Synaptic mechanisms for basic auditory processing

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

SYNAPTIC MECHANISMS FOR BASIC AUDITORY PROCESSING

by
Guangying Wu

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(NEUROSCIENCE)

May 2009

Copyright 2009 Guangying Wu

ii
Acknowledgement

! % "'
( " !
! ! ! % !! ( #% #
! % %( ) % ! " # # ( ! ' !% ! $
%%' # " "! ! % ( #' % # # % ! #! ' # ! # % # # ! " " ( ! " % ' ' '
( %) ! " ! %' # ! # % %

#!) !!%(
! " % ! " ' "

#( " # # ! & #
! %%( ! ' " % ' ' # #% % ( % #% " ! % ' !
%! #!(

iii
Table of Contents

Acknowledgement
Table of Contents
List of Figures
Abstract
Introduction
Chapter 1: Audition
1.1. Overview of Auditory System
1.1.1. External Ear, Middle Ear and Inner Ear
1.1.2. Brain Stem
1.1.3. Midbrain
1.1.4. Thalamus
1.1.5. Auditory Cortex
1.2. Organization of Auditory Cortex
1.2.1 Tonotopic Map
1.2.2 Laminar Organization
1.3. Fundamental Response Properties of Auditory Cortical Neuron
1.3.1. Tonal Receptive Field and Characteristic Frequency
1.3.2. Frequency Tuning
1.3.3. Intensity Tuning
1.3.4. Direction Selectivity
1.3.5. Duration Tuning
1.4. Functional Circuits underlying Response Properties
Chapter 2: Contributions of Thalamocortical Excitation and Intracortical
Excitation on Frequency Tuning
2.1. Background and Introduction
2.2. Methods and Materials
2.2.1. Animal Preparation and Extracellular Recording
2.2.2. Local Cortical Silencing
2.2.3. In vivo Whole-cell Recording
2.2.4. Data Analysis
2.3. Results
2.3.1. Silencing Cortex with a Cocktail of Muscimol and SCH50911
ii
iii
vi
viii
1
3
3
3
4
5
6
7
8
8
9
11
11
11
11
12
12
12

14
14
17
17
18
19
21
22
22

iv
2.3.2. Thalamocortical Input Determines the Area of Synaptic TRFs
2.3.3. Weakly Tuned Thalamic and Sharply Tuned Intracortical Inputs
2.3.4. Recurrent Excitation Largely Defines Frequency Tuning
2.4. Discussion
2.5. Summary
Chapter 2 Endnote
Chapter 3: Contribution of Cortical Inhibition on Frequency Tuning
3.1. Background and Introduction
3.2. Methods and Materials
3.2.1. Animal Preparation and Extracellular Recording
3.2.2. In vivo Whole-cell Recording
3.2.3. Cell-attached and Current-clamp Recordings, Juxtacellular Labeling
3.2.4. Data Analysis
3.3. Results
3.3.1. Frequency Tuning Curves of Excitatory and Inhibitory Inputs
3.3.2. Lateral Inhibitory Sharpening of Frequency Tuning
3.3.3. Fast-spike Inhibitory Neuron and Regular-spike Excitatory Neuron
3.3.4. Spike TRFs of Fast-spike Inhibitory Neurons
3.3.5. TRF of Membrane Potential Responses in Fast-spike Neurons
3.4. Discussion
3.4.1. Approximately Balanced Excitation and Inhibition
3.4.2. Inhibitory Mechanisms for Shaping Frequency Representation
3.4.3. Properties of Auditory Cortical Inhibitory Neurons
3.4.4. Implication on Cortical Circuitry
3.5. Summary
Chapter 3 Endnote
Chapter 4: Synaptic Mechanisms underlying Intensity Tuning
4.1. Background and Introduction
4.2. Methods and Materials
4.2.1. Extracellular Recording
4.2.2. In vivo Whole-cell Recording
4.2.3. Data Analysis
4.3. Results
4.3.1. Distribution of Nonmonotonic Neurons in the Rat Auditory Cortex
4.3.2. Excitatory and Inhibitory Synaptic Receptive Fields
4.3.3. Imbalanced Inhibition underlying Intensity Tuning
4.3.4. Inhibitory Contribution to Nonmonotonic Intensity Tuning
4.3.5. Temporal Shaping of Intensity Tuning by Synaptic Inhibition
4.3.6. Synaptic Mechanisms for Cortical Intensity Tuning
4.4. Discussion
4.4.1. Nonmonotonic Neurons in the Auditory Cortex
26
30
31
35
38
40
41
41
44
44
45
47
48
51
51
55
60
64
68
71
71
72
74
75
76
77
78
78
80
80
81
82
85
85
88
92
95
96
99
100
100

v
4.4.2. Nonmonotonic Excitation Primarily Determines Intensity Tuning
4.4.3. Temporal Shaping of Intensity Tuning by Inhibitory Inputs
4.4.4. Potential Synaptic Circuits underlying Nonmonotonic Neurons
4.5. Summary
Chapter 4 Endnote
Chapter 5: How Is Stimulus Duration Encoded?
5.1. Background and Introduction
5.2. Methods and Materials
5.2.1. Extracellular Recording
5.2.2. In vivo Whole-cell Recording
5.2.3. Data Analysis
5.3. Results
5.3.1. Tonotopic Organization of Dorsal Cochlear Nuclei
5.3.2. Two Types of Neuronal Responses to Sound Duration
5.3.3. Synaptic Mechanisms underlying Off-set Responses
5.4. Discussion
5.4.1. How Sound Duration Is Encoded in Auditory System?
5.4.2. Phasic Neurons Responds to the Sound Offset with Less Variability
5.4.3. Firing of Sustained neurons Correlates the Sound Duration
5.5. Summary
References
Appendix A
103
104
105
107
108
109
109
112
112
113
114
117
117
117
121
125
125
126
127
128
129
147

vi
List of Figures

Figure 1.1 Reconstruction of morphology and laminar location of an A1
neuron.

Figure 2.1 Specific silencing of local intracortical connections with a cocktail
pharmacological method.

Figure 2.2 Changes in excitatory synaptic TRF after local cortical silencing.

Figure 2.3 Intracortical inputs are more sharply tuned than thalamocortical
inputs.

Figure 2.4 Similarly tuned intracortical inputs sharpen the frequency tuning
curve in layer 4.

Figure 3.1 Frequency tuning of synaptic inputs in an example A1 excitatory
neuron.

Figure 3.2 frequency tuning of synaptic inputs in ten other excitatory neurons
in A1.

Figure 3.3 Summary of frequency tunings of synaptic inputs and membrane
potential responses.

Figure 3.4 Cell-attached reordings from fast-spike and regular-spike neurons
and juxtacelluar labeling.

Figure 3.5 Spike TRFs of FS and RS neurons.

Figure 3.6 Spike TRFs of FS neurons recorded with loose-patch technique.

Figure 3.7 Spike TRFs of RS neurons recorded with loose-patch technique.

Figure 3.8 Spike TRFs of regular-spike and fast-spike neurons.

Figure 3.9 Suprathreshold and subthrehold regions of synaptic TRFs in FS
and RS neuron.

Figure 4.1 Multiunit intensity-tuning in the rat auditory cortex.

9

24

28

32

34

53

56

59

61

63

65

66

67

70

86

vii

Figure 4.2 Synaptic TRFs of intensity-tuned and non-intensity-tuned neurons.

Figure 4.3 Intensity-tuning of recorded and derived membrane potential
responses.

Figure 4.4 Intensity-tuning of synaptic conductance’s evoked by CF tones.

Figure 4.5 Summary for intensity-tuned and untuned neurons.

Figure 4.6 An A1 nonmonotonic cell.

Figure 5.1 Tonotopic organization in the rat dorsal cochlear nuclei.

Figure 5.2 Two distinct types of firing pattern of DCN neurons.

Figure 5.3 Raster of the onset/offset type neuron.

Figure 5.4 Raster of the onset/offset type neuron.

Figure 5.5 Synaptic inputs and membrane potential changes of an example
DCN neuron.

Figure 5.6 Synaptic inputs and membrane potential changes of another DCN
neuron.

Figure 5.7 Synaptic inputs to an auditory cortical neuron.

90

92

94

98

102

116

119

120

121

123

124

124

viii
Abstract

Neurons are organized into circuits to process various information in the
brain. To understand how information is processed, it’s fundamental to investigate
the patterns of excitatory and inhibitory inputs underlying neuron’s response
properties. Through the four independent but closely related studies described in my
dissertation, I investigated the synaptic mechanisms underlying basic response
properties of rat auditory neurons, i.e. frequency tuning, intensity tuning and
temporal coding. Firstly, we developed a novel method to silencing the cortex and
dissected the excitatory input from thalamic neurons and that from cortical excitatory
neurons. The results demonstrated that thalamic input had a flattened frequency
tuning curve. In contrast, intracortical excitatory input was sharply tuned with a
tuning curve that closely matched that of suprathreshold responses. It suggests the
recurrent excitatory circuits define the cortical frequency tuning. Secondly, to study
the cortical inhibitory neurons, we combined single-unit cell-attached recording,
juxtacellular labeling and whole-cell recording together on the same neuron in vivo.
We discovered that the frequency tuning curve of inhibitory input was broader than
that of excitatory input. So a relatively stronger inhibition was flanked to the
preferred frequencies of the cell and laterally sharpened the frequency tuning of
membrane responses. The less selective inhibition can be attributed to a broader
bandwidth and lower threshold of spike tonal receptive field of fast-spike inhibitory

ix
neurons than nearby excitatory neurons. Thirdly, we uncovered the synaptic
mechanisms underlying cortical intensity tuning. The results demonstrated that
excitatory inputs to those intensity-tuned neurons have already shown a weak tuning,
while inhibitory inputs suppressed the excitation at higher intensities to strengthen
the tuning through a dynamic temporal control. Finally, we turned to cochlear nuclei
neurons in the brain stem to study the mechanisms underlying duration properties.
The results demonstrated neurons here had sustained excitatory and inhibitory inputs.
But the imbalanced excitation and inhibition in temporal aspect generated a more
precise response to represent duration information. Bringing all the previous studies
together, the interaction and dynamics between excitation and inhibition can work
together to shape the response properties and guarantee a faithful representation of
sensory world.

1
Introduction

Millions of years’ evolution endowed humans with a powerful system to
receive, process and perceive sensory information encoded in sound, which allows us
to locate objects, feel environments, comprehend languages and appreciate music.
In physical definition, sound is the mechanical waves generated by the vibration of
air molecules. Since sound is a kind of wave, it possesses the basic properties of a
typical wave: phase, frequency, amplitude and waveform. So it can convey
temporal, spectral and level information to us. However, how those fundamental
properties are represented and processed in the auditory system is still mysterious.
Modern neurobiology blooms since last century. Typically, in vitro studies
using cultured cells or sliced tissues provide us tremendous knowledge of channels
and synapses, but they fail to explain how the system works as a whole, which is the
goal of current neuroscience. So far, we still don’t know how information is
represented and processed in a systems-level. Choosing the auditory system, we’re
hoping to get some clues although it’s just a beginning. In Chapter 1, I will
introduce the background information about the sound, the auditory system, and the
response properties. Since sound has four basic properties, Chapter 2 and 3 will
discuss the mechanisms underlying frequency tuning, specifically, the contributions

2
of thalamocortical circuits, intracortical excitatory circuits and local inhibitory
circuits to frequency tuning. How intensity tuning is shaped in the auditory cortex
will be discussed in Chapter 4. In the last chapter, Chapter 5 will present our recent
work on how temporal information is processed at the first stage of central auditory
processing. As for mechanisms underlying more complex processing, such as
waveform, they are left for future investigation.

3
Chapter 1
Audition

1.1. Overview of Auditory System
Millions of years’ evolution designed and engineered the extreme
sophisticated, elegant and powerful system for animals to detect and process the
sound information arising from their neighborhood and themselves. Sound is
normally referred to pressure waves generated by the vibration of molecules. It
reaches to the ear and undergoes a series of transduction and transformation in the
auditory system. The external ear, middle ear and inner ear are periphery. The central
auditory system consists of nuclei in the brain stem, midbrain, thalamus and cortex
along the auditory pathway. The auditory information is transmitted from periphery
system to central system through auditory nerve, which belongs to cranial nerve VIII.
Here, a general and brief introduction to this amazing system is provided.

1.1.1. External Ear, Middle ear and Inner ear
The external ear consists of pinna, concha, auditory meatus, and eardrum
(tympanic membrane). Sound transmits into the external ear, and sound waves cause
the vibration of the eardrum. The vibration is transmitted from the relatively larger

4
tympanic membrane to the small oval window of the cochlea through the lever action
of three small bones, malleus, incus and stapes. Such anatomical structure
powerfully boosts the pressure on the eardrum about 200-fold when sound energy
transmits to the cochlea (Purves et al., 2004). The physical characteristics allow the
middle ear to dynamically convert air-borne vibration to aqueous vibration in the
cochlea.
The major transduction from mechanical vibration to electrical activities
happens in the cochlea of the inner ear. The cochlea is snail-shaped, with a basal end
and apical end. Hair cells residing in the cochlea are the sensory receptors. Inner hair
cells are the major receptors to convert mechanical energy to glutamate release. The
physical characteristics enable the hair cells in a specific location along the cochlear
axis to selectively respond to a specific range of frequency. Spiral ganglion neurons
respond to the glutamate released from those hair cells. Their axons form auditory
nerve. Through auditory nerves, auditory information is relayed to the neurons in the
brain stem, the first stage of the central auditory processing.

1.1.2. Brain Stem
Each auditory nerve bifurcates into an ascending axon terminal and a
descending axon terminal. The ascending pathway innervates the neurons in
anterioventral cochlear nuclei (AVCN), while descending pathway innervates the

5
neurons in posterioventral cochlear nuclei (PVCN) and dorsal cochlear nuclei (DCN).
Other than those three major nuclei, the organization of auditory nuclei is much more
complex in the brain stem. AVCN send excitatory input to both contralateral and
ipsilateral medial superior olivary (MSO). This kind of arrangement makes MSO a
perfect coincident detector to measure the interaural temporal difference, which is
crucial for animals to locate the sound source, e.g. their preys or predators. AVCN
also innervates ipsilateral lateral superior olivary (LSO) and medial nucleus of the
trapezoid body (MNTB). LSO neuron receives excitation from ipsilateral AVCN and
inhibition from contralateral MNTB. So LSO is believed to play a role on measure
the interaural level difference. DCN can directly innervate the neurons in infereior
colliculus (IC) in the midbrain, which is the fastest auditory pathway.

1.1.3. Midbrain
Inferior Colliculus (IC) is the auditory center in the midbrain. Neurons in IC
integrate inputs from different sources, including olivary complex, lemniscal
complex and cochlear nuclei. They send their axons to medial geniculate body
(MGB) in the thalamus. One good example of its integration is its role on
localizing sound. Barn owls are extraordinary nocturnal animal at localizing sounds.
Experiments in the barn owl show that the convergence of ITD and ILD in the
midbrain generates a computed auditory space map. Individual neurons within this

6
map only respond best to sounds from a specific region of space with a preferred
elevation and a preferred horizontal location, or azimuth.
Other than integration, IC is able to process sounds with complex temporal
features. Many neurons in IC can respond to frequency-modulated sounds, while
others can respond to sounds with specific durations. Such kinds of sounds normally
have biologically-related meanings in natural environment, for example,
communication and danger avoidance.

1.1.4. Thalamus
Medial geniculate body (MGB) is the auditory center in the thalamus. It
conveys all ascending auditory information to the cortex. The MGB consists of
several subdivisions, including the ventral, medial and dorsal parts, i.e. MGBv,
MGBm, MGBd. MGBv provides the major innervations to the primary auditory
cortex. Most input to the MGBv are coming from IC. MGB has two featured
properties. First, the MGB is the earliest place in the auditory pathway where
neurons can selectively respond to different combinations of frequencies. Second,
neurons in the MGB can selectively respond to specific time intervals between the
two frequencies, which are in a millisecond-scale instead of shorter ranges. So
neurons in the MGB receive inputs from distinct pathways from IC. One is mainly
related to spectral information; the other is mainly temporal information.

7
1.1.5. Auditory Cortex
Auditory cortex is the final destination of afferent auditory information. It
can also be divided into different subfields according to heterogeneous response
properties and connectivity. The primary auditory cortex (A1) receives point-to-point
input from MGBv. So it is tonotopically organized. The anterior part of A1 responds
best to higher frequency sounds, while the posterior part respond best to lower
frequency sounds. For other subdivisions, such as ventral auditory field, posterior
field and anterior auditory field, they receive more diffuse input from MGBm,
MGBd and even A1.
Although higher-order processing such as perception is not well known,
recent studies have been able to provide some clues. For example, in the study on
marmosets, a small primate with a complex vocal repertoire, it showed there’s a
small region, in which neurons are selective to pitch (Bendor and Wang, 2005). Pitch
is an important component for us to perform some higher tasks, such as appreciating
music and communicating with others. Some behavioral studies showed that the
auditory cortex is important for processing temporal information of sound. For
example, when the auditory cortex is ablated in these animals, they cannot
discriminate the sounds with the same frequency components but different temporal
sequence (Scharlock et al., 1965).

8
1.2. Organization of Auditory Cortex
1.2.1 Tonotopic Map
In the auditory system, the cochlea has already decomposed the spectrum of
sound as a Fourier transformation device, so that the hair cells are arrayed
tonotopically along the length of the basilar membrane. This tonotopic organization
is well preserved through the whole auditory pathway, especially in the major
auditory centers at different stages of auditory processing. So the primary auditory
cortex also has a topographical map of the cochlea. Since the axons of neurons has
already projected contralaterally as early as in brain stem, auditory system is
prominent in its binaural processing. In the cortex, there exists a striped arrangement
of neurons with different binaural properties. Such organization is orthogonal to the
frequency axis of the tonotopic map. For example, both ears can excite the neurons
in one stripe, while in the neighboring stripe neurons are excited by one ear but
inhibited by the other (Purves et al., 2004).
As shown in an example auditory cortical map (Figure 4.1), three major
fields can be identified according to the tonotopic organization of frequency
representations: the primary auditory cortex (A1) which exhibits a clear tonotopic
gradient along the anterior-posterior axis, a small anterior auditory field (AAF)
which exhibits a reversed tonotopic gradient compared to A1, and a ventral auditory

9
field (VAF) which has an apparent dorsal-ventral CF gradient, consistent with
previous reports (Bao et al., 2003; Kalatsky et al., 2005). In these regions, the
majority of sampling sites exhibited increased spike responses at high intensity
levels.

1.2.2 Laminar Organization

Figure 1.1 Reconstruction of morphology and laminar location of an A1 neuron.
Parts of five sequential Cryostat sections of the auditory cortex. The section number is labeled. All
sections were processed with the avidin-biotin-peroxidase method to stain the cell loaded with 0.5%
biocytin. Each of sec.12-14 contains part of the stained cell, indicated by the black arrows. White
arrows indicate a continuing blood vessel. Sec.11, 15 were then processed with Nissle-staining of cell
bodies to exhibit cortical layers. The boundaries for each layer and white matter (WM) are roughly
indicated. The pictures were taken with a Nikon imaging system (E1000 + Camera DXM 1200) at a
magnification of 40x. b. A complete picture of the stained cell by aligning sec.12-14. Magnification:
200x. Courtesy of Dr. Li I Zhang

d

10
As the other sensory cortices, auditory cortex can be divided into six layers
from the cortical surface to the white matter. Layer I is the surface layer, usually
called molecular layer, which is most close to the pial and dura. It consists mainly of
apical dendrites of pyramidal cells and horizontal axons of local neurons. Layer II
and Layer III contains many small and medium-size pyramidal neurons, as well as
non-pyramidal neurons including local interneurons. Layers I through III are the
main target of afferents which is from layer IV and layer V neurons. At the same time,
layer III is also the main source of corticocortical efferent’s. Layer IV contains
different types of pyramidal neurons. Neurons here receive direct innervations from
thalamic neurons and innervations from other cortical neurons. Layer V contains
large pyramidal neurons. It sends information to the subcortical nuclei, such as IC.
Layer VI neurons also receive thalamic inputs directly, but only a small portion of
thalamocortical innervation, and send efferent fibers back to the thalamus,
establishing a very precise reciprocal interconnection between the cortex and the
thalamus (Creutzfeldt, 1995).

11
1.3. Fundamental Response Properties of Auditory Cortical Neuron
1.3.1. Tonal Receptive Field and Characteristic Frequency
The receptive field of a sensory neuron is normally referred as a map in
which the stimulus with different parameters will evoke responses. In the auditory
system, receptive fields can be the fields in auditory space, or can be domains of
auditory frequencies. Here, we only discuss the receptive field of auditory neurons to
the sounds in the frequency and intensity domain. Typically, the auditory cortical
neuron showed a “V” shape receptive field in frequency-intensity map. After getting
the receptive field, the neuron’s characteristic frequency is defined as the frequency
at which even the lowest intensity can evoke responses.

1.3.2. Frequency Tuning
Each neuron is only sensitive to a specific range of frequencies. This kind of
selectivity is referred as frequency tuning. Strictly speaking, even within this range,
the response level is not uniform to the different frequency. So a tuning curve is used
to demonstrate a cell’s response level as the function of frequency.

1.3.3. Intensity Tuning
Similar to the frequency, each neuron is only sensitive to a specific range of
intensities. This kind of selectivity is referred as intensity tuning. We can also get a

12
tuning curve to demonstrate a cell’s response level as the function of intensity, to
understand how selective the neuron is to a specific intensity.

1.3.4. Direction Selectivity
Recent study showed the neurons in auditory cortex also respond to the
sweeps. They are frequency modulated sounds with downward or upward frequency
components. Those properties are analogue to the direction selectivity in visual
system. (Zhang et al., 2003)

1.3.4 Duration Tuning
Along the auditory pathway, duration-tuned neurons have been described in
inferior colliculus (IC) and auditory cortex (Potter, 1965; Feng et al., 1990; Pinheiro
et al., 1991; He et al., 1997; Casseday et al., 1994). Those duration-tuned neurons
respond to their ‘best duration’ with the maximum number of spikes.

1.4. Functional Circuits underlying Response Properties
Feedforward Excitation: Principal cells in the recepient layer of auditory cortex
receive feedforward excitation from thalamic neurons.
Feedforward Inhibition: Principal cells also receive inhibitory inputs from local

13
inhibitory neurons. Since those inhibitory neurons are innervated by thalamic
neurons. This kind of inhibition is referred as feedforward inhibition.
Recurrent Excitation: Principal cells can receive other excitatory inputs from
intracortical excitatory neurons. They provide recurrent excitation to the principal
cells.

14
Chapter 2
Contributions of Thalamocortical Excitation and
Intracortical Excitation on Frequency Tuning

2.1. Background and Introduction
In the recipient layer of sensory cortices, i.e. visual cortex, auditory cortex
and somatosensory cortex, principal neurons receive two major sources of excitatory
inputs: one from thalamic neurons, the other from intracortical excitatory neurons
(Figure 1.3). Previous studies on the sensory representation and processing of
principal neurons reflect a converging thalamocortical pathway (Reid and Alonso,
1995; Ferster et al., 1996; Chung and Ferster, 1998; Miller et al., 2001; Bruno and
Sakmann, 2006). However, the functional contributions and the underlying patterns
of thalamocortical and intracortical excitatory inputs remain elusive (Douglas et al.,
1995; Somers et al, 1995; Miller et al., 2001; Alonso and Swadlow, 2005).
Previously, two types of methodology have been developed to understand
the thalamocortical contribution to cortical responses. One is based on the direct
comparison of the response properties between simultaneously recorded neurons in
the thalamus and cortex (Reid and Alonso, 1995; Miller et al., 2001; Miller et al.,
2002; Martinez et al., 2005); and the other is based on the isolation of
thalamocortical input by preventing spiking of cortical neurons (Ferster et al., 1996;

15
Fox et al., 2003; Kaur et al., 2004; Zhang and Suga, 1997; Chung and Ferster, 1998;
Volgushev et al., 2000). The first type of studies mostly used extracellular recordings
and identified putatively connected thalamic and cortical units on the basis of the
temporal correlation between their spikes. This approach provides information on the
tuning properties of individual thalamic and cortical neurons, as well as the nature of
the connection between them. A recent study in the somatosensory cortex (Bruno and
Sakmann, 2006), which paired extracellular recording of thalamic neurons with
intracellular recording of cortical cells, suggests that cortical neurons receive a
number of weak, but synchronously activated, thalamic inputs, which show tuning
properties similar to the recorded cortical neuron. However, as the pattern underlying
divergent output connections made by a single thalamic neuron or convergent
thalamic inputs made on a single cortical neuron remains largely unknown, it is
difficult to determine the respective functional roles of thalamocortical and
intracortical inputs. The second type of study depends on an effective silencing of the
cortex without affecting thalamocortical transmission. Three methods have been
previously used to silence the cortex: (i) cortical application of muscimol, an agonist
of GABA
A
receptors, to prevent neuronal spiking (Fox et al., 2003; Kaur et al., 2004;
Zhang and Suga, 1997), (ii) cooling the cortex (4–14°C) to block spike generation in
neurons (Ferster et al., 1996; Villa et al., 1991; Volgushev et al., 2000), and (iii)
electrical stimulation of the cortex to produce a long inhibition widow (4100 ms)
following excitation, during which spikes cannot be generated (Chung and Ferster,
1998). However, all of these methods are expected to have impacts on

16
thalamocortical presynaptic transmission. Electrical stimulation can result in
complex presynaptic effects such as short-term depression or facilitation. Although
the mechanism underlying cooling-induced action-potential block is not yet clear, it
is probable that both the action potential spread in axons and presynaptic vesicle
release will also be affected by a marked temperature decrease. Microinjection,
iontophoresis or perfusion of muscimol have more often been applied to silence
intracortical connections (Fox et al., 2003; Kaur et al., 2004; Zhang and Suga, 1997).
It was assumed that muscimol was a highly specific agonist to GABA
A
receptors.
However, this view has been challenged by recent findings that muscimol can
activate GABA
B
receptors at relatively low concentrations and can reduce synaptic
transmission through presynaptic GABA
B
receptors (Yamauchi et al., 2000).
In this work, we developed a new pharmacological method for silencing the
cortex. By simultaneously blocking GABA
B
receptors with a specific antagonist, we
were able to largely prevent the nonspecific effect of muscimol on presynaptic
transmission. By applying in vivo whole-cell voltage-clamp recording in the rat
primary auditory cortex (A1), we examined tone-evoked synaptic responses in layer
4 neurons before and after local cortical silencing. We found that thalamocortical
inputs determine the area of the synaptic frequency-intensity tonal receptive field
(TRF), whereas intracortical excitatory inputs largely define the frequency tuning by
selectively amplifying responses at preferred frequencies of the cortical cell.

17
2.2. Methods and Materials
2.2.1. Animal Preparation and Extracellular Recording
All experimental procedures used in this study were approved by the
Animal Care and Use Committee of the University of Southern California.
Experiments were carried out in a soundproof booth (Acoustic Systems) as described
previously (Zhang et al., 2001; Tan et al., 2004; Wu et al., 2006). Female Sprague-
Dawley rats ~3 months old and weighing 250–300 g were anaesthetized with
ketamine and xylazine (see discussion in Appendix A.2). Pure tones (0.5–64 kHz at
0.1-octave intervals, 25-ms duration, 3-ms ramp) at eight 10-dB–spaced sound
intensities were delivered to the contralateral ear. Multiunit spike responses were
recorded with parylene-coated tungsten microelectrodes (2 M , FHC) placed 500–
600 μm below the pial surface. The number of tone-evoked spikes was counted in a
window of 10–30 ms from the onset of tone stimulus. Auditory cortical mapping was
carried out by sequentially recording from an array of cortical sites. The location of
A1 was identified as previously described (Zhang et al., 2001; Tan et al., 2004; Wu et
al., 2006). For recording in the auditory thalamus, we systematically mapped the
MGB with extracellular recordings in a three-dimensional manner by varying the
depth and xy coordinates of the electrode. We identified the MGBv, which projects to
A1 (Winer et al., 2005), according to its tonotopy of frequency representation and the
relatively sharper spike TRFs seen there than in other MGB divisions (Calford and
Webster, 1981).

18
2.2.2. Local Cortical Silencing
Muscimol can activate both GABA
A
(EC
50
= 1.7 μM) and GABA
B
(EC
50
=
25 μM) receptors (Yamauchi et al., 2000). For effective silencing of the cortex with
minimum impact on presynaptic transmission, we derived an optimized
concentration ratio for muscimol and SCH50911 based on their competitive binding
to GABA
A
or GABA
B
receptors:
(1)
(2)
(3)
(4)
where A: SCH90511; B: muscimol; Gb: GABA
B
receptor; Ga: GABA
A
receptor.
Here, EC
50
or IC
50
values are used to calculate binding constants to reflect
the functional effects of binding on channel opening or blocking: Kb1=1/(1 μM),
Kb2 = 1/(25 μM), Ka1 1/(900 μM), Ka2=1/(1.7 μM). We consider 5% receptors
bound as no significant effect, and 95% as fully effective. Under these
conditions, a ratio of 1.5:1 (SCH50911 : muscimol) was chosen (see Appendix A.1
for detailed calculation). A high concentration (6mM: 4mM) was used to effectively
silence a relatively large cortical region. Pharmacological reagents (dissolved in
ACSF solution containing Fast Green) or control solution (ACSF containing Fast
Green) was injected through a glass micropipette with a tip opening of about 2-3 μm

19
in diameter, attached via polyethylene tubing to a syringe. The pressure inside the
tubing was monitored with a pressure gauge. After premapping A1, the pipette was
inserted to 500-600 μm beneath the cortical surface at around the center of A1,
controlled by a motorized micromanipulator. Injection was under a pressure of 3-4
p.s.i. and continued for 5 minutes. The injected volume was estimated to be around
10-20 nl, as measured in mineral oil. Without applying pressure, there was no
apparent leakage of intrapipette solution since there was no leakage of green color
and no change in cortical responses. The staining by fast green was monitored under
the surgical microscope, which spread fast under the injection pressure and covered a
cortical area with a radius of 500-600 μm at the end of the injection. This also
means that the initial concentration of injected cocktail will be quickly diluted by
about 50 folds, as estimated from the change in volume. Experiments are normally
completed within 30 minutes after drug injection with one drug experiment
performed in each animal preparation. Cortical responses largely recovered 7 hours
after the injection, likely due to the slow diffusion of drugs in the cortex (Fox et al.,
2003).

2.2.3. In vivo Whole-cell Recording
Whole-cell recordings (Zhang et al., 2003; Wehr and Zador, 2003; Tan et
al., 2004; Wu et al., 2006; Moore and Nelson, 1998; Margrie et al., 2002) were
obtained from neurons located at 500–700 μm beneath the cortical surface,
corresponding to the input layers of the auditory cortex (Games and Winer, 1988).

20
For voltage-clamp recording, the patch pipette (4-7 M ) contained (in mM): 125 Cs-
gluconate, 5 TEA-Cl, 4 MgATP, 0.3 GTP, 10 phosphocreatine, 10 HEPES, 1 EGTA,
2 CsCl, 2 QX-314, pH 7.2, and 0.5% biocytin. The whole-cell and pipette
capacitance were completely compensated and the initial series resistance (20 50M ) was compensated for 50-60% to achieve effective series resistance of 10-25
M . Signals were filtered at 5 kHz and sampled at 10 kHz. For current clamp
recording to examine spikes, the same patch pipette was used, containing (in mM):
125 K-gluconate, 4 MgATP, 0.3 GTP, 10 phosphocreatine, 10 HEPES, 1 EGTA, pH
7.2, and 0.5% biocytin. Histological staining of the recorded cells (Horikawa and
Armstrong, 1988; Zhu et al., 2004; Hirsch et al., 1998) after recording indicates that
the whole-cell recording method under our current condition biasedly sampled
pyramidal neurons. In this study, the measured membrane potentials of the recorded
neurons ranged from -61 to -72 mV with a mean of -63.8 mV .
To obtain tone-evoked excitatory inputs, the cells were clamped directly at -
70mV, which is around the reversal potential of inhibitory currents (Ei) as also
described in our previous studies (Zhang et al., 2003; Tan et al., 2004; Wu et al.,
2006). Cortical cells can be reasonably clamped before and after cocktail application
with clamping deviation within around ± 5 mV(see Appendix A.2 and Appendix
Figures A.5A.7). For a few cases when both excitatory and inhibitory TRFs were
obtained before cortical silencing, we derived the excitatory synaptic conductance
Ge(t) according to I(t, V) = Gr(V-Er) + Ge(t)(V-Ee) + Gi(t)(V-Ei), where V is the
clamping voltage, Gr is the resting conductance, Er is the resting potential; Ee and Ei

21
are the reversal potentials for excitatory and inhibitory synaptic currents,
respectively; and I(t, V) is the current amplitude under V, and V(t) is given by V(t) =
Vc – Rs*I(t), where Rs was the effective series resistance and Vc is the clamping
voltage applied (Zhang et al., 2003; Wehr and Zador, 2003; Tan et al., 2004;
Anderson et al., 2000; Wu et al., 2006; Hirsch et al., 1998; Borg-Graham et al.,
1998). The liquid junction potential is estimated to be 12 mV. We found that
varying Ei values between -65 and -75 mV, did not change the conclusion of this
study.

2.2.4. Data Analysis
Tone-evoked excitatory synaptic or membrane potential responses were
identified according to their onset latencies and peak amplitudes. Only responses
with the onset and peak occurring within 7-30 ms from the onset of tone stimulus,
and with peak amplitude of at least three folds of standard deviation of baseline
fluctuation were considered tone-evoked responses. The response onset latency was
taken as the time point in the rising phase of the response curve, where the amplitude
change was two folds of standard deviation of baseline fluctuation (Zhang et al.,
2003). The boundaries of synaptic TRFs were defined according to the consistency
of tone-evoked responses in 2-4 repetitions, and the continuity of responses with the
change of frequency and intensity. For color maps of synaptic TRFs, repetitions
were averaged, and only the pixels within the determined TRF boundary were
labeled (Tan et al., 2004; Wu et al., 2006).

22

The estimated membrane potential responses (Vest) for the voltage-clamp
recordings before the drug application were derived using Vest (t) = (GrVr + Ge(t)Ee
+ Gi(t)Ei)/(Gr(t) + Ge(t) + Gi(t)), where Vr is the resting membrane potential. Since
Gi was not determined for most of the cells in this study, we derived the membrane
potential changes caused by excitatory inputs alone, and estimated the spike
threshold at 20 mV above the resting membrane potential.
To correct the non-specific effect of the cocktail application, we averaged
tone-evoked responses at the intensity threshold of synaptic TRFs before and after
cortical silencing. We have assumed that these responses were mostly contributed by
monosynaptic thalamic inputs. The relative reduction in their amplitudes (r) can be
largely explained by the changes in input and series resistances after cocktail
application (see Appendix A.2 and Appendix Table A.1 for detailed calculation), and
thus is used as a reflection of non-specific effects of cocktail application. The
amplitude of each response was then normalized by a correction factor of 1/(1-r). It
is worth noting that varying the correction factor by ± 25% does not qualitatively
change the conclusion of this study.

2.3. Results
2.3.1. Silencing Cortex with a Cocktail of Muscimol and SCH50911
To understand how thalamocortical and intracortical synaptic inputs
contribute to the processing of individual cortical neurons, we developed a cocktail

23
pharmacological method to effectively silence intracortical connections, while
largely preserving thalamocortical synaptic transmission. Previously, cortical
application of muscimol has been used to prevent spiking of cortical neurons (Fox et
al., 2003; Kaur et al., 2004; Zhang and Suga, 1997). However, recent studies suggest
that muscimol can activate GABA
B
receptors at a relatively low concentration (EC
50

= 25 μM) (Yamauchi et al., 2000). Because GABA
B
receptors exist on
thalamocortical axons, cortical muscimol application will result in a dramatic
reduction of evoked transmitter release from these axons (Porter and Nieves, 2004;
Figure 2.1A). Cortical microinjection of muscimol or baclofen, a specific agonist of
GABA
B
receptors, largely eliminated tone-evoked field potentials recorded in the
cortex. To overcome the nonspecific effect ofmuscimol on presynaptic transmission,
we applied SCH50911, a specific antagonist of GABAB receptors, together with
muscimol. This substantially restored the magnitude of toneevoked field potentials
(Figure 2.1A). Application of SCH50911 alone slightly increased the amplitude of
field-potential responses (Figure 2.1A) and prolonged tone-evoked spiking activity
(Appendix Figure A.1), but did not change the shape of spike TRFs of cortical
neurons (Appendix Figure A.1). This is consistent with the late, prolonged inhibition
that is mediated by postsynaptic GABA
B
receptors.
Considering the competitive binding between muscimol and SCH50911 and
GABA
A
and GABA
B
receptors, respectively, we derived an optimized concentration
ratio for the coapplied SCH50911 and muscimol (1.5:1) to achieve an effective
activation of GABA
A
receptors and a blockade of GABA
B
receptors (see Methods &

24

Figure 2.1. Specific silencing of local intracortical connections with a cocktail pharmacological
method. (A) Tone-evoked field potentials recorded in A1 before (top) and after (bottom) cortical
injection of muscimol (1 mM), baclofen (1 mM), SCH50911 (1.5mM) or a cocktail of muscimol (1
mM) and SCH50911 (1.5 mM). Small arrow marks the onset of tone stimulus. (B) Effective
blocking of cortical spikes by the muscimol and SCH50911 (4mM: 6mM) cocktail in both layer 4 and
layer 6 within a horizontal distance of 500 m from the injection site (see Methods). Multi-unit tone-
evoked spikes were detected by extracellular recordings. Red dot line indicates 90% reduction in
spike count. (C) Example excitatory synaptic TRFs of A1 neurons obtained shortly after muscimol
injection (left) and cocktail injection (right). Each small trace represents the response (recorded at -
70mV) to a tone of a particular frequency and intensity. (D) Average bandwidth of synaptic TRF
measured at 60dB (BW60) in A1 injected with muscimol, cocktail, or vehicle solution (ACSF). Bar is
s.d. (E) Spike TRF (average of four repetitions) for a recording site in the MGBv before and after
cortical injection of the cocktail. Color represents the number of spikes evoked by a tone stimulus.
(F) Percentage change in the bandwidth and spike count for tone evoked spikes (measured at 60 dB)
in the MGBv before and after cortical cocktail application (n = 6 sites). Bar = s.d.

25
Materials and Appendix A.1 for detailed calculation). After slowly injecting about
10–20 nl of the cocktail solution into the cortex, firing of cortical cells in a horizontal
distance of 500 μm was effectively blocked (spike count was reduced by > 95%), as
indicated by the extracellular multiunit recordings (Figure 2.1B). This effect can last
for at least 3 h (data not shown). Because the highest density of local intracortical
excitatory input comes from neurons within a 500- μm radius (Douglas et al., 1995;
Roerig and Chen, 2002; Marino et al., 2005) and because the size of rat A1 is about
1–2 mm
2
(Zhang et al., 2003), we believe that this local injection method is effective
at silencing the majority of intracortical connections, although some long-distance
connections may not be affected. As a control, when the cortex was injected with the
vehicle solution (artificial cerebrospinal fluid, ACSF), no significant effect was
observed (n = 5, t-test, P > 0.5) on tone-evoked cortical field potentials (data not
shown).
To examine the effect of local cortical silencing on the receptive field
properties of single cortical neurons, we carried out in vivo whole-cell voltage-clamp
recordings on excitatory pyramidal neurons in layer 4 of A1 (premapped with
extracellular recording, see Methods) shortly after cortical injection of muscimol or
the mixture of muscimol and SCH50911. Excitatory synaptic responses evoked by
pure tones of various frequencies and intensities were recorded with the neuron
clamped at –70 mV. The excitatory synaptic TRF was reconstructed after the
recording. In muscimol-treated cortices, only traces of synaptic responses were
observed in a small tonal-responsive area (Figure 2.1C, left). On the contrary, in

26
cocktail-treated cortices, excitatory synaptic responses with large amplitudes were
observed (Figure 2.1C, right). The bandwidth of an excitatory synaptic TRF
measured at a 60-dB sound pressure level (SPL) was not significantly different (t-
test, P > 0.3) from that observed in normal A1 or control A1 where ACSF was
injected (Figure 2.1D). The local silencing of A1 under our experimental condition
did not affect the response properties in the auditory thalamus that projects to A1, as
multiunit spike TRFs in the ventral division of the medial geniculate body (MGBv)
did not change noticeably after cortical injection of the cocktail (Figures 2.1E, F).
This wellrestricted drug effect in the cortex may be attributed to the small volume of
drug application in our experiments. The above population studies suggest that the
shapes of the excitatory synaptic TRFs of neurons in the input layers of A1 are
primarily determined by thalamocortical input. The apparently reduced bandwidth of
excitatory synaptic TRFs in the presence of muscimol (Figure 2.1D) may be largely
attributed to the nonspecific effect of muscimol on presynaptic GABA
B
receptors
(Kaur et al., 2004).

2.3.2. Thalamocortical Input Determines the Area of Synaptic TRFs
To further examine the pattern of thalamocortical and intracortical
excitatory inputs and their roles in determining frequency tuning, we recorded
synaptic TRFs from the same A1 neuron before and after cortical silencing (for
example, Figure 2.2). The cell was clamped at –70 mV and then at 0 mV to record
tone-evoked excitatory and inhibitory currents, respectively (Figure 2.2A). The

27
linearity of the I-V curve suggests that the cell was reasonably clamped (Figure 2.2B,
see Appendix A.2 for detailed discussion). The injection of the muscimol and
SCH50911 cocktail was made at a cortical site about 500 μm below the pial surface
and had a horizontal distance from the recorded cell of 100 μm. Tone-evoked
inhibitory currents were eliminated after the injection (Figure 2.2A), consistent with
silencing of intracortical inhibitory connections. The amplitude of excitatory currents
was substantially reduced (Figure 2.2A), which can be attributed to silencing of
intracortical excitatory connections and nonspecific effects of cocktail application,
including those caused by changes in series resistance and input resistance
(Appendix Table A.1, see Appendix A.2 for detailed calculation).
We assumed that tone-evoked excitatory responses at the subthreshold
intensity threshold (20 dB in this particular cell) mostly originated from
monosynaptic thalamocortical synapses. This assumption is supported by three
observations. First, the reduction in the amplitude of excitatory currents was the
smallest at the subthreshold intensity threshold. Second, the multiunit spike TRFs
recorded from the same site before silencing had a higher intensity threshold (30 dB,
Appendix Figure A.2), suggesting that the synaptic currents at the subthreshold
intensity threshold are unlikely to have originated from local intracortical inputs.
Third, the kinetics of the rising phase of response currents at the intensity threshold
remained mono-phased after cocktail application, whereas those of response currents
to best-frequency tones above the intensity threshold changed from bi-phased to
mono-phased, consistent with synaptic inputs of two sources that have different

28
onsets (Figure 2.2D, bottom). Thus, on the basis of the relative change inthe
amplitude of average response at the intensity threshold after the cocktail

Figure 2.2. Changes in excitatory synaptic TRF after local cortical silencing.
(A) Excitatory (left) and inhibitory (right) synaptic currents evoked by a tone of 1.5 kHz and 70 dB
before (gray) and after (black) cocktail application. (B) Left, synaptic currents (average of five
repeats) evoked by a tone of 1.9 kHz and 70 dB recorded at different holding potentials. Right, I V
curves (V is corrected) for synaptic currents averaged within a 20-22.5 ms window after the stimulus

29
onset (black) and 0-1 ms window after the response onset (red). (C) Morphology of this recorded
cell. Bar, 20 m. (D) TRF of excitatory synaptic currents before (average of two repeats) and after
(four repeats) silencing. Blue dots mark the responses at the intensity threshold (20 dB). The color
maps show the average amplitudes. Number in the bracket indicates the original scale before
correction. Bottom, I-IV, the rising phase of average synaptic response to a 1.3 kHz tone at 60 dB
(I/II) or a 5.6 kHz tone at 20 dB (III/IV) before (I/III) and after (II/IV) cortical silencing. (E) Color
map of onset latencies of evoked excitatory currents. (F) Onset latencies (at 70 dB) before (blue) and
after (red) cocktail application. Triangle represents the difference. (G) Amplitudes of responses
before (blue) and after (red) cocktail application at 70 dB. (H) Tuning curves of excitatory currents at
four different tone intensities. The black line represents the tuning curve of subtracted responses
(before minus after).

application (–29%), we estimated that the cocktail application had caused a 29%
nonspecific reduction of the response amplitude in this cell (see Appendix A.2 for
more discussion). The amplitude of each excitatory response after cortical silencing
was then corrected by a factor of 1.41 (see Methods and Materials).
From the excitatory TRF after the correction (shown by the bottom color
map in Figure 2.2D), it is apparent that after cortical silencing there is no noticeable
change in the range of responding frequencies at various testing intensities, or in the
intensity threshold (Figure 2.2D). This result further supports the notion that
thalamocortical input primarily defines the shapes of excitatory synaptic TRFs. In
addition, the graded amplitude of thalamocortical responses indicates cortical neuron
is innervated by a number of thalamic neurons, with each possessing spike TRF in
the frequency-intensity range defined by the excitatory synaptic TRF of this cell.
We next examined the onset latency for each response in the TRF before
and after cortical silencing (Figure 2.2E). A clear pattern of onset latency values was
observed in the TRF, with the shortest latencies appearing at high intensities and
clustering around preferred frequencies, and the longest latencies appearing at the

30
periphery of the TRF. This is reminiscent of a similar finding in visual cortical
receptive fields (Bringuier et al., 1999), from which it was proposed that long-
latency responses are a result of intracortical spread of visual activity. In the present
study, however, there was no significant change (n = 47, paired t-test, P > 0.9) in
short or long onset latencies after cortical silencing (Figure 2.2F), suggesting that the
onset latency is determined mostly by thalamocortical input. The variation in onset
latency is likely attributable to variation in the conduction velocity subcortically and
in the integration time for action potential generation in subcortical neurons.

2.3.3. Weakly Tuned Thalamic and Sharply Tuned Intracortical Inputs
The analysis above indicates that thalamocortical input primarily
determines the area of the synaptic TRF. We next examined the role of thalamic and
intracortical inputs in determining cortical frequency tuning. Intracortical
connections provide both excitatory and inhibitory inputs. Previous data have
suggested that intracortical inhibitory input can sharpen spike tuning curves through
an analogous ‘iceberg’ effect (Zhang et al., 2003; Wehr and Zador, 2003; Tan et al.,
2004) and that it may also increase the temporal precision of spike responses by
temporally following excitatory input with a brief delay (Wehr and Zador, 2003). In
contrast, the functional role of intracortical excitatory input remains largely
unknown. Here, we compared the frequency tuning curves before and after cortical
silencing, which are depicted by the envelope of amplitudes of responses at certain
intensity. It appears that the shape of the tuning curves became more flattened after

31
cortical silencing, as reflected by an increase in the half-peak bandwidth of the
tuning curves (Figure 2.2G,H). The flat or plateau peak of the thalamocortical tuning
curve (Figure 2.2G) indicates that thalamocortical inputs are, in fact, weakly tuned.
A total of five cortical neurons were examined in this manner (another four
cells are shown in Figure 2.3). In all of these neurons, there was no noticeable
change in the range of responding frequencies or in the intensity threshold after
cortical silencing (Figure 2.3A,C,E,G,J and Appendix Figure A.3). At the mean time,
the frequency tuning curve became flattened, showing both a broad plateau peak that
covered the preferred frequencies of the cell and a broader half-peak bandwidth than
that of the total inputs (Figure 2.3B,D,F,H–J). We further quantified the pattern of
intracortical excitatory inputs by subtracting the thalamocortical component from the
total excitatory responses (Figures 2.2H and 2.3B,D,F,H, black lines). The tuning
curves for excitatory intracortical inputs showed the same preferred frequencies as
those for total excitatory inputs, but their half-peak bandwidths were significantly
narrower than those for both the total and thalamocortical inputs (n = 5, paired t-test,
P < 0.01, Figure 2.3I). These results demonstrate that local intracortical excitatory
input is more sharply tuned than thalamocortical input and has a greater contribution
at the preferred frequencies of the cell.

2.3.4. Recurrent Excitation Largely Defines Frequency Tuning
What is the role of intracortical input in determining the frequency tuning of
cortical cells? The preferred frequencies of intracortical excitatory input closely

32
Figure 2.3. Intracortical inputs are more sharply tuned than thalamocortical inputs.

33
(A-H) Change in excitatory synaptic TRF in another four cells. A, C, E, G, Top, color maps represent
the excitatory synaptic TRF before and after cocktail application. Bottom, kinetics of the rising phase
of synaptic currents (1-IV). The curved line outlines the boundary of synaptic TRF before (blue) and
after (red) cocktail application (V). Data are presented in the same way as in Fig.2.2. B, D, F, H
,

Top, excitatory synaptic currents evoked by tones (at 70 dB) of different frequencies before and after
cocktail application. The amplitudes of currents after application are corrected. Bottom, excitatory
tuning curves at 70 dB before (blue) and after (red) cortical silencing. Data are presented in a similar
manner as in Fig.2. Black lines are for the subtracted inputs. (I) Half-peak bandwidths of tuning
curves at 70 dB for total excitatory inputs (before), thalamic inputs (after) and intracortical inputs
(subtracted). Data points from the same cell are connected with lines (n = 5, paired t-test, * P < 0.01).
(J) Average ratio of onset latency, intensity threshold of excitatory synaptic TRF, bandwidth at 10 dB
above the intensity threshold (TRF BW), and half-peak bandwidth of tuning curve at 70 dB (half-peak
BW) between after and before values. Bar = s.d.

matched the plateau peak of the thalamocortical tuning curve (Figures 2.2G and
2.3B,D,F,H), suggesting that the intracortical inputs arose from a group of similarly
tuned neurons. We carried out further whole-cell current-clamp recordings to
compare subthreshold and spike TRFs in the same cortical neuron. We found that the
frequency range for spike responses was narrower than that for excitatory synaptic
input (Figure 2.4A,B). The average frequency range for spike responses measured at
60 dB was about 54 ± 7% (mean ± s.d., n = 10) of that for excitatory synaptic input,
consistent with previous findings that spike threshold sharpens neuronal tuning for
many stimulus attributes (Tan et al., 2004; Carandini and Ferster, 2000; Anderson et
al., 2000; Priebe and Ferster, 2005; Wilent and Contreras, 2005). As a result, each
intracortical excitatory input, which depends on a cortical cell’s firing, will
inevitably have a narrower frequency range than direct thalamocortical input. When
excitatory intracortical inputs with similar tuning properties are pooled, they can
selectively amplify excitatory responses at their preferred frequencies.

34

Figure 2.4. Similarly tuned intracortical inputs sharpen the frequency tuning curve in layer 4. (A)
Membrane potential responses to tones of various frequencies and intensities, recorded under current
clamp from a representative A1 neuron. Blue dashed line delineates the boundary for the TRF of tone-
evoked membrane depolarizations. The yellow and red lines indicate the frequency range (at 70 dB) for
subthreshold and spike responses respectively. (B) Left, frequency range for subthreshold (yellow) and
spike (red) responses at 70 dB of 10 cells detected with current-clamp recording. Middle, percentage
frequency range for spike responses within the range for membrane potential responses. Each cross
represents one cell. The square represents the average of all cells (± s.d.). Right, for 24 cells in which
both voltage-clamp and current-clamp recordings were obtained, the frequency range (at 70 dB) of
membrane depolarizations (ordinate) matches well with that of excitatory synaptic currents (abscissa). (C)
Left, normalized tuning curves after thresholding within the estimated suprathreshold response range for
total excitatory input (black), thalamocortical (red) and intracortical (blue) input in a cell. Right, average
bandwidths at 50% and 80% peak amplitude of “suprathreshold” tuning curves (n = 5 cells, paired t-test, *
P < 0.03). Bar = s.d.

35
We next compared the contribution of thalamocortical and intracortical
inputs to the synaptic tuning curve over the frequency range of suprathreshold
responses, which reflects the spiking probability of the cell under fluctuating
membrane potentials. The suprathreshold response range was estimated by the
derived membrane-potential responses (determined by synaptic responses before
cortical silencing, seeMethods) that showed an increase of >20mV. The tuning
curves of the total, the thalamocortical and the derived intracortical inputs were then
thresholded in the defined suprathreshold frequency range and normalized. The
bandwidths of these thresholded tuning curves were compared. In the suprathreshold
frequency range, the tuning curve of intracortical input closely matched that of total
excitatory input (Figure 2.4C). In contrast, the thalamocortical tuning curve was
significantly broader (n = 5, paired t-test, P < 0.03). This finding further supports the
notion that, although thalamocortical input determines the subthreshold responding
range, the frequency tuning of the cortical neuron is largely defined by more
narrowly tuned intracortical excitatory input.

2.4. Discussion
Extensive efforts have been made to address the role of feedforward
thalamocortical input in determining the response properties or the representation
and processing functions of layer 4 neurons in the sensory cortex (Reid and Alonso,
1995; Ferster et al., 1996; Chung and Ferster, 1998; Miller et al., 2001; Bruno and
Sakmann, 2006). Experimental studies in the primary visual, somatosensory and

36
auditory cortices all suggest that the response properties of layer 4 cortical neurons
can be explained by the convergence of thalamic inputs (Reid and Alonso, 1995;
Ferster et al., 1996; Chung and Ferster, 1998; Miller et al., 2001; Bruno and
Sakmann, 2006; Alonso and Swadlow, 2005; Swadlow and Gusev, 2002). However,
both the extent to which response properties of cortical neurons represent those of
thalamic inputs and the functional pattern of these inputs has not been fully
addressed. Moreover, the contribution of recurrent intracortical excitatory
connections to cortical processing remains largely unclear. Modeling studies suggest
that they may function as a cortical amplifier for the feedforward excitation6, and
may account for the emergence of contrast-invariant orientation selectivity in the
visual cortex (Somers et al., 1995). In contrast, a recent study in the somatosensory
cortex suggests that cortical amplification may not be required, as layer 4 neurons
receive a large number of synchronous sensory-driven thalamic inputs, which
together can be strong enough to trigger spike responses (Bruno and Sakmann,
2006).
In the present study, by using a cortical silencing method that leaves
thalamocortical transmission largely unaffected, we were able to isolate the
thalamocortical component underlying the synaptic TRFs of cortical neurons. We
conclude that thalamocortical input determines the range of subthreshold responses,
as synaptic TRFs cover the same area before and after cortical silencing, although the
amplitude of each excitatory response is reduced. This reduction can be attributed to
the silencing of intracortical excitatory connections and nonspecific effects caused by

37
the cocktail application (see Appendix A.2 for detailed discussion). By taking into
account the nonspecific reduction and correcting the response amplitudes
accordingly, we estimate that for the largest tone-evoked excitatory response
(saturating response) in the TRF, about 61 ± 11% (mean ± s.d.) of the response has a
thalamocortical origin and 39 ± 11% has a cortical origin (n = 5 cells). These values
are comparable to the estimation in the cat primary visual cortex that thalamic input
comprises about 46% of the total excitatory input (Chung and Ferster, 1998). The
smallest tone-evoked responses after cortical silencing presumably originate from
single thalamic inputs, and they can generate membrane depolarizations of about
0.5–1 mV. We can thus estimate that each saturating response evoked by a pure tone
stimulus consists of 18 ± 6 synchronous thalamic inputs (averaged from the five
cells), consistent with the findings in the somatosensory cortex that many
synchronous thalamic inputs are underlying sensory-evoked responses of single
cortical neurons (Bruno and Sakmann, 2006). Although we could not infer the total
number of thalamic projections made onto an A1 neuron, the excitatory TRF after
cortical silencing has revealed a functional pattern of thalamic inputs.
The comparison between the tuning curves of total excitatory,
thalamocortical and derived intracortical input showed that weakly tuned
thalamocortical input has been remarkably sharpened by intracortical input. The peak
of the thalamocortical tuning curve is broad and flat, suggesting a low level of
selectivity if responses at the peak are suprathreshold. The tuning curve of
intracortical input is sharper, and the tuning curve of total excitatory input in the

38
spiking frequency range more closely resembles that of intracortical input than that
of thalamocortical input. If we consider the synaptic tuning curve to be a distribution
of spiking probability, we can conclude that intracortical input defines the shape of
the spike tuning curve, in a manner similar to adding a pyramid on top of a flat base.
In other words, intracortical input reconstitutes the sharpness of frequency tuning in
the cortex. The similar preferences of thalamocortical and intracortical tunings imply
a recurrent circuitry in which local similarly tuned neurons excite each other. These
intracortical inputs selectively amplify the thalamocortical signal and determine the
optimal stimulus of the cortical cell. In conclusion, our results are consistent with
models in which intracortical recurrent excitation determines stimulus selectivity of
cortical neurons. We propose that by combining the breadth of feedforward
excitation and selectivity of recurrent excitation, a reliable and faithful conveyance
of subcortically processed sensory information to the cortex can be ensured.

2.5. Summary
Neurons in the recipient layers of sensory cortices receive excitatory input
from two major sources: the feedforward thalamocortical and recurrent intracortical
inputs. To address their respective functional roles, we developed a new method for
silencing cortex by competitively activating GABA
A
while blocking GABA
B

receptors. In the rat primary auditory cortex, in vivo whole-cell recording from the
same neuron before and after local cortical silencing revealed that thalamic input
occupied the same area of frequency-intensity tonal receptive field as the total

39
excitatory input, but showed a flattened tuning curve. In contrast, excitatory
intracortical input was sharply tuned with a tuning curve that closely matched that of
suprathreshold responses. This can be attributed to a selective amplification of
cortical cells’ responses at preferred frequencies by intracortical inputs from
similarly tuned neurons. Thus, weakly tuned thalamocortical inputs determine the
subthreshold responding range, whereas intracortical inputs largely define the tuning.
Such circuits may ensure a faithful conveyance of sensory information.

40
Chapter 2 Endnote

The work presented in this chapter appeared in the following publication:
Liu, B.*, Wu, G.K.*, Arbuckle, R., Tao, H.W., and Zhang, L.I. (2007). Defining
Cortical Frequency Tuning with Recurrent Excitatory Circuitry. Nat. Neurosci. 10,
1594-1600. (*Equal contribution)

41
Chapter 3
Contributions of Cortical Inhibition on Frequency Tuning

3.1. Background and Introduction
Cortical inhibition plays an important role in shaping receptive field
properties of neurons in sensory cortices (Sillito, 1977, 1979; Kyriazi et al., 1996;
Wang et al., 2002). The underlying synaptic mechanisms remain controversial. This
is partially due to technical limitations, which make it difficult to characterize the
structure of cortical inhibitory circuits and the functional properties of cortical
inhibitory neurons. The recent application of in vivo whole-cell voltage-clamp
recording in the cortex provides a powerful approach to unraveling excitatory and
inhibitory synaptic input circuits underlying functions of cortical neurons. In the
auditory cortex, two apparently conflicting models have been proposed for inhibitory
sharpening of cortical tonal receptive fields (TRFs). First, recent in vivo whole-cell
voltage-clamp recordings from individual auditory cortical neurons indicate that the
frequency preferences as well as the responding frequency ranges are similar for
tone-evoked excitatory and inhibitory synaptic input. This suggests a synaptic input
network with balanced excitation and inhibition (Zhang et al., 2003; Wehr and Zador,
2003; Tan et al., 2004; Wu et al., 2006). Because inhibitory input always follows
excitatory input with a brief temporal delay, it is proposed that inhibitory input can
scale down excitation and thus narrow the frequency range for spike responses in a

42
so-called “iceberg” effect (Wehr and Zador, 2003). Second, it was proposed
previously that cortical inhibitory input may have broader tuning than excitatory
input, resulting in lateral inhibition in the surround of TRFs (Suga and Manabe,
1982; Shamma, 1985; Calford and Semple, 1995; Sutter and Loftus, 2003; Oswald et
al., 2006). This second model is primarily based on extracellular recording
experiments of two-tone suppression, in which one tone modifies (usually
suppresses) the response to a later tone (Suga and Manabe, 1982; Calford and
Semple, 1995; Sutter and Loftus, 2003). Although the response properties of
auditory cortical inhibitory neurons are largely unknown, several extracellular
studies in the somatosensory cortex suggest that putative cortical inhibitory neurons
may possess less selective representational properties than principal neurons (Simons
and Carvell, 1989; Swadlow, 1989), supporting the second model.
In previous intracellular studies (Wehr and Zador, 2003; Zhang et al., 2003;
Tan et al., 2004), the patterns of tone-evoked excitatory and inhibitory inputs have
not been examined in sufficient detail. Although inhibitory input has been shown to
be able to scale down the level of membrane excitation, no direct comparison has
been made between the frequency tuning curves of membrane potential responses in
the presence and absence of inhibition. Thus, the existence of lateral inhibition
effects cannot be excluded by the findings on the apparently balanced excitation and
inhibition. In this study, we tested the possibility that tone-evoked synaptic
excitation and inhibition in a single cortical neuron do not match precisely, and that

43
the fine structure in their distribution patterns can result in an equivalent lateral
inhibitory sharpening of TRFs.
Studies on the pattern of inhibitory input to a cortical neuron alone cannot
fully address the response properties of presynaptic inhibitory neurons, except that
the spike TRFs of the presynaptic inhibitory neurons will be no larger than the tonal
responsive area defined by the inhibitory inputs to the cell. Because only 15 –25%
of neurons are inhibitory in many cortical areas (Peters and Kara, 1985; Hendry et
al., 1987; Priet et al., 1994) and it remains difficult to identify and target these
neurons in vivo, our knowledge of their functional properties has lagged far behind
that of excitatory neurons. There have been limited studies on the functional
properties of cortical inhibitory neurons. In the rabbit somatosensory cortex,
extracellular recordings from suspected inhibitory neurons (SINs), identified
according to their spike features, suggest that SINs lack directional preference,
unlike principal cells, and exhibit high sensitivity to sensory stimuli (Swadlow,
2003). In the cat visual cortex, pioneering studies with intracellular recording and
biocytin labeling have reported both simple and complex-cell like inhibitory neurons,
with their orientation tuning properties ranging from unselective to tightly tuned
(Azouz et al., 1997; Hirsch et al., 2003). By using Ca
2+
imaging, a recent study in
superficial layers of mouse visual cortex suggest that GABAergic neurons exhibit
much weaker orientation selectivity compared to non-GABAergic neurons (Sohya et
al., 2007). Despite limited studies in visual and somatosensory cortices, functional
properties of inhibitory neurons in auditory cortex have rarely been examined.

44
Because the sampling bias in the classic “blind” whole-cell recording method (which
prefers larger cells such as pyramidal) and the sparse distribution of inhibitory
neurons prevent effective sampling of these neurons, in this study, we combined cell-
attached recording with juxtacellular labeling or subsequent intracellular recording to
selectively target fast-spike inhibitory neurons, the major source of local inhibitory
input to pyramidal neurons. Our results suggest that fast-spike inhibitory neurons
exhibit broader frequency tuning than excitatory neurons, and this property may
contribute to the equivalent lateral-inhibition effect.

3.2. Methods and Materials
3.2.1 Animal Preparation and Extracellular Recording
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-proof booth (Acoustic Systems) as
described previously (Zhang et al., 2001; Tan et al., 2004; Wu et al., 2006). Female
Sprague-Dawley rats (about 3 months old and weighing 250–300g) were
anaesthetized with ketamine and xylazine (ketamine: 45mg/kg; xylazine: 6.4mg/kg;
i.p.). The right auditory cortex was exposed and the right ear canal was plugged.
Multi-unit spikes were recorded with parylene-coated tungsten microelectrodes (2
M , FHC) at 500–600μm below the pial surface. Electrode signals were amplified
(Plexon Inc.), band-pass filtered between 300 and 6,000 Hz and then thresholded by
a custom-made LabView software (National Instrument) to extract the spike times.

45
Pure tones (0.5–64 kHz at 0.1 octave intervals, 25-ms duration, 3ms ramp) at eight
10 dB-spaced sound intensities were delivered through a calibrated free-field speaker
facing the left ear. The number of tone-evoked spikes was counted within a window
of 10-30 ms from the onset of tone stimulus. The characteristic frequency (CF) of a
recording site was defined as the tone frequency at the intensity threshold for spike
responses. Auditory cortical mapping was carried out by sequentially recording from
an array of cortical sites to identify the location and frequency representation of A1
as previously described. During mapping, the cortical surface was slowly perfused
with pre-warmed artificial cerebrospinal fluid (ACSF; in mM: NaCl 124, NaH2PO4
1.2, KCl 2.5, NaHCO3 25, Glucose 20, CaCl2 2, MgCl2 1) to prevent it from drying.

3.2.2. In vivo Whole-cell Recording
After premapping of A1, whole-cell recordings (Moore and Nelson, 1999;
Margrie et al., 2002; Zhang et al., 2003; Wehr and Zador, 2003; Tan et al., 2004; Wu
et al., 2006) were obtained from neurons located at 500–650 μm beneath the cortical
surface, corresponding to the input layers of the auditory cortex (Games and Winer,
1988). We prevented cortical pulsation with 4% agarose. For voltage-clamp
recording, the patch pipette (4-7 M ) contained (in mM): 125 Cs-gluconate, 5 TEA-
Cl, 4 MgATP, 0.3 GTP, 10 phosphocreatine, 10 HEPES, 1 EGTA, 2 CsCl, 2 QX-314,
pH 7.2, and 0.5% biocytin. Recordings were made with an Axopatch 200B amplifier
(Axon Instruments). The whole-cell and pipette capacitance were completely
compensated and the initial series resistance (20 50M ) was compensated for 50-

46
60% to achieve an effective series resistance of 10-25 M . Signals were filtered at 5
kHz and sampled at 10 kHz. Only neurons with resting membrane potentials lower
than -55 mV and stable series resistance were used for further analysis.
To obtain tone-evoked synaptic conductances, the cells were clamped at -
70mV and 0mV respectively, which are around the reversal potentials of inhibitory
and excitatory currents, as also described in our previous studies (Zhang et al., 2003;
Tan et al., 2004; Wu et al., 2006). The linearity of I-V curve (Figure 3.1F) suggests
that cortical cells can be reasonably clamped, which is further supported by the fact
that when cells were clamped at 0 mV, no significant excitatory currents were
observed (Figure 3.2A), except the outward Cl- currents. This may be attributed to
the use of intracellular cesium, TEA, QX-314, and ketamine anesthesia, which
together block most voltage-dependent currents (through K+ and Na+ channels, and
NMDA receptors).
Histological staining of the recorded cells after recording (Horikawa and
Armstrong, 1988; Hirsch et al., 2003; Zhu et al., 2004) indicates that the whole-cell
voltage-clamp recording method under our current condition biasedly sampled
pyramidal neurons. All twenty-five successfully reconstructed morphologies after
whole-cell voltage-clamp recordings showed typical pyramidal cells, consistent with
previous work (Moore and Nelson, 1999; Margrie et al., 2002; Wu et al., 2006).

47
3.2.3. Cell-attached Recording Followed by Current-clamp Recording or
Juxtacellular Labeling
For cell-attached recording, pipettes with smaller tip openings
(impedance:10-14 M ) were used to overcome the recording bias towards cells with
larger cell bodies. Pipettes were filled with ACSF solution containing 0.5% biocytin,
or intracellular solution containing 0.5% biocytin when subsequent whole-cell
current-clamp recordings were to follow. Loose seal (0.2-1 giga Ohm) was made
from neurons, allowing spikes only from the patched cell to be recorded. Recording
was under voltage-clamp mode without applying a holding voltage. Spike responses
are reflected by the current spikes (Figure 3.4A). Signals were filtered at 0.1-10
kHz. Spike shapes were determined online by custom-developed LabView software.
The chance of encountering a fast-spike neuron is around 10% in our recording
experience. The online spike sorting enabled us to specifically search for fast-spike
neurons.
Once the spike TRF was mapped (normally three repetitions), we applied
current pulses of 0.25–1 nA for 200 ms ON and 200 ms OFF for up to 20 min (Joshi
and Hawken, 2006; Turner et al., 2005; Otmakhova et al., 2002). During and after
the protocol, tone-evoked spike responses were monitored to make sure that there
was no damage to cell or drifting of the recording pipette. After the recording,
animals were perfused with 4% paraformaldehyde for histological procedures.
Normally, juxtacellular labeling only revealed the somatic and dendritic
morphologies. Six FS and ten RS cells were successfully reconstructed, and their

48
morphologies are all consistent with that of FS inhibitory neurons and excitatory
pyramidal neurons, respectively.
To determine the subthreshold TRFs of fast-spike neurons, the same small-
tipped pipettes were used, containing (in mM): 125 K-gluconate, 4 MgATP, 0.3 GTP,
10 phosphocreatine, 10 HEPES, 1 EGTA, 2 QX-314, pH 7.2, and 0.5% biocytin.
After identifying the fast-spike cell type and obtaining the complete spike TRF, the
cell-attached recording was followed by breaking in the cell membrane. Normal
histological procedures were carried out following the current-clamp recording. It
should be noted that it remains extremely difficult to achieve high-quality whole-cell
recordings under this searching strategy to dissect excitatory and inhibitory synaptic
inputs to fast-spike neurons. Nonetheless, the current technique allows us to define
the subthreshold and spike response ranges.

3.2.4. Data Analysis
Synaptic conductances: Excitatory and inhibitory synaptic conductance
were derived according to I (t, V) = Gr(V-Er) + Ge(t)(V-Ee) + Gi(t)(V-Ei). (Borg-
Graham et al., 1998; Anderson et al., 2000; Zhang et al., 2003; Wehr and Zador,
2003; Wu et al., 2006). I is the amplitude of synaptic current at any time point; Gr
and Er are the resting conductance and resting membrane potential, and were derived
from the baseline currents of each recording; 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. In this study, a corrected clamping voltage was used,

49
instead of the holding voltage applied (Vh). V(t) is corrected by V(t) = Vh – Rs*I(t),
where Rs was the effective series resistance. A 12 mV junction potential was
corrected. By holding the recorded cell at two different voltages, Ge and Gi were
calculated from the equation. Ge and Gi reflect the strength of pure excitatory and
inhibitory synaptic inputs, respectively. Under holding potentials of –70 mV,
activation of NMDA receptors can be ignored (Hestrin et al., 1990; Jahr and Stevens,
1990a; Jahr and Stevens, 1990b; Pinault, 1996). Thus the tone-evoked synaptic
currents are primarily mediated by AMPA and GABAA receptors.
Membrane potential responses: Membrane potential was calculated
according to: Vest (t)=(GrEr + Ge(t)Ee + Gi(t)Ei)/(Gr + Ge(t) + Gi(t)). Vest is the
estimated membrane potential change. To estimate the spiking responses of the
pyramidal cell from synaptic conductances, the spike threshold is set at 20 mV above
the resting membrane potential, according to results from our current-clamp
recordings (19.86 ± 4.12 mV, mean ± SD, n = 4). No significant difference in spike
threshold was observed for inhibitory neurons (17.51 and 19.11 mV, n=2). We noted
that by using a different equation with consideration of the cell capacitance (Wehr
and Zador, 2003), no qualitative difference in the membrane potential tuning curves
was observed. The reversal potential for inhibitory conductance is determined by the
ratio of Cl concentration in the intra-pipette solution and in the cerebrospinal fluid.
In our condition, the estimated Cl reversal potential is -70 mV after correction of
pipette junction potential. In our analyses, we also derived synaptic conductances
based on three different presumptive reversal potentials (-60, -70 and -80mV) and

50
verified that our conclusion was not sensitive to the variation in Cl reversal
potential.
Frequency tuning curves: A) The amplitude envelops for excitatory and
inhibitory inputs. After deriving excitatory and inhibitory conductances at a desired
testing intensity, the peak amplitudes of both conductances at each testing frequency
were determined. The envelop for peak amplitudes along frequency domain was
derived by using cubic spline interpolation algorithm in a custom-made software in
MATLAB. B) Normalization of tuning curves in Figure 3.3B and 3.3E. First, all
conductance values in each tuning curve were normalized to the maximum response
value. Next, the excitatory tuning curves or the membrane potential tuning curves
based on excitation alone were extended or compressed along the frequency axis by
a scaling factor to obtain the same half-peak bandwidth (i.e. bandwidth from f-50 to
f50). These normalized tuning curves were then aligned according to the half-peak
bandwidth before averaging. The corresponding inhibitory tuning curve or
membrane potential tuning curve based on both excitation and inhibition was
normalized by the same scaling factors, and was shifted by the same frequency
distance.
Tone-evoked responses: A) Spike responses. With cell-attached recording,
spikes can be detected without ambiguity because their amplitudes are normally
higher than 100pA, while the baseline fluctuation is within 10pA. Tone-driven spikes
were identified within a 15ms time window from a peri-stimulus-spike-time
histogram (PSTH) generated from all the response traces. In anaesthetized A1,

51
spontaneous firing in a single cell is lower than 10 Hz, suggesting that the error in
defining tone-evoked spikes caused by spontaneous activity is minor. The
characteristic frequency (CF) for the spike TRF (either from loose-patch or multi-
unit extracellular recording) was defined as the logarithmic center of the responding
frequency range at the intensity threshold (e.g. Zhang, et al, 2001, 2002). B)
Synaptic current and membrane potential responses. These responses were identified
according to their onset latencies and peak amplitudes. All the response traces
evoked by the same test stimulus were averaged, and the onset latency of this
average trace was identified at the time point in the rising phase of response wave
form, which was 3 folds of standard deviation of baseline. Only responses with
onset latencies within 7-30 ms from the onset of tone stimulus were considered in
this study.
Estimation of CF in cortical sites from premapping: Similar as previous
described (Zhang, et al., 2003; Tan et al., 2004), fifteen to twenty extracellular
recordings were made in each animal to roughly define the tonotopicity of the A1.
The CFs of recorded sites were plotted according to the relative cortical coordinates
of those sites along the tonotopic axis, and a fitted line was derived to determine a
frequency representation gradient (see Zhang, et al., 2002). CFs of unrecorded
cortical sites in the same A1 were then estimated according to their coordinates in the
tonotopic map.

52
3.3. Results
3.3.1. Frequency Tuning Curves of Excitatory and Inhibitory Inputs
To examine fine structures in the spectral patterns of excitatory and
inhibitory input to auditory cortical neurons, we applied blind in vivo whole-cell
voltage-clamp recordings in the recipient layer (layer 4) of the rat primary auditory
cortex (A1). The blind whole-cell recording method used under our experimental
conditions resulted in recording exclusively from excitatory neurons, as described
previously (Wu et al., 2006). The auditory stimuli were pure tones of various
frequencies and intensities, which were presented in a pseudo-random sequence.
Excitatory and inhibitory synaptic currents in response to tones were recorded under
the clamping voltages of -70 mV and 0 mV respectively, the potential levels close to
the reversal potential for GABAA receptor and glutamate receptor mediated currents
respectively. The excitatory and inhibitory synaptic conductances were derived from
synaptic currents (see Methods and Materials). As shown by an example cell in
Figure 3.1A and 3.1B, the excitatory and inhibitory synaptic TRFs largely overlap
with each other, consistent with previous reports (Zhang et al., 2003; Tan et al., 2004;
Wu et al., 2006). Interestingly, after deriving the envelope of response peaks, i.e. the
frequency tuning curve, for both excitatory and inhibitory synaptic input at 70dB
sound pressure level (SPL), we observed that inhibitory frequency tuning curve was
more flattened than that of excitatory input, as reflected by the faster saturation of
conductance magnitude along the frequency domain and the more plateau-like peak
of the tuning curve (Figure 3.1C). To quantify this effect, we measured the

53

Figure 3.1 Frequency Tuning of Synaptic Inputs in an example A1 Excitatory Neuron
(A and B) Left, excitatory (A) and inhibitory (B) synaptic currents recorded in an example neuron at -
70 mV and 0mV, respectively, in response to pure tones of various frequencies and intensities. Each
F

54
small trace represents response to a tone (averaged from two repeats). Middle, color maps represent
the TRFs of synaptic responses, with the color of each pixel indicating the peak amplitude of synaptic
currents. Inset below the color map shows individual traces of synaptic currents responding to a best-
frequency tone at 40, 50, 60 dB sound pressure level (SPL). Right, TRFs of excitatory and inhibitory
synaptic conductances, which were derived from the averaged synaptic currents.
(C) The enlarged profile of excitatory (upper) and inhibitory (middle) conductances at 70 dB SPL for
the same cell (cell1). Envelops, i.e. frequency tuning curves, were calculated from the peak
amplitudes of synaptic conductances and were indicated by the dashed lines. Bottom, the inhibitory
tuning curve was superimposed with the excitatory curve. The black and red scale values are for the
excitatory and inhibitory curve respectively.
(D) Dotted lines depict the boundary of TRF of excitatory (black) and inhibitory (red) input for the
same cell. Colored solid lines indicate the frequency ranges for responses with amplitudes larger than
60% of maximum value at each testing intensity. Black: excitatory; red: inhibitory.
(E) Frequency tuning curve of derived peak membrane potential responses when only excitatory input
was considered (black) or when both excitatory and inhibitory inputs were considered (red). The
tuning curves are normalized. The subtraction between the two curves is shown by the blue line.
Dashed red vertical line indicates the frequency for the peak response, and two solid red vertical lines
indicate the estimated frequency range of spike responses.
(F) I-V relation from the same neuron. Left, average synaptic currents (five repeats) evoked by a tone
at 5.2kHz and 60 dB at different holding potentials as indicated. Right: I-V curves for currents
averaged within the 20-22.5ms window after the stimulus onset. “r” is the correlation coefficient.

bandwidth of the tuning curve at 60% of maximum value (60% BW) for both
excitatory and inhibitory input. At all test intensities above the subthreshold
intensity threshold (20dB in this cell), the 60% BWs of inhibitory input (Figure
3.1D, red lines) were consistently broader than those of excitatory input (Figure
3.1D, black lines), suggesting that inhibitory input is less selectively tuned.
The less selective inhibitory tuning will presumably generate relatively
more inhibition in frequency regions flanking the peak of the excitatory tuning curve,
and result in a narrowing of frequency tuning in a manner analogous to lateral
inhibition. To demonstrate this effect, we derived tone-evoked membrane potential
changes for the same cell with and without considering inhibitory input (see Methods
and Materials). As shown in Figure 3.1E, the tuning curve of derived membrane

55
potential responses at 70dB SPL was sharper when excitatory and inhibitory input
were integrated than when only excitatory input was considered. The relative
sharpening effect of inhibition was demonstrated by subtracting the normalized
membrane potential tuning curve derived without considering inhibition from that
when inhibition was present (Figure 3.1E, blue). Apparently stronger inhibitory
effect is generated at the flanks of the peak of membrane excitation, which
determines the best frequency of the cell (Figure 3.1E).

3.3.2. Lateral Inhibitory Sharpening of Frequency Tuning
For other ten cells in which both excitatory and inhibitory currents were
recorded, we derived the excitatory and inhibitory frequency tuning curves (Figure
3.2A and 3.2B), as well as the tuning curves of membrane potential changes in the
absence and presence of inhibition (Figure 3.2C). In all these cells, inhibitory tuning
curve exhibited broader bandwidths than the excitatory tuning curve around its peak,
although the responding frequency range of inhibitory input was similar as or
slightly narrower than those of excitatory input (Figure 3.2B). In result, the derived
membrane potential tuning curves in the presence of inhibition were narrower around
the peak than those in the absence of inhibition (Figure 3.2C). By comparing the two
normalized membrane potential tuning curves for each cell, we estimated the relative
suppression effect of inhibition at different tone frequencies. In all the cases,
suppression tended to increase from the center of the best frequency on both sides,
consistent with the concept of lateral inhibition or inhibitory sidebands.

56

Figure 3.2 Frequency tunings of synaptic inputs in ten other excitatory neurons in A1
(A) Left, derived excitatory (‘E’) and inhibitory (‘I’) synaptic conductances within the responding
frequency range (labeled by numbers below) at 70 dB SPL. Right, example traces of excitatory

57
(upper) and inhibitory responses (lower) to a best-frequency tone. Traces are normalized to have the
same peak amplitude. Vertical line indicates the onset of excitatory response. Arrowheads indicate
that excitatory currents normally exhibit two phases in their arising kinetics. The onsets of inhibitory
inputs are roughly at the transition between these two phases.
(B) Tuning curves of excitatory (black) and inhibitory (red) conductances were superimposed for
comparing their shapes. The black and red scale values are for excitatory and inhibitory curves
respectively.
(C) Normalized tuning curves of membrane potential changes derived for the cells shown in (A) and
(B), with (red) and without (black) considering inhibition. Data are presented in a similar manner as in
Figure 3.1E.

The percentage difference in the bandwidth of excitatory and inhibitory
tuning curves at various levels were summarized for all the eleven cells (Figure 3.3).
We found that the responding frequency range (i.e. 0%BW) of inhibitory input was
on average slightly narrower than that of excitatory input (Figure 3.3A). In contrast,
inhibitory tuning curve was significantly broader at the levels of 40%, 60% and 80%
of maximum amplitude (Figure 3.3A). This was further demonstrated by averaging
all the normalized synaptic tuning curves (Figure 3.3B; see Methods and Materials).
The broader inhibitory bandwidths at the flanks of the excitatory tuning peak can
result in relatively stronger inhibition around the excitation peak, as shown by the
subtraction of the two averaged tuning curves. By examining the bandwidths at all
levels (at 1% step), we found that between the levels of 54% and 88% of maximum
amplitude, the bandwidths of the average inhibitory tuning curve were significantly
broader than those of the excitatory tuning curve (n = 11, two-way ANOVA, p <
0.05).
The effect of inhibition on the membrane potential tuning curve was
quantified by comparing the tuning curves of membrane potential responses

58
generated with and without inhibitory inputs. In the presence of inhibition, the
bandwidths of tuning curves were significantly reduced at various levels (e.g. 40%,
60% and 80% of maximum amplitude) (Figure 3.3C and 3.3D). The percentage
reduction of bandwidth appeared to increase as the measurement moved closer to the
peak of membrane response tuning curve (Figure 3.3D). To further demonstrate this
effect, the membrane potential tuning curves generated with and without inhibition
were normalized (in both amplitude and frequency domains) for all the cells and then
were averaged (Figure 3.3E; see Methods and Materials). There was a significant
reduction of bandwidth due to inhibition between the levels of 40% and 78% of peak
response (Figure 3.3E, right). The greatest reduction appears at the level rather close
to the peak (Figure 3.3E, right). Here, we also estimated the spike response range in
the averaged membrane potential tuning curve, since the spike threshold was at the
level of about 60% of maximum response (Figure 3.2C; 59.6% ± 5.8%, mean ± SD,
n = 11; see Methods and Materials). Within the estimated frequency range of spike
responses, the relative suppression increased from the peak of the tuning curves on
both sides. Together, these results demonstrate that the balance of excitation and
inhibition is only approximate, and that in addition to generally reducing excitation,
the more broadly-tuned inhibitory input can further sharpen the frequency
representation of cortical neurons by laterally narrowing the tuning curve especially
at around the peak.

59

Figure 3.3 Summary of frequency tunings of synaptic inputs and membrane potential responses
(A) Average percentage difference in bandwidth between excitatory and inhibitory tuning curves ((In-
Ex)/In) at different amplitude levels. Data are from the 11 cells shown in Figure 3.1 and 3.2. Bar
represents SEM. The numbers in the parenthesis indicate the “p” value for the statistics (paired t-test,
n = 11).
(B) Left, average frequency tuning curve of excitatory (black) and inhibitory (red) conductances after
normalization (see Experimental Procedures). f-50 and f50 are two reference points for aligning
normalized tuning curves. Right, percentage difference in bandwidth between the two averaged
tuning curves at different amplitude levels (step size = 1%). Data point labeled with red color
indicates significant difference (p < 0.05; two-way ANOVA).
(C) Bandwidths of membrane potential tuning curves without and with inhibition at 60% of maximum
amplitude (60% BW). *, paired t-test, P < 0.0001, n = 11 cells.
(D) Percentage reduction in bandwidth of membrane potential tuning curves after integration of
inhibition at the levels of 20%, 40%, 60%, 80% of maximum response, summarized for the 11 cells.
Bar = SEM. “p” values are indicated in the parenthesis (one-group t-test, n = 11).
(E) Left, average frequency tuning curve of peak membrane potential changes with (red) or without
inhibition (black). Blue line is plotted similarly as that in Figure 3.1E. Solid red lines indicate the
estimated spiking frequency range. Right, percentage reduction in bandwidth along the tuning curves
(one percent step) after integrating inhibition, analyzed in the same way as in (B).

60
3.3.3. Fast-spike Inhibitory Neurons and Regular-spike Excitatory Neurons
Why does inhibitory input exhibit less selective frequency tuning than
excitatory input? Our recent study suggests that the shape of membrane potential
tuning curve (especially in the suprathreshold frequency ranges) is largely defined by
recurrent cortical excitatory inputs (Liu et al., 2007). Thus, if cortical inhibitory
neurons have broader frequency tuning than cortical excitatory neurons, this can
cause broader tuning of inhibitory input. Previous studies in cortical slices
demonstrate that cortical pyramidal neurons in the input layer primarily receive
inhibitory input from nearby fast-spike inhibitory neurons (Agmon et al., 1992; Gil
and Amitai, 1996; Gibson et al.,1999; Inoue and Imoto, 2006; Sun et al., 2006). To
examine the frequency tuning and spike TRF of inhibitory neurons, we applied cell-
attached recording and juxtacellular labeling (see Methods and Materials) as to
distinguish neuronal types according to the spiking property and morphology of the
recorded cells. With cell-attached recording, only spikes from the targeted neuron
are recorded. Because the pipette capacitance was completely compensated in our
experiments, distortion of spike waveform was minimized, and thus we could
compare the spike waveforms from different recordings. Figure 3.4A (upper panel)
shows two typical types of spikes, either spontaneously generated or evoked by tone
stimuli. The first type (observed in about 10% of encountered neurons; see
Experimental Procedures) exhibits a relatively large upward peak and a short peak-
to-peak interval, which is defined as the time interval between the trough and the
upward peak of the spike waveform. The second type has a smaller upward peak and

61

Figure 3.4 Cell-Attached Recordings from Fast-spike (FS) and Regular-spike (RS) Neurons and
Juxtacellular Labeling
(A) Upper panel: example spike waveform and tone-evoked spike response in a FS (left) and a RS
neuron (right). Two dashed lines indicate the peak-to-peak interval. Arrowhead indicates the onset of
tone stimulus. Lower panel: example reconstructed dendritic morphologies of neurons labeled by
juxtacellular methods following cell-attached recordings. The values of their peak-to-peak interval are
indicated in the parenthesis.
(B) Distribution of average peak-to-peak intervals from 111 recorded cells. The dashed line indicates
the separation between the groups of FS and RS neurons in this study. Sampling here was not random
because we specifically searched for fast-spike neurons.
(C) Correlation between the CFs of neurons determined from their spike TRFs and the CFs predicted
from their positions in the tonotopic map, which was determined with multi-unit extracellular
recordings.

62
a longer peak-to-peak interval. According to the distribution of peak-to-peak
intervals (Figure 3.4B), we arbitrarily categorized the recorded neurons into two
groups: fast-spike (FS, with peak-to-peak interval <0.6 ms) and regular-spike (RS,
with peak-to-peak interval 0.6 ms) neurons. The average peak-to-peak interval is
0.35 0.09 ms (mean ± SD; n = 42) for fast-spike neurons, and 0.83 ± 0.12 ms (n =
69) for regular-spike neurons. The spiking property of FS neurons is consistent with
previous reports of fast-spiking inhibitory neurons (Mountcastle et al., 1969;
Swadlow, 1989; Azouz et al., 1997). The fast-spike neurons often exhibit a train of
action potentials when stimulated with a brief tone, while regular-spike neurons
usually exhibit single-spike responses (Figure 3.4A). Based on the morphology of
the recorded cells successfully reconstructed after juxtacellular labeling or
intracellular labeling (see Experimental Procedures), we found that the fast-spike
neurons exhibited locally-constrained and smooth dendritic arbors, while typical
pyramidal-cell morphology with spiny dendritic arbors were mostly observed for
regular-spike neurons (Figure 3.4A, lower panel). Previous studies have shown that
fast-spiking neurons are parvalbumin-positive GABAergic neurons, occupying about
70% inhibitory neuron population in layer 4 (Kawaguchi and Kubota, 1997; Gonchar
and Burkhalter, 1997). It is reasonable to assume that the fast-spike neurons
recorded under our condition were inhibitory and that the regular-spike neurons were
primarily excitatory. Since the characteristic frequencies (CFs) of the recorded fast-
spike and regular-spike neurons correlate equally well with the estimated CFs for
these cells (predicted according to their positions in the A1 tonotopic map, see

63
Methods and Materials), we conclude that inhibitory neurons are organized into the
same tonotopic map as excitatory neurons (Figure 3.4C). The same tonotopic map
for both excitatory and inhibitory neurons suggests that the topographic organization
of thalamocortical innervations is independent of cortical neuronal types.

Figure 3.5 Spike TRFs of FS and RS Neurons
(A, B) Example spike TRF of a FS neuron (A) and a nearby RS neuron (B). The enlarged recording
trace indicates the spike wave form (right). Inset shows tone-evoked spikes in single trials. Arrowhead
indicates the onset of tone stimulus.
(C) Left, distribution of bandwidths of spike TRFs at 60 dB SPL (BW60) plotted against the cells’
CFs. Right, average spike BW60 of FS and RS cells with CFs below or above 6 kHz. Bars are SEM.
*, two-sample t-test, P<0.02.
(D) Average intensity threshold of spike TRFs of the same group of FS (open) and RS (filled) neurons
with CFs below or above 6 kHz. Bars are SEM. *, two-sample t-test, P<0.02.
(E) Average onset latency of the first tone-evoked spike in FS and RS neurons. Bars are SEM. *, two-
sample t-test, P<0.001. The number of cells is indicated.

64
3.3.4. Spike TRFs of Fast-spike Inhibitory Neurons
Because the majority of inhibitory neurons in layer 4 are fast-spike neurons,
we specifically examined the spike TRFs of this type of inhibitory neurons. In
Figure 3.5A and 3.5B, two nearby fast-spike and regular-spike neurons in the same
preparation were recorded with the cell-attached recording method. Their complete
spike TRFs were reconstructed and compared. The fast-spike neuron was tuned with
a CF close to that of the regular-spike neuron (2.8 and 1.9 kHz respectively, see
Methods and Materials). It appeared that the TRF area of the fast-spike neuron was
larger than that of the regular-spike neuron, with a lower intensity threshold and
broader responding frequency ranges at all testing intensities. To compare the TRF
properties between fast-spike and regular-spike neurons, a group of FS and RS
neurons were randomly recorded (Figure 3.6 and Figure 3.7). We arbitrarily divided
these cells into two groups according to their CFs (CF < 6kHz and CF 6 kHz)
since TRF properties may also depend on the cell’s CF (Zhang et al., 2001; Polley et
al., 2007). As shown in Figure 3.5C and Figure 3.8, the responding frequency ranges
at all testing intensities were significantly broader in fast-spike neurons, and the
intensity threshold of their TRFs was significantly lower than that of regular-spike
neurons (Figure 3.5D). The broader bandwidth and lower intensity threshold predict
that the fast-spike neurons possess significantly larger TRFs than nearby regular-
spike neurons. In other words, at a given intensity, fast-spike inhibitory neurons
exhibit less-selective frequency-tuning than nearby excitatory neurons. Neurons in
layer 4 receive excitatory input primarily through thalamocortical and local

65
intracortical excitatory projections, and inhibitory input mainly from local fast-spike
neurons. The broader spike TRFs of fast-spike neurons than those of nearby
excitatory neurons will account, at least partially, for the less selective frequency
tuning of inhibitory input.

Figure 3.6 Spike TRFs of FS neurons recorded with loose-patch technique
TRFs of sixteen cells (1-16) from eighteen recordings were shown in color maps. Each TRF was
averaged from 2-5 repetitions. The max scales of the color bar are: 4 in (1), 1.67 in (2), 6 in (3), 6 in
(4), 3.3 in (5), 2.7 in (6), 1.75 in (7), 3 in (8), 2.75 in (9), 3 in (10), 3 in (11), 5 in (12), 3 in (13), 3 in
(14), 4in (15), 1.75 in (16).

66

Figure 3.7 Spike TRFs of RS neurons recorded with loose-patch technique.
TRFs of twenty cells (1-20) from twenty-two recordings were shown in color maps. Each TRF was
averaged from 2-4 repetitions. The max scales of the color bar are: 2 in (1), 2 in (2), 1.5 in (3), 1.5 in
(4), 0.8 in (5), 1.333 in (6), 2 in (7), 2 in (8), 1.5 in (9), 2.5 in (10), 0.667 in (11), 2 in (12), 1 in (13), 1
in (14), 1 in (15), 1.25 in (16), 1.667 in (17), 2 in (18), 1.333 in (19), 1.5 in (20).
Among those cells, FS cells (15) and (11) were recorded from cortical sites close to RS (18) and (7) in
the same animal preparations, respectively.

67

Figure 3.8 Spike TRFs of regular-spike (RS) and fast-spike (FS) neurons
(A) Bandwidths at various testing intensities (40-70 dB SPL) of loose-patch recorded cells were
plotted according to their characteristic frequencies (CF). The dash or solid lines indicate the averaged
bandwidths of FS or RSneurons, respectively.
(B) Summary of the averaged bandwidths. Right, maximum number of evoked spikes within 10-40
ms time window after the onset of the tone stimulus. (n=18 for FS and 22 for RS; *: two-sample t-test,
p<0.005 ) .
(C) Summary of the averaged bandwidths into two sampling groups: CF < 6kHz and CF >6kHz. (n=8
for FS & CF<6kHz, 8 for RS & CF<6kHz, 10 for FS & CF>6kHz. and 14 for RS & CF>6kHz, *:
two-sample t-test, p<0.005). Bar=SE.

68
3.3.5. TRF of Membrane Potential Responses in Fast-spike Neurons
We next examined whether the broader TRFs of fast-spike inhibitory
neurons are due to broader ranges of their excitatory inputs, or due to their more
efficient conversion of synaptic inputs to spike outputs. Because of the sparse
distribution of inhibitory neurons, cell-attached recording followed by whole-cell
current-clamp recording was applied to select fast-spike neurons. Their TRFs of
membrane depolarization responses were examined (Figure 3.9A). Here, tone-
evoked membrane depolarizations were identified according to the amplitude and
onset latency of membrane potential changes following the stimulus onset (see
Methods and Materials). We then compared the spike TRF (determined with initial
cell-attached recording) and the TRF of membrane depolarizations (determined with
subsequent current-clamp recording) of the same cell. In the fast-spike inhibitory
neuron, the spike TRF largely overlapped with the TRF of membrane depolarization
responses, with the same intensity threshold and slightly narrower responding
frequency ranges above the intensity threshold (Figure 3.9A and 3.9C). In
comparison, a regular-spike neuron which had a similar CF exhibited a smaller spike
TRF, leaving larger subthreshold regions at the periphery of its membrane potential
TRF (Figure 3.9B and 3.9C). A total of 6 fast-spike neurons and 9 regular-spike
neurons were recorded with both spike and membrane potential TRFs obtained. The
percentage occupancy of spike response region in the membrane response region was
measured at 60 dB SPL (Figure 3.9D, left). The fast-spike neurons possessed
relatively larger spike response regions (77.3 ± 8.5%, mean ± SD, n = 6) than

69
regular-spike neurons (54.3 ± 7.6%, n = 9; two-sample t-test, P < 0.001). In the
meantime, the frequency range of membrane depolarization responses at 60 dB SPL
was not significantly different between the two groups of neurons (Figure 3.9D,
right). Thus, our results demonstrate that the frequency range of excitatory inputs is
similar between fast-spike and regular-spike neurons. However, fast-spike neurons
can convert a broader range of synaptic inputs into spike outputs. This may be
partially due to stronger thalamocortical synapses made on fast-spike inhibitory
neurons than on pyramidal cells, as suggested by several recent studies in cortical
slices (Cruikshank et al., 2007; Daw et al., 2007). Further technical development
will be needed to address the synaptic mechanisms underlying the high-sensitivity of
fast-spike neurons (see Methods and Materials). Nonetheless, the less selective
outputs from fast-spike neurons may contribute to the more broadly-tuned inhibitory
input to cortical excitatory neurons, and result in the lateral sharpening of frequency
representation of these cells.

70

Figure 3.9 Suprathreshold and Subthrehold Regions of Synaptic TRFs in FS and RS Neurons
(A, B) Left, spike TRF of a FS (A) and a RS (B) neuron mapped with cell-attached recordings.
Enlarged spike waveform is shown in the inset. Right, TRF of membrane potential responses mapped
with subsequent whole-cell current-clamp recording. Insets show the reconstructed morphology of the
recorded FS and RS cells labeled with intracellular loading of biocytin.
(C) Left, frequency ranges of synaptic responses (red) and spike responses (blue) at different
intensities for cells shown in (A) and (B).
(D) Left, percentage frequency range of spike responses relative to synaptic responses at 60 dB SPL in
FS (n = 6) and RS neurons (n = 9). Two groups are significantly different (P<0.001, two-sample t-
test). Right, frequency range of synaptic responses at 60 dB SPL in FS and RS neurons.

71
3.4. Discussion
In this study, we examined the detailed patterns of excitatory and inhibitory
input to individual cortical excitatory neurons. We have revealed that under
approximately balanced excitation and inhibition, inhibitory input exhibits
significantly broader frequency tuning than excitatory input. This results in relatively
stronger inhibition in frequency regions flanking the preferred frequencies of the cell
and an effective lateral sharpening of its frequency tuning. By further examining the
spike TRFs of fast-spike inhibitory neurons, our results suggest that the observed
lateral sharpening may be attributed to the broader spike TRFs of fast-spike neurons
compared to regular-spike neurons. In addition, intracellular recordings indicate that
the broader spike TRFs of fast-spike neurons are not due to broader ranges of
excitatory inputs to these neurons, but their higher efficiency of converting inputs to
spike outputs.

3.4.1. Approximately Balanced Excitation and Inhibition
Previous in vivo whole-cell recording studies have led to the conclusion that
balanced excitation and inhibition underlie the frequency-intensity tonal receptive
field in the rat auditory cortex. This is evidenced by: first, excitatory and inhibitory
synaptic receptive fields are largely matched and exhibit similar preferred
frequencies (Zhang et al., 2003; Wehr and Zador, 2003; Tan et al., 2004; Wu et al.,
2006); second, excitatory and inhibitory conductances activated by the same tone
stimulus have similar amplitudes and exhibit significant linear covariation under

72
different tone stimuli (Wehr and Zador, 2003; Zhang et al., 2003; Tan et al., 2004);
third, inhibition closely follows excitation, resulting in a scaling down of excitation
(Ojima and Murakami, 2002; Wehr and Zador, 2003; Tan et al., 2004; Wu et al.,
2006). However, the “balance” between excitation and inhibition has not been
quantified with sufficient measuring resolutions in previous studies. In this study our
data are largely consistent with the concept of balanced excitation and inhibition in
that excitatory and inhibitory inputs have similar TRF areas and preferred
frequencies. However, more detailed examination of the tuning patterns of
excitatory and inhibitory inputs reveals that the balance is in fact only approximate.
Inhibition has less-selective frequency tuning than excitation, as reflected by the
broader 60% BW and 80% BW, which contribute significantly to the spike frequency
range of the cell. Our analysis further suggests that under the approximate balance,
the fine structures in the excitatory and inhibitory tuning patterns can endow extra
processing power to cortical neurons.

3.4.2. Inhibitory Mechanisms for Shaping Frequency Representation
Previous pharmacological experiments in sensory cortices have shown that
blocking cortical inhibition results in a reduction of representational selectivity, e.g.
orientation selectivity in the visual cortex (Sillito, 1977; Sillito, 1979; Cook et al.,
1997), and frequency selectivity in the auditory cortex (Chen and Jen, 2000; Wang et
al., 2000; Wang et al., 2002). Previous studies have demonstrated that synaptic
inhibition closely follows the excitation evoked by the same tone stimulus, and can

73
interact with excitation to shape the spike response. In our work, the onset of
inhibitory input was measured at 2.7 ± 0.5 ms (mean ± sd, n = 15 neurons; evoked by
CF tones at 70 dB) after that of excitatory input, while the onset of the first tone-
evoked spike was 4.5 ± 0.8 ms (n = 8 neurons; obtained in current-clamp recordings)
after that of excitatory input (also see Figure 3.5E). The significant later onset of the
first spike compared to inhibitory input (independent t-test, p<0.001) indicates that
inhibitory input can affect spike responses. This is consistent with previous findings
in the thalamocortical circuit of various sensory cortices that the temporally delayed
inhibitory input can control the threshold for the generation of spike responses (e.g.
Douglas and Martin 1991; Somers et al., 1995; Anderson et al., 2000; Zhang et al.,
2003; Wehr and Zador, 2003; Tan et al., 2004; Higley and Contreras, 2006). Because
inhibition is roughly balanced with excitation in the rat auditory cortex, it is
proposed that the inhibitory sharpening of frequency tuning can be simply achieved
through generally scaling down the level of excitation (Wehr and Zador, 2003; Tan,
et al., 2004). In this study, our data have revealed an additional mechanism for
inhibitory sharpening, i.e. an equivalent lateral inhibitory effect. Since inhibitory
tuning curve exhibits a broader peak than excitation, it generates relatively stronger
inhibition in frequency regions at the flanks of the excitation peak, and thus narrows
the tuning curve of membrane potential responses round the peak. Taken together,
our results have united the two models of balanced excitation-inhibition and lateral
inhibition, and demonstrate that fine structures in the pattern of synaptic inputs can
play significant roles in determining cortical representation and processing functions.

74
3.4.3. Properties of Auditory Cortical Inhibitory Neurons
In many cortical areas, only 15 –25% of total neurons are GABAergic
inhibitory neurons (Peters and Kara,1985; Hendry et al., 1987; Priet et al., 1994).
Because of their sparseness and the difficulty in identifying them in vivo, our
knowledge of functional properties of cortical inhibitory neurons is scarce; neither do
we know much about the structure of inhibitory circuits. Extracellular recordings
from suspected inhibitory neurons in the input layer of rabbit somatosensory cortex
suggest that they respond unselectively to the direction of whisker displacement,
while principal cells are known to exhibit direction selectivity (Swadlow, 2003). In
layer 4 of cat primary visual cortex, two functional populations of inhibitory cells
(simple and complex, similar to pyramidal neurons) were identified by intracellular
recording and biocytin labeling (Hirsch et al., 2003), while a recent calcium imaging
study in mice suggest that in layer 2/3, GABAergic neurons exhibit much weaker
orientation selectivity compared to non-GABAergic neurons (Sohya et al., 2007).
There are different subtypes of inhibitory neurons according to their distinct
physiological, morphologically and neurochemical properties (Kawaguchi and
Kubota, 1997; Gonchar and Burkhalter, 1997; Gupta et al., 2000). In this study, by
combining cell-attached spike recording with juxtacellular labeling or with
subsequent whole-cell recording and labeling, we specifically targeted fast-spike
inhibitory neurons in the rat A1. Fast-spike inhibitory neurons are parvalbumin
positive, and occupy about 70% of inhibitory neuron population in layer 4. The
major type of fast-spike neurons is basket cells. The minor type is chandelier cells,

75
which normally are not driven by sensory inputs under physiological conditions (Zhu
et al., 2004). Thus, tone-evoked inhibitory inputs to layer 4 neurons are likely
provided primarily by nearby fast-spike basket cells. Our results indicate that fast-
spike inhibitory neurons have broader spike TRFs and exhibit lower selectivity but
higher sensitivity in response to tonal stimuli, consistent with the conventionally
assumed role of cortical inhibitory neurons.

3.4.4. Implication on Cortical Circuitry
The onsets of tone-evoked spike responses in fast-spike inhibitory neurons
are slightly earlier (13.05 ± 0.37 ms, mean ± SEM, n = 18) than those of regular-
spike neurons (14.82 ± 0.26 ms, n = 22; Figure 3.5E). Since fast-spike inhibitory
neurons are the major source of cortical inhibition in layer 4 (Agmon et al., 1992; Gil
and Amitai, 1996; Gibson et al.,1999; Inoue and Imoto, 2006; Sun et al., 2006), this
suggests that inhibitory inputs to layer 4 neurons could be primarily feedforward
(Tan et al., 2004). In addition, our data indicate that neighboring fast-spike and
excitatory neurons have similar preferred frequencies and frequency ranges of
excitatory inputs (as indicated by the tone-evoked membrane depolarizations). Thus,
a simplified circuitry model can be proposed here: the neighboring excitatory
neurons and fast-spike inhibitory neurons are innervated by a similar set of
thalamocortical axons, and the excitatory neurons also receive feedforward inhibition
from the fast-spike neurons. The slightly narrower frequency range of inhibitory
input than that of excitatory input in pyramidal neurons (Figure 3.3A) can be

76
attributed to the fact that the frequency range of spike output of fast-spike neurons is
slightly narrower than that of their excitatory input (Figure 3.9C and 3.9D). Because
the amplitudes of inhibitory inputs are graded at any given testing intensity, the
recorded pyramidal neuron likely receives inhibitory inputs from a group of cortical
inhibitory neurons. Further understanding of the distribution pattern of inhibitory
neurons that project to a single pyramidal neuron will be needed for a more realistic
model of cortical inhibitory circuitry.

3.5. Summary
Cortical inhibition plays an important role in shaping neuronal processing.
The underlying synaptic mechanisms remain controversial. Here, in vivo whole-cell
recordings from neurons in the rat primary auditory cortex revealed that the
frequency tuning curve of inhibitory input was broader than that of excitatory input.
This results in relatively stronger inhibition in frequency domains flanking the
preferred frequencies of the cell and a significant sharpening of the frequency tuning
of membrane responses. The less selective inhibition can be attributed to a broader
bandwidth and lower threshold of spike tonal receptive field of fast-spike inhibitory
neurons than nearby excitatory neurons, although both types of neurons receive
similar ranges of excitatory input, and are organized into the same tonotopic map.
Thus, the balance between excitation and inhibition is only approximate, and
intracortical inhibition with high sensitivity and low selectivity can laterally sharpen
the frequency tuning of neurons, ensuring their highly selective representation.

77
Chapter 3 Endnote

The work presented in this chapter appeared in the following publication:
Wu G.K.*, Arbuckle R.A.*, Liu B., Tao H.W. and Zhang L.I. (2008) Lateral
Sharpening of Cortical Frequency Tuning by Approximate Balanced Inhibition.
Neuron 58, 132-143. (*Equal contribution)

78
Chapter 4
Synaptic Mechanisms Underlying Intensity Tuning

4.1. Background and Introduction
Intensity-tuned neurons are characterized by their nonmonotonic responses
to tone intensities (Greenwood and Murayama 1965). Such neurons (also named
nonmonotonic neurons) have been observed along the central auditory pathway,
including the cochlear nucleus (Greenwood and Maruyama, 1965; Young and
Brownell, 1976), inferior colliculus (Aitkin 1991; 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). The response properties of cortical
intensity-tuned neurons (Phillips et al.,1995; Heil and Irvine, 1998) and their
susceptibility to specific changes after training animals with sound magnitude
discrimination task (Polley et al.,2004 and 2006) suggest that these neurons may play
important roles in the encoding of sound loudness and envelop transients. Because
auditory nerve fibers, the inputs to the central auditory system, have monotonically
increasing response-versus-intensity functions (Kiang et al., 1965), the generation of
intensity tuning in the central auditory system must rely on neural inhibition to
reduce activity preferentially at high intensities. Studies using extracellular
recordings with two-tone masking paradigms (Suga and Manabe, 1982; Calford and

79
Semple, 1995; Sutter and Loftus, 2003), with GABA receptor blockade (Faingold et
al., 1991; Pollak and Park, 1993; Wang et al., 2002; Sivaramakrishnan et al., 2004),
as well as using intracellular recordings (Ojima and Murakami, 2002) suggest that
intensity tuning may be produced by the spatial and/or temporal interaction of the
inhibition and excitation. However, without direct examination of sound-activated
synaptic inputs in individual intensity-tuned neurons, the synaptic mechanisms or
neuronal biophysical properties (Durstewitz and Sejnowski, 2000) that may underlie
the nonmonotonic response-intensity function or the conversion from monotonic to
nonmonotonic function remain elusive.
Recently, several studies on synaptic inputs underlying tone-evoked
responses indicate that the frequency tuning and the frequency-intensity tonal
receptive fields (TRFs) of cortical neurons are shaped by balanced excitatory and
inhibitory synaptic inputs (Zhang et al., 2003; Wehr and Zador, 2003; Tan et al.,
2004). This is evidenced by the covariation of the amplitudes of excitatory and
inhibitory synaptic conductances evoked by the same tone stimulus (Zhang et al.
2003, Tan et al., 2004) and a relatively stable temporal interval between them (Wehr
and Zador 2003). However, those data were mostly acquired from the primary
auditory cortex (A1) of rats where the majority of neurons do not exhibit intensity
tuning or exhibit weak tuning (Phillips and Kelly, 1989; Zhang et al., 2001; Doron et
al., 2002; Polley et al., 2004; Polley et al., 2006). In the present study, using in vivo
whole-cell voltage-clamp recording technique, we examined the excitatory and
inhibitory synaptic TRFs in two distinct classes of cortical neurons, intensity-tuned

80
and non-intensity-tuned. We quantified the amplitude and temporal relationship
between the excitatory and inhibitory inputs evoked by tone stimuli of various
intensities at characteristic frequencies (CFs) of the cells. Our data indicate that
cortical intensity tuning is determined by the interplay between tone-evoked
imbalanced excitatory and inhibitory synaptic inputs. In intensity-tuned neurons,
excitatory inputs already exhibit intensity tuning, whereas the inhibitory inputs
increase monotonically in their strength and quickly saturate with intensity
increments. In addition, the temporal delay of inhibitory inputs relative to excitatory
inputs is reduced with the increase of intensity, resulting in an enhanced suppression
of excitation at high intensities and a significant sharpening of intensity tuning.
These findings also imply that by controlling the relative timing of excitation and
inhibition, synaptic circuits can achieve a de novo construction of representational
properties.

4.2. Methods and Materials
4.2.1. Extracellular Recording
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-proof booth (Acoustic Systems). Female
Sprague-Dawley rats about 3 months old and weighing 250–300 g were
anaesthetized with ketamine and xylazine (ketamine: 45 mg/kg; xylazine: 6.4 mg/kg;
i.p.). The right auditory cortex was exposed and the right ear canal was plugged. The

81
body temperature was maintained at 37.5ºC by a feed-back heating system (Harvard
Apparatus, MA). Multiunit spike responses were recorded with parylene-coated
tungsten microelectrodes (FHC, ME) at 500–600 μm below the pial surface (Zhang
et al., 2001; Bao et al., 2003; Polley et al., 2004). Electrode signals were amplified
(Plexon Inc, TX), band-pass filtered between 300 and 6,000 Hz and then thresholded
in a custom-made software (LabView, National Instrument) to extract the spike
times. Pure tones (0.5–64 kHz at 0.1-octave intervals, 25-ms duration, 3 ms ramp) at
eight 10-dB-spaced sound intensities were delivered through a calibrated free-field
speaker facing the left ear. A time window from 10-25 ms from the onset of tone
stimulus was used for tone-evoked spike responses. The threshold of the spike TRF
was chosen to be the minimum stimulus intensity included in the TRF, and the
characteristic frequency (CF) was the tone frequency that evoked a response at
threshold. Bandwidth at 10 dB or 30 dB (BW10 or BW30) above threshold was the
frequency width (in octaves) of the TRF at that intensity level. Auditory cortical
mapping was carried out by sequentially recording from an array of cortical sites
with an average grid size of 120 μm for dense mappings and about 200 μm for rough
mappings (Kilgard and Merzenich, 1999; Zhang et al., 2001; Bao et al., 2003; Polley
et al., 2004; Tan et al., 2004).

4.2.2. In vivo Whole-cell Recording
Whole-cell recordings (Moore and Nelson, 1998; Margrie et al., 2002;
Zhang et al., 2003; Wehr and Zador, 2003; Tan et al., 2004) were obtained from

82
neurons located 500–700 m beneath the cortical surface. We prevented cortical
pulsation with 4% agarose. For voltage-clamp recording, the patch pipette (4-7 M )
contained (in mM): 125 Cs-gluconate, 5 TEA-Cl, 4 MgATP, 0.3 GTP, 10
phosphocreatine, 10 HEPES, 0.5 EGTA, 2 CsCl, pH 7.2. 5mM QX-314 was
included to improve the whole-cell clamping of the cell (Nelson et al., 1994).
Recordings were made with an Axopatch 200B amplifier (Axon Instruments). The
whole cell recording method under our conditions shows a sampling bias towards
relatively large pyramidal neurons, consistent with results from other laboratories
with similar experimental settings (Moore and Nelson, 1998; Margrie et al., 2002).
The whole-cell and pipette capacitances were completely compensated and the initial
series resistance (20 50M ) was compensated for 50-60% to achieve effective
series resistances of 10-25 M . Signals were filtered at 5 kHz and sampled at 10
kHz. Only neurons with resting membrane potentials lower than -55 mV and stable
series resistance (less than 10% change from the beginning of the recording) were
used for further analysis. The CFs of the synaptic TRFs of recorded neurons
matched their positions in the tonotopic map determined by extracellular recordings.

4.2.3. Data Analysis
The excitatory synaptic conductance Ge(t) and inhibitory synaptic
conductance Gi(t) at time t were derived (Borg-Graham et al., 1998, Anderson et al.,
2000) using I(t, V) = Gr(V-Er)+Ge(t)(V-Ee)+Gi(t)(V-Ei) where V is the clamping
voltage, Gr is the resting conductance, Er is the resting potential; Ee and Ei are the

83
reversal potentials for excitatory and inhibitory synaptic currents, respectively; and
I(t, V) is the current amplitude under V. Currents into the neuron were assigned a
negative value. The resting or leak conductance Gr was derived using Ir(V) = Gr(V –
Er) where Er is the resting potential, and Ir(V) is the resting current. Measurement
of I(V) at two voltages will solve the value of Ge and Gi in the equation. In this
study, a corrected clamping voltage V was used, instead of the clamping voltage
applied (Vc). V(t) is given by V(t) = Vc – Rs*I(t), where Rs was the effective series
resistance. Synaptic currents were obtained with the cell clamped at the reversal
potentials for inhibitory and excitatory currents respectively, for each of the 568 test
tone stimuli. For some of the experiments, the reversal potentials of glutamatergic
and GABAergic (Cl ) currents were roughly measured at the beginning by
examining the reversal of spontaneous glutamatergic and GABAergic currents
respectively, as the holding potential was changed. Under experimental condition in
this study, the reversal potential was found to be 0-8 mV for glutamatergic inputs,
and around -70 mV for GABAergic inputs, consistent with the values of Ee and Ei
determined by considering the ionic composition of the pipette solution and the
cerebrospinal fluid. In some cases, Ei values of -65 and -75 mV were also tested,
and this did not change the conclusion of the study.
Estimated membrane potential response Vest was derived from synaptic
conductances using Vest(t)=(GrEr+Ge(t)Ee+Gi(t)Ei) /(Gr+ Ge(t)+ Gi(t)), where Er
is the resting membrane potential, which was determined for each recorded neuron
under current-clamp recording at the beginning of the experiment. If only the

84
excitatory synaptic conductance was taken into account, Gi(t) was set to zero. The
spike threshold was set around -45 mV for auditory cortical neurons, an observation
from a previous study (Tan et al., 2004).
In this study, we have assumed linear, isopotential neurons in deriving
excitatory and inhibitory synaptic conductances, same as in previous studies (Zhang
et al., 2003; Wehr and Zador, 2003 and 2005; Tan et al., 2004). However, deviations
due to space clamp error and cable attenuation for synaptic inputs at the distal
dendrites (Spruston et al., 1993) should be kept in mind, as extensively discussed in
several recent studies (Wehr and Zador, 2003; Tan et al., 2004). Nevertheless, the
two major observations for intensity-tuned neurons, the nonmonotonic excitatory and
monotonic inhibitory inputs as well as the nonmonotonic change of temporal delay
between excitation and inhibition, are unlikely to be affected. First, the linearity of
synaptic IV curves (Figure 4.2E) suggested that synaptic conductances were not
strongly affected by nonlinearities of cortical neurons. This may be attributed to the
use of intracellular cesium, TEA, QX-314 and ketamine anaesthesia, which together
block most voltage-dependent currents. The relative accuracy of derived excitatory
reversal potential (Figure 4.2E) also suggests reasonably accurate voltage-clamp for
those recorded synaptic inputs, because errors in space clamp will result in apparent
deviations from the actual real reversal potential (Shu et al., 2003). In addition, the
use of cesium and TEA (blockers of potassium channels) will reduce the membrane
permeability and thus decrease the cable attenuation (Spruston et al., 1993). Second,
we make comparisons between responses to CF stimuli in the same recorded cell,

85
and between intensity-tuned and non-tuned neurons. Although space clamp errors or
cable attenuation will cause an underestimation of absolute conductances, our
conclusions are not dependent on the absolute level of conductances or the ratio
between excitation and inhibition. Finally, simulation based on a compartmental
neuron model suggests that the relative timing between excitatory and inhibitory
inputs at 50% peak is not significantly affected by the cable effects (Wehr and Zador,
2003).

4.3. Results
4.3.1. Distribution of Nonmonotonic Neurons in the Rat Auditory Cortex
To effectively investigate nonmonotonic neurons, we first determined the
spatial distribution of such neurons in the adult rat auditory cortex by high-density
mappings with multiunit extracellular recordings (100-180 sampling sites for each
map, see Methods and Materials). The frequency-intensity tonal receptive field for
spike responses (spike TRF) was reconstructed for each recorded sampling site. The
change of tone-evoked spike response in the function of tone intensity was examined
at the characteristic frequency (CF), which is the frequency the neuron is most
sensitive to. As shown in an example auditory cortical map (Figure 4.1A), three
major fields can be identified according to the tonotopic organization of frequency
representations: the primary auditory cortex (A1) which exhibits a clear tonotopic
gradient along the anterior-posterior axis, a small anterior auditory field (AAF)
which exhibits a reversed tonotopic gradient compared to A1, and a ventral auditory

86

Figure 4.1 Multiunit intensity-tuning in the rat auditory cortex
(A) An example map of frequency representation in the adult auditory cortex. The color indicates the
characteristic frequency (CF) for the sampling site located in the center of each polygon. Polygons
(Voronoi tessellations) were generated so that every point on the cortical surface was assumed to have
the characteristics of its closest neighbors. A1, the primary auditory cortex; AAF, anterior auditory
field; VAF, ventral auditory field; NM, non-monotonic zone. A, anterior; D, dorsal; P, posterior; V,

87
ventral. “x” indicates a sampling site which exhibited an intensity threshold of higher than 60 dB or
no clear frequency tuning.
(B and C) Left, tonal receptive field of spike responses (spike TRF) for a non-intensity-tuned site “I”
and an intensity-tuned site “II”, as labeled in the map in (A). Color represents the number of spikes
evoked by a tone stimulus. Right, normalized spike counts as a function of intensity level for CF-tone
stimuli. Intensity level “0” was set as where the highest spike rate was evoked. Data were from
fourteen randomly-chosen non-intensity-tuned sampling sites in A1 near NM and nine intensity-tuned
sites recorded within the NM zone, respectively. Error bar = SEM.
(D) Bandwidths at 10 dB (BW10) and 30 dB (BW30) above intensity threshold measured for A1 units
and NM units. Bar = SEM. “*” indicates significant difference, p < 0.03, ANOVA test.
(E) Distribution of intensity-tuning indices of A1 (red, n=65) and NM (green, n=24) units.

field (VAF) which has an apparent dorsal-ventral CF gradient, consistent with
previous reports (Bao et al., 2003; Kalatsky et al., 2005). In these regions, the
majority of sampling sites exhibited increased spike responses at high intensity levels
(Figure 4.1B, left). A typical monotonic function is shown after averaging spike
response-level functions at CFs of fourteen similar sites (Figure 4.1B, right).
Interestingly, sampling sites in a small posterior zone (named nonmontonic auditory
zone: “NM”) located between A1 and VAF consistently exhibited nonmonotonic
response-level functions, i.e. markedly reduced responses at high intensity levels
(Figure 4.1C). In addition, the bandwidth of spike TRF at 30 dB above threshold
(BW30) at those nonmonotonic sites was significantly narrower than at the
monotonic sites (Figure 4.1D). In all of nine high-density mapping experiments, we
observed a similar organization of frequency representation and the existence of a
NM zone. Between sampling sites of the NM zone and those of nearby A1 area
which have similar CFs, no significant difference was observed in either the response
onset latency (NM,15.99 ± 0.32 ms (SEM); A1, 15.72 ± 0.29 ms (SEM); p>0.5,
ANOV A test), or the threshold of spike TRFs (p > 0.5, ANOV A test).

88
We used an intensity-tuning index to quantify the level of intensity tuning at
CF. The index is defined as the ratio between the spike counts at the preferred
intensity (with the highest level of response) and at 30 dB above the preferred
intensity (or the highest intensity tested). Sampling sites with an index < 0.6 were
considered to be strongly intensity-tuned. In A1, about 5% of sampling sites
exhibited strong intensity tuning, while in the NM zone, about 80% of sampling sites
were strongly intensity-tuned (Figure 4.1E). Our data are consistent with those from
previous extracellular recordings (Phillips and Kelly, 1989; Doron et al., 2002; Bao
et al., 2003; Polley et al., 2004 and 2006), which already suggest that nonmonotonic
neurons are more abundant in the posterior part of the rat auditory cortex.

4.3.2. Excitatory and Inhibitory Synaptic Receptive Fields
To examine the synaptic mechanisms underlying intensity tuning of cortical
neurons, we applied in vivo whole-cell voltage-clamp recordings (see Methods and
Materails) to neurons in A1 and the NM zone. By voltage-clamping the cell’s
membrane potential at -70 mV and 0 mV, the reversal potentials for GABAA
receptor-mediated Cl currents and glutamate receptor-mediated excitatory currents
respectively, we obtained TRFs for both excitatory and inhibitory inputs in the
recorded cell. Synaptic TRFs for an example cell in the NM zone were shown in
Figures 4.2A and 4.2B, and for a cell in A1 were shown in Figures 4.2C and 4.2D.
Linear current-voltage relationship (I-V curve) was observed for the recorded
synaptic currents evoked by CF tones at 70 dB SPL (Figure 4.2E). The derived

89
reversal potential for the early component of these currents (mainly excitatory) was 0
± 4mV (SD), close to the known reversal potential for glutamatergic currents. These
data suggest that under our voltage-clamp recording conditions, those synaptic inputs
that contributed to the recorded tone-evoked currents were detected with a
reasonable accuracy (see Methods and Materials). The excitatory and inhibitory
synaptic TRFs obtained from voltage-clamp experiments provide a basis for
determining the synaptic mechanisms underlying the intensity-tuning properties of
cortical neurons.
Intensity tuning is usually defined at CFs of cortical neurons according to
their spike responses (Schreiner et al., 1992; Phillips et al., 1995; Heil and Irvine,
1998; Polley et al., 2004). Because synaptic TRF usually exhibits a lower intensity
threshold and is broader than the spike TRF of the same cell (Tan et al., 2004), to be
consistent with previous studies, the CF of the cell was estimated according to the
TRF of membrane potential responses, which were derived from the excitatory and
inhibitory synaptic conductances evoked by each tonal stimulus (see Methods and
Materials). Based on the resting membrane potential of the cell and the threshold for
spike generation (-45 mV; Tan, et al., 2004), the frequency-intensity responsive area
for spike responses was then estimated (Figure 4.2G and 4.2H, hatched area). The
CF was defined as the frequency at the threshold intensity of the spike TRF. For the
example cell recorded from the NM zone (Figure 4.2G), the intensity threshold was
estimated to be 20 dB SPL and the CF to be 2.14 kHz ± 0.1 octaves. At this CF, the
membrane potential response changed nonmonotonically with the increase of

90

Figure 4.2 Synaptic TRFs of intensity-tuned and non-intensity-tuned neurons
(A, B) An example intensity-tuned neuron recorded from the NM zone. TRFs of synaptic currents
evoked by pure tone stimuli at various frequencies and intensities were obtained, with the neuron
clamped at -70 mV (A) and 0 mV (B), respectively. The colormaps on the right indicate the
amplitudes of individual synaptic currents.
(C, D) An example non-intensity-tuned neuron recorded from A1. Data are presented as in (A, B).
(E) Left, synaptic currents (average of five repeats) evoked by a CF tone at 70 dB, recorded at
different holding potentials from the same neuron shown in (A, B). Arrow head indicates the onset of

91
tone stimuli. Right: I-V curves for synaptic currents averaged within 0-1 ms (red) and 20-22.5ms
(black) windows after the response onsets.
(F) Jitters of responses to repeated stimuli (70dB CF tone). Upper panel, small white box on the color
line indicates the time point at the onset of the response of the same color shown on top; small black
bar indicates the time point at 50% peak response. Lower panel, average latencies and peak
conductance. Bar = SD.
(G) TRF of peak membrane potential responses derived from each pair of synaptic inputs recorded at
-70 mV (A) and 0 mV (B). Color represents the postsynaptic potential change in mV. The resting
membrane potential of the cell was -65 mV. Arrow indicates the spike threshold for required
membrane potential change. Hatched area represents the estimated spike TRF and roughly covers all
the stimuli that could trigger suprathreshold responses. Note that the estimated spike TRF of this cell
is circumscribed.
(H) TRF of peak membrane potential responses derived for the cell shown in (C, D). The resting
membrane potential of this cell was -61 mV.

intensity, with the strongest response evoked at an intensity level close to the
intensity threshold (Figure 4.2G), suggesting that this cell represents a typical
intensity-tuned neuron. In contrast, for the example cell recorded from A1 (Figure
4.2H), the derived membrane potential response at the CF changed monotonically
with intensity increments, consistent with multiunit recording results for monotonic
sites (Figure 4.1B). In addition, although both the intensity-tuned and untuned
neurons exhibited broad synaptic TRFs, the spike TRF of the intensity-tuned neuron
was much smaller (Figures 4.2G, hatched area) and became circumscribed,
consistent with multiunit recording results for many nonmonotonic sites (Figure
4.1C). The response-intensity functions of calculated membrane potential responses
correlated well with those of directly recorded membrane potential responses, as
shown in five recorded neurons in which both voltage-clamp and current-clamp
recordings were obtained (Figure 4.3). Thus, by integrating excitatory and inhibitory

92

Figure 4.3 Intensity-tuning of recorded and derived membrane potential responses
(A) Left, average excitatory (recorded at -70 mV) and inhibitory (recorded at 0 mV) synaptic currents
and recorded membrane potential changes (PSPs) at CF (1.07 kHz) and 0-70dB intensities for cell 1.
Note that due to QX-314 which was included intracellularly, spikes of the recorded neuron were
blocked. Right, maps in gray scale depict the peak amplitudes of the recorded PSPs (Recorded, ‘R’)
and the derived PSPs (Derived, ‘D’) for cell 1 and another 4 cells. Monotonic neurons: cell 2, 3;
nomonotonic neurons: cell 1, 4, 5. Gray scale bar: white, 0mV; black, maximal value: 18mV, 22mV
for cell 1; 15mV, 17mV for cell 2; 17mV, 20mV for cell 3; 15mV, 16mV for cell 4; 26mV, 32mV for
cell 5.
(B) Intensity-tuning indices for the recorded PSPs and derived PSPs from the same cell. There is no
significant difference between these two groups: p > 0.9, paired t-test.

synaptic conductances, we were able to determine the monotonic or nonmonotonic
response properties of the neurons.

4.3.3. Nonmonotonic Excitation and Imbalanced Inhibition underlying Intensity
Tuning
We next determined the relative contribution of excitatory and inhibitory
inputs to intensity tuning of cortical cells by examining the amplitude of CF tone-
evoked excitatory and inhibitory synaptic conductances as a function of intensity

93
level, i.e. the amplitude-level function. In both the intensity-tuned and untuned
neurons, the excitatory and inhibitory synaptic TRFs exhibited an identical shape
(Figures 4.2A and 4.2B; 4.2C and 4.2D). However, for the intensity-tuned neuron,
there was a difference in the amplitude-level function between excitatory and
inhibitory conductances. It appeared that the amplitudes of excitatory currents
evoked by CF-tones were smaller at high intensity levels (Figure 4.2A, shaded
column), while those of inhibitory currents remained more or less similar (Figure
4.2B, shaded column). To quantify this phenomenon, after determining the CF, the
excitatory and inhibitory synaptic conductances evoked by repeated CF-tones were
averaged for various intensities. The same tone stimulus evoked relatively consistent
synaptic responses with small variations in amplitude (SD < 5% mean amplitude)
and small jitters in response latency (SD < 0.3 ms; Figure 4.2F). As shown in Figure
4.4A for the intensity-tuned neuron, the amplitude of the averaged excitatory
conductance changed nonmonotonically, with the peak amplitude evoked at the
threshold intensity (20 dB SPL). The amplitude of the inhibitory conductance
instead increased monotonically with the increase of intensity. Thus, the excitatory
and inhibitory inputs were imbalanced in this intensity-tuned neuron, similar as a
recent observation reported online (Tan, et al., 2006). In contrast, for the example
non-intensity-tuned neuron (Figure 4.4B), excitatory and inhibitory inputs exhibited
similar monotonic tuning curves, consistent with the previous findings of balanced
excitatory and inhibitory inputs that underlie the tone-evoked responses of A1
neurons (Zhang et al., 2003; Wehr and Zador, 2003; Tan et al., 2004). These data

94
suggest that the intensity tuning of the cortical neuron is primarily inherited from the
nonmonotonic excitatory inputs.

Figure 3.4 Intensity-tuning of synaptic conductances evoked by CF tones
(A) Left, traces of average excitatory (“E”) and inhibitory (“I”) synaptic conductances responding to
different CF-tone intensities for the intensity-tuned neuron shown in Figure 4.2A, 4.2B and 4.2G.
Traces are average from five repetitions. Right, average peak excitatory and inhibitory conductances
as a function of tone intensity. Bar = SEM.

95
(C) Derived membrane potential changes from (A) as a function of tone intensity, by considering
excitatory inputs only (filled triangle), by integrating excitatory and inhibitory inputs (filled square),
or by integrating inhibitory inputs and modified excitatory inputs (open circle). The excitatory inputs
were modified as such that amplitudes of all responses to tones above 20 dB were scaled to that at 20
dB, which was the peak value. The error bars were generated for derived PSPs by randomly pairing
excitatory input and inhibitory input in different repeats.
(E) Left column, 50% peak delay (filled square) and onset delay (open circle) of inhibitory
conductances relative to the associated excitatory conductances, plotted as a function of tone intensity.
50% peak delays were extracted from waveforms at half maximal amplitude (inhibition minus
excitation). Error bars were the summation of the variations (SD) in the timing of excitatory and
inhibitory inputs. Right column, pairs of average excitatory (black) and inhibitory (gray) synaptic
conductances from (A) plotted at a higher temporal resolution.
(B, D, F) Same presentation as in (A, C, E), but for the non-intensity-tuned neuron shown in Figure
4.2C, 4.2D and 4.2H.

4.3.4. Inhibitory Contribution to Nonmonotonic Intensity-tuning
To determine whether the monotonic inhibitory inputs contribute to
intensity tuning, we compared the intensity-tuning curves of membrane potential
responses derived by considering excitatory conductances only and by integrating
excitatory and inhibitory conductances. As shown in Figure 4.4C, for the same
intensity-tuned neuron, excitatory conductances alone can not fully account for the
intensity-tuning curve, because the amplitude of membrane potential changes (PSPs)
derived from excitatory conductances alone was reduced from the peak by only
about 25% at the highest intensity (Figure 4.4C, filled triangles), and even the
weakest excitatory input (at 70 dB SPL) was capable of triggering spikes.
Integrating the inhibitory conductances (averaged value) not only reduced the
amplitude of membrane potential responses in general but also significantly
enhanced the intensity tuning, as indicated by a steeper reduction in the membrane
potential responses with intensity increments (Figure 4.4C, filled squares). At the
highest intensity, the PSP was reduced by 75% from the peak, leaving only a narrow

96
range of intensities (20-30 dB SPL) at which spikes can be generated. In
comparison, for the non-intensity-tuned neuron, the inhibitory conductances only
scaled down the membrane potential responses without changing the shape of the
tuning curve (Figure 4.4B and 4.4D). These results indicate that cortical inhibition is
actively involved in shaping the intensity tuning of cortical neurons.

4.3.5. Temporal Shaping of Nonmonotonic Intensity-tuning by Synaptic
Inhibition
Cortical inhibitory inputs can shape spike responses through their temporal
interaction with excitatory inputs (Zhang et al., 2003; Wehr and Zador, 2003; Zhu et
al., 2004). For example, cortical synaptic inhibition enhances the direction
selectivity of cortical responses to frequency-modulated sound sweeps through a
larger suppression of synaptic excitation under stimuli of non-preferred direction
than of preferred direction (Zhang et al., 2003). This is achieved by a larger
temporal overlap between excitatory and inhibitory inputs evoked by non-preferred
stimuli, due to an asymmetric integration of inputs sequentially activated by sound
sweeps (Zhang et al., 2003). Here we examined the level of temporal overlap
between CF-tone evoked excitatory and inhibitory conductances at different
intensities. The average excitatory and inhibitory conductances evoked by the same
tone stimulus were plotted together (Figure 4.4E, right). Consistent with the
previous results (Zhang et al., 2003; Wehr and Zador, 2003; Tan et al., 2004),
inhibitory inputs followed the excitatory inputs with a brief temporal delay. To

97
quantify the relative delay of the inhibitory inputs, a delay index was used, which
was defined as the interval between the time points at which 50% peak amplitude
was reached in the rising phase of the average excitatory and inhibitory conductance
traces (inhibition minus excitation; Wehr and Zador, 2003). Interestingly, for the
example intensity-tuned neuron, the delay index reduced with the increase of
intensity (Figure 4.4E, left, filled squares), whereas for the example non-intensity-
tuned neuron, it remained more or less the same across different intensity levels
(Figure 4.4F). We also measured the difference in the onset latencies of excitatory
and inhibitory inputs, which were defined as the time points at which the amplitude
of evoked conductance became larger than three times of the standard deviation of
the baseline fluctuation. The result was consistent with the measurement of delay
indices (Figure 4.4E and 4.4F, open circles). The reduced relative delay of inhibition
as intensity goes higher leads to an increased temporal overlap between excitatory
and inhibitory inputs, and thus a larger suppression of tone-evoked excitation at high
intensities. Surprisingly, such nonmonotonic change of inhibitory delay can be
sufficient for the generation of intensity tuning. This is demonstrated by the
nonmonotonic tuning of membrane potential responses achieved even after removing
the nonmonotonicity of the excitatory inputs by keeping their amplitudes always at
the peak value (at 20 dB) (Figure 4.4C, open circles).

98

Figure 4.5 Summary for intensity-tuned (nonmonotonic) and untuned neurons
(A) A group of intensity-tuned neurons. Color maps depict average peak excitatory and inhibitory
conductances (from 3-5 repetitions), and membrane potential changes at 0-70 dB intensities of CF-
tones for each of 10 recorded NM neurons. Scale bar: dark blue, 0; dark red, maximal value: from cell
1 to cell 10, 4 nS, 3 nS, 5nS, 3 nS, 4nS, 5nS, 4nS, 7nS, 6nS, 2nS for excitatory conductance; 5nS,
2nS, 6nS, 2nS, 3nS, 7nS, 4nS, 5nS, 6nS, 2nS for inhibitory conductance; 16mV, 15mV, 16mV, 21mV,
33mV, 18mV, 28mV, 17mV, 14mV, 23mV for PSP. Right, scatter plot of intensity-tuning indices for
excitatory conductance, inhibitory conductance and PSP of each recorded neuron.

99
(B) Normalized evoked synaptic conductances of intensity-tuned neurons recorded from the NM
zone, as a function of relative intensity of CF-tones. Black triangle, excitatory; red square, inhibitory.
(C) Normalized membrane potential changes, as a function of relative intensity. Red square, with
consideration of inhibition; black triangle, without consideration of inhibition. “*”, p < 0.03, paired t-
test.
(D) Relative delay of inhibition, as a function of relative intensity. Data from the same cell are
represented by the same color and symbol. Black squares are average results. Bar = SEM.
(E-H) Same as (A-D) respectively, but for the group of non-intensity-tuned neurons (n = 9) recorded
in A1. Scale bar for the color maps in (E): dark blue, 0; dark red, maximal value: 2nS, 5nS, 2nS, 4nS,
4nS, 3nS, 4nS, 4nS, 2nS for excitatory conductance; 2nS, 3nS, 2nS, 2nS, 3nS, 3nS, 5nS, 5nS, 4nS for
inhibitory conductance; 22mV, 31mV, 18V, 26mV, 21mV, 19mV, 24mV, 28mV, 16mV for PSP.

4.3.6. Synaptic Mechanisms for Cortical Intensity Tuning
A total of thirteen intensity-tuned neurons were recorded from the NM
zone. In ten of them, complete excitatory and inhibitory synaptic TRFs were
obtained (Figure 4.5A). Data from these neurons were summarized. Here, the
intensity-tuning index was defined as the ratio between the response amplitudes
(either synaptic conductances or membrane potential changes) at the preferred
intensity and at the highest intensity tested. All of these neurons had an intensity-
tuning index <0.6 for their derived membrane potential responses (PSPs), and were
tuned to intensities close to their intensity thresholds (Figures 4.5A and 4.5B). In all
of them, excitatory inputs were clearly tuned, as indicated by their intensity-tuning
indices (<0.8; Figure 4.5A), and their nonmonotonically changing amplitudes with
intensity increments (Figure 4.5B). In contrast, inhibitory inputs were not tuned to
intensity and their monotonically changing amplitudes quickly saturated (Figures
4.5A and 4.5B). Consistent among all the intensity-tuned neurons, cortical inhibition
played an important role in sharpening the intensity tuning, as indicated by the

100
significant difference between the tuning curves obtained with excitatory inputs
alone and with both inputs taken into account (Figures 4.5A and 4.5C). This effect
can be largely attributed to the significantly reduced temporal delay of inhibitory
inputs relative to the excitatory inputs at high intensities (Figure 4.5D). As a
comparison, for nine non-intensity-tuned neurons recorded from A1, excitatory and
inhibitory inputs exhibited similar monotonic changes in the amplitude (Figures 4.5E
and 4.5F). The intensity-tuning indices for both the synaptic conductances and
derived PSPs were larger than 0.9. The inhibitory delay did not exhibit an intensity-
dependent change (Figure 4.5H). The end result is that synaptic inhibition in non-
intensity-tuned neurons only scales down the membrane potential responses, without
changing the shape of the intensity-tuning curve (Figure 4.5G). Taken together,
although the extent to which synaptic excitation is thalamic or cortical in origin
cannot be inferred from our data, we conclude that intensity tuning of cortical
neurons is primarily inherited from their excitatory inputs relayed from intensity-
tuned neurons, and is further enhanced by cortical inhibitory inputs.

4.4. Discussion
4.4.1. Nonmonotonic Neurons in the Auditory Cortex
The distribution of nonmonotonic neurons has been extensively studied in
the cat auditory cortex with extracellular recording methods. Nonmonotonic neurons
have been observed throughout cat A1: besides mildly tuned units, about 25% of
sampling units exhibit strong intensity tuning with the firing rate reduced by >50% at

101
high intensities compared to that at the preferred intensity (Heil et al. 1994; Sutter
and Schreiner, 1995). In addition, in the posterior auditory field (PAF or Field P),
roughly 80% of the neurons exhibit intensity tuning, and about 40% of them show
circumscribed spike TRFs (Heil and Irvine, 1998; Kitzes and Hollrigel, 1996;
Phillips and Orman, 1984). Despite the difference in location, experimental data
suggested that common neural mechanisms may underlie the nonmonotonic
response-level function of both A1 and PAF neurons (Sutter and Loftus, 2003; Ojima
and Murakami, 2002). In the present study of rat auditory cortex, both the location
and the response properties of neurons of the NM zone (more than 80% of neurons
exhibiting nonmonotonicity ) are very similar to those of Field P in the cat auditory
cortex (Read et al., 2002). However, in A1 only about 5% of neurons are
significantly intensity-tuned, with intensity-tuning indices < 0.6 (Figure 4.1E).
Because of the sparseness of nonmonotonic cells in rat A1, only one such cell was
recorded in our experiment (Figure 4.6A and 4.6B). In this cell, the tuning curves for
excitatory and inhibitory inputs were similar to those in NM neurons, suggesting that
the synaptic mechanisms revealed in this study may represent common mechanisms
for nonmonotonic response-level functions. As discussed below, these mechanisms
can also account for neural models proposed for nonmonotonic cells in cat auditory
cortex.
Previous studies indicate that major forms of social vocalization in rats are
ultrasonic (16-64 kHz), also termed as 22 kHz calls (Sales and Pye, 1974; Nyby and
Whitney, 1978; Kaltwasser, 1990; Brudzynski et al., 1993 and 2002; Hashimoto et

102
al., 2004). This is consistent with the relatively broad cortical representation of high
frequencies in the rat auditory cortex (Figure 4.1A). Interestingly, it appears that
intensities of low-frequency sounds are more broadly represented since neurons in

Figure 4.6 An A1 nonmonotonic cell
(A) A cell recorded in A1 exhibited similar characteristics of intensity tuning as neurons in the NM
zone. Average excitatory (with cell clamped at -70 mV) and inhibitory (clamped at 0 mV) synaptic
responses, and peak amplitudes of derived membrane potential changes at CF (2.14 kHz) and 0-70dB
intensities are shown. The amplitudes of excitatory inputs as well as the membrane potential
responses exhibit nonmonotonic tuning, while those of inhibitory inputs change monotonically with
intensity increments.
(B) The relative onset delays of inhibitory inputs at various intensities for the same cell in (A). Bar =
SD. Note that this cell is not included for summary in Figure 5.

the NM zone normally have CFs of lower than 8 kHz (Figure 4.1, and observations
not shown). While intensities of high-frequency sounds could be represented by
sparsely distributed nonmonotonic cells in A1 (Figure 4.1E; Zhang et al., 2001; Bao
et al., 2003; Polley et al., 2004 and 2006), the relationship between the distribution of

103
nonmonotonic cortical cells and the ability of rats to discriminate intensities of
sounds at various frequencies remains to be investigated.

4.4.2. Nonmonotonic Excitation Primarily Determines Intensity Tuning
Previously lateral inhibition has been proposed to be responsible for
producing intensity tuning at higher levels of the auditory system (Shamma, 1985;
Suga and Manabe, 1982; Phillips et al., 1995; Calford and Semple, 1995; Sutter and
Loftus, 2003). This model has been supported by extracellular recording
experiments using forward-masking or simultaneous two-tone masking paradigms
(Suga and Manabe, 1982; Calford and Semple, 1995; Sutter and Loftus, 2003).
These studies suggested that the intensity tunings of excitatory and suppressive
domains were negatively correlated, and that inhibitory sidebands could be involved
in intensity tuning. However, due to the nature of masking protocol, lateral
suppressive domains may reflect complex temporal interaction between various
synaptic inputs evoked by the testing sounds, and may not reflect the frequency-
intensity range of inhibitory synaptic inputs per se. By deriving excitatory and
inhibitory synaptic conductances, we demonstrated that in nonmonotonic neurons
inhibitory inputs always follow excitatory inputs evoked by the same tone stimulus,
indicating that a pure inhibition domain in the TRF is not necessary for generating
intensity tuning. Nevertheless, the nonmonotonic tuning of excitatory inputs and the
monotonic tuning of inhibitory inputs agree with the negatively correlated excitation
and inhibition suggested in previous studies. Since both intensity-tuned and untuned

104
cortical neurons have monotonic inhibition, the intensity tuning is thus primarily
determined by nonmonotonic excitatory inputs.

4.4.3. Temporal Shaping of Intensity Tuning by Inhibitory Inputs
The proposed lateral inhibition underlying the intensity tuning may reflect
in part differential temporal interaction between excitatory and inhibitory synaptic
inputs at various intensities. The temporal shaping effect of inhibitory inputs has
recently been suggested in an intracellular study, which examined tone-evoked
membrane depolarization and hyperpolarization, and showed that the onset latency
of hyperpolarization became shorter as sound intensity increased, resulting in a
shortening of the duration of the preceding depolarization (Ojima and Murakami,
2002). Because the membrane potential response is determined by the temporal
integration of both excitatory and inhibitory inputs, the property of each input cannot
be directly inferred, e.g. the onset of hyperpolarization cannot be simply considered
as the onset of inhibitory inputs. By dissecting tone-evoked pure excitatory and
inhibitory synaptic conductances in well defined intensity-tuned neurons, our
findings extended the previous observations. Instead of simply scaling down the
membrane potential responses (Wehr and Zador, 2003; Tan et al., 2004), inhibitory
inputs have significantly sharpened intensity tuning, an effect that can be largely
attributed to a nonmonotonic tuning of the relative delay of these inputs.
How does nonmonotonic tuning of inhibitory delay arise? Although the
absolute onset latencies of both excitatory and inhibitory inputs decrease with

105
intensity increments in both intensity-tuned and untuned neurons (Heil, 2004;
Figures 4.4A and 4.4B), the relative delay of inhibitory inputs becomes shorter at
high intensities only in intensity-tuned neurons. We speculate that this change of
temporal delay can be partially attributed to an intrinsic synaptic mechanism, i.e. the
integration time to bring the neuron to firing threshold is determined by the slope and
amplitude of the evoked excitatory postsynaptic potential (PSP) according to the
integrate-and-fire model. In our recorded neurons, stronger PSPs are accompanied
by stronger excitatory conductances with steeper slopes (Figures 4.4A and 4.4B),
indicating that a monotonic increase in PSPs will result in a monotonic reduction in
the integration time for the initial spike, and vice versa. If this holds for earlier
stages of the auditory pathway, the nonmonotonic excitatory inputs to the cortical
intensity-tuned neurons will inherit a longer integration time at high intensities, and
the monotonic inhibitory inputs will inherit a shorter integration time instead,
resulting in a shortening of the relative time interval between the two inputs. It is
possible that the latter effect alone could be exploited by a neural network to convert
a monotonic tuning curve to nonmonotonic. The implication of our results is that by
controlling the relative timing of excitation and inhibition, synaptic circuits can
achieve a de novo construction of representational properties.

4.4.4. Potential Synaptic Circuits underlying Nonmonotonic Neurons
Synaptic circuits underlying nonmonotonic cortical neurons appear to be
different from that of monotonic cells. Synaptic TRFs of A1 monotonic neurons are

106
marked by covaried tone-evoked excitatory and inhibitory synaptic inputs (Zhang et
al., 2003; Wehr and Zador, 2003; Tan et al., 2004). This suggests a feed-forward
inhibition circuit associated with A1 neurons, in which the presynaptic GABAergic
neurons may be innervated by the same set of thalamocortical afferents as the
recorded A1 cell, similar as previously proposed for other sensory cortices (Miller et
al., 2001). In the present study, as recordings were made in the major thalamo-
recipient layers of the auditory cortex, the non-covaried excitatory and inhibitory
inputs in the intensity-tuned neurons suggest that the inhibitory inputs are from
cortical GABAergic neurons innervated by a group of monotonic neurons in the
thalamus, whereas excitatory inputs are likely mainly from thalamic nonmonotonic
neurons. The latter is supported by the existence of nonmonotonic neurons in the
medial geniculate body (Aitkin and Webster, 1972), and by the evidence that CF tone
evoked excitation can be largely attributed to thalamocortical inputs (Cruikshank et
al., 2002; Kaur et al., 2004).
Finally, in this study, we followed previous studies (Schreiner et al., 1992;
Phillips et al., 1995; Heil and Irvine, 1998; Polley et al., 2004) in characterizing
responses to pure tones of various intensities at the CFs of the recorded neurons.
Under more complex stimuli, e.g. broadband noise, the spike response-intensity
function of intensity-tuned neurons will be complicated by the spectrotemporal
integration of synaptic inputs activated at different frequencies and time points.
Understanding of intensity tuning under these stimuli will require knowledge on

107
more details of synaptic circuits, e.g. the origins and properties of presynaptic
neurons for each synaptic input, that underlie cortical neurons’ responses.

4.5. Summary
Intensity-tuned neurons, characterized by their nonmonotonic response-
level function, may play important roles in the encoding of sound intensity-related
information. The synaptic mechanisms underlying intensity tuning remain yet
unclear. Here, in vivo whole-cell recordings in rat auditory cortex revealed that
intensity-tuned neurons, mostly clustered in a posterior zone, receive imbalanced
tone-evoked excitatory and inhibitory synaptic inputs. Excitatory inputs exhibit
nonmonotonic intensity-tuning, whereas with tone intensity increments, the
temporally-delayed inhibitory inputs increase monotonically in strength. In addition,
this delay reduces with the increase of intensity, resulting in an enhanced suppression
of excitation at high intensities and a significant sharpening of intensity tuning. In
contrast, non-intensity-tuned neurons exhibit covaried excitatory and inhibitory
inputs and the relative time interval between them is stable with intensity increments,
resulting in monotonic response-level function. Thus, cortical intensity tuning is
primarily determined by excitatory inputs, and shaped by cortical inhibition through
a dynamic control of excitatory and inhibitory timing.

108
Chapter 4 Endnote

The work presented in this chapter appeared in the following publication:
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.

109
Chapter 5
How Is Stimulus Duration Coded?

5.1. Background and Introduction
Other than frequency and intensity information, temporal information is
also an important component of sound. To study how temporal information is
processed, it is necessary to start with its fundamental features first. The duration of
a sound is a biologically important feature. Previous studies show animals perceive
the same sounds differently depending on the duration (Repp et al., 1978). It also
suggests sound duration might not simply represented by the neurons’ responses
throughout the sound, since most auditory neurons above the brainstem respond only
transiently (Pollak and Casseday, 1989). Along the auditory pathway, duration-tuned
neurons have been described in inferior colliculus (IC) and auditory cortex (Potter,
1965; Feng et al., 1990; Pinheiro et al., 1991; He et al., 1997; Casseday et al., 1994).
Those duration-tuned neurons respond to their ‘best duration’ with the maximum
number of spikes. The finding of duration-tuned neurons suggests that the duration
tuning is constructed through the interacttion of excitation and inhibition in IC
(Casseday et al., 1994). Other studies demonstrate that many neurons in the cortex
respond with long latency and are selective for long or short sound duration (He et
al., 1997).

110
Although duration-tuned neurons are prominent in the auditory nuclei above
brainstem, there’s no sufficient evidence ruling out the possibility that duration is
represented or encoded at an earlier stage. The axons of the principal neurons entered
the dorsal acoustic stria and eventually projected to the contralateral inferior
colliculus (Osen, 1972; Ryugo et al., 1981). It will be no surprise that the duration-
tuned neurons in IC might inherit duration information from the brain stem. Previous
studies are only focused on the duration that is encoded by firing rate. However, in
the brain stem, many neurons have sustained firing pattern in responding to
stimulation (Young and Brownell, 1979; Smith, 1989). The neural representation of
sound duration can still be achieved by the sustained firing of those neurons
responding throughout the sound. Moreover, earlier psychophysical experiments
showed that sound offset serves as an important acoustic cue in some phenomena
such as perception of sound duration (Schlauch et al., 2001). The presence of cortical
responses to sound offset was clearly demonstrated by the auditory-evoked potential
(Takahashi et al., 2004), neuromagnetic response (Gutschalk et al., 2002; Hamada et
al., 2004; Hari et al., 1987), and functional magnetic resonance imaging activation
(Harms et al., 2005; Okada et al., 2004). In brain stem, auditory on and off-potentials
has been found by measuring brainstem auditory evoked potentials (BAEP) in
responding to noise or tone stimulation (Brinkmann and Scherg, 1979; Henry, 1985),
or by electrode recording (Pérez-Abalo, 1988). The general waveform and interpeak
latencies suggest this offset response is generated within the cochlea and auditory

111
brainstem (Henry, 1985b). All the above evidences suggest offset responses to the
stimulation might be a more universal candidate for duration coding.
In a recent study, the comparison of offset and onset responses of primiary
auditory cortical neurons in awake cats suggests offset responses are precise and
salient for effectively encoding sound offsets and the offset responses may be elicited
as active spike responses to sound offset rather than simple rebound facilitation (Qin
et al., 2007). In the present project, using in vivo whole-cell voltage-clamp recording
technique, we examined the excitatory and inhibitory synaptic inputs in two distinct
classes of dorsal cochlear nuclei (DCN) neurons, sustained type and phasic type. We
quantified the amplitude and temporal relationship between the excitatory and
inhibitory inputs evoked either by tone stimuli of various durations at characteristic
frequencies (CFs) of the cells or by white noise with various durations. Our data
indicate that phasic type and sustained type neurons have both sustained excitatory
and inhibitory inputs. But phasic type neuron received imbalanced strong inhibition,
while sustained type neuron might receive comparable inhibition and excitation.
Moreover, the hyperpolarization caused by the huge inhibtion might activate
hyperpolarization-activated channels in the DCN neuron, which can elicit the offset
spike responses for the phasic neurons. The reliability and precision of offset
responses of phasic neurons, together with the linearity of spike number to sound
duration from sustained neurons, suggest that duration coding might not rely on a
specialized group of duration-tuned neurons, and duration representation might be
more universal.
112
5.2. Methods and Materials
5.2.1. Extracellular Recording
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-proof booth (Acoustic Systems). Female
Sprague-Dawley rats about 3 months old and weighing 250–300 g were
anaesthetized with ketamine and xylazine (ketamine: 45 mg/kg; xylazine: 6.4 mg/kg;
i.p.). The rat ear canal was plugged. The left brain stem was exposed after the left
hemisphere of cerebellum was removed by vacuum suction. The body temperature
was maintained at 37.5ºC by a feed-back heating system (Harvard Apparatus, MA).
Multiunit spike responses were recorded with parylene-coated tungsten
microelectrodes (FHC, ME) at 0–1200 μm below the pial. Electrode signals were
amplified (Plexon Inc, TX), band-pass filtered between 300 and 6,000 Hz and then
thresholded in a custom-made software (LabView, National Instrument) to extract
the spike times. Pure tones (0.5–64 kHz at 0.1-octave intervals, 25-ms duration, 3
ms ramp) at eight 10-dB-spaced sound intensities were delivered through a calibrated
free-field speaker facing the left ear. A time window from 3-30 ms from the onset of
tone stimulus was used for tone-evoked spike responses. The threshold of the spike
TRF was chosen to be the minimum stimulus intensity included in the TRF, and the
characteristic frequency (CF) was the tone frequency that evoked a response at
threshold. Bandwidth at 10 dB or 30 dB (BW10 or BW30) above threshold was the
frequency width (in octaves) of the TRF at that intensity level. Auditory nuclei
113
mapping was carried out by sequentially recording from an array of sites in 3-
dimension.

5.2.2. In vivo Whole-cell Recording
Whole-cell recordings (Moore and Nelson, 1998; Margrie et al., 2002;
Zhang et al., 2003; Wehr and Zador, 2003; Tan et al., 2004) were obtained from
neurons located 200–500 m beneath the brain stem surface. We prevented pulsation
with 4% agarose. For voltage-clamp recording, the patch pipette (6-9 M )
contained (in mM): 125 Cs-gluconate, 5 TEA-Cl, 4 MgATP, 0.3 GTP, 10
phosphocreatine, 10 HEPES, 0.5 EGTA, 2 CsCl, pH 7.2. 5mM QX-314 was
included to improve the whole-cell clamping of the cell (Nelson et al., 1994).
Recordings were made with an Axopatch 200B amplifier (Axon Instruments). The
whole cell recording method under our conditions shows a sampling bias towards
relatively large fusiform or giant neurons (data not shown). The whole-cell and
pipette capacitances were completely compensated and the initial series resistance
(20 50M ) was compensated for 50-60% to achieve effective series resistances of
10-25 M . Signals were filtered at 5 kHz and sampled at 10 kHz. Only neurons
with reasonable resting membrane potentials and stable series resistance (less than
10% change from the beginning of the recording) were used for further analysis.
The CFs of the synaptic TRFs of recorded neurons matched their positions in the
tonotopic map determined by extracellular recordings.

114
5.2.3. Data Analysis
The excitatory synaptic conductance Ge(t) and inhibitory synaptic
conductance Gi(t) at time t were derived (Borg-Graham et al., 1998, Anderson et al.,
2000) using I(t, V) = Gr(V-Er)+Ge(t)(V-Ee)+Gi(t)(V-Ei) where V is the clamping
voltage, Gr is the resting conductance, Er is the resting potential; Ee and Ei are the
reversal potentials for excitatory and inhibitory synaptic currents, respectively; and
I(t, V) is the current amplitude under V. Currents into the neuron were assigned a
negative value. The resting or leak conductance Gr was derived using Ir(V) = Gr(V –
Er) where Er is the resting potential, and Ir(V) is the resting current. Measurement
of I(V) at two voltages will solve the value of Ge and Gi in the equation. In this
study, a corrected clamping voltage V was used, instead of the clamping voltage
applied (Vc). V(t) is given by V(t) = Vc – Rs*I(t), where Rs was the effective series
resistance. Synaptic currents were obtained with the cell clamped at the reversal
potentials for inhibitory and excitatory currents respectively, for each of the 568 test
tone stimuli. For some of the experiments, the reversal potentials of glutamatergic
and GABAergic (Cl ) currents were roughly measured at the beginning by
examining the reversal of spontaneous glutamatergic and GABAergic currents
respectively, as the holding potential was changed. Under experimental condition in
this study, the reversal potential was found to be 0-8 mV for glutamatergic inputs,
and around -70 mV for GABAergic inputs, consistent with the values of Ee and Ei
determined by considering the ionic composition of the pipette solution and the

115
cerebrospinal fluid. In some cases, Ei values of -65 and -75 mV were also tested,
and this did not change the conclusion of the study.
Estimated membrane potential response Vest was derived from synaptic
conductances using

where Er is the resting membrane potential, which was determined for each recorded
neuron under current-clamp recording at the beginning of the experiment (Wehr and
Zador, 2003). If only the excitatory synaptic conductance was taken into account,
Gi(t) was set to zero. The spike threshold was set around -45 mV for auditory
cortical neurons, an observation from a previous study (Tan et al., 2004).
In this study, we have assumed linear, isopotential neurons in deriving
excitatory and inhibitory synaptic conductances, same as in previous studies (Zhang
et al., 2003; Wehr and Zador, 2003 and 2005; Tan et al., 2004). However, deviations
due to space clamp error and cable attenuation for synaptic inputs at the distal
dendrites (Spruston et al., 1993) should be kept in mind, as extensively discussed in
several recent studies (Wehr and Zador, 2003; Tan et al., 2004). However, the
linearity of synaptic IV curves suggested that synaptic conductances were not
strongly affected by nonlinearities of cortical neurons (data not shown). This may be
attributed to the use of intracellular cesium, TEA, QX-314 and ketamine anaesthesia,
which together block most voltage-dependent currents. In addition, the use of
cesium and TEA (blockers of potassium channels) will reduce the membrane
116
permeability and thus decrease the cable attenuation (Spruston et al., 1993).
Secondly, simulation based on a compartmental neuron model suggests that the
relative timing between excitatory and inhibitory inputs at 50% peak is not
significantly affected by the cable effects (Wehr and Zador, 2003).

Figure 5.1 Tonotopic organization in the rat dorsal cochlear nuclei
(A) A view of coronal section of rat brain. Color area indicates the dorsal cochlear nuclei. The color
gradient represents characteristic frequency (CF). Blue indicates low CF, Red indicates high CF. D,
dorsal ; V, ventral.
(B) A view of sagittal section. Representations are similar to (A)
(C) A view of horizontal sections. Representations are similar to (A)

117
5.3. Results
5.3.1. Tonotopic Organization of Dorsal Cochlear Nuclei
To precisely locate the dorsal cochlear nuclei and effectively investigate
response properties of neurons in the dorsal cochlear nuclei, we first determined the
tonotopic organization of dorsal cochlear nuclei in the adult rat brain stem by
mappings with multiunit extracellular recordings (100-120 sampling sites for each
map, see Methods and Materials). The frequency-intensity tonal receptive field for
spike responses (spike TRF) was reconstructed for each recorded sampling site. The
tonotopic organization is represented according to the recording site’s characteristic
frequency (CF), which is the frequency the neuron is most sensitive to. As shown in
an example DCN auditory map (Figure 5.1), the tonotopic organization of frequency
representations is in three dimensional: the high-CF-to-low-CF axis are from dorsal
to ventral (Figure 5.1A), caudal to rostral (Figure 5.1B), and medial to lateral (Figure
5.1C). In the depth of 900-1000 μm, a boundary separating dorsal cochlear nuclei
and ventral cochlear nuclei was distinguished by a sudden jump of sampling site’s
CF from low to high (data not shown).

5.3.2. Two Types of Neuronal Responses to Sound Duration
Within the dorsal cochlear nuclei, in vivo cell-attached recording (loose
patch recording) was applied. The spike responses from a single neuron were
recorded. After getting the neuron’s tonal spiking receptive field, its CF was defined.
Then either white noise or CF tones with various durations were delivered as
118
stimulation. 25 ms, 50 ms, 100 ms and 200 ms sound stimulation were used, in some
cases, longer durations were also applied, e.g., 250 ms and 300 ms. Loose patch
recordings revealed two distinctive types of neuron with different firing pattern. For
sustained type neuron, the neurons’ firing was sustained through the duration of
sound stimulation (Figure 5.2A). Another type, ‘phasic type’, those neurons only
respond to the onset and/or offset of the sound stimulation (Figure 5.2B). Among all
the 22 cells, 12 neurons showed sustained type (54.5%), 6 showed phasic type
(27.3%), 5 showed no evoked responses and spontaneous activities only (18.2%).
For a typical phasic type in Figure 5.3, the raster plotting and PSTH showed
a clear offset responses and a reduction of spike activities following onset and offset
responses. Especially, the suppression area following the offset responses last more
than 100 ms. The onset spike seemed less reliable in comparison with the offset
spike. For 51 testing trials with sound duration of 25 ms, the chance of onset spike
was only 25.4%, while that of offset spike was 94.1% (Figure 5.3A). The chance of
onset spike in responding to longer durations were even smaller (25.4% for 50 ms
duration; 23.5% for 100 ms duration; 13.7% for 200 ms duration; Figure 5.3B, C, D).
The time variance of the offset responses (last spike before the suppression) was also
extremely low (STD=1.44 ms for 25 ms duration). The results suggest that the offset
responses from phasic type neurons might be reliably and accurately represent the
offset of the sound, which is important for representing stimulation duration.
As for sustained type neurons, typically, they fired a train of spikes during
the duration of sound stimulation (Figure 5.4). They were not like phasic type neuron
119
that only generated one or two spikes at the onset and/or the offset of the stimulation.
The firing rate for the stimulation with 25 ms duration was 3.8 spikes per stimulus,
5.7 spikes per stimulus for 50 ms duration, 11.3 spikes per stimulus, 21.8 spikes per
stimulus (Figure 5.4A, B, C, D). The linearity of duration v.s. spike number suggest
that the stimulation duration might be represented by the total spike number in this
type of neuron (data not shown). Moreover, sustained type neuron might be able to
convey more information to the afferent auditory pathway.

Figure 5.2 Two distinct types of firing patthern of DCN neurons
(A) Sustained firing pattern. Black lines under the response traces indicate sound stimulation.
(B) Onset and/or offset firing pattern. Black lines under the response traces indicate sound
stimulation.

120

Figure 5.3 Raster of the onset/offset type neuron
(A) Upper panel: firing pattern in responding to 25 ms sound stimulation, in 51 trials; lower panel:
PSTH of the spike responses, black bar indicates sound stimulation.
(B) Same as (A) except in responding to 50 ms sound stimulation.
(C) Same as (A) except in responding to 100 ms sound stimulation.
(D) Same as (A) except in responding to 200 ms sound stimulation.

121

Figure 5.4 Raster of the onset/offset type neuron
(A) Upper panel: firing pattern in responding to 25 ms sound stimulation, in 51 trials; lower panel:
PSTH of the spike responses, black bar indicates sound stimulation.
(B) Same as (A) except in responding to 50 ms sound stimulation.
(C) Same as (A) except in responding to 100 ms sound stimulation.
(D) Same as (A) except in responding to 200 ms sound stimulation.

122
5.3.3. Synaptic Mechanisms underlying Offset Responses
Since the auditory nerve shows sustained firing pattern, the sustained type
neuron might lack inhibition or only receive weak inhibition, while the response
properties of phasic type neurons must be shaped by the local circuits in DCN. To
examine how both types of response patterns were generated, we applied in vivo
whole-cell voltage-clamp recordings (see Methods and Materails) to neurons in
dorsal cochlear nuclei. By voltage-clamping the cell’s membrane potential at -70
mV and 0 mV, the reversal potentials for glycine receptor-mediated Cl currents and
glutamate receptor-mediated excitatory currents respectively, we obtained TRFs for
both excitatory and inhibitory inputs in the recorded cell (data not shown). Then the
CF for this cell was determined. Either white noise or CF tones with various
durations were delivered as stimulation. 25 ms, 50 ms, 100 ms and 200 ms sound
stimulation were used. Synaptic inputs for two example cells in the dorsal cochlear
nuclei were shown in Figures 5.5 and 5.6. Linear current-voltage relationship (I-V
curve) was observed for the recorded synaptic currents at 70 dB SPL (data not
shown). These data suggest that under our voltage-clamp recording conditions, those
synaptic inputs that contributed to the recorded tone-evoked currents were detected
with a reasonable accuracy (see Methods and Materials). The excitatory and
inhibitory synaptic inputs obtained from voltage-clamp experiments provide a basis
for determining the synaptic mechanisms underlying duration representation in the
dorsal cochlear nuclei.

123
In responding to the sound, the DCN neuron showed large inhibitory
synaptic inputs in contrast to the smaller excitatory synaptic inputs (Figure 5.5 A).
Both inhibitory and excitatory inputs demonstrated sustained inputs during the sound
stimulation (Figure 5.5 C). Such sustained inputs were not observed in the auditory
cortex in which only transient responses were seen (Figure 5.7). We derived the
neuron’s membrane potential changes according to its excitatory and inhibitory
conductances (Figure 5.5B, D). The membrane potentials for both durations showed
one major depolarization phase at the onset of the stimulation, and one at the offset
of the stimulation. It suggests the dynamics between excitation and inhibition, either
in input strengths or in temporal, might create such phasic responses in the principal
neurons in the dorsal cochlear nuclei.

Figure 5.5 Synaptic inputs and membrane potential changes of an example DCN neuron
(A) (C) Synaptic inputs evoked by a 25ms tone (A) and a 100 ms tone (C)stimulations. Blue,
excitatory conductance; Red inhibitory conductance;
(B) (D) Derived membrane potential changes by integrating synaptic inputs in (A) and (C)
respectvely.

124

Figure 5.6 Synaptic inputs and membrane potential changes of another DCN neuron
(A) (B) (C) (D) The representation is similar to Figure 5.5.

Figure 5.7 Synaptic inputs to an auditory cortical neuron
(A) Upper panel: inhibitory synaptic currents evoked by a 25 ms tone stimulation; lower panel:
excitatory synaptic currents evoked by the same duration stimulation. Green bar, sound stimulation.
(B) Same as (A), except for a 100 ms sound stimulation.

125
5.4. Discussion
5.4.1. How Sound Duration Is Encoded in Auditory System?
Sound duration is a signature of biological importance. Many studies have
made efforts to investigate how it is represented in the nervous system. One
possibility is that the neural representation of sound duration can still be achieved by
the sustained firing of those neurons responding throughout the sound. By this
strategy, neurons can convey maximal information to the next level without losing
duration information. Our data demonstrated a type of neurons with sustained firing
pattern. For those neurons, the spike number is linearly changing with the increase of
the sound duration. That suggests neurons might use the simplest way, spike number,
to represent the fundamental component of a sound: duration. Another mechanism is
through a specialized type of ‘duration-tuned’ neurons. Those duration-tuned
neurons respond to their ‘best duration’ with the maximum number of spikes.
Neurons tuned to sound duration were first found in the midbrain of the frog and
later in the bat inferior colliculus. (Potter, 1965; Feng et al., 1990; Pinheiro et al.,
1991; Casseday et al.,1994). Long-duration-selective neurons and short-duration-
selective neurons were also found in the cat’s auditory cortex (He et al., 1997). The
finding of duration-tuned neurons suggests that the duration tuning is construted
through the interatction of excitation and inhibition in IC (Casseday, 1994).
However, duration-tuned neurons normally have a broad tuning curve and have large
trial-to-trail variation, which undermines the cell’s ability to accurately represent
duration information (He et al., 1997). Earlier psychophysical experiments showed
126
that sound offset serves as an important acoustic cue in some phenomena such as
perception of sound duration (Schlauch et al. 2001). Our data shows the neurons in
dorsal cochlear nuclei have both sustained firing pattern and on-off pattern. For
phasic type neuron, they respond to the sound offset with reliable and precise
spiking. This kind of response pattern might be more efficient and accurate to reflect
and extract sound duration information. Although duration-tuned neurons were not
found in cochlear nuclei (Casseday, 1994), that does not mean cochlear nuclei
neurons cannot represent sound duration in other format.

5.4.2. Phasic Neurons Responds to the Sound Offset with Less Variability
Offset responses have been frequently reported previously in the midbrain,
the thalamus, and in the cortex in various species of mammals (Aitkin and Prain,
1974; Calford and Webster, 1981; Rhode and Smith, 1986; Bordi and LeDoux, 1994;
Grothe, 1994). Some of those studies in awake animals described the presence of
offset responses in addition to onset responses with an incidence of 10–30% of
recorded neurons (Brugge and Merzenich 1973; Chimoto et al. 2002; Evans and
Whitfield 1964; Pfingst and O’Connor 1981; Recanzone 2000). It has been
suggested that the Offset response may be formed by a rebound after the offset of
inhibitory input (Calford and Webster, 1981), and that duration-tuned Offset
responding neurons may take advantage of this rebound to create a coincidence
mechanism that in turn produces duration tuning (Casseday et al., 1994). But our
data demonstrated that inhibitory and excitatory inputs are both sustained.
127
Especially, the inhibitory inputs are normally longer lasting compared with
excitatory inputs. So the rebound after the offset of inhibitory input is ruled out for
generating off responses. From our data, it implies that the imbalance of dynamics
between excitatory and inhibitory inputs might underlie the generation of off
responses because the decay time constant of excitation is larger than that of
inhibition.
Another possibility of generating offset responses may rely on the
hyperpolarization-activated cation channels (HCV channels). Studies show several
subtypes of HCV channels express in dorsal cochlear principal neurons (Pal et al.,
2003). Since there’s a hyperpolarization following the onset depolarization, the
hyperpolarization phase might open such kind of HCV channels to regulate offset
membrane depolarization. Future experiments with blockades of HCV channels will
provide intuitive information about whether they are involved in the generation of
offset responses so as to modulate duration processing.

5.4.3. Firing of Sustained Neurons Correlates the Sound Duration
Sustained type neurons have been arguable through the years about whether
they are related to duration coding. One claim is they are not possible because in the
auditory nuclei above brain stem, most neurons respond to sound transiently (Pollak
and Casseday, 1989). However, neurons in brain stem might still use sustained firing
to convey duration information. Recent studies have shown there are neurons with
sustained firing pattern in higher auditory nuclei in awake animals (Qin et al., 2007;
128
Barbour and Wang, 2003 a,b; Brugge and Merzenich, 1973; Chimoto et al., 2002;
Evans and Whitfield, 1964; Pfingst and O’Connor, 1981; Qin et al., 2003;
Recanzone, 2000; Shamma and Symmes, 1985; Volkov et al., 1985). Our data
suggests the spike number is linearly correlated with the sound duration. But what’s
the synaptic mechanisms underlying the sustained firing is still mysterious. The
future experiment will be required to get the recordings that can show a sustained
firing pattern or sustained depolarization after the integration of excitatory and
inhibitory inputs. Those sustained type neurons might receive weaker inhibitory
inputs instead of lack of inhibition.

5.5. Summary
Duration is a biologically important feature of sound. The present study
examined the neurons in the dorsal cochlear nuclei, among which the spiking
responses and synaptic responses depended on the duration of noise or tone
stimulation. Two types of neurons have been revealed according to their distinct
firing pattern: sustained type and phasic type. Both types might be underlying the
duration coding. However, the synaptic mechanisms underlying those two types of
response properties are not clear. In my project, in vivo whole cell recordings have
been applied to those dorsal cochlear nuclei neurons. The onset and offset responses
of phasic type neurons might be due to the imbalance of the dynamics between
synaptic excitation and inhibition. While the mechanism underlying sustained firing
pattern is left for the future investigation.
129 Reference
Agmon, A., and Connors, B.W. (1992). Correlation between intrinsic firing patterns
and thalamocortical synaptic responses of neurons in mouse barrel cortex. J.
Neurosci. 12, 319-329.
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.
Aitkin, L.M., Prain, S.M. (1974). Medial geniculate body: unit responses in the
awake cat. J. Neurophysiol. 37, 512-521.
Alonso, J.M. & Swadlow, H.A. Thalamocortical specificity and the synthesis of
sensory cortical receptive fields. J. Neurophysiol. 94, 26-32 (2005).
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.
Azouz, R., Gray, C.M., Nowak, L.G., and McCormick, D.A. (1997). Physiological
properties of inhibitory interneurons in cat striate cortex. Cereb. Cortex. 7, 534-545.
Bao, S., Chang, E.F., Davis, J.D., Gobeske, K.T., and Merzenich, M.M. (2003).
Progressive degradation and subsequent refinement of acoustic representations in the
adult auditory cortex. J. Neurosci. 23,10765-10775.
Barbour, D.L., and Wang, X. (2003a). Auditory cortical responses elicited in awake
primates by random spectrum stimuli. J. Neurosci. 23, 7194-7206.
130 Barbour, D.L., and Wang, X. (2003b). Contrast tuning in auditory cortex. Science
299, 1073-1075.
Bendor, D.A., and X. Wang. (2005). The neuronal representation of pitch in primate
auditory cortex. Nature 436, 1161-1165.
Bordi, F., and LeDoux, J.E. (1994). Response properties of single units in areas of rat
auditory thalamus that project to the amygdala. I. Acoustic discharge patterns and
frequency receptive fields. Exp. Brain Res. 98:, 261-274.
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.
Bringuier, V., Chavane, F., Glaeser, L. and Fregnac, Y. (1999). Horizontal
propagation of visual activity in the synaptic integration field of area 17 neurons.
Science 283, 695-699.
Brinkmann, R.D., Scherg, M. (1979). Human auditory on- and off-potentials of the
brainstem. Scand Audiol. 8, 27-32.
Brudzynski, S.M., and Pniak, A. (2002). Social contacts and production of 50-kHz
short ultrasonic calls in adult rats. J. Comp. Psychol. 116, 73-82.
Brudzynski, S.M., Bihari, F., Ociepa, D., and Fu, X.W. (1993). Analysis of 22 kHz
ultrasonic vocalization in laboratory rats: long and short calls. Physiol. Behav. 54,
215-221.
Brugge, J.F., and Merzenich, M.M. (1973). Responses of neurons in auditory cortex
of the macaque monkey to monaural and binaural stimulation. J. Neurophysiol. 36,
1138-1158.
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.
131 Bruno, R.M., and Sakmann, B. (2006). Cortex is driven by weak but synchronously
active thalamocortical synapses. Science 312, 1622-1627.
Calford, M. B., and Webster, W. R. (1981). Auditory representation within principal
division of cat medial geniculate body: an electrophysiology study. J. Neurophysiol.
45, 1013-1028.
Calford, M.B., and Semple, M.N. (1995). Monaural inhibition in cat auditory cortex.
J. Neurophysiol. 73, 1876-1891.
Carandini, M., and Ferster, D. (2000). Membrane potential and firing rate in cat
primary visual cortex. J. Neurosci. 20, 470-484.
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.
Chen, Q.C., and Jen, P.H. (2000). Bicuculline application affects discharge patterns,
rate-intensity functions, and frequency tuning characteristics of bat auditory cortical
neurons. Hear. Res. 150,161-174.
Chimoto, S., Kitama, T., Qin, L., Sakayori, S., and Sato, Y. (2002). Tonal response
patterns of primary auditory cortex neurons in alert cats. Brain Res 934, 34-42.
Chung, S., and Ferster, D. (1998). Strength and orientation tuning of the thalamic
input to simple cells revealed by electrically evoked cortical suppression. Neuron 20,
1177-1189.
Crook, J.M., Kisvarday, Z.F., and Eysel, U.T. (1997). GABA-induced inactivation of
functionally characterized sites in cat striate cortex: effects on orientation tuning and
direction selectivity. Vis. Neurosci. 14, 141-158.
Cruikshank, S.J., Lewis, T.J., and Connors, B.W. (2007). Synaptic basis for intense
thalamocortical activation of feedforward inhibitory cells in neocortex. Nat.
Neurosci. 10, 462-468.
132 Cruikshank, S.J., Rose, H.J. and Metherate, R. (2002). Auditory thalamocortical
synaptic transmission in vitro. J. Neurophysiol 87, 361-384.
Davies, P.W., Erulkar, S.D., and Rose, J.E. (1956). Single unit activity in the
auditory cortex of the cat. Bull. Johns Hopkins Hosp. 99, 55-86.
Daw, M.I., Ashby, M.C., and Isaac, J.T. (2007). Coordinated developmental
recruitment of latent fast spiking interneurons in layer IV barrel cortex. Nat.
Neurosci. 10, 453-461.
Doron, N.N., Ledoux, J.E., and Semple, M.N. (2002). Redefining the tonotopic core
of rat auditory cortex: physiological evidence for a posterior field. J. Comp Neurol.
453, 345-360.
Douglas, R.J., and Martin, K.A. (1991). A functional microcircuit for cat visual
cortex. J. Physiol. 440, 735-69.
Douglas, R.J., Koch, C., Mahowald, M., Martin, K.A., and Suarez, H.H. (1995).
Recurrent excitation in neocortical circuits. Science 269, 981-985.
Durstewitz, D., and Sejnowski, T.J. (2000). Flexible functional connectivity in
working memory networks with non-monotonic neural response functions. Annual
Meeting of Society for Neuroscience, Program Abstract No. 711.1.
Evans, E.F., and Whitfield, I.C. (1964). Classification of unit responses in the
auditory cortex of the unanesthetized and unrestrained cat. J. Physiol. Lond. 171,
476-493.
Faingold, C.L, Boersma Anderson, C.A., and Caspary, D.M. (1991). Involvement of
GABA in acoustically-evoked inhibition in inferior colliculus neurons. Hear. Res.
52, 201-216.
Feng, A.S., Hall, J.C., and Goller, D.M. (1990). Neural basis of sound pattern
recognition in anurans. Prog. Neurobiol. 34, 313-329.
133 Ferster, D., Chung, S., and Wheat, H. (1996). Orientation selectivity of thalamic
input to simple cells of cat visual cortex. Nature 380, 249-252.
Fox, K., Wright, N., Wallace, H., and Glazewski, S. (2003). The origin of cortical
surround receptive fields studied in the barrel cortex. J. Neurosci. 23, 8380-8391.
Games, K.D., and Winer, J.A. (1988). Layer V in rat auditory cortex: projections to
the inferior colliculus and contralateral cortex. Hear. Res. 34, 1-25.
Ghosh, K., Kowal, D., Dawson, L.A., and Tasse, R. (1999). Design and models for
estimating antagonist potency (pA2, Kd and IC50) following the detection of
antagonism observed in the presence of intrinsic activity. Neuropharmacology 38,
361-373.
Gibson, J.R., Beierlein, M., and Connors, B.W. (1999). Two networks of electrically
coupled inhibitory neurons in neocortex. Nature 402, 75-79.
Gil, Z., and Amitai, Y. (1996). Properties of convergent thalamocortical and
intracortical synaptic potentials in single neurons of neocortex. J. Neurosci. 16,
6567–6578.
Gonchar, Y., and Burkhalter, A. (1997). Three distinct families of GABAergic
neurons in rat visual cortex. Cereb. Cortex. 7, 347-358.
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.
Gupta, A., Wang, Y., and Markram, H. (2000). Organizing principles for a diversity
of GABAergic interneurons and synapses in the neocortex. Science 287, 273-278.
134 Gutschalk, A., Patterson, R.D., Rupp, A., Uppenkamp, S., and Scherg, M. (2002).
Sustained magnetic fields reveal separate sites for sound level and temporal
regularity in human auditory cortex. Neuroimage 15, 207-216.
Hamada, T., Iwaki, S., and Kawano, T. (2004). Speech offsets activate the right
parietal cortex. Hear Res 195, 75-78.
Hari, R., Pelizzone, M., Makela, J.P., Hallstrom, J., Leinonen, L., and Lounasmaa,
O.V. (1987). Neuromagnetic responses of the human auditory cortex to on- and
offsets of noise bursts. Audiology 26, 31-43.
Harms, M.P., Guinan, J.J. Jr., Sigalovsky, I.S., and Melcher, J.R. (2005). Short-term
sound temporal envelope characteristics determine multisecond time patterns of
activity in human auditory cortex as shown by fMRI. J Neurophysiol 93, 210-222.
Hashimoto, H., Moritani, N., Aoki-Komori, S., Tanaka, M., and Saito, T.R. (2004)
Comparison of ultrasonic vocalizations emitted by rodent pups. Exp. Anim. 53, 409-
416.
He, J., Hashikawa, T., Ojima, H., and Kinouchi, Y. (1997). Temporal integration and
duration tuning in the dorsal zone of cat auditory cortex. J. Neurosci. 17, 2615-2625
Heil, P. (2004). First-spike latency of auditory neurons revisited. Curr. Opin.
Neurobiol. 14, 461-467.
Heil, P., and Irvine, D.R. (1998). The posterior field P of cat auditory cortex: coding
of envelope transients. Cereb. Cortex 8, 125-141.
Heil, P., Rajan, R., Irvine, D.R. (1994). Topographic representation of tone intensity
along the isofrequency axis of cat primary auditory cortex. Hear Res 76,188-202
Hendry, S.H., Schwark, H.D., Janes, E.G., and Yan, J. (1987). Numbers and
proportions of GABA-immunoreactive neurons in different areas of monkey cerebral
cortex. J. Neurosci. 7,1503-1519.
135 Henry, K.R. (1985). ON and OFF components of the auditory brainstem response
have different frequency- and intensity-specific properties. Hear Res. 18, 245-51.
Henry, KR. (1985). Tuning of the auditory brainstem OFF responses is
complementary to tuning of the auditory brainstem ON response. Hear Res. 19, 115-
125.
Hestrin, S., Nicoll, R.A., Perkel, D.J., and Sah, P. (1990). Analysis of excitatory
synaptic action in pyramidal cells using whole-cell recording from rat hippocampal
slices. J. Physiol 422, 203-225.
Higley, M.J., and Contreras, D. (2006). Balanced excitation and inhibition
determine spike timing during frequency adaptation. J. Neurosci. 26, 448-57.
Hines, M. NEURON--a program for simulation of nerve equations. In: Neural
Systems: Analysis and Modeling, edited by F. Eeckman. Norwell, MA: Kluwer,
1993, p. 127-136
Hirsch, J.A., Alonso, J.M., Reid, R.C., and Martinez, L.M. (1998). Synaptic
integration in striate cortical simple cells. J. Neurosci. 18, 9517-9528.
Hirsch, J.A., Martinez, L.M., Pillai, C., Alonso, J.M., Wang, Q., and Sommer, F.T.
(2003). Functionally distinct inhibitory neurons at the first stage of visual cortical
processing. Nat. Neurosci. 6, 1300-1308.
Horikawa, K., and Armstrong, W.E. (1988). A versatile means of intracellular
labelling: injection of biocytin and its detection with avidin conjugates. J. Neurosci.
Methods 25, 1-11.
Inoue, T., and Imoto, K. (2006). Feedforward inhibitory connections from multiple
thalamic cells to multiple regular-spiking cells in layer 4 of the somatosensory
cortex. J. Neurophysiol. 96, 1746-1754.
Jahr, C.E., and Stevens, C.F. (1990a). A quantitative description of NMDA receptor-
channel kinetic behavior. J. Neurosci. 10, 1830-1837.
136 Jahr, C.E., and Stevens, C.F. (1990b).Voltage dependence of NMDA-activated
macroscopic conductances predicted by single-channel kinetics. J. Neurosci. 10,
3178-3182.
Joshi, S., and Hawken, M. J. (2006). Loose-patch-juxtacellular recording in vivo-A
method for functional characterization and labeling of neurons in macaque V1. J.
Neurosci. Methods. 156, 37-49.
Kalatsky, V.A., Polley, D.B., Merzenich, M.M., Schreiner, C.E., and Stryker, M.P.
(2005). Fine functional organization of auditory cortex revealed by Fourier optical
imaging. Proc. Natl. Acad. Sci. USA. 102, 13325-13330.
Kaltwasser M.T. (1990) Acoustic signaling in the black rat (Rattus rattus). J. Comp.
Psychol. 104, 227-232.
Kaur, S., Lazar, R., and Metherate, R. (2004). Intracortical pathways determine
breadth of subthreshold frequency receptive fields in primary auditory cortex. J
Neurophysiol. 91, 2551-2567.
Kawaguchi, Y., and Kubota, Y. (1997). GABAergic cell subtypes and their synaptic
connections in rat frontal cortex. Cereb. Cortex. 7, 476-486.
Kemp, J.A., Marshall, G.R., and Woodruff, G.N. (1986). Quantitative evaluation of
the potencies of GABA receptor agonists and antagonists using the rat hippocampal
slice preparation. Br. J. Pharmacol. 87, 677-684.
Kiang, N.Y., Watanabe,T., Thomas, E.C., and Clark, L.F. (1965). Discharge patterns
of single fibers in the cat’s auditory nerve (Cambridge, MA: MIT Press).
Kilgard, M.P., and Merzenich, M.M. (1999). Distributed representation of spectral
and temporal information in rat primary auditory cortex. Hear Res. 134, 16-28.
Kitzes, L.M, and Hollrigel, G.S. (1996). Response properties of units in the posterior
auditory field deprived of input from the ipsilateral primary auditory cortex. Hear
Res. 100, 120-30.
137 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.
Kyriazi, H.T., Carvell, G.E., Brumberg, J.C., and Simons, D.J. (1996). Quantitative
effects of GABA and bicuculline methiodide on receptive field properties of neurons
in real and simulated whisker barrels. J Neurophysiol 75, 547-560.
Liu, B.H., Wu, G.K., Arbuckle, R., Tao, H.W., and Zhang, L.I. (2007). Defining
cortical frequency tuning with recurrent excitatory circuitry. Nat. Neurosci. 10,
1594-1600.
Margrie, T.W., Brecht, M., and Sakmann, B. (2002). In vivo, low-resistance, whole-
cell recordings from neurons in the anaesthetized and awake mammalian brain.
Pflugers Arch. 444, 491-498.
Marino, J., Schummers, J., Lyon, D.C., Schwabe, L., Beck, O., Wiesing, P.,
Obermayer, K., and Sur, M. (2005). Invariant computations in local cortical
networks with balanced excitation and inhibition. Nat. Neurosci. 8, 194-201.
Martinez, L.M., Wang, Q., Reid, R.C., Pillai, C., Alonso, J.M., Sommer, F.T., and
Hirsch, J.A. (2005). Receptive field structure varies with layer in the primary visual
cortex. Nat. Neurosci. 8, 372-379.
Mayer, M.L., Westbrook, G.L., and Guthrie, P.B. (1984). Voltage-dependent block
by Mg2+ of NMDA responses in spinal cord neurones. Nature 309, 261-263.
Miller, K.D., Pinto, D.J., and Simons, D.J. (2001). Processing in layer 4 of the
neocortical circuit: new insights from visual and somatosensory cortex. Curr. Opin.
Neurobiol. 11, 488-497.
Miller, L.M., Escabi, M.A., and Schreiner, C.E. (2001). Feature selectivity and
interneuronal cooperation in the thalamocortical system. J. Neurosci. 21, 8136-8144 .
Miller, L.M., Escabi, M.A., Read, H.L., and Schreiner, C.E. (2001). Functional
convergence of response properties in the auditory thalamocortical system. Neuron
32, 151-160.
138 Miller, L.M., Escabi, M.A., Read, H.L., and Schreiner, C.E. (2002). Spectrotemporal
receptive fields in the lemniscal auditory thalamus and cortex. J. Neurophysiol. 87,
516-527.
Moore, C. I., and Nelson, S. B. (1998). Spatio-temporal subthreshold receptive fields
in the vibrissa representation of rat primary somatosensory cortex. J. Neurophysiol.
80, 2882–2892.
Mountcastle, V.B., Talbot, W.H., Sakata, H., and Hyvarinen, J. (1969). Cortical
neuronal mechanisms in flutter-vibration studied in unanesthetized monkeys.
Neuronal periodicity and frequency discrimination. J. Neurophysiol. 32, 452-484.
Nelson, S., Toth, L., Sheth, B., and Sur. M. (1994). Orientation selectivity of cortical
neurons during intracellular blockade of inhibition. Science 265, 774-777.
Nowak, L., Bregestovski, P., Ascher, P., Herbet, A., and Prochiantz, A. (1984).
Magnesium gates glutamate-activated channels in mouse central neurones. Nature
307, 462-465.
Nyby, J., and Whitney, G. (1978). Ultrasonic communication of adult myomorph
rodents. Neuroscience and Biobehavioral Reviews 2, 1-14.
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.
Okada, T., Honda, M., Okamoto, J., and Sadato, N. (2004). Activation of the primary
and association auditory cortex by the transition of sound intensity: a new method for
functional examination of the auditory cortex in humans. Neurosci Lett 359, 119-
123.
Ong, J., Marino, V., Parker, D.A., Kerr, D.I., and Blythin, D.J. (1998). The
morpholino-acetic acid analogue Sch 50911 is a selective GABAb receptor
antagonist in rat neocortical slices. Eur. J. Pharmacol. 362, 35-41.
Osen, K.K. (1972). Projection of the cochlear nuclei on the inferior colliculus in the
cat. J. Comp. Neurol. 144, 355-372.
139 Oswald, A.M., Schiff, M.L., and Reyes, A.D. (2006). Synaptic mechanisms
underlying auditory processing. Curr Opin Neurobiol. 16, 371-376.
Otmakhova, N.A., Otmakhov, N., and Lisman, J.E. (2002). Pathway-specific
properties of AMPA and NMDA-mediated transmission in CA1 hippocampal
pyramidal cells. J. Neurosci. 22, 1199-1207.
Pal, B., Por, A., Szucs, G., Kovacs, I., and Rusznak, Z. (2003). HCN channels
contribute to the intrinsic activity of cochlear pyramidal cells. Cell. Mol. Life Sci.
60, 2189-2199.
Pérez-Abalo, M.C., Valdés-Sosa, M.J., Bobes, M.A., Galán, L., and Biscay, R.
(1988). Different functional properties of on and off components in auditory
brainstem responses to tone bursts. Audiology. 27, 249-59.
Peters, A., and Kara, D.A. (1985). The neuronal composition of area 17 of rat visual
cortex. II. The nonpyramidal cells. J Comp Neurol. 234, 242-63.
Pfingst, B.E., and O’Connor, T.A. (1981). Characteristics of neurons in auditory
cortex of monkeys performing a simple auditory task. J. Neurophysiol. 45, 16-34.
Phillips, D.P., and Kelly, J.B. (1989). Coding of tone-pulse amplitude by single
neurons in auditory cortex of albino rats (Rattus norvegicus). Hear. Res. 37, 269-279.
Phillips, D.P., and Orman, S.S. (1984) Responses of single neurons in posterior field
of cat auditory cortex to tonal stimulation. J. Neurophysiol. 51,147-63.
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.
Pinault, D. (1996). A novel single-cell staining procedure performed in vivo under
electrophysiological control: morpho-functional features of juxtacellularly labeled
thalamic cells and other central neurons with biocytin or Neurobiotin. J Neurosci
Methods. 65, 113-136.
140 Pinheiro, A.D., Wu, M., and Jen, P.H.S. (1991). Encoding repetition rate and
duration in the inferior colliculus of the big brown bat, Eptesicus fuscus. J Comp.
Physiol. A 169, 69–85.
Pollak, G.D., and Casseday, J.H. (1989). The Neural Basis of Echolocation in Bats
(Springer-Verlag, Berlin)
Pollak, G.D., and Park, T.J. (1993). The effects of GABAergic inhibition on
monaural response properties of neurons in the mustache bat's inferior colliculus.
Hear. Res. 65, 99-117
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. Proc. Natl. Acad. Sci. USA. 101, 16351-16356.
Polley, D.B., Read, H.L., Storace, D.A., and Merzenich, M.M. (2007).
Multiparametric auditory receptive field organization across five cortical fields in the
albino rat. J. Neurophysiol. 97, 3621-3638.
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.
Porter, J.T., and Nieves, D. (2004). Presynaptic GABAb receptors modulate thalamic
excitation of inhibitory and excitatory neurons in the mouse barrel cortex. J.
Neurophysiol. 92, 2762-2770.
Potter, H.D. (1965). Patterns of acoustically evoked discharges of neurons in the
mesencephalon of the bullfrog. J. Neurophysiol. 28, 1155-1184.
Priebe, N., and Ferster, D. (2005). Direction Selectivity of Excitation and Inhibition
in Simple Cells of the Cat Primary Visual Cortex. Neuron 45, 133-145.
Priet, J.J., Peterson, B.A., and Winer, J.A. (1994). Morphology and spatial
distribution of GABAergic neurons in cat primnary auditory cortex (A1). J. Comp.
Neurol. 334, 349 -382.
141 Purves, et al. (2004) Neuroscience. 3
rd
edition, Sinauer Associates, MA.
Qin, L., Chimoto, S., Sakai, M., Wang, J.Y., and Sata, Y. (2007). Comparison
Between offset and onset responses of primary auditory cortex on-off neurons in
awake cats. J. Neurophysiol. 97, 3421-3431.
Qin, L., Kitama, T., Chimoto, S., Sakayori, S., and Sato, Y. (2003). Time course of
tonal frequency-response-area of primary auditory cortex neurons in alert cats.
Neurosci. Res. 46, 145-152.
Read, H.L., Winer, J.A., and Schreiner, C.E. (2002). Functional architecture of
auditory cortex. Curr. Opin. Neurobiol. 12, 433-440.
Recanzone, G.H. (2000). Response profiles of auditory cortical neurons to tones and
noise in behaving macaque monkeys. Hear. Res 150, 104 -118.
Reid, R.C., and Alonso, J.M. (1995). Specificity of monosynaptic connections from
thalamus to visual cortex. Nature 378, 281-284.
Rennaker, R.L., Carey, H.L., Anderson, S.E., Sloan, A.M., and Kilgard, M.P. (2007).
Anesthesia suppresses nonsynchronous responses to repetitive broadband stimuli.
Neuroscience 145, 357-369.
Repp, B.H., Liberman, A.M., Eccardt, T., and Pesetsky, D. (1978). Perceptual
integration of acoustic cues for stop, fricative, and affricate manner. J. Exp. Psychol.
Hum. Percept. Perform 4, 621-637.
Rhode, W.S., and Smith, P.H. (1986). Encoding timing and intensity in the ventral
cochlear of the cat. J. Neurophysiol. 56, 261-286.
Roerig, B., and Chen, B. (2002). Relationships of Local Inhibitory and Excitatory
Circuits to Orientation preference maps in ferret visual cortex. Cereb. Cortex 12,
187-198.
142 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. Hear.
Res. 11, 235-247.
Ryugo, D.K., Willard, F.H., and Ekete, D.M. (1981). Differential afferent projections
to the IC from the cochelar nucleus in the albino mouse. Brain Res. 210, 342-349.
Sales, G. D., and Pye, D. (1974). Ultrasonic communication by animals. London:
Chapman and Hall.
Scharlock, D.P., Neff, W.D., and Strominger, N.L. (1965). Discrimination of tone
duration after bilateral ablation of cortical auditory areas. J. Neurophysiol. 28, 673-
681.
Schlauch, R.S., Ries, D.T., and DiGiovanni, J.J. (2001). Duration discrimination and
subjective duration for ramped and damped sounds. J. Acoust. Soc. Am. 109, 2880-
2887.
Schreiner, C.E., Mendelson, J.R., and Sutter, M.L. (1992). Functional topography of
cat primary auditory cortex: representation of tone intensity. Exp. Brain Res. 92,105-
122.
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.
J. Acoust. Soc. Am. 78, 1622-1632.
Shamma, S.A., and Symmes, D. (1985). Patterns of inhibition in auditory cortical
cells in awake squirrel monkeys. Hear. Res. 19, 1-13.
Shu, Y, Hasenstaub, A, and McCormick, D.A. (2003). Turning on and off recurrent
balanced cortical activity. Nature 423, 288-293.
Sillito, A.M. (1977). Inhibitory processes underlying the directional specificity of
simple, complex and hypercomplex cells in the cat's visual cortex. J. Physiol. 271,
699-720.
143 Sillito, A.M. (1979). Inhibitory mechanisms influencing complex cell orientation
selectivity and their modification at high resting discharge levels. J. Physiol. 289, 33-
53.
Simons, D.J., and Carvell, G.E. (1989). Thalamocortical response transformation in
the rat vibrissa/barrel system. J Neurophysiol. 61, 311-30.
Sivaramakrishnan, S., Sterbing-D'Angelo, S.J., Filipovic, B., D'Angelo, W.R.,
Oliver, D.L., and Kuwada, S. (2004). GABA( A) synapses shape neuronal responses
to sound intensity in the inferior colliculus. J. Neurosci. 24, 5031-5043.
Smith, P.H., and Rhode, W.S. (1989) Structural and functional properties distinguish
two types of multipolar cells in the ventral cochlear nucleus. J Comp Neurol. 282,
595-616.
Sohya, K., Kameyama, K., Yanagawa, Y., Obata, K., and Tsumoto, T. (2007).
GABAergic neurons are less selective to stimulus orientation than excitatory neurons
in layer II/III of visual cortex, as revealed by in vivo functional Ca2+ imaging in
transgenic mice. J. Neurosci. 27, 2145-2149.
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.
Spruston, N., Jaffe, D.B., Williams, S.H. and Johnston, D. (1993). Voltage- and
space-clamp errors associated with the measurement of electrotonically remote
synaptic events. J. Neurophysiol. 70, 781-802.
Suga, N., and Manabe, T. (1982). Neural basis of amplitude-spectrum representation
in auditory cortex of the mustached bat. J. Neurophysiol. 47, 225-255.
Sun, Q.Q., Huguenard, J.R., and Prince, D.A. (2006).Barrel cortex microcircuits:
thalamocortical feedforward inhibition in spiny stellate cells is mediated by a small
number of fastspiking interneurons. J. Neurosci. 26, 1219-1230.
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.
144 Sutter, M.L., and Schreiner, C.E. (1995). Topography of intensity tuning in cat
primary auditory cortex: single-neuron versus multiple-neuron recordings. J.
Neurophysiol. 73, 190-204.
Swadlow, H.A. (1989). Efferent neurons and suspected interneurons in S-1 vibrissa
cortex of the awake rabbit: receptive fields and axonal properties. J. Neurophysiol.
62, 288-308.
Swadlow, H.A. (2003). Fast-spike interneurons and feedforward inhibition in awake
sensory neocortex. Cereb. Cortex. 13, 25-32.
Swadlow, H.A., and Gusev, A.G. (2002). Receptive-field construction in cortical
inhibitory interneurons. Nat. Neurosci. 5, 403-404.
Takahashi, H., Nakao, M., and Kaga, K. (2004). Cortical mapping of auditory-
evoked offset responses in rats. Neuroreport 15, 1565-1569.
Tan, A.Y., Atencio, C.A., Polley, D.B., Merzenich, M.M., and Schreiner, C.E.
Unbalanced synaptic inhibition can create intensity-tuned auditory cortex neurons.
arXiv: q-bio.NC/0607036 21 Jul 2006.
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.
Thomson, A.M., West, D.C., and Lodge, D. (1985). An N-methylaspartate receptor-
mediated synapse in rat cerebral cortex: a site of action of ketamine? Nature 313,
479-481.
Turner, J.G., Hughes, L.F., and Caspary, D.M. (2005). Divergent response properties
of layer-V neurons in rat primary auditory cortex. Hear. Res. 202, 129-140.
Villa, A.E., Rouiller, E.M., Simm, G.M., Zurita, P., de Ribaupierre, Y., and de
Ribaupierre, F. (1991). Corticofugal modulation of the information processing in the
auditory thalamus of the cat. Exp. Brain. Res. 86, 506-517.
145 Volgushev, M., Vidyasagar, T.R., Chistiakova, M., Yousef, T., and Eysel, U.T.
(2000). Membrane properties and spike generation in rat visual cortical cells during
reversible cooling. J. Physiol. 522, 59-76.
Volkov, I.O., Dembnovetskii, O.F., and Galaziuk, A.V. (1985). Characteristics of the
responses of auditory cortex neurons in the cat to tonal stimulation during nembutal
anesthesia and after recovery from it. Neirofiziologiia 17, 728-737.
Wang, J., Caspary, D., and Salvi, R.J. (2000). GABA-A antagonist causes dramatic
expansion of tuning in primary auditory cortex. Neuroreport. 11, 1137-1140.
Wang, J., McFadden, S.L., Caspary, D., and Salvi, R. (2002). Gamma-aminobutyric
acid circuits shape response properties of auditory cortex neurons. Brain Res. 944,
219-231.
Wehr, M., and Zador, A. M. (2003). Balanced inhibition underlies tuning and
sharpens spike timing in auditory cortex. Nature 426, 442-446.
Wehr, M., and Zador, A.M. (2005). Synaptic mechanisms of forward suppression in
rat auditory cortex. Neuron 47, 437-45.
Wilent, W.B., and Contreras, D. (2005). Stimulus-dependent changes in spike
threshold enhance feature selectivity in rat barrel cortex neurons. J. Neurosci. 25,
2983-2991.
Winer, J.A., Miller, L.M., Lee C.C., and Schreiner, C.E. (2005). Auditory
thalamocortical transformation: structure and function. Trends Neurosci. 28, 255-
263.
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.
Yamauchi, T., Hori, T., and Takahashi, T. (2000). Presynaptic inhibition by
muscimol through GABAb receptors. Eur. J. Neurosci. 12, 3433-3436.
146 Young, E.D., and Brownnell, W.E. (1976). Response to tones and noise of single
cells in dorsal cochlear nucleus of unanesthetized cats. J. Neurophysiol. 39, 282-300.
Zhang, L.I., Bao, S., and Merzenich, M.M. (2001). Persistent and specific influences
of early acoustic environments on primary auditory cortex. Nat. Neurosci. 4, 1123-
1130.
Zhang, L.I., Bao, S., and Merzenich, M.M. (2002). Disruption of primary auditory
cortex by synchronous auditory inputs during a critical period. Proc. Natl. Acad. Sci.
U. S. A. 99, 2309-2314.
Zhang, L.I., Tan, A.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, Y., and Suga, N. (1997). Corticofugal amplification of subcortical responses
to single tone stimuli in the mustached bat. J. Neurophysiol. 78, 3489-3492.
Zhu, Y., Stornetta, R.L., and Zhu, J.J. (2004). Chandelier cells control excessive
cortical excitation: characteristics of whisker-evoked synaptic responses of layer 2/3
nonpyramidal and pyramidal neurons. J Neurosci. 24, 5101-5108.

147
Appendix A
Defining Cortical Frequency Tuning with Recurrent
Excitatory Circuitry

A.1. Derive the application concentration for mixed SCH50911 and
muscimol
A.1.1. Equations for the competitive binding with GABA receptors
A +Gb Kb
1
A Gb A Gb
[]
= A
[]
Gb
[]
Kb
1
[] (1)
B +Gb Kb
2
B Gb B Gb
[]
= B
[]
Gb
[]
Kb
2
[] (2)
A +Ga Ka
1
A Ga A Ga [] = A [] Ga [] Ka
1
[]
(3)
B +Ga Ka
2
B Ga B Ga [] = B [] Ga [] Ka
2
[]
(4)
A: SCH90511; B: muscimol; Gb: GABA
B
receptor; Ga: GABA
A
receptor. Here, Ka
and Kb refer to functional binding constants. EC
50
or IC
50
values were recovered
from the literatures, and used to calculate K values, as we considered functional
effects of bindings: opening or blocking channels:
Kb
1
=
1
1 μM
(Ong et al., 1998; Ghosh et al., 1999);
Kb
2
=
1
25 μM
(Yamauchi et al., 2000);
Ka
2
=
1
1.7 μM
(Kemp et al., 1986)

148
Ka
1
was estimated according to the instruction that SCH50911 has no binding
affinity to GABA
A
receptors at concentration up to 100 μM (Tocris). To estimate
the value of Ka
1
, we assumed that less than 10% of GABA
A
receptors are bound by
SCH50911 at 100 μM. According to equation (3), and setting [Ga
total
] = [Ga] +
[A·Ga],

A Ga
[]
Ga
total
[]
=
Ka
1
A
[]
1+Ka
1
A
[]
=
Ka
1
100
1+Ka
1
100
0.1
Ka
1
1
900 μM

A.1.2. Estimate the activation effect of muscimol application alone
Derived from the two reactions (2 and 4):
B Gb
[]
=
Gb
[]
Kb
2
B
[]
1+Kb
2
B
[]
(5)
B Ga
[]
=
Ga
[]
Ka
2
B
[]
1+Ka
2
B
[]
(6)
At the effective silencing concentration of muscimol ( 25 M, Fox et al., 2003),

B Gb
[]
Gb
[]
=
Kb
2
B
[]
1+Kb
2
B
[]
=
1
1
Kb
2
B
[]
+1
=
1
25 μM
25 μM
+1
50%

B Ga
[]
Ga
[]
=
Ka
2
B
[]
1+Ka
2
B
[]
=
1
1
Ka
2
B
[]
+1
=
1
1.7 μM
25 μM
+1
94%

149
Thus, considerable portion of GABA
B
receptors will be activated at the effective
concentration of muscimol for activating GABA
A
receptors.

A.1.3. To derive the ratio of SCH50911 and muscimol for effectively activating
GABA
A
receptors, while blocking GABA
B
receptors
Derived from equations (1) (4), we can obtain:
A Gb
[]
=
A
[]
Gb
total
[]
1
Kb
1
+ A
[]
+
Kb
2
Kb
1
B
[]
(7) ( Gb
total
[]
= Gb
[]
+ A Gb
[]
+ B Gb
[]
)
B Gb
[]
=
B
[]
Gb
total
[]
1
Kb
2
+ B
[]
+
Kb
1
Kb
2
A
[]
(8)
A Ga
[]
=
A
[]
Ga
total
[]
1
Ka
1
+ A
[]
+
Ka
2
Ka
1
B
[]
(9) ( Ga
total
[]
= Ga
[]
+ A Ga
[]
+ B Ga
[]
)
B Ga
[]
=
B
[]
Ga
total
[]
1
Ka
2
+ B
[]
+
Ka
1
Ka
2
A
[]
(10)
(i) To achieve that less than 5% of GABA
B
receptors are activated by muscimol.
From eq (8), we can derive

B Gb []
Gb
total
[]
=
B []
1
Kb
2
+ B [] +
Kb
1
Kb
2
A []
=
B []
25 + B [] +25 A []

150
If
B Gb
[]
Gb
total
[]
0.05, then
B
[]
25 + B
[]
+25 A
[]
0.05, or 0.95 B [] 1.25 1+ A [] ()

As [A]>>1 μM in its application,
A
[]
B
[]
0.95
1.25
=0.76 (11)
(ii) To achieve that the majority GABA
A
receptor will not be blocked by SCH50911.
In other words, less than 5% of GABA
A
receptors are bound with SCH50911.
From eq (9), we can derive

A Ga []
Ga
total
[]
=
A []
1
Ka
1
+ A [] +
Ka
2
Ka
1
B []
=
A []
900 + A [] +
900
1.7
B []
0.05
0.95 A
[]
45 +26 B
[]
As effective [B] (μM) is around 25, 26 B
[]
>>45, then 0.95 A
[]
26 B
[]

A
[]
B
[]
26
0.95
=27.4 (12)
Combining eq (11) and eq (12) together, we get
0.76 A
[]
B
[]
27.4 (13)

A.1.4. Activation of GABA
A
receptors under our experimental condition
At the ratio used in the current project ( A
[]
B
[]
=3/2), and considering muscimol
concentration at 25 μM, the effects on GABA
A
receptors would be

151

B Ga []
Ga
total
[]
=
B []
1
Ka
2
+ B [] +
Ka
1
Ka
2
A []
=
B []
1.7 μM + B [] +
1.7 μM
900 μM
1.5 B []
= 93%
i.e. more than 93% of GABA
A
receptors will be activated.

A.1.5. Experiment on the application of SCH50911 alone
Application of SCH50911 alone slightly increased the amplitude of
field-potential responses and prolonged tone-evoked spiking activity, but did not
change the shape of spike TRFs of cortical neurons (Figure 2.1A, Figure A.1).

Figure A.1 Tone-evoked spiking activity and spike TRF recorded from the same cortical site before
(A) and after (B) cortical injection of SCH50911 alone. Left, Peri-stimulus spike-timing histogram
(PSTH). Right, spike TRF presented as a gray-scale map. (C) Percentage change in the bandwidth
of spike TRF (measured at 60dB) and in the number of tone-evoked spikes (at 60 dB) after SCH50911
application. N = 5 sites. Bar = SD. Paired t-test, * P<0.01.

152
A.2. Space clamp and effects of cortical muscimol application on
synaptic responses
Here, we want to provide a quantitative estimation of nonspecific effects of
drug application on voltage clamp and synaptic currents.

A.2.1. Correcting amplitude of synaptic responses according to changes in
presumptive pure thalamic inputs
In this study, we estimated the level of nonspecific reduction in the recorded
synaptic responses after drug application according to changes in the presumptive
pure thalamocortical inputs at the subthreshold intensity threshold. The assumption
is supported by two lines of evidence: first, the multiunit TRF recorded at the same
site before silencing has higher intensity threshold (Figure A.2), suggesting that the
synaptic inputs at the subthreshold intensity threshold are unlikely contributed by
local intracortical connections; second, the kinetics of response currents (Figure 2.2D
and Figure 2.3A, C, E, G) also supports this assumption. For all the recorded
neurons, there was no noticeable change in the range of responding frequencies or in
the intensity threshold after cortical silencing (Figure A.3). We analyzed the
kinetics of the rising phase of response currents before and after cocktail
microinjection. The kinetics of an input consisting of both thalamocortical and

153
intracortical components is likely characterized by two or multiple phases due to
differences in the onset timing: an early fast rising phase representing the

Figure A.2 Multi-unit spike TRF before cortical silencing recorded from the same site as the
example cell in Figure 2.2. Arrow indicates the level of intensity threshold for subthreshold responses.

Figure A.3 Frequency range for excitatory synaptic responses before (blue) and after
(red) the cocktail application. Data are grouped according to the tone intensity. N= 5 cells.
Bar = SD.

154
monosynaptic thalamic inputs, and later phases for integrated local and thalamic
inputs. On the other hand, there will be one phase for pure thalamic inputs. For
all the five cells obtained, we averaged inputs evoked by best frequencies at 60 dB
SPL, as well as at the intensity threshold. As shown in the figures, the inputs at 60
dB are clearly bi phased before cortical silencing and become mono phased after
drug application, suggesting that the later phase can be attributed to intracortical
inputs. On the other hand, those at the intensity threshold remain mono phased
before and after silencing, consistent with the assumption.

A.2.2. Can isopotential of the cell for the recorded synaptic inputs be achieved
in our voltage clamp recordings?
In this study, we have assumed linear, isopotential neurons for the recorded
excitatory synaptic inputs, as described in previous studies (Zhang et al., 2003; Wehr
and Zador, 2003; Tan et al., 2004; Wu et al., 2006). Potential deviations due to
space clamp error and cable attenuation for synaptic inputs at the distal dendrites
were discussed in several recent studies (Wehr and Zador, 2003; Tan et al., 2004; Wu
et al., 2006). Here, we provide further estimation of our space clamp before and
after drug application.

155
First, cortical cells appear to be “reasonably” clamped under our
experimental condition. This is supported by the linearity of I V curves (Figure
2.2B) and the fact that when cells were clamped at 0 mV, no significant excitatory
currents were observed (Figure 2.2B), except the outward Cl
-
mediated currents. In
addition, the derived reversal potential for the early component of tone evoked
currents (0 1 ms window after response onset, mainly excitatory) was 0 ± 5mV
(Figure 2.2B), close to the known reversal potential for glutamatergic currents.
These data suggest that under our voltage clamp recording conditions, the
tone evoked synaptic currents can be detected with a “reasonable” clamping
accuracy, with a deviation within 5 mV. This may be attributed to the use of
intracellular cesium, TEA, QX 314, and ketamine anesthesia, which together block
most voltage dependent currents (though K
+
and Na
+
channels, and NMDA
receptors). Furthermore, the overlap of the frequency range for excitatory response
with that for tone evoked membrane potential depolarization (Figure 2.4B, right)
suggests that excitatory responses recorded at -70mV reflect the effective inputs
made onto the recorded cell, consistent with our previous observation (Wu et al.,
2006).
Second, the reasonable voltage clamp can be achieved in the dendrites
close to soma (<200 μm), as simulated with a Neuron model (Hines, 1993).

156

Figure A.4 Equivalent cylinder neuron model

Model description (Figure A.4): We used simplified equivalent cylinder
model, which consists of a cylinder of 10 micron length and 10 micron diameter as
soma and a cable of 200 micron length and 0.8 micron diameter as dendrite.
Other physical parameters used in the simulation were the specific intracellular
resistivity R
a
= 200 •cm, the specific membrane capacitance C
m
= 1 μF/cm
2
, the
specific leakage conductance at normal condition g
m
= 2e-5 S/cm
2
, resting potential
E
r
= -60mV. The soma was clamped at -70mV, which is the holding potential used
in the experiment. We used alpha synapse, whose conductance has the form of
g = g
max
(t/ ) e
(1 t/ )
, to simulate the synaptic input. was set as 4ms, g
max

as 0.005 μS, and the excitatory reversal potential (E
e
) as 0mV. In this model, we
estimated the effects of increasing leakage conductance by muscimol, by assuming
g
m
= 1e-4 S/cm
2
after drug application.
As shown in Figure A.5A, the synapses located 200 micron away from the
soma can be clamped well, which is not significantly affected by the increase of
membrane leakage. This model study suggests that the tone evoked synaptic

157
inputs we examined may be from dendritic regions close to soma, which had been
reasonably clamped.

Figure A.5 The modeling of cable effect on recorded synaptic currents and holding potentials
(A) The actual holding potential along the dendrite V
act
is not significantly changed by the increase of
leakage conductance; the reasonable holding condition is maintained for voltage clamping recording.
(B) The attenuation (I
syn1
/I
syn2
) caused by the increase of leakage conductance changes only slightly
with the increase of synaptic current amplitude. g
max
is the maximum conductance for alpha synapse.

Figure A.6 The current voltage (I V) relationship of tone evoked synaptic currents after
the cocktail application
(A) synaptic currents (average of four repeats) evoked by a tone of 1.9 kHz and 70 dB, recorded at
different holding potentials: 0 mV (cyan), -30 mV (red), -70 mV (green) and -100 mV (blue). (B) I V
curves for synaptic currents averaged within a 20 22.5 ms window after the stimulus onset. Bar = SD.
A B
A B

158
Third, I V curve remains linear and crosses the origin (0 0). An example
I V curve obtained from a neuron after cortical silencing is shown (Figure A.6).
In this case, clamping voltages, synaptic currents, and a junction potential of 12mV
were corrected according to the discussion below.

A.2.3 The potential effects of cocktail drug application on voltage clamp and
synaptic currents
Two effects were observed for cocktail drug application: 1) the drug
microinjection caused gentle mechanical effects, resulting in an increase of series
resistance (R
s
) after application (Table A.1), which then remained stable (with change
<10%); 2) Muscimol in the cocktail increases the leakage of the cell by constitutively
activating GABAa receptors, as reflected by an increase in leakage current (g
leak
).
Both effects will affect I V curve. However, they can be corrected given that
isopotential condition for voltage-clamping is maintained, which is supported by
linear I V curve after drug application and the simulation with increased membrane
permeability.
Here, we first derived the effects of drug application on clamping voltage
(V
h
) and synaptic currents (I
syn
):

159
i. Derivation of equations

Based on the isopotential neuron model, the actual voltage applied on the
soma V
act
can be expressed by equation (1). V
h
is the holding potential provided by
the amplifier; I
e
is the recorded current through the electrode, which is the sum of
real synaptic currents I
syn
and leakage currents I
r
; R
s
is the series resistance. The
amplitudes of I
syn
and I
leak
(inward current) are determined by the equation (3) and (4)
respectively. g
syn
is the synaptic conductance and g
r
is the leakage conductance;
E
syn
is the synaptic reversal potential (here 0 mV for excitatory synapse) and E
r
is the
resting potential; R
in
is the input resistance.
V
act
(t) =V
h
I
e
(t) R
s
(1)
I
e
(t) = I
syn
(t) + I
leak
(t) (2)
I
syn
(t) = g
syn
(t) V
act
(t) E
syn
()
(3)
I
leak
(t) = g
r
* V
act
(t) E
r
()
g
r
=1/R
in
(4)

Simplified isopotential neuron model

160
Thus, the V
act
in the presence of synaptic inputs will be significantly
different from that in the absence of synaptic inputs, and both the clamping voltage
and the measured synaptic currents need to be corrected according to the change of
V
act
. Without correction, the measured synaptic current is simply derived by:
I’
syn
(t)

= I
e
(t) – I
leak
(t=0), which does not consider that I
leak
changes with input
current.
Using the equations (1) (4), we can derive the relationship between I
syn

and I’
syn
as following.
I
syn
(t) =
R
in
+ R
s
R
in
I'
syn
(t) (5)
And the relationship between V
h
and I’
syn
,
V
h
= I'
syn
(t) R
in
+ R
s
R
in
*R
s
+
R
in
+ R
s
R
in

2
g
syn
(t)

R
s
R
in
E
r
+
R
in
+ R
s
R
in
E
syn
(6)
As E
syn
= 0 mV, comparing equation (3) and (6), the curve (V
act
I
syn
) will cross the
origin (0 0), while the curve (V
h
I’
syn
) will not.
ii. Calculation of input resistance (R
in
) from our experiments
Based on the value of I
leak
, we can estimate leakage conductance and input
resistance according to Equation (4). By holding at two different voltages V
act1
and
V
act2
, and measuring the corresponding leakage currents I
leak1
and I
leak2
, then we can
calculate R
in
.

161
g
r
= I
leak1
I
leak2
()
V
act1
V
act2
()
(7)
A.2.4 Correction of synaptic currents by changes in R
in
and R
s
(Table A.1)
For three of the five cells in this study, we applied two clamping voltages
after cocktail application. I’
syn
can then be corrected by the factor (R
in
+ R
s
)/R
in
.
For these three cells, the correction factor based on presumptive pure thalamic inputs
(which is what we used) is 1.66 ± 0.30, while it is 1.36 ± 0.07 if based on the
changes in R
in
and R
s
, which contributes to 73% of the correction we used (see A.2.6
below for further discussion on other potential non specific effects).
012707 111506 111606
g
r
(before) 6.96 nS 8.9 nS 9.4 nS
g
r
(after)

9.44 nS 9.9 nS 17.1 nS
g
r
2.48 nS 1 nS 7.7 nS
R
in
(before)

144 M 112 M 106 M
R
in
(after)

106 M 101 M 58 M
R
in
38 M 11 M 48 M
11 M 17 M 8 M

30 M 40 M 23 M
R
s
19 M 23 M 15 M

Table A.1 Estimated parameters based on the recording data

162
A.2.5 Correction of I V curves after drug application
According to equations (1) and (5), we calculated V
act
and I
syn
, and replotted
the curves of V
act
vs I
syn
(Figure A.7, right column), and compared them with the V
h

I’
syn
curves (Figure A.7, left column). After corrections, the I V curves almost
pass the origin, which also confirms that 1) reasonable clamping is conserved; and 2)
isopotential model is valid.

A.2.6 Other potential nonspecific effects of drug application
Change in cable effects may be due to increase in membrane permeability.
The attenuation of the recorded synaptic currents caused by the increase of
membrane permeability is a function of synaptic position, and increases as the
synapse becomes farther away from the soma. When the synapse is at 200 micron
distance from the soma, this gives about 10% decrease in measured peak current.
Moreover this attenuation is only slightly affected by the increase of amplitude of
synaptic currents (Figure A.5B).
The pharmacological methods are highly dependent on the specificity of the
pharmacological agents used. However, our current understanding of these agents
may still be limited as they may not have been tested in a broad spectrum for
potential side effects. For example, the effects of muscimol on presynaptic GABA
B

163

Figure A.7 I V curve correction. Left column are the I V curves before correction, which do not
pass the origin. Right column are the I V curves after correction, whose x intercepts are very close
to the origin, which indicate that the cells were reasonably clamped under our experimental condition.
Each row represents one cell. Bar = SD.

164
receptors are only described recently (Yamauchi et al., 2000). There may be other
non specific effects of the cocktail drugs, which can contribute to the correction
factors. In addition, the application of the cocktail at the ratio in our experiments,
still results in the activation of GABAb receptors (~ 5%). Its potential effects on
presynaptic transmission cannot be excluded.
Taken together, we think that the changes in Rs and Rin caused the major
non specific effect of the cocktail application, since they largely explain the
observed reduction of excitatory synaptic responses. It should also be noted that this
study was carried out in anesthetized animals. The anesthesia will reduce the general
activity level since ketamine blocks NMDA receptors (Thomson et al., 1985), and
can result in a scaling down of the tuning curves. This type of effects may not
qualitatively affect our conclusion. Although ketamine anesthesia will also effect the
temporal pattern of activity, e.g. long-latency (>100ms) and oscillatory responses
(Rennaker et al., 2007), these effects normally will not affect the short-latency
responses examined in this study, as NMDA receptors are mostly blocked by Mg
2+

before spikes are generated (Nowak et al., 1984; Mayer et al., 1984).

Asset Metadata

Creator Wu, Guangying (author)

Core Title Synaptic mechanisms for basic auditory processing

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Neuroscience

Publication Date 02/13/2009

Defense Date 12/02/2008

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

Tag auditory processing,duration coding,frequency tuning,in vivo whole cell recording,intensity tuning,OAI-PMH Harvest

Language English

Advisor Zhang, Li I. (committee chair), Chen, Jeannie (committee member), Chow, Robert HP. (committee member), Langen, Ralf (committee member), Sampath, Alapakkam P. (committee member)

Creator Email guangying@gmail.com,guangyiw@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1975

Unique identifier UC1231233

Identifier etd-Wu-2449 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-149405 (legacy record id),usctheses-m1975 (legacy record id)

Legacy Identifier etd-Wu-2449.pdf

Dmrecord 149405

Document Type Dissertation

Rights Wu, Guangying

Type texts

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

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email uscdl@usc.edu

Abstract (if available)

Abstract Neurons are organized into circuits to process various information in the brain. To understand how information is processed, it's fundamental to investigate the patterns of excitatory and inhibitory inputs underlying neuron's response properties. Through the four independent but closely related studies described in my dissertation, I investigated the synaptic mechanisms underlying basic response properties of rat auditory neurons, i.e. frequency tuning, intensity tuning and temporal coding. Firstly, we developed a novel method to silencing the cortex and dissected the excitatory input from thalamic neurons and that from cortical excitatory neurons. The results demonstrated that thalamic input had a flattened frequency tuning curve. In contrast, intracortical excitatory input was sharply tuned with a tuning curve that closely matched that of suprathreshold responses. It suggests the recurrent excitatory circuits define the cortical frequency tuning. Secondly, to study the cortical inhibitory neurons, we combined single-unit cell-attached recording, juxtacellular labeling and whole-cell recording together on the same neuron in vivo. We discovered that the frequency tuning curve of inhibitory input was broader than that of excitatory input. So a relatively stronger inhibition was flanked to the preferred frequencies of the cell and laterally sharpened the frequency tuning of membrane responses. The less selective inhibition can be attributed to a broader bandwidth and lower threshold of spike tonal receptive field of fast-spike inhibitory neurons than nearby excitatory neurons. Thirdly, we uncovered the synaptic mechanisms underlying cortical intensity tuning. The results demonstrated that excitatory inputs to those intensity-tuned neurons have already shown a weak tuning, while inhibitory inputs suppressed the excitation at higher intensities to strengthen the tuning through a dynamic temporal control.