (SNR) was calculated
for CF-tone evoked responses as the evoked firing rate within a 50 ms time window following
the tone onset divided by the average spontaneous firing rate.
Synaptic response traces evoked by the same test stimuli were averaged separately for
each behavior state. Synaptic onset latency was determined at the time point where the evoked
current exceeded the average baseline by 2 s.d.. Peak amplitude was determined by averaging
within a 5 ms window centered at the response peak after subtracting baseline current. Charge
transfer was calculated by summing up the current values within the evoked response time
window after subtracting baseline current. Excitatory and inhibitory synaptic conductances were
derived according to I = G e * (V Ee) + Gi * (V Ei). I is the amplitude of the synaptic
current at any time point after subtracting the average baseline current specific to each behavior
state; Ge and Gi are the excitatory and inhibitory synaptic conductance; V is the holding voltage,
and Ee(0 mV) and Ei(- 70 mV) are the reversal potentials. The clamping voltage V was
corrected from the applied holding voltage (V h): V = Vh Rs * I, where Rs is the effective series
resistance. By holding the recorded cell at two different voltages (the reversal potentials for
excitatory and inhibitory current respectively), Ge and Gi could be resolved from the equation
(Borg-Graham, et al., 1998; Anderson, et al., 2000). Resting conductance was calculated based
on the average baseline currents within a 50 ms window before the onset of evoked currents
recorded under two different voltages (- 70 mV and 0 mV).
To determine frequency tuning, peak amplitudes of synaptic inputs at different
frequencies were fit with an envelope curve using a MATLAB software Envelope 1.1 (developed
by Lei Wang, MathWorks), as previously described(Sun et al., 2009). Total frequency range
75
and BW50% were defined as the bandwidths of the fitted envelope curve at 10% and 50% of
maximum level respectively. BF was defined as the frequency corresponding to the maximum
of the fitted envelope curve.
5.2.7 Statistics
Shapiro-Wilk test were first applied to exam whether samples had a normal distribution. In the
case of a normal distribution, t-test or ANOVA test was applied. Otherwise, a non-parametric
test (Wilcoxon signed-rank test in this study) was applied. Data were presented as mean ± s.d. if
not otherwise specified. In this study, all the representative cases are followed by a population
summary to demonstrate the repeatability. No statistical test was run to determine sample size a
priori. The sample sizes we chose are similar to those used in previous publications. Since for
many cells, responses of the same recorded neuron in different states were tested for multiple (>
10) times, paired t-test or Wilcoxon signed-rank test was also performed on an individual-cell
basis. The results were generally consistent with the group comparison. For the linear
regression, both the correlation coefficient (r) and P value were calculated to evaluate the
strength and significance of the linear correlation. r values were indicated for individual linear
regressions and summarized in Figure 5.2G and Figure 5.5M. For all the L2/3 neurons in
Figure 5.2, P values for the correlation between responses in two states (similar as in Figure
5.2B) were all lower than 1×10
- 9
. P values for the correlation between excitatory and inhibitory
responses of the neurons in Figure 5.4(similar as in Figure 5.4G) were all lower than 1×10
- 8
. P
values for the correlation between synaptic responses in two different states in Figure 5.5 were
all lower than 1×10
- 5
.
76
5.3 Results
5.3.1 Laminar-specific down-regulation of auditory responses
We first examined whether and how auditory cortical responses are modulated by changes of
behavioral state in awake head-fixed mice habituated to rest or run on a flat rotatable plate (see
Methods). The behavior of the animal was monitored with a video camera, and the speed of the
rotation of the plate was recorded in real time (Figure 5.1A). The animal displayed three
identifiable behavioral states (Figure 5.1B): quiescence (Q, quiet resting), active without
locomotion (A- L, whisking and/or facial/jaw/paw movements), and locomotion (L,
running). During locomotion the mouse also whisked. These behavioral states correlated well
with different speeds of plate rotation (Figure 5.1B, 5.1C). A- L state caused small back and
forth movements of the plate, the speed of which was clearly distinguished from that caused by
locomotion (Figure 5.1C). The power spectrum of the local field potential (LFP) recorded in the
A1 (Figure 5.1D) showed an increase in the power of high frequency oscillations (20 80 Hz) while a decrease in the power of low frequency oscillations (1 10 Hz) during both the A- L
and L states as compared to the Q state ( Figure 5.1E, 5.1F), consistent with previous reports
that locomotion or whisking can result in a desynchronized brain state (Crochet and petersen,
2006; Poulet and Petersen, 2008; Niell and Stryker, 2010). After determining the location of the
primary auditory cortex (A1) with extracellular recordings, we performed in vivo cell-attached
loose-patch recordings from individual A1 neurons (see Methods). Both the spontaneous and
sound-evoked spikes of the cells were recorded. All the cells in this recorded population were
presumed excitatory cells, since they all exhibited broad spike waveforms (see insets in Figure
5.1G, 5.1J for example spike waveforms). Compared to the Q state, A - L and L states
77
similarly reduced the spontaneous firing rate in layer 2/3 (L2/3) neurons (Figure 5.1G-I), but did
not affect it in layer 4 (L4) neurons (Figure 5.1J-L). Such L2/3-specific decrease in spontaneous
activity is reminiscent of a previous report in rat auditory cortex that spontaneous activity of
78
superficial neurons is suppressed during cortical desynchronization(Sakata and Harris, 2012).
Furthermore, A- L and L states similarly reduced the spiking response to the characteristic
frequency (CF) tone in L2/3 neurons as compared to the Q state ( Figure 5.1M-O), whereas in
L4 neurons the CF-tone evoked spiking activity was not affected by the changes of behavioral
state (Figure 5.1P-R). Thus, different from the observations in the visual cortex(Niell and
Stryker, 2010; Polack et al., 2013; Bennett et al., 2013), where locomotion results in enhanced
sensory-evoked activity in both layer 2/3 and layer 4, the auditory cortex exhibited a laminar-
specific down-regulation of sensory-evoked responses. In addition, it is worth mentioning that
both the spontaneous and evoked firing rates are higher in awake quiescence than urethane-
anesthetized states (spontaneous: 2.29 ± 0.89 Hz for anesthesia, 4.06 ± 1.1 Hz for awake, P <
Figure 5.1 Behavioral state-dependent modulation of spike responses in the mouse A1
A, Experimental setup. R, recording electrode; P, head-fixation post; S, sound stimulation; C, camera; v, velocity
meter. B, Sample records of plate rotation speed in different behavioral states. Q, quiescence; A L, active
without locomotion; L, locomotion.C, Distribution of average speeds (within a 1 sec epoch) in randomly
sampled 2000 epochs (N = 3 animals). D, Top, sample records of LFP in the A1. Scale: 250 µV and 0.5 s.
Middle, simultaneously recorded plate rotating speed. Arrow indicates speed at 0. Scale: 20 cm s
1
and 0.5 s.
Bottom, power spectrum of LFP. E, Top, percentage change in power of low-frequency (1 10 Hz) and high-
frequency(20 80 Hz) components of LFP relative to quiescence, for the recording in D. Power spectrums were
generated for each 3.3 sec segment of LFP records. Bar = s.d. N = 5 segments. From left to right, t-test (**P =
0.0035, t = 5.118), Wilcoxon signed rank test (*P = 0.0313, Z = 1.888), t-test (***P = 0.0006, t = 8.268), t-test
(**P = 0.0012, t = 6.865). Bottom, the ratio of power of the low versus high-frequency component. One-way
ANOVA (P = 6.68×10
6
, F = 37.73) and post hoc test (**P < 0.01, ***P < 0.001, same for the below). F,
Summary of recordings in 6 animals. Power ratio was normalized by the average value in the Q state. Top, t-
test (**P = 0.0011, 0.0016, ***P = 0.0003, 0.0006; t = 5.802, 5.302, 7.453, 6.637 respectively). Bottom, one-
way ANOVA (P = 1.75×10
-5
, F = 24.80) and post hoc test. G-I, Spontaneous firing in L2/3 neurons in different
states. G, Top, records of spontaneous spikes of a L2/3 excitatory cell. Bottom, simultaneously recorded plate
rotation speed. Arrow indicates speed at 0. Right inset, superimposed 500 individual spikes. H, Average
spontaneous spike rates in the same cell. Bar = s.d. One-way ANOVA (P = 1.26×10
- 5
, F = 12.88, N = 30 5-sec
segments.) and post hoc test. I, Summary of 17 recorded L2/3 excitatory cells. Spike rate was normalized by
the average value in the Q state. One -way ANOVA (P = 2.66×10
- 7
, F = 21.10) and post hoc test. J-L,
Spontaneous spikes recorded in L4 excitatory cells. Data are presented similarly as in G-I. K, One-way ANOVA
(P = 0.9841, F = 0.0160, N = 28 segments) and post hoc test. L, One-way ANOVA (P = 0.1542, F = 1.955, N =
15) and post hoc test. M, Peri-stimulus spike time histogram (PSTH, bin size = 1 ms) for the responses of a L2/3
excitatory cell to CF tones (black lines) in different states. Inset, sample record of evoked spikes by the tone. N,
Average evoked spike number per stimulus trial plotted for the same cell. Bar = s.d. One-way ANOVA (P =
1.21×10
- 5
, F = 12.93, N = 25 trials) and post hoc test. O, Summary of average evoked spike numbers for 17
similarly recorded L2/3 excitatory cells. One-way ANOVA (P = 1.89×10
- 5
, F = 13.76) and post hoc test. P, Q,
Evoked spike responses of a L4 excitatory cell. N = 25 trials. One-way ANOVA (P = 0.7415, F = 0.3004) and
post hoc test. R, Summary of average evoked spike numbers for 15 recorded L4 cells. One-way ANOVA (P =
0.1708, F = 1.844) and post hoc test.
79
0.05, t-test; evoked: 50.2 ± 12.7 Hz for anesthesia, 60.2 ± 11.7 Hz for awake, P < 0.05, t-test, N
= 55 and 32 respectively).
5.3.2 Behavioral state-dependent gain modulation
To determine the nature of the behavioral state-dependent modulation of auditory processing, we
continuously mapped the frequency-intensity tonal receptive field (TRF) of spiking response in
an individual cell by applying tone pips of different frequencies and intensities (see Methods).
Trials during active (A) states were separated from those in quiescence. Since A- L state
produced similar effects as locomotion (Figure 5.1N, 5.1Q), we did not further separate the trials
in these two states. The spike TRF was reconstructed from at least ten complete sets of spike
responses to 41 testing frequencies and 8 testing intensities. In the example L2/3 cell in Figure
5.2A, A state apparently reduced the firing rates without affecting the overall shape of the
spike TRF. Within the TRF, the average evoked spike number in an A state trial strongly
correlated with that in the corresponding Q state trial(Pearsons correlation coefficient r =
0.94) (Figure 5.2B), indicating that firing rates to different tone stimuli were reduced in
proportion. In other words, firing rates were scaled down. The slope of the linear regression line
(0.71) indicated an about 30% reduction in response gain (Figure 5.2B). Concurrently, the onset
of the evoked spike response was slightly delayed in A than Q state ( Figure 5.2C, arrows),
together with a reduced level of spontaneous activity. In comparison, in the example L4 cell, A
state did not apparently affect evoked or spontaneous firing rates or the shape of the TRF
(Figure 5.2D-F), with the slope of the linear regression line close to 1 (0.98, Figure 5.2E). The
onset delay of the evoked spike response was neither changed apparently (Figure 5.2F).
80
In all the 17 recorded L2/3 excitatory cells, the evoked spike rate in Q s tate correlated
well with that in A state ( Figure 5.2G). The slopes of the linear regression line for A -state
versus Q -state responses were nearly all below 1, with a mean ± s.d. of 0.77 ± 0.14 (Figure
5.2H). This indicates that the response gain of L2/3 excitatory cells was reduced in active states.
Figure 5.2 Gain modulation of auditory responses by behavioral state
A, TRFs of spike responses of a L2/3 cell in quiescence (Q) and active (A) states. Each element in the array represents
the PSTH of evoked spikes (60 ms time window, bin size = 2 ms, 10 repeats) to the corresponding tone stimulus.
Color map depicts the average spike number evoked by tones. B, Evoked spike number by a tone in active state
plotted against that by the same stimulus in quiescence, for the cell shown in A. The best-fit linear regression line is
shown. C, Normalized PSTHs for all the tone responses in two states, for the cell shown in A. Arrows indicate
response onsets.D t is the onset difference. D-F, TRFs of a L4 cell. Data are presented in the same way as in A-C.
G, Distribution of correlation coefficients (r) for all the recorded L2/3 cells. H, Distribution of slopes of the linear
regression (i.e. the gain value). Arrow points to the mean value. I, Distribution of differences in onset latency (A
Q) of spike responses. J, Characteristic frequency (CF) of spike TRF in A versus Q state. The best -fit linear
regression line is shown. K, Bandwidth at 20 dB above the intensity threshold (BW20) of spike TRF in A versus
Q state. L, Intensity threshold of spike TRF in A versus Q state. M-R, Summary for L4 cells. Data are
presented in the same way as in G-L.
81
Most of these cells showed a 1 2 ms (1.5 ± 0.7 ms, mean ± s.d.) delay of spiking response onset
in A state relative to Q state ( Figure 5.2I). In 10 of these cells, complete TRFs for both A
and Q s tates were obtained. For these TRFs, no significant difference was observed for the
characteristic frequency (Figure 5.2J), the bandwidth at 20 dB above the intensity threshold
(BW20, Figure 5.2K), or the intensity threshold of TRF (Figure 5.2L) between the two different
states, indicating that the shape and sharpness of TRF remained the same despite the change in
response gain. No correlation was found between the scaling factor and the cortical depth of
recorded L2/3 cells. In comparison, in all the 15 recorded L4 cells, the slopes were close to 1
(1.02 ± 0.09, mean ± s.d.), indicating no change in response gain (Figure 5.2M, 5.2N). The onset
latency of evoked spiking response was overall unaffected by changes of behavioral state
(latency = 0.1 ± 0.8 ms,Figure 5.2O), nor was the shape of spike TRF changed (Figure 5.2P-
R).
The absence of response changes in L4 suggests that activity of thalamic neurons may not
change during active behaviors. To confirm this, we performed loose-patch recordings in the
ventral part of the medial geniculate body (MGBv), the thalamic nucleus that provides direct
feedforward input into the A1. Indeed, neither the spontaneous (Figure 5.3A) nor the evoked
spike responses of thalamic neurons (Figure 5.3B) changed significantly from the Q to A
state, indicating that auditory thalamic neurons were not directly affected by the behavioral
changes.
Although the evoked firing rate of L2/3 cells was reduced in active states, their
spontaneous activity was relatively more suppressed so that the signal-to-noise ratio (SNR), as
defined by the ratio of evoked firing rate over spontaneous firing rate, was in fact increased
(Figure 5.3C). This increased SNR, together with the unchanged intensity threshold (Figure
82
5.2L), indicates that the sensitivity of auditory processing was not reduced in active states
despite the decreased response level.
5.3.3 Balanced excitation and inhibition in quiescent state
It remains unknown what are the properties of sound-driven synaptic excitation and inhibition to
auditory cortical neurons in the awake brain, since previous studies have been mostly carried out
in anesthetized preparations(Wehr and Zador, 2003; Zhang et al., 2003; Tan et al., 2004; Wu et
al., 2008). To understand the synaptic mechanisms underlying the observed behavioral state-
dependent modulation of response gain, we performed whole-cell voltage-clamp recordings to
reveal excitatory and inhibitory synaptic inputs to A1 neurons (see Methods). We first examined
properties of synaptic responses in quiescence. As shown by an example cell, a best-frequency
Figure 5.3 Activity of thalamic neurons and
summary of SNR under different behavioral
states
A, Left, sample records of spontaneous spikes in
an MGBv neuron in Q and A states. Right,
summary of average spontaneous firing rates (N =
12 cells). Data points from the same cell are
connected with a line. Solid symbol represents
mean ± s.d. (paired t-test, P = 0.4795, t = 0.7320,
N = 12 cells). B, Left, PSTH for CF-tone evoked
spikes in different states in the same MGBv neuron
as shown in A. Right, summary of average evoked
spike numbers by CF tones in different states
(paired t-test, P = 0.7640, t = 0.3078, N = 12
cells). C, SNR in A relative to Q state. There
is a significant increase in L2/3 cells. From left to
right, Wilcoxon signed rank test (**P = 0.0017, Z
= 3.435), t-test (P = 0.5834, t = 0.5614), t-test (P =
0.6282, t = 0.4982).
83
Figure 5.4 Properties of synaptic responses in quiescence
A, Average traces of excitatory and inhibitory currents evoked by a best-frequency (BF) tone at 40 dB SPL in an
example pyramidal neuron. Black line indicates tone duration. Red line marks 50% duration of the current. Scale:
50 (exc) /100 (inh) pA, 100 ms. B, Durations of BF-tone evoked synaptic currents. Short, 50 ms; long, 200 or 500
ms. Wilcoxon signed-rank test, P = 0.81, 0.73; Z = 0.2439, 0.4024; N = 9, 3 for exc and inh respectively. C,
Comparison of peak amplitudes (paired t-test, P = 0.4, 0.89; t = 0.8876, 0.1520; N = 9, 3 for exc and inh respectively).
D, Differences in excitatory and inhibitory onset latency (Inh Exc). Arrow points to the mean value. E, Comparison
of peak amplitudes. Wilcoxon signed-rank test, ***P = 0.0001, Z = 4.464, N = 21. F, Distribution of E/I ratios of
peak conductances. G, Frequency tuning of excitation and inhibition in an example cell. Left, average excitatory and
inhibitory currents to tones at different frequencies (interval, 0.2 octave). Scale: 30 (exc) / 100 (inh) pA, 200 ms.
Right, peak amplitude of excitatory versus inhibitory current evoked by the same stimulus. Top inset, superimposed
normalized frequency tuning curves for excitation (red) and inhibition (blue). H, Another example cell plotted
similarly as in G. Scale: 100 (exc) / 500 (inh) pA, 200 ms. I-K, Comparison of frequency range (I), half-peak
bandwidth (J), and BF (K) between excitation and inhibition in the same cell (P = 0.52, 0.56, 0.46, respectively, paired
t-test, N = 15 cells). The unity line is shown.
84
(BF) tone elicited robust excitatory and inhibitory currents in the same cell (Figure 5.4A). The
major component of these currents could be best described as transient, as its temporal duration,
as measured at the half-peak level (i.e. 50% duration), did not increase with increasing tone
durations (Figure 5.4B). In addition, the peak amplitude of synaptic currents did not increase
either with increasing stimulus durations (Figure 5.4C). The onset latency of inhibitory current
was mostly delayed by 1 3 ms (1.95 ± 1.03 ms, mean ± s.d.) relative to that of the excitatory
current evoked by the same stimulus (Figure 5.4D), similar as what has been observed in
anesthetized animals(Wehr and Zador, 2003; Zhang et al., 2003; Tan et al., 2004; Wu et al.,
2008; Zhou et al., 2010), and suggesting a fast feedforward nature of the inhibition(Wu et al.,
2011; Zhou et al., 2012). The amplitude of inhibitory conductance was larger than that of the
corresponding excitatory conductance (Figure 5.4E), with a mean E/I ratio of 0.42 ± 0.12
(Figure 5.4F), which is in line with our previous studies in anesthetized animals (Wu et al.,
2006; Sun et al., 2010; Zhou et al., 2010). We further compared the frequency tuning of
excitation and inhibition at a moderate intensity level (40 50 dB sound pressure level, SPL). As
shown by two example cells (Figure 5.4G,5.4H), the total frequency ranges for evoked
excitatory and inhibitory currents were about the same, and the amplitude of excitatory current
linearly correlated with that of the inhibitory current evoked by the same stimulus, resulting in
very similar excitatory and inhibitory frequency tuning curves (Figure 5.4G, 5.4H, inset). This
finding suggests a functional balance of excitation and inhibition, i.e. the strength of inhibition
co-varies with that of excitation. In a total of 15 recorded cells, we observed that the frequency
range of excitation was not different from that of inhibition (Figure 5.4I). The bandwidth of
excitatory frequency tuning curve, as measured at the half-peak level (BW 50%), was not
different from that of the inhibitory tuning curve of the same cell (Figure 5.4J). Excitatory and
85
inhibitory tuning curves also exhibited the same best frequency (Figure 5.4K). Together, these
results indicate that in awake A1 excitation and inhibition are balanced in the frequency domain,
with inhibition temporally delayed briefly relative to excitation.
5.3.4 Scaling down of excitation and inhibition in active states
To examine how behavioral state regulates synaptic inputs, we applied best-frequency tones and
varied their intensity in a random order. Responses were parsed into active state (including
locomotion) and quiescence trials. As shown by an example L2/3 excitatory cell, the average
excitatory current in response to BF tones increased in amplitude with increasing tone intensities
(Figure 5.5A). In the A state the excitatory currents to all tone intensities became smaller
(Figure 5.5A) in an approximately proportional manner (Figure 5.5B, left and middle panel).
The onset latencies of evoked excitatory currents were not changed significantly (Figure 5.5B,
right panel). The evoked inhibitory currents to different tone intensities were also reduced by a
similar factor from the Q to A state, while the timing of inhibitory currents was not affected
(Figure 5.5C, 5.5D). In comparison, in a L4 excitatory cell, neither the excitatory nor inhibitory
currents changed significantly in their amplitudes, nor were the onset latencies of these currents
affected (Figure 5.5E-H).
To examine state-dependent changes of synaptic responses in the frequency domain, we
applied tone stimuli of different frequencies at a moderate intensity (40 or 50 dB SPL). As
shown by an example L2/3 excitatory cell (Figure 5.5I), tone-evoked excitatory and inhibitory
currents were both reduced in amplitude from the Q to A state. The strong linear relationship
between synaptic responses in the two states indicated a scaling of both excitatory and inhibitory
inputs (Figure 5.5J). Such coordinated modulation resulted in apparently unchanged frequency
86
87
tuning of synaptic inputs, as demonstrated by the superimposed normalized synaptic tuning
curves for A and Q stat es (Figure 5.5I, inset). Consistent with the results of spiking
response, in layer 4 we did not observe apparent effects of A state on the amplitude of either
evoked excitatory or inhibitory synaptic responses (Figure 5.5K, 5.5L). Together, the above
results suggest that tone-evoked excitatory and inhibitory inputs to L2/3 cells were scaled down
in behaviorally active states, thus preserving the functional balance between excitation and
inhibition.
In all the 13 recorded L2/3 cells, we observed a strong linear correlation between evoked
synaptic amplitudes in A and Q states for both excitation and inhibition ( Figure 5.5M),
further demonstrating a scaling of excitatory and inhibitory responses by changes of behavior
state. To determine the scaling factor, we calculated the ratio of peak response amplitude in A
versus Q state for each tone -evoked response. As shown by the distribution of ratios for all the
recorded neurons (Figure 5.5N), in L2/3 the scaling factors were nearly all lower than 1 for both
excitation and inhibition, with a mean ± s.d. of 0.64 ± 0.18 and 0.70 ± 0.18 respectively. In
Figure 5.5 Modulation of synaptic responses by behavioral state
A, Average evoked excitatory currents to BF tones at different intensities in quiescence and active states in a L2/3
excitatory cell. Scale: 200 pA, 100 ms. B, Left, peak excitatory amplitudes in active versus quiescence state with
the best-fit line shown. The near zero point depicts the responses to 10 dB tones not shown in A. Middle, ratio of
response amplitudes (Q/A) for all testing intensities (zero responses excluded). Solid symbol represents mean ± s.d.
Right, difference in onset latency of evoked synaptic currents (A Q). C, D, Inhibitory responses in Q and A
states recorded in the same L2/3 neuron. Scale: 500 pA, 100 ms. E-H, Excitatory and inhibitory responses of a L4
excitatory cell. Scale: 100 pA and 100 ms in E; 200 pA and 100 ms in G. I, Excitatory and inhibitory currents to
tones of different frequencies in a L2/3 cell. Scale: 40 (Exc)/ 80 (Inh) pA, 200 ms. Inset, superimposed normalized
excitatory (top) and inhibitory (bottom) frequency tuning curves in quiescence (black) and active (red) states. J,
Peak response amplitude in active versus quiescence state for the same cell shown in I. Top, excitation; bottom,
inhibition. K, L, An example L4 cell plotted in the same manner as in I, J. Scale: 50 (E)/ 150 (I) pA, 200 ms. M,
Cumulative distribution of correlation coefficients for synaptic responses in A versus Q state for L2/3 neurons.
N, Distribution of ratios between peak synaptic amplitudes in A and Q states (A/Q) for all tone stimuli in all
recorded cells. Arrows indicate mean values. Left, excitation; right, inhibition. O, Slopes (scaling factors) of the
linear regression for A versus Q peak synapti c responses in all recorded cells. No significant difference
between scaling factors of excitation and inhibition in L2/3 cells (P = 0.5481, t = 0.6136, unpaired t-test, N = 11, 7
for excitatory and inhibitory respectively). P, The scaling factors for excitatory responses plotted against that for
inhibitory responses in the same cell. For L2/3, P = 0.6177, t = 0.5402, paired t-test, N = 5.
88
contrast, in L4 the scaling factors were 1.03 ± 0.11 and 1.01 ± 0.10 for excitation and inhibition
respectively (Figure 5.5N). We also quantified the scaling factor by the slope of the linear
regression for peak response amplitudes in A versus Q state ( Figure 5.5O). The average
slope was 0.65 ± 0.11 for excitation and 0.69 ± 0.13 for inhibition in L2/3 cells, while 1.02 ±
0.06 and 1.00 ± 0.02 respectively in L4 cells (Figure 5.5O). Similar results were obtained when
the charge transfer of synaptic currents was measured. When examined in the same cell, the
slope for excitatory responses was similar as that for inhibitory responses (Figure 5.5P).
Therefore evoked excitation and inhibition in a L2/3 cell were reduced by a similar factor from
the Q to A state, while excitation and inhibition in a L4 cell were not affected by changes of
behavioral state (Figure 5.5O). Finally, we argued that the reduction in recorded synaptic
responses was not due to compromised recording quality during animal movements, as the linear
current-voltage relationship of the cell remained as good in active states, and there was no
obvious change in series resistance with animal movements. In addition, we found that the
response amplitude to the same stimulus remained essentially unchanged before and after an
epoch of animal movements and remained relatively stable from the start till the end of the
recording sessions.
5.3.5 Change of membrane properties in active states
In visual cortex, locomotion is shown to cause a depolarization of neuronal membrane potential,
which contributes to the increased visual responses(Polack et al., 2013; Bennett et al., 2013).
Here, the reduced auditory responses suggested that the neurons membrane potential might be
hyperpolarized instead. We thus recorded resting membrane potentials in current-clamp mode
using a K
+
- based internal solution (see Methods) from L2/3 excitatory neurons. In the absence
of sound stimulation, the membrane potential (V m) fluctuated largely with occasional
89
appearance of spontaneous spikes (Figure 5.6A), but the distribution of potentials was
statistically unimodal (Figure 5.6B, P > 0.05, Hartigans dip test). In active states, the
distribution shifted toward more hyperpolarized values (Figure 5.6B), resulting in a more
hyperpolarized mean Vm. In all similarly recorded cells (N = 7), the distribution of potentials was
Figure 5.6 Modulation of resting membrane potential and resting conductance by behavioral state
A, Sample current-clamp recording records (top) together with the simultaneously recorded speeds (bottom) for an
example L2/3 neuron. Arrow labels the level for 70 mV. B, Normalized distribution of membrane voltages during
quiescent and active states for the cell shown in (a). Arrow points to the average spike threshold of the cell. C-F,
Comparison of mean resting membrane potential (C), s.d. of resting V m(D), spike threshold (E) and percentage of
V m values near spike threshold (not lower than the threshold by more than 10 mV) (F) under different behavioral
states for 7 recorded L2/3 neurons. Bar = SEM, paired t-test (***P = 0.0002, **P = 0.0083, P = 0.5496, ***P =
0.0002; t = 8.114, 3.862, 0.6338, 8.139 for (C-F) respectively). G, Average excitatory currents of an example cell
in response to BF tones (70 dB SPL) in different states. Thick black line marks tone duration. H, Average inhibitory
currents of the same cell. I, Resting conductances right before, during and after an epoch of active behaviors for
five L2/3 cells. Data points from the same cell are connected by lines. N = 5, bar = s.d., paired t-test (***P =
0.0004, **P = 0.0015; t = 10.65, 7.810)
90
unimodal in both states, and the transition from the Q to A state resulted in a more
hyperpolarized mean Vm(from 61.67 ± 1.83 mV to 64.75 ± 1.84 mV, Figure 5.6C), as well as
a reduction of Vm variability (Figure 5.6D). In the meanwhile, the level of spike threshold
remained constant (Figure 5.6E). As a consequence, the probability of instantaneous V m being
within 10 mV of the spike threshold was reduced in A compared to Q state ( Figure 5.6F).
This can contribute to the reduction of both spontaneous and evoked spiking activity of L2/3
neurons. Additionally, we found a reduction of baseline conductance from the Q to A state
(Figure 5.6G-I), which would result in an increase of input resistance of the cell. The decrease
of baseline conductance may be partly attributed to reduced spontaneous synaptic events as
evidenced by the reduction of spontaneous spiking activity of L2/3 neurons (Figure 5.1I).
5.3.6 Modulation of PV neuron activity
The reduced auditory evoked inhibition to L2/3 neurons in active states likely reflects reduced
inhibitory neuron activity. It is known that the major source of inhibition to L2/3 excitatory cells
is from the same layer (Dantzker et al., 2000). Since parvalbumin (PV) expressing neurons most
likely contribute to the feedforward inhibition in L2/3 excitatory cells(Li et al., 2014) and play a
major role in controlling the network gain because of their high firing rates and strong synaptic
connections (Pfeffer et al., 2013; Li et al., 2014), we specifically examined PV neuron activity in
different behavioral states. By injecting an AAV viral vector encoding Cre-dependent
channelrhodopsin2 (ChR2) in PV-Cre mice (Figure 5.7A and see Methods), we were able to
identify PV neurons with a previously described optogenetic method(Lima et al., 2009). With
loose-patch recordings, we actively searched for PV neurons, identification of which was based
on their trains of spikes in response to a pulse of blue LED light applied to the A1 surface
(Figure 5.7B, left panel). These neurons responded robustly to tone stimuli (Figure 5.7B, right
91
panel). Consistent with previous reports with two-photon imaging guided loose-patch recordings
(Ma et al., 2010; Li et al., 2014), PV neurons exhibited shorter trough-to-peak intervals in their
spike waveforms as compared to excitatory neurons (Figure 5.7C, 5.7D). They also tended to
have higher peak/trough amplitude ratios than excitatory cells (Figure 5.7D). In L2/3, we found
that both the spontaneous and evoked spiking activity of PV neurons was reduced from the Q
to A state ( Figure 5.7E, 5.7F), whereas in L4 PV cell activity was not affected (Figure 5.7G,
Figure 5.7 Changes of activity of PV
neurons between different behavioral
states
A, Confocal image of a brain section
showing ChR2-EYFP expression in the A1
region. B, Left, sample loose-patch
recording trace showing a train of evoked
spikes of a PV cell to blue LED illumination
(marked by the blue bar). Right, sample
spike response of the same cell to CF-tone
stimulation. Top inset, schematic drawing
of loose-patch of a cell. Bottom inset,
superimposed 500 individual spikes (black) and the average spike shape (red) of the cell.
C, Spike waveforms of four more PV and
excitatory cells. D, Plot of peak to trough
amplitude ratio versus trough-to-peak
interval for average spike waveforms of PV
and excitatory cells recorded with loose-
patch methods. Optogenetically identified
PV cells were labeled by open circles. E,
Summary of spontaneous firing rates of
recorded L2/3 PV cells in different states. N
= 8 cells. **P = 0.0044, t = 4.140, paired t-
test. F, CF-tone evoked spike numbers for
the cells in (e). **P = 0.0047, t = 4.086,
paired t-test. G, Summary of spontaneous
firing rates for PV cells recorded in L4. P =
0.5013, t = 0.7243, paired t-test, N = 6
cells. H, CF-tone evoked spike numbers for
the cells in G. P = 0.7286, t = 0.3671,
paired t-test.
92
5.7H). These results suggest that the L2/3 networks comprising both excitatory and PV
inhibitory neurons are generally suppressed by active behaviors.
5.3.7 Contribution of L1-mediated suppression
It has been shown previously that layer 1 plays a role in modulating activity in layer 2/3(Zagha
et al., 2013; Jiang et al., 2013). We then examined whether L1, which contains only inhibitory
neurons(Winer and Larue, 1989; Hestrin and Armstrong, 1996), played a role in the behavioral
state-dependent modulation of L2/3 activity. With loose-patch recordings, we found that L1
neurons were modulated by behavioral state differently from L2/3 and L4 cells: their
spontaneous (Figure 5.8A) as well as evoked (Figure 5.8B) firing rates were both increased
instead of decreased from the Q to A state. Since L1 neurons inhibit both excitatory and
inhibitory cells in layer 2/3(Shlosberg et al., 2006; Wozny and Williams et al., 2011; Jiang et al.,
2013), the increased activity of L1 neurons may generally increase the inhibitory tone in the L2/3
network, leading to the reduced activity of L2/3 cells. To further confirm the involvement of L1,
we silenced L1 spiking by applying 5 µM tetrodotoxin (TTX) to the A1 surface, following a
previously published method (Shlosberg et al., 2006). We monitored auditory evoked multiunit
spike responses in different layers before and after TTX application. As shown in Figure 5.8C,
within a limited time window (~ 150 s) after the topical application of TTX, the firing rate of L1
neurons gradually reduced to zero. Concurrently the evoked firing rate of L2/3 neurons increased
to a stable level while that of L4 neurons remained unchanged. These results indicate that L1
spiking activity tonically suppressed L2/3 but not L4 neurons so that removing L1 inhibition
increased firing rates of L2/3 cells. Beyond the time window, firing rates in L2/3 gradually
reduced, indicating that L2/3 cells were also progressively affected by TTX. L4 neurons also
progressively reduced their firing rates but with a delay. Therefore, the 150 s window provided a
93
good opportunity to examine the effect of silencing L1 spikes, while leaving spikes in L2/3 and
L4 unaffected by the drug. We next examined CF-tone evoked spikes in individual neurons with
loose-patch recordings. Before TTX application, the L2/3 cells exhibited a normal reduction of
response level from the Q to A state by 20.1 ± 6.9% ( Figure 5.8D, 5.8B). Within the 150 s
window after the drug application, the response level in the Q state was increased by 19.2 ±
8.6% compared to that before the drug application (Figure 5.8D). From the Q t o A state,
response level was reduced by only 4.1 ± 3.4%. Thus the behavioral state-dependent gain
reduction was largely blocked when L1 spiking was silenced, supporting the notion that the
increase of L1 firing rates in active states was responsible, at least partially, for the reduced
Figure 5.8 Role of L1 in the behavioral state-dependent L2/3-specific gain modulation
A, Left, spontaneous spikes of a L1 neuron in different states. Right, summary of average spontaneous firing
rates for 6 L1 neurons. Solid symbol represents mean ± s.d. **P = 0.0081, t = 4.252, paired t-test. B, Left,
PSTH for CF-tone evoked spikes of the same cell as in A. Right, summary of evoked spike numbers for 6 L1
neurons. **P = 0.0019, t = 5.085, paired t-test. C, Time courses of CF-tone evoked multiunit spike responses
in different A1 layers after the topical application of TTX (at time zero). Shaded area denotes the analysis time
window during which L2/3 responses were increased to a stable level while L4 responses remained unaffected.
N = 4 animals for each layer. Bar = s.d. D, Summary of evoked firing rates of individual L2/3 excitatory cells in
different behavioral states before and after TTX application. Spike rates were normalized by Q state before
TTX application. N = 10 cells. Among these 10 cells, 6 were also recorded in active states after TTX
application. **P < 0.01, ***P < 0.001, paired t-test. E, Summary of relative response levels (A/Q) before and
after TTX application. **P = 0.0035, t = 5.193, paired t-test, N = 6. F, Summary of normalized evoked spike
numbers in different states before and after TTX application for L4 neurons. N = 6 cells.
94
response level in L2/3. As a control, the response level of L4 neurons was not affected by the
drug application within the analysis time window, nor by the change of behavioral state (Figure
5.8F), which is consistent with our earlier observations.
5.4 Discussion
How sensory processing in cerebral cortex is modulated by behavioral and cognitive states has
been an important question for understanding the integrative function of the brain. To address
this question, it is critical to examine how sensory-evoked responses in individual cortical
neurons are modulated at cellular and synaptic levels. With the high-quality in vivo whole-cell
voltage-clamp recordings, we were able to reveal the excitatory and inhibitory synaptic inputs to
a cortical neuron under different behavioral states. Our results indicated a robust functional
balance between synaptic excitation and inhibition in the awake A1, as manifested by the
covariation of inhibitory and excitatory response amplitudes across different tone frequencies.
Relative to quiescence, behaviorally active states scaled down excitatory and inhibitory inputs at
a similar level in L2/3 but not L4 neurons, resulting in a proportional reduction of their spike
responses to different tone stimuli. As a consequence, the sensory tuning of spike response as
well as the functional balance between excitation and inhibition was preserved.
5.4.1 Behavioral state-dependent gain modulation
The observed suppression of auditory responses during active behaviors is reminiscent of a
previous report that the animals engagements in an auditory task result in reduc ed auditory
cortical responses (Otazu et al., 2009). In the current study, the magnitude of the modulation is
relatively small. The average effect on evoked spike responses is about 20% (Figure 5.1O),
while for some individual neurons it can be as high as 50%. Nevertheless, the moderate
95
modulation effect may possibly change the information transfer in the A1 and consequently
affect sound-dependent behaviors. This notion is supported by a study in the visual cortex
showing that a moderate reduction in evoked firing rates caused by optogenetically activating PV
inhibitory neurons can lead to a significant change in performance in visual detection tasks
(Glickfeld et al., 2013). Our results contrast with observations in visual and somatosensory
cortices that behaviorally active states depolarize the membrane potential of cortical neurons
(Gentet et al., 2010; Zagha et al., 2013; Polack et al., 2013; Bennett et al., 2013), and with the
general thoughts that active states are characterized by an increase in excitatory and inhibitory
conductances even in baseline conditions without sensory stimuli(Destexhe et al., 2013). It is
possible that the modulatory effect of behavioral state is specific to sensory modality. Or it may
depend on whether the specific sensory processing is engaged in the behavior that provides the
modulation. Possibly during locomotion and whisking, the animals exploration of the external
environment depends more on visual and tactile than auditory perception. In accordance with this
change in task demands, the relative salience of auditory information may be reduced.
Despite the reduction of response level, there is no change in the shape or size of TRFs of
L2/3 neurons from quiescence to active states. And the intensity threshold and frequency tuning
are well preserved. This is a result of gain modulation of spike responses, i.e. responses to
different tone stimuli are scaled by a similar factor. Counterintuitively, the signal-to-noise ratio
of auditory information in this output layer (L2/3) of A1 is in fact increased by about 35% in
active states (Figure 5.3C). The enhanced SNR is due to relatively more suppressed spontaneous
activity than evoked activity, which is different from the observations in visual cortex that
locomotion elevates SNR by enhancing sensory evoked responses(Polack et al., 2013; Bennett
et al., 2013), resulting in enhanced visual discrimination(Bennett et al., 2013). While the
96
functional significance of this enhanced SNR in A1 remains to be examined with behavioral
studies, our current results suggest that the sensitivity as well as the quality of auditory
processing is at least maintained from quiescence to active states.
5.4.2 Balanced excitation and inhibition in awake cortex
Spectrotemporally balanced excitation and inhibition has been demonstrated previously in
auditory cortical neurons of anesthetized animals, characterized by a similar frequency tuning of
excitation and inhibition, a roughly constant ratio between excitatory and inhibitory response
amplitudes across different stimuli, and a stereotypic temporal sequence of excitation briefly
followed by inhibition(Wehr and Zador, 2003; Zhang et al., 2003; Tan et al., 2004; Wu et al.,
2008). The short interval (~ 2 3 ms) between the onsets of excitation and inhibition is consistent
with a synaptic circuit dominant with feedforward inhibition(Wu et al., 2011). Recently, an
awake recording study in the visual cortex reports that while balanced excitation and inhibition is
prevalent in the anesthetized cortex, in the awake cortex inhibition is much more broadly tuned
than excitation in terms of spatial tuning(Haider et al., 2013). Our current study in the auditory
cortex however indicates that the functional balance between excitation and inhibition is
ubiquitous across different brain states, and that this balance is actively preserved through a
specific modulation of excitation and inhibition. The evoked synaptic excitation and inhibition
are reduced by a similar factor from quiescence to active states. Such balanced scaling down of
excitatory and inhibitory inputs would result in reduced output responses(Liu et al., 2011; Xiong
et al., 2013), as well as a longer integration time for spike generation(Zhou et al., 2012)(see
Figure 5.2I). In addition, the scaled excitatory and inhibitory inputs suggest that the observed
gain modulation of spike responses could be largely attributed to a network effect of suppression
of L2/3 circuits, which results in a reduction of total evoked synaptic conductance. Together, our
97
results strongly suggest that balanced excitation and inhibition is a fundamental synaptic circuit
basis for auditory cortical processing in awake conditions.
5.4.3 L1 mediated suppression of L2/3 activity
Modulation of cortical activity may be achieved through bottom-up or top-down pathways
(Harris, 2013). Previously, it has been shown in mice that action potential firing in the
somatosensory thalamus increases during whisking, which drives the desynchronized state in the
somatosensory cortex(Poulet et al., 2012). In this study, the absence of changes in spiking
responses in L4 and MGBv neurons as well as in synaptic inputs to L4 cells argues that the
Figure 5.9 Our proposed model for state-dependent suppression in L2/3 of A1
In our proposed model, state-dependent signals from unknown sources activate L1 inhibitory neurons, which
in turn increase their inhibitory tone onto both excitatory pyramidal neurons and parvalbumin inhibitory
neurons in L2/3. The decrease of inhibitory inputs from parvalbumin neurons can override the increase of
inhibitory inputs from L1 neurons. The net results is a scaling down of balanced excitation and inhibition by
active behavior states.
98
behavioral state-dependent suppression of sensory responses in L2/3 of A1 is unlikely due to a
modulation of neuronal activity in subcortical nuclei along the ascending auditory pathway. On
the other hand, previous studies have suggested that behavior can affect network state through
corticocortical inputs(Nelson et al., 2013; Zagha et al., 2013). Corticocortical projections are
known to ramify their axons in layer 1(Felleman and Van Essen, 1991; Cauller et al., 1998;
Gonchar et al., 2003; Petreanu et al., 2009), which is in a good position to mediate state-
dependent modulations of cortical activity in a top-down control. In this study, we found that L1
activity is increased from quiescence to active states. Since L1 activity can inhibit both
excitatory and inhibitory cells in L2/3(Shlosberg et al., 2006; Wozny et al., 2011; Jiang et al.,
2013), the increased spiking of L1 neurons may generally enhance the inhibitory tone in the L2/3
network(Figure 5.9). This is evidenced by the increase of spike responses of L2/3 excitatory
cells when spiking of L1 neurons is suppressed. Silencing of L1 spiking activity largely blocks
the reduction of sensory evoked responses of L2/3 neurons when animals transition from
quiescence to active states. This result indicates that the behavioral state-dependent gain
modulation in the L2/3 network can be attributed, at least partially, to a direct regulation of L1-
mediated inhibition.
Another possible way of modulating cortical activity is through neuromodulatory
systems. In the mouse cortex, neuromodulatory projections such as cholinergic and
noradrenergic fibers are distributed diffusely in all cortical layers without clear patterns (see the
Allen Brain Atlas data portal at www.brain-map.org). In the visual cortex, the locomotion-
induced depolarization of membrane potential and increase of firing rate is attributed to an effect
of noradrenergic input, and is more or less uniform across L2/3 and L4(Polack et al., 2013),
which is consistent with the diffuse pattern of noradrenergic fibers. In this study, the suppression
99
of spontaneous and evoked responses induced by active behaviors is observed in L2/3 but not in
L4. To our knowledge, there have not been reports about laminar-specific expression patterns of
receptors for neuromodulators that are consistent with our current observations. Nevertheless our
results do not exclude the possibility that the observed activity changes, including the
enhancement of L1 activity, are mediated by effects of specific neuromodulators.
In summary, our study has demonstrated that a balanced scaling down of excitatory and
inhibitory inputs underlies the suppressive gain modulation of sensory responses of L2/3
excitatory neurons induced by active behaviors. We postulate that scaling of synaptic inputs may
be a simple strategy employed by brain circuits to maintain the quality of sensory processing
while optimizing the level of salience of sensory information according to momentary behavioral
demands.
100
Bibliography
Abeles, M., and Goldstein, M.H., Jr. (1972). Responses of single units in the primary auditory
cortex of the cat to tones and to tone pairs. Brain research 42, 337-352.
Aitkin, L. (1991). Rate-level functions of neurons in the inferior colliculus of cats measured with
the use of free-field sound stimuli. J Neurophysiol 65, 383-392.
Aitkin, L.M., and Webster, W.R. (1972). Medial geniculate body of the cat: organization and
responses to tonal stimuli of neurons in ventral division. J Neurophysiol 35, 365-380.
Anderson, J.S., Carandini, M., and Ferster, D. (2000). Orientation tuning of input conductance,
excitation, and inhibition in cat primary visual cortex. J Neurophysiol 84, 909-926.
Apostolides, P.F., and Trussell, L.O. (2014). Superficial stellate cells of the dorsal cochlear
nucleus. Frontiers in neural circuits 8, 63.
Bennett, C., Arroyo, S., and Hestrin, S. (2013). Subthreshold mechanisms underlying state-
dependent modulation of visual responses. Neuron 80, 350-357.
Borg-Graham, L.J., Monier, C., and Fregnac, Y. (1998). Visual input evokes transient and strong
shunting inhibition in visual cortical neurons. Nature 393, 369-373.
Bregman, A.S., Ahad, P.A., and Kim, J. (1994). Resetting the pitch-analysis system. 2. Role of
sudden onsets and offsets in the perception of individual components in a cluster of overlapping
tones. The Journal of the Acoustical Society of America 96, 2694-2703.
Brown, M.C., and Ledwith, J.V., 3rd (1990). Projections of thin (type-II) and thick (type-I) auditory-nerve fibers into the cochlear nucleus of the mouse. Hearing Res 49, 105-118.
Brugge, J.F., Dubrovsky, N.A., Aitkin, L.M., and Anderson, D.J. (1969). Sensitivity of single
neurons in auditory cortex of cat to binaural tonal stimulation; effects of varying interaural time
and intensity. J Neurophysiol 32, 1005-1024.
Calford, M.B., and Semple, M.N. (1995). Monaural inhibition in cat auditory cortex. J
Neurophysiol 73, 1876-1891.
Callaway, E.M. (1998). Local circuits in primary visual cortex of the macaque monkey. Annual
review of neuroscience 21, 47-74.
Casseday, J.H., Ehrlich, D., and Covey, E. (1994). Neural tuning for sound duration: role of
inhibitory mechanisms in the inferior colliculus. Science 264, 847-850.
Casseday, J.H., Ehrlich, D., and Covey, E. (2000). Neural measurement of sound duration:
control by excitatory-inhibitory interactions in the inferior colliculus. J Neurophysiol 84, 1475-
1487.
101
Cauller, L.J., Clancy, B., and Connors, B.W. (1998). Backward cortical projections to primary
somatosensory cortex in rats extend long horizontal axons in layer I. J Comp Neurol 390, 297-
310.
Constantinople, C.M., and Bruno, R.M. (2011). Effects and mechanisms of wakefulness on local
cortical networks. Neuron 69, 1061-1068.
Crochet, S., and Petersen, C.C. (2006). Correlating whisker behavior with membrane potential in
barrel cortex of awake mice. Nature neuroscience 9, 608-610.
Dantzker, J.L., and Callaway, E.M. (2000). Laminar sources of synaptic input to cortical
inhibitory interneurons and pyramidal neurons. Nature neuroscience 3, 701-707.
Davies, P.W., Erulkar, S.D., and Rose, J.E. (1956). Single unit activity in the auditory cortex of
the cat. Bulletin of the Johns Hopkins Hospital 99, 55-86.
Davis, K.A., Ding, J., Benson, T.E., and Voigt, H.F. (1996). Response properties of units in the
dorsal cochlear nucleus of unanesthetized decerebrate gerbil. J Neurophysiol 75, 1411-1431.
de la Rocha, J., Marchetti, C., Schiff, M., and Reyes, A.D. (2008). Linking the response
properties of cells in auditory cortex with network architecture: Cotuning versus lateral
inhibition. J Neurosci 28, 9151-9163.
Destexhe, A., Rudolph, M., and Pare, D. (2003). The high-conductance state of neocortical
neurons in vivo. Nature reviews Neuroscience 4, 739-751.
Ding, J., and Voigt, H.F. (1997). Intracellular response properties of units in the dorsal cochlear
nucleus of unanesthetized decerebrate gerbil. J Neurophysiol 77, 2549-2572.
Douglas, R.J., and Martin, K.A. (2004). Neuronal circuits of the neocortex. Annual review of
neuroscience 27, 419-451.
Ehret, G., and Merzenich, M.M. (1988). Complex sound analysis (frequency resolution, filtering
and spectral integration) by single units of the inferior colliculus of the cat. Brain research 472,
139-163.
Evans, E., and Nelson, P. (1973). The responses of single neurones in the cochlear nucleus of the
cat as a function of their location and the anaesthetic state. Experimental Brain Research 17, 402-
427.
Evans, E.F., and Whitfield, I.C. (1964). Classification of Unit Responses in the Auditory Cortex
of the Unanaesthetized and Unrestrained Cat. The Journal of physiology 171, 476-493.
Faingold, C.L., Gehlbach, G., and Caspary, D.M. (1986). Decreased effectiveness of GABA-
mediated inhibition in the inferior colliculus of the genetically epilepsy-prone rat. Experimental
neurology 93, 145-159.
102
Fanselow, E.E., and Nicolelis, M.A. (1999). Behavioral modulation of tactile responses in the rat
somatosensory system. J Neurosci 19, 7603-7616.
Felleman, D.J., and Van Essen, D.C. (1991). Distributed hierarchical processing in the primate
cerebral cortex. Cereb Cortex 1, 1-47.
Ferezou, I., Haiss, F., Gentet, L.J., Aronoff, R., Weber, B., and Petersen, C.C. (2007).
Spatiotemporal dynamics of cortical sensorimotor integration in behaving mice. Neuron 56, 907-
923.
Fujino, K., and Oertel, D. (2003). Bidirectional synaptic plasticity in the cerebellum-like
mammalian dorsal cochlear nucleus. Proc Natl Acad Sci U S A 100, 265-270.
Gentet, L.J., Avermann, M., Matyas, F., Staiger, J.F., and Petersen, C.C. (2010). Membrane
potential dynamics of GABAergic neurons in the barrel cortex of behaving mice. Neuron 65,
422-435.
Glickfeld, L.L., Histed, M.H., and Maunsell, J.H. (2013). Mouse primary visual cortex is used to
detect both orientation and contrast changes. J Neurosci 33, 19416-19422.
Goard, M., and Dan, Y. (2009). Basal forebrain activation enhances cortical coding of natural
scenes. Nature neuroscience 12, 1444-1449.
Godfrey, D.A., Kiang, N.Y., and Norris, B.E. (1975). Single unit activity in the dorsal cochlear
nucleus of the cat. J Comp Neurol 162, 269-284.
Golding, N.L., and Oertel, D. (1997). Physiological identification of the targets of cartwheel cells
in the dorsal cochlear nucleus. J Neurophysiol 78, 248-260.
Gonchar, Y., and Burkhalter, A. (2003). Distinct GABAergic targets of feedforward and
feedback connections between lower and higher areas of rat visual cortex. J Neurosci 23, 10904-
10912.
Greenwood, D.D., and Maruyama, N. (1965). Excitatory and inhibitory response areas of
auditory neurons in the cochlear nucleus. J Neurophysiol 28, 863-892.
Grothe, B. (1994). Interaction of excitation and inhibition in processing of pure tone and
amplitude-modulated stimuli in the medial superior olive of the mustached bat. J Neurophysiol
71, 706-721.
Grothe, B., Pecka, M., and McAlpine, D. (2010). Mechanisms of sound localization in mammals.
Physiological reviews 90, 983-1012.
Guo, W., Chambers, A.R., Darrow, K.N., Hancock, K.E., Shinn-Cunningham, B.G., and Polley,
D.B. (2012). Robustness of cortical topography across fields, laminae, anesthetic states, and
neurophysiological signal types. J Neurosci 32, 9159-9172.
103
Haider, B., Hausser, M., and Carandini, M. (2013). Inhibition dominates sensory responses in the
awake cortex. Nature 493, 97-100.
Hancock, K.E., and Voigt, H.F. (2002). Intracellularly labeled fusiform cells in dorsal cochlear
nucleus of the gerbil. I. Physiological response properties. J Neurophysiol 87, 2505-2519.
Hancock, K.E., and Voigt, H.F. (2002). Intracellularly labeled fusiform cells in dorsal cochlear
nucleus of the gerbil. II. Comparison of physiology and anatomy. J Neurophysiol 87, 2520-2530.
Harris, K.D. (2013). Top-down control of cortical state. Neuron 79, 408-410.
He, J. (2001). On and off pathways segregated at the auditory thalamus of the guinea pig. J
Neurosci 21, 8672-8679.
Hestrin, S., and Armstrong, W.E. (1996). Morphology and physiology of cortical neurons in
layer I. J Neurosci 16, 5290-5300.
Hines, M. (1993). NEURON a program for simulation of nerve equations. Neural Systems:
Analysis and Modeling (MA.: Kluwer academic).
Jahr, C.E., and Stevens, C.F. (1990). A Quantitative Description of Nmda Receptor-Channel
Kinetic-Behavior. J Neurosci 10, 1830-1837.
Jahr, C.E., and Stevens, C.F. (1990). Voltage Dependence of Nmda-Activated Macroscopic
Conductances Predicted by Single-Channel Kinetics. J Neurosci 10, 3178-3182.
Jiang, X., Wang, G., Lee, A.J., Stornetta, R.L., and Zhu, J.J. (2013). The organization of two new
cortical interneuronal circuits. Nature neuroscience 16, 210-218.
Kadner, A., Kulesza, R.J., Jr., and Berrebi, A.S. (2006). Neurons in the medial nucleus of the
trapezoid body and superior paraolivary nucleus of the rat may play a role in sound duration
coding. J Neurophysiol 95, 1499-1508.
Kaltenbach, J.A., and Lazor, J. (1991). Tonotopic maps obtained from the surface of the dorsal
cochlear nucleus of the hamster and rat. Hearing Res 51, 149-160.
Kandel, E.R., schwartz, J. H., Jessell, T. M., Siegelbaum, S. A., Hudspeth, A. J. (2013).
Principles of neural science.
Kanold, P.O., and Manis, P.B. (1999). Transient potassium currents regulate the discharge
patterns of dorsal cochlear nucleus pyramidal cells. J Neurosci 19, 2195-2208.
Kanold, P.O., and Manis, P.B. (2005). Encoding the timing of inhibitory inputs. J Neurophysiol
93, 2887-2897.
Kiang, N.Y., Watanabe, T., Thomas, E.C., and Clark, L.F. (1965). Discharge patterns of single
fibers in the cats auditory nerve (Cambridge, MA. : MIT Press).
104
Kopp-Scheinpflug, C., Tozer, A.J., Robinson, S.W., Tempel, B.L., Hennig, M.H., and Forsythe,
I.D. (2011). The sound of silence: ionic mechanisms encoding sound termination. Neuron 71,
911-925.
Kuo, S.P., Lu, H.W., and Trussell, L.O. (2012). Intrinsic and synaptic properties of vertical cells
of the mouse dorsal cochlear nucleus. J Neurophysiol 108, 1186-1198.
Kuo, S.P., and Trussell, L.O. (2011). Spontaneous spiking and synaptic depression underlie
noradrenergic control of feed-forward inhibition. Neuron 71, 306-318.
Kuwabara, N., and Suga, N. (1993). Delay lines and amplitude selectivity are created in
subthalamic auditory nuclei: the brachium of the inferior colliculus of the mustached bat. J
Neurophysiol 69, 1713-1724.
Kuwada, S., and Batra, R. (1999). Coding of sound envelopes by inhibitory rebound in neurons
of the superior olivary complex in the unanesthetized rabbit. J Neurosci 19, 2273-2287.
Kuwada, S., Batra, R., Yin, T.C., Oliver, D.L., Haberly, L.B., and Stanford, T.R. (1997).
Intracellular recordings in response to monaural and binaural stimulation of neurons in the
inferior colliculus of the cat. J Neurosci 17, 7565-7581.
Lee, S., Carvell, G.E., and Simons, D.J. (2008). Motor modulation of afferent somatosensory
circuits. Nature neuroscience 11, 1430-1438.
Li, L.Y., Li, Y.T., Zhou, M., Tao, H.W., and Zhang, L.I. (2013). Intracortical multiplication of
thalamocortical signals in mouse auditory cortex. Nature neuroscience 16, 1179-1181.
Li, L.Y., Xiong, X.R., Ibrahim, L.A., Yuan, W., Tao, H.W., and Zhang, L.I. (2014). Differential
Receptive Field Properties of Parvalbumin and Somatostatin Inhibitory Neurons in Mouse
Auditory Cortex. Cereb Cortex.
Li, Y.T., Ibrahim, L.A., Liu, B.H., Zhang, L.I., and Tao, H.W. (2013). Linear transformation of
thalamocortical input by intracortical excitation. Nature neuroscience 16, 1324-1330.
Li, Y.T., Ma, W.P., Pan, C.J., Zhang, L.I., and Tao, H.W. (2012). Broadening of cortical
inhibition mediates developmental sharpening of orientation selectivity. J Neurosci 32, 3981-
3991.
Lima, S.Q., Hromadka, T., Znamenskiy, P., and Zador, A.M. (2009). PINP: a new method of
tagging neuronal populations for identification during in vivo electrophysiological recording.
PloS one 4, e6099.
Liu, B.H., Li, Y.T., Ma, W.P., Pan, C.J., Zhang, L.I., and Tao, H.W. (2011). Broad inhibition
sharpens orientation selectivity by expanding input dynamic range in mouse simple cells. Neuron
71, 542-554.
Liu, B.H., Wu, G.K., Arbuckle, R., Tao, H.W., and Zhang, L.I. (2007). Defining cortical
frequency tuning with recurrent excitatory circuitry. Nature neuroscience 10, 1594-1600.
105
Ma, W.L., and Brenowitz, S.D. (2012). Single-neuron recordings from unanesthetized mouse
dorsal cochlear nucleus. J Neurophysiol 107, 824-835.
Ma, W.P., Liu, B.H., Li, Y.T., Huang, Z.J., Zhang, L.I., and Tao, H.W. (2010). Visual
representations by cortical somatostatin inhibitory neurons--selective but with weak and delayed
responses. J Neurosci 30, 14371-14379.
Manis, P.B. (1990). Membrane properties and discharge characteristics of guinea pig dorsal
cochlear nucleus neurons studied in vitro. J Neurosci 10, 2338-2351.
Manis, P.B., Spirou, G.A., Wright, D.D., Paydar, S., and Ryugo, D.K. (1994). Physiology and
morphology of complex spiking neurons in the guinea pig dorsal cochlear nucleus. J Comp
Neurol 348, 261-276.
Mugnaini, E., Warr, W.B., and Osen, K.K. (1980). Distribution and light microscopic features of
granule cells in the cochlear nuclei of cat, rat, and mouse. J Comp Neurol 191, 581-606.
Navawongse, R., and Voigt, H.F. (2009). Single neuron recordings in dorsal cochlear nucleus
(DCN) of awake gerbil. Hearing Res 255, 44-57.
Nelken, I., and Young, E.D. (1994). Two separate inhibitory mechanisms shape the responses of
dorsal cochlear nucleus type IV units to narrowband and wideband stimuli. J Neurophysiol 71,
2446-2462.
Nelson, A., Schneider, D.M., Takatoh, J., Sakurai, K., Wang, F., and Mooney, R. (2013). A
circuit for motor cortical modulation of auditory cortical activity. J Neurosci 33, 14342-14353.
Niell, C.M., and Stryker, M.P. (2010). Modulation of visual responses by behavioral state in
mouse visual cortex. Neuron 65, 472-479.
Oertel, D., Bal, R., Gardner, S.M., Smith, P.H., and Joris, P.X. (2000). Detection of synchrony in
the activity of auditory nerve fibers by octopus cells of the mammalian cochlear nucleus. Proc
Natl Acad Sci U S A 97, 11773-11779.
Oertel, D., and Young, E.D. (2004). What's a cerebellar circuit doing in the auditory system?
Trends in neurosciences 27, 104-110.
Ojima, H., and Murakami, K. (2002). Intracellular characterization of suppressive responses in
supragranular pyramidal neurons of cat primary auditory cortex in vivo. Cereb Cortex 12, 1079-
1091.
Olsen, S.R., Bortone, D.S., Adesnik, H., and Scanziani, M. (2012). Gain control by layer six in
cortical circuits of vision. Nature 483, 47-52.
Ostapoff, E.M., Benson, C.G., and Saint Marie, R.L. (1997). GABA- and glycine-
immunoreactive projections from the superior olivary complex to the cochlear nucleus in guinea
pig. J Comp Neurol 381, 500-512.
106
Oswald, A.M., Schiff, M.L., and Reyes, A.D. (2006). Synaptic mechanisms underlying auditory
processing. Current opinion in neurobiology 16, 371-376.
Otazu, G.H., Tai, L.H., Yang, Y., and Zador, A.M. (2009). Engaging in an auditory task
suppresses responses in auditory cortex. Nature neuroscience 12, 646-654.
Parham, K., and Kim, D.O. (1995). Spontaneous and sound-evoked discharge characteristics of
complex-spiking neurons in the dorsal cochlear nucleus of the unanesthetized decerebrate cat. J
Neurophysiol 73, 550-561.
Petersen, C.C., and Crochet, S. (2013). Synaptic computation and sensory processing in
neocortical layer 2/3. Neuron 78, 28-48.
Petreanu, L., Mao, T., Sternson, S.M., and Svoboda, K. (2009). The subcellular organization of
neocortical excitatory connections. Nature 457, 1142-1145.
Pfeffer, C.K., Xue, M., He, M., Huang, Z.J., and Scanziani, M. (2013). Inhibition of inhibition in
visual cortex: the logic of connections between molecularly distinct interneurons. Nature
neuroscience 16, 1068-1076.
Phillips, D.P. (1988). Effect of tone-pulse rise time on rate-level functions of cat auditory cortex
neurons: excitatory and inhibitory processes shaping responses to tone onset. J Neurophysiol 59,
1524-1539.
Phillips, D.P., and Kelly, J.B.(1989). Coding of tone-pulse amplitude by single neurons in
auditory cortex of albino rats (Rattus norvegicus). Hearing Res 37, 269-279.
Phillips, D.P., Semple, M.N., and Kitzes, L.M. (1995). Factors Shaping the Tone Level
Sensitivity of Single Neurons in Posterior Field of Cat Auditory-Cortex. J Neurophysiol 73, 674-
686.
Pickles, J.O. (2008). An introduction to the physiology of hearing (UK: Emerald).
Polack, P.O., Friedman, J., and Golshani, P. (2013). Cellular mechanisms of brain state-
dependent gain modulation in visual cortex. Nature neuroscience.
Polley, D.B., Heiser, M.A., Blake, D.T., Schreiner, C.E., and Merzenich, M.M. (2004).
Associative learning shapes the neural code for stimulus magnitude in primary auditory cortex. P
Natl Acad Sci USA 101, 16351-16356.
Polley, D.B., Steinberg, E.E., and Merzenich, M.M. (2006). Perceptual learning directs auditory
cortical map reorganization through top-down influences. J Neurosci 26, 4970-4982.
Portfors, C.V., and Roberts, P.D. (2007). Temporal and frequency characteristics of cartwheel
cells in the dorsal cochlear nucleus of the awake mouse. J Neurophysiol 98, 744-756.
Poulet, J.F., Fernandez, L.M., Crochet, S., and Petersen, C.C. (2012). Thalamic control of
cortical states. Nature neuroscience 15, 370-372.
107
Poulet, J.F., and Petersen, C.C. (2008). Internal brain state regulates membrane potential
synchrony in barrel cortex of behaving mice. Nature 454, 881-885.
Qin, L., Chimoto, S., Sakai, M., Wang, J., and Sato, Y. (2007). Comparison between offset and
onset responses of primary auditory cortex ON-OFF neurons in awake cats. J Neurophysiol 97,
3421-3431.
Recanzone, G.H. (2000). Response profiles of auditory cortical neurons to tones and noise in
behaving macaque monkeys. Hearing Res 150, 104-118.
Reynolds, J.H., and Chelazzi, L. (2004). Attentional modulation of visual processing. Annual
review of neuroscience 27, 611-647.
Rhode, W.S. (1999). Vertical cell responses to sound in cat dorsal cochlear nucleus. J
Neurophysiol 82, 1019-1032.
Rhode, W.S., and Kettner, R.E. (1987). Physiological study of neurons in the dorsal and
posteroventral cochlear nucleus of the unanesthetized cat. J Neurophysiol 57, 414-442.
Rhode, W.S., and Smith, P.H. (1986). Encoding timing and intensity in the ventral cochlear
nucleus of the cat. J Neurophysiol 56, 261-286.
Rhode, W.S., and Smith, P.H. (1986). Physiological studies on neurons in the dorsal cochlear
nucleus of cat. J Neurophysiol 56, 287-307.
Rhode, W.S., Smith, P.H., and Oertel, D. (1983). Physiological response properties of cells
labeled intracellularly with horseradish peroxidase in cat dorsal cochlear nucleus. J Comp Neurol
213, 426-447.
Rouiller, E., de Ribaupierre, Y., Morel, A., and de Ribaupierre, F. (1983). Intensity functions of
single unit responses to tone in the medial geniculate body of cat. Hearing Res 11, 235-247.
Rudy, B., Fishell, G., Lee, S., and Hjerling-Leffler, J. (2011). Three groups of interneurons
account for nearly 100% of neocortical GABAergic neurons. Developmental neurobiology 71,
45-61.
Ryugo, D.K., and May, S.K. (1993). The projections of intracellularly labeled auditory nerve
fibers to the dorsal cochlear nucleus of cats. J Comp Neurol 329, 20-35.
Sachs, M.B., and Abbas, P.J. (1974). Rate versus level functions for auditory-nerve fibers in cats:
tone-burst stimuli. The Journal of the Acoustical Society of America 56, 1835-1847.
Sakata, S., and Harris, K.D. (2012). Laminar-dependent effects of cortical state on auditory
cortical spontaneous activity. Frontiers in neural circuits 6, 109.
Schofield, B.R. (1991). Superior paraolivary nucleus in the pigmented guinea pig: separate
classes of neurons project to the inferior colliculus and the cochlear nucleus. J Comp Neurol 312,
68-76.
108
Scholl, B., Gao, X., and Wehr, M. (2010). Nonoverlapping sets of synapses drive on responses
and off responses in auditory cortex. Neuron 65, 412-421.
Schreiner, C.E., Mendelson, J.R., and Sutter, M.L. (1992). Functional topography of cat primary
auditory cortex: representation of tone intensity. Experimental brain research Experimentelle
Hirnforschung Experimentation cerebrale 92, 105-122.
Semple, M.N., and Kitzes, L.M. (1985). Single-unit responses in the inferior colliculus: different
consequences of contralateral and ipsilateral auditory stimulation. J Neurophysiol 53, 1467-1482.
Shamma, S.A. (1985). Speech processing in the auditory system. II: Lateral inhibition and the
central processing of speech evoked activity in the auditory nerve. The Journal of the Acoustical
Society of America 78, 1622-1632.
Shlosberg, D., Amitai, Y., and Azouz, R. (2006). Time-dependent, layer-specific modulation of
sensory responses mediated by neocortical layer 1. J Neurophysiol 96, 3170-3182.
Shofner, W.P., and Young, E.D. (1985). Excitatory/inhibitory response types in the cochlear
nucleus: relationships to discharge patterns and responses to electrical stimulation of the auditory
nerve. J Neurophysiol 54, 917-939.
Somers, D.C., Nelson, S.B., and Sur, M. (1995). An Emergent Model of Orientation Selectivity
in Cat Visual Cortical Simple Cells. J Neurosci 15, 5448-5465.
Spirou, G.A., Davis, K.A., Nelken, I., and Young, E.D. (1999). Spectral integration by type II
interneurons in dorsal cochlear nucleus. J Neurophysiol 82, 648-663.
Stuart, G., and Spruston, N. (1998). Determinants of voltage attenuation in neocortical pyramidal
neuron dendrites. J Neurosci 18, 3501-3510.
Sun, Y.J., Wu, G.K., Liu, B.H., Li, P., Zhou, M., Xiao, Z., Tao, H.W., and Zhang, L.I. (2010).
Fine-tuning of pre-balanced excitation and inhibition during auditory cortical development.
Nature 465, 927-U928.
Sutter, M.L., and Loftus, W.C. (2003). Excitatory and inhibitory intensity tuning in auditory
cortex: Evidence for multiple inhibitory mechanisms. J Neurophysiol 90, 2629-2647.
Tan, A.Y., Zhang, L.I., Merzenich, M.M., and Schreiner, C.E. (2004). Tone-evoked excitatory
and inhibitory synaptic conductances of primary auditory cortex neurons. J Neurophysiol 92,
630-643.
Tan, A.Y.Y., Atencio, C.A., Polley, D.B., Merzenich, M.M., and Schreiner, C.E. (2007).
Unbalanced synaptic inhibition can create intensity-tuned auditory cortex neurons. Neuroscience
146, 449-462.
Voigt, H.F., and Young, E.D. (1980). Evidence of inhibitory interactions between neurons in
dorsal cochlear nucleus. J Neurophysiol 44, 76-96.
109
Wehr, M., and Zador, A.M. (2003). Balanced inhibition underlies tuning and sharpens spike
timing in auditory cortex. Nature 426, 442-446.
Winer, J.A., and Larue, D.T. (1989). Populations of GABAergic neurons and axons in layer I of
rat auditory cortex. Neuroscience 33, 499-515.
Winter, I.M., Robertson, D., and Yates, G.K. (1990). Diversity of Characteristic Frequency Rate-
Intensity Functions in Guinea-Pig Auditory-Nerve Fibers. Hearing Res 45, 191-202.
Wouterlood, F.G., and Mugnaini, E. (1984). Cartwheel neurons of the dorsal cochlear nucleus: a
Golgi-electron microscopic study in rat. J Comp Neurol 227, 136-157.
Wozny, C., and Williams, S.R. (2011). Specificity of synaptic connectivity between layer 1
inhibitory interneurons and layer 2/3 pyramidal neurons in the rat neocortex. Cereb Cortex 21,
1818-1826.
Wu, G.K., Arbuckle, R., Liu, B.H., Tao, H.W., and Zhang, L.I. (2008). Lateral sharpening of
cortical frequency tuning by approximately balanced inhibition. Neuron 58, 132-143.
Wu, G.K., Li, P., Tao, H.W., and Zhang, L.I. (2006). Nonmonotonic synaptic excitation and
imbalanced inhibition underlying cortical intensity tuning. Neuron 52, 705-715.
Wu, G.K., Tao, H.W., and Zhang, L.I. (2011). From elementary synaptic circuits to information
processing in primary auditory cortex. Neuroscience and biobehavioral reviews 35, 2094-2104.
Xiong, X.R., Liang, F., Li, H., Mesik, L., Zhang, K.K., Polley, D.B., Tao, H.W., Xiao, Z., and
Zhang, L.I. (2013). Interaural level difference-dependent gain control and synaptic scaling
underlying binaural computation. Neuron 79, 738-753.
Yajima, Y., and Hayashi, Y. (1989). Response properties and tonotopical organization in the
dorsal cochlear nucleus in rats. Experimental brain research Experimentelle Hirnforschung
Experimentation cerebrale 75, 381-389.
Yang, L., and Pollak, G.D. (1997). Differential response properties to amplitude modulated
signals in the dorsal nucleus of the lateral lemniscus of the mustache bat and the roles of
GABAergic inhibition. J Neurophysiol 77, 324-340.
Young, E.D., and Brownell, W.E. (1976). Responses to tones and noise of single cells in dorsal
cochlear nucleus of unanesthetized cats. J Neurophysiol 39, 282-300.
Young, E.D., and Davis, K.A. (2001). Circuitry and Function of the DCN (New York: Springer-
Verlag).
Young, E.D., and Voigt, H.F. (1982). Response properties of type II and type III units in dorsal
cochlear nucleus. Hearing Res 6, 153-169.
110
Zagha, E., Casale, A.E., Sachdev, R.N., McGinley, M.J., and McCormick, D.A. (2013). Motor
cortex feedback influences sensory processing by modulating network state. Neuron 79, 567-
578.
Zhang, L.I., Tan, A.Y.Y., Schreiner, C.E., and Merzenich, M.M. (2003). Topography and
synaptic shaping of direction selectivity in primary auditory cortex. Nature 424, 201-205.
Zhang, S., and Oertel, D. (1994). Neuronal circuits associated with the output of the dorsal
cochlear nucleus through fusiform cells. J Neurophysiol 71, 914-930.
Zhou, M., Tao, H.W., and Zhang, L.I. (2012). Generation of intensity selectivity by differential
synaptic tuning: fast-saturating excitation but slow-saturating inhibition. J Neurosci 32, 18068-
18078.
Zhou, Y., Liu, B.H., Wu, G.Y.K., Kim, Y.J., Xiao, Z.J., Tao, H.Z.W., and Zhang, L.I. (2010).
Preceding Inhibition Silences Layer 6 Neurons in Auditory Cortex. Neuron 65, 706-717.
Zhou, Y., Mesik, L., Sun, Y.J., Liang, F., Xiao, Z., Tao, H.W., and Zhang, L.I. (2012).
Generation of spike latency tuning by thalamocortical circuits in auditory cortex. J Neurosci 32,
9969-9980.

Abstract (if available)

Abstract Sound information is transmitted into electrical signals in the inner ear and these signals, i.e. action potentials, are passed and processed along the central auditory pathway. For perception of sound to occur, the basic attributes of sound (e.g. location, intensity, duration, et al.) need to be extracted and represented by neurons in the central auditory pathway. The synaptic circuit mechanisms underlying the extracting of these sound information attributes, however, remain poorly understood. During my PhD studies, I applied in vivo whole cell voltage clamp recording technique in studying how fusiform neurons in the Dorsal Cochlear Nucleus (DCN) of rat integrate excitatory and inhibitory synaptic inputs and generate novel spike patterns that presumably can encode intensity and temporal properties of sound information. In an independent but related project, I used head-fixed awake mice as a model to study how sound information processing could be modulated by different animal behavioral states. ❧ In my first project, we carried out in vivo whole-cell recordings from pyramidal neurons in the rat DCN, where intensity selectivity first emerges along the auditory pathway. Our results revealed that intensity-selective cells received fast-saturating excitation but slow-saturating inhibition with intensity increments, whereas in intensity-nonselective cells excitation and inhibition were similarly slow-saturating. The differential intensity tuning profiles of the non-intensity-tuned excitation and inhibition qualitatively determined the intensity selectivity of output responses. In addition, the selectivity was further strengthened by significantly lower excitation/inhibition ratios at high intensity levels compared to intensity-nonselective neurons. Our results demonstrate that intensity-selectivity in the DCN is generated by extracting the difference between tuning profiles of intensity-nonselective excitatory and inhibitory inputs, which we propose can be achieved through a differential circuit mediated by feedforward inhibition. ❧ In my second project, we studied how different temporal firing patterns (primary-like, pauser and buildup) were generated in pyramidal neurons in the rat DCN. We found that “primary-like” neurons received strong fast-rising excitation, whereas “pauser” or “buildup” neurons received accumulating excitation with a relatively weak fast rising phase followed by a slow rising phase. Pauser cells had stronger fast-rising excitation than buildup cells. On the other hand, inhibitory inputs to the three types of cells exhibited similar temporal patterns with a strong fast-rising phase. Dynamic-clamp recordings demonstrated that the differential temporal patterns of excitation could primarily account for the different discharge patterns. In addition, discharge patterns in a single neuron varied in a stimulus-dependent manner, which could be attributed to a modulation of excitation/inhibition ratio. Further studies of excitatory inputs to vertical and cartwheel cells suggested that fast-rising and accumulating excitation are separately conveyed by auditory nerve and parallel fibers respectively. The differential summation of excitatory input from two sources may thus contribute to the generation of response diversity. ❧ In my third project, we studied how transient offset response was generated in DCN fusiform cells. We found that in off responding neurons, their excitatory input started to decay later than inhibitory input did. This sharp increase of net excitation at sound offset caused the strong offset depolarization. Shortly after this offset depolarization, a second phase of inhibitory input accounted for the transient feature of the offset depolarization. Our results demonstrate that a fine temporal interaction of excitatory and inhibitory inputs can generate precisely timed off response in the rat DCN. ❧ In my fourth project, we developed in vivo whole cell voltage-clamp recording techniques in the auditory cortex of head-fixed awake mouse. We reported that sensory-evoked spike responses of layer 2/3 (L2/3) excitatory cells were scaled down with preserved sensory tuning when animals transitioned from quiescence to active behaviors, while L4 and thalamic responses were unchanged. Whole-cell voltage-clamp recordings further revealed that tone-evoked synaptic excitation and inhibition exhibited a robust functional balance. Changes of behavioral state caused scaling down of excitation and inhibition at an approximately equal level in L2/3 cells, but no synaptic changes in L4 cells. This laminar-specific gain control could be attributed to an enhancement of L1-mediated inhibitory tone, with L2/3 parvalbumin inhibitory neurons suppressed as well. Thus, L2/3 circuits can adjust the salience of output in accordance with momentary behavioral demands while maintaining the sensitivity and quality of sensory processing.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Functional synaptic circuits in primary visual cortex

PDF

Synaptic mechanism underlying development and function of neural circuits in rat primary auditory cortex

PDF

Synaptic mechanisms for basic auditory processing

PDF

Cortical synaptic circuitry underlying visual processing in the primary visual cortex

PDF

From sensory processing to behavior control: functions of the inferior colliculus

PDF

Axon guidance cues in development of the mammalian auditory circuit

PDF

Functional properties of the superficial cortical interneurons

PDF

Contextual modulation of sensory processing via the pulvinar nucleus

PDF

Physiology of the inner ear: the role of the biophysical properties of spiral ganglion neurons in encoding sound intensity information at the auditory nerve

Asset Metadata

Creator Zhou, Mu (author)

Core Title Synaptic circuits for information processing along the central auditory pathway

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Physiology and Biophysics

Publication Date 09/10/2014

Defense Date 08/25/2014

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag central auditory pathway,dorsal cochlear nucleus,in vivo whole cell recording,OAI-PMH Harvest,primary auditory cortex,synaptic circuits,voltage clamp

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Zhang, Li I. (committee chair), Bottjer, Sarah W. (committee member), Farley, Robert A. (committee member), Tao, Huizhong W. (committee member)

Creator Email zomehoh@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-471827

Unique identifier UC11287507

Identifier etd-ZhouMu-2908.pdf (filename),usctheses-c3-471827 (legacy record id)

Legacy Identifier etd-ZhouMu-2908.pdf

Dmrecord 471827

Document Type Dissertation

Format application/pdf (imt)

Rights Zhou, Mu

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA