Provided for non-commercial research and educational use.

Not for reproduction, distribution or commercial use.

This article was originally published in the The Senses: A Comprehensive Reference,
published by Elsevier, and the attached copy is provided by Elsevier for the author’s benefit
and for the benefit of the author’s institution, for non-commercial research and educational
use including without limitation use in instruction at your institution,

sending it to specific colleagues who you know, and providing a copy to your institution’s
administrator.

All other uses, reproduction and distribution, including without limitation commercial reprints,
selling or licensing copies or access, or posting on open internet sites, your personal or
institution’s website or repository, are prohibited. For exceptions, permission may be sought
for such use through Elsevier’s permissions site at:

http://www.elsevier.com/locate/permissionusematerial

M A Rutherford and W M Roberts, Afferent Synaptic Mechanisms. In: Allan I. Basbaum,

Akimichi Kaneko, Gordon M. Shepherd and Gerald Westheimer, editors The Senses:
A Comprehensive Reference, Vol 3, Audition, Peter Dallos and
Donata Oertel. San Diego: Academic Press; 2008. p. 365-396.
 Author's personal copy

3.22 Afferent Synaptic Mechanisms

M A Rutherford and W M Roberts, University of Oregon, Eugene, OR, USA
ª 2008 Elsevier Inc. All rights reserved.

3.22.1 Introduction 366

3.22.2 Afferent Neurons of Acousticolateralis Sensory Systems 366
3.22.2.1 Spiral Ganglion Neurons 367
3.22.2.2 Auditory Nerve Fibers 368
3.22.2.3 Innervation Patterns and the Function of Auditory Organs 371
3.22.3 Hair Cells 372
3.22.3.1 Hair Cell Shape in Relation to Synaptic Signaling 372
3.22.3.2 The Synaptic Body and Synaptic Vesicle Pools 372
3.22.3.3 Calcium Buffering in Hair Cells 375
3.22.4 Ribbon Synapses – A Small but Diverse Class 377
3.22.4.1 The Synaptic Body (Ribbon) and Its Vesicles 377
3.22.4.2 Molecular Specializations of Ribbon Synapses 378
3.22.4.3 Anatomy of the Synaptic Body: Heterogeneity of Structure 380
3.22.5 The Synaptic Vesicle Cycle in Hair Cells 381
3.22.5.1 Endocytosis 381
3.22.5.2 Vesicle Mobility 381
3.22.5.3 Docking and Priming 382
3.22.5.4 Transmitter Release and Clearance 382
3.22.5.5 Relating Anatomical and Functional Vesicle Pools 383
3.22.6 Physiology of the Hair Cell Afferent Synapse 384
3.22.6.1 Encoding of Phase 385
3.22.6.2 Encoding of Frequency 386
3.22.6.3 Baseline and Evoked Synaptic Activity 386
3.22.6.4 Tuning Mechanisms in Hair Cells 388
3.22.6.4.1 Frequency-selective augmentation of membrane potential oscillation 388
3.22.6.4.2 Synaptic frequency selectivity 388
3.22.7 Auditory Nerve Fiber Firing Patterns Mirror Presynaptic Vesicle Pool Dynamics 389
3.22.7.1 Adaptation of the Auditory Nerve Response 389
3.22.7.2 The Role of the Synaptic Body in Auditory Nerve Fiber Spike Timing 390
3.22.8 Synapse to Psychophysics 391
References 391

Glossary
active zone (AZ) The area where synaptic vesicles
fuse with the presynaptic membrane. The AZ is
juxtaposed to the postsynaptic terminal across the
synaptic cleft.
afferent The flow of signals from the periphery
toward the central nervous system (CNS).
auditory nerve fiber (ANF) The axon of an audi-
tory afferent neuron that projects to the brainstem
in the vestibulocochlear (eighth cranial) nerve. In
mammals, ANFs have cell bodies in the spiral
ganglion.
characteristic frequency (CF) The frequency of
maximum sensitivity to sound.
cytomatrix at the active zone (CAZ) The cytos-
keletal meshwork comprised of multidomain
proteins that aggregate through specific protein–
protein and protein–lipid interactions. The CAZ
creates a scaffold upon which the presynaptic
machinery is assembled.
hair cells The sensory receptors of the auditory
system and related sense organs of the ear and
lateral line of vertebrate animals.

365
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
 Author's personal copy
366 Afferent Synaptic Mechanisms
inner hair cell (IHC) One of the highly innervated
hair cells in the mammalian ear that are primarily
responsible for the perception of sound. They are
located in a single row near the modiolar (inner)
edge of the organ of Corti.
root mean square (RMS) A type of average com-
puted by first squaring a set of values, then
computing the mean of the squared values, then
taking the square root of the mean. The RMS
average deviation of N samples from their mean is
p
(N  1) times the standard error of the mean.
sound pressure level (SPL) The RMS pressure
change during a sound, expressed on the logarith-
mic decibel (dB) scale, measured in the vicinity of
the listener during a specified time period.
spiral ganglion The cell bodies of the afferent
neurons that innervate the organ of Corti; housed in
the bony spiral lamina of the mammalian cochlea.
spontaneous rate (SR) Frequency of spikes in an
auditory nerve fiber in the absence of an applied
sound stimulus.
synaptic body or ribbon (SB) The most obvious
presynaptic structure at hair cell afferent synapses,
and the subject of many carefully designed experi-
ments to reveal its function(s). There is general
3.22.1 Introduction
We will begin with an overview of the hair cell
afferent synapse and related structures, with a focus
on mammalian inner hair cells (IHCs) and spiral
ganglion neurons. Our review will then turn to elec-
trophysiological studies and discussions of the
specializations that may endow these first auditory
synapses with their ability to maintain fast and reli-
able, information-rich signaling to the brain. Readers
unfamiliar with the basics of cellular neuroscience
will better appreciate the unique properties of audi-
tory afferent synapses after reviewing the properties of
conventional chemical synapses described in the many
excellent neuroscience textbooks (e.g., Delcomyn, F.,
1998; Kandel, E. R. et al., 2000; Hille, B., 2001;
Nicholls, J. G. et al., 2001). Additional information spe-
cific to hair cell afferent synapses can be found in recent
literature reviews (Nouvian, R. et al., 2006; Moser, T.
et al., 2006a; 2006b).
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
agreement that the SB serves an important pur-
pose, which is not presently known. It harbors
synaptic vesicles attached to it via fine filaments
that are thought to either facilitate or impede the
movement of vesicles along its surface.
synaptic vesicle (SV) Spherical lipid bilayer that
contains neurotransmitter.
synaptic vesicle cycle The recurrent processes
by which SVs are continuously created, used and
recreated. Exocytosis of the neurotransmitter is
triggered by calcium and involves fusion of the SV
membrane with the plasmalemma and/or the for-
mation of a transient aqueous connection between
the SV and the extracellular space. After exocyto-
sis, SVs are reformed, refilled, and reenter the
exocytic pathway.
threshold The SPL needed to produce a criterion
response. Defined in this way, threshold is a mea-
sure of sensitivity to sound: low threshold means
high sensitivity. Sensitivity is usually expressed in
units of response/stimulus, but in auditory physiol-
ogy threshold is traditionally defined in units of
stimulus only, omitting explicit reference to the
response criterion used in a particular experiment.
3.22.2 Afferent Neurons of
Acousticolateralis Sensory Systems
The primary afferent neurons in the auditory and
vestibular organs of all vertebrate species, and in
related mechanosensory and electrosensory organs
of the lateral line system of aquatic vertebrates are
bipolar neurons that have cell bodies located in or
near the ear. They extend one process peripherally to
the sensory epithelium, the other centrally to the
brainstem. The peripheral processes have often
been called dendrites, but most are myelinated and
can properly be called axons. This distinction is
likely to be important for the assembly and mainte-
nance of afferent synapses because axons and
dendrites use different protein-targeting signals
(Burack, M. A. et al., 2000). The afferent neurons in
auditory and vestibular organs have cell bodies
located close to the corresponding sensory epithe-
lium, and therefore have short peripheral axons.
 Author's personal copy
Lateral line afferent neurons have long peripheral
axons that extend from cell bodies in the lateral line
ganglia near the ear to sensory neuromasts on the
head and trunk (Dambly-Chaudiere, C. et al., 2003).
3.22.2.1 Spiral Ganglion Neurons
In mammals, the cell bodies of auditory afferent
neurons reside in the spiral ganglion, which parallels
the inner (modiolar) edge of the organ of Corti
(Figure 1(a)). Type I spiral ganglion neurons are
responsive to acoustic stimulation and constitute
90–95% of cochlear nerve afferents (Spoendlin, H.,
1969; Liberman, M. C., 1982). Intracellular labeling
with horseradish peroxidase and subsequent histolo-
gical processing revealed that they contact only the
IHCs (Robertson, D., 1984). Both the peripheral and
the central axons of a type I afferent neuron are
myelinated and have a cluster of voltage-gated
Naþ channels (Nav 1.6) at each node of Ranvier.
Naþ channels are also found in the unmyelinated
axon segment between the postsynaptic terminal
and the most distal peripheral node, suggesting that
spikes may be initiated at or near the site of synaptic
input (Hossain, W. A. et al., 2005), as occurs in verte-
brate skeletal muscle (Kandel, E. R. et al., 2000).
Minimizing the distance from the synapse to the
spike initiation site is expected to minimize attenua-
tion and low-pass filtering of excitatory postsynaptic
potentials (EPSPs). The cell bodies of type I spiral
ganglion neurons are also wrapped in a loose myelin-
like sheath that may speed conduction, reduce tem-
poral jitter, or decrease the probability of conduction
failure across the cell body during high-frequency
firing (Hossain, W. A. et al., 2005).
In mammals, the peripheral axons of type I affer-
ents (also known as radial fibers) remain unbranched
and contact only a single IHC. Unlike most neurons,
which receive synaptic inputs from hundreds or
thousands of separate sites of synaptic contact (active
zones; AZs), each type I afferent neuron typically
receives sensory input from a single hair cell AZ
(Spoendlin, H., 1969; Liberman, M. C., 1980;
Spoendlin, H., 1985; Hashimoto, S. et al., 1990). In
other vertebrates, most auditory afferent neurons
make multiple postsynaptic contacts on one or more
hair cells (Lewis, E. R. et al., 1982).
Each mammalian IHC provides the sole input to
5–30 type I afferent axons, which vary in number by
species and hair cell position on the cochlear tonoto-
pic axis (Kiang, N. Y. S. et al., 1982; Liberman, M. C.,
1982; Fuchs, P. A. et al., 2003). The mammalian
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 367
cochlea thus represents a case of synaptic divergence
in which each IHC controls many private communi-
cation channels to the brain. The function of this
unusual arrangement is only partially understood,
but it is easy to speculate that the dynamic membrane
potential of the hair cell contains more information
than can be transmitted across one synapse and
encoded in a single spike train.
Information is degraded by events that alter a
signal in an unpredictable way (noise). The synaptic
noise introduced by the stochastic nature of exocy-
tosis can be reduced by having synaptic currents from
many AZs converge on the spike initiation zone of
one afferent axon, as occurs at the calyx synapses of
vestibular hair cells (Goldberg, J. M., 1996); but the
mammalian auditory system uses a different strategy
in which the maximum possible separation of infor-
mation channels to the brain is maintained by having
each synaptic AZ control action potential generation
in a separate afferent axon. If separate information
channels later converge onto a single higher-order
neuron in the brain stem (e.g., a globular bushy cell in
the cochlear nucleus that detects coincident input),
this arrangement can reduce the effects of variability
in the spike conduction time from the ear to the
central nervous system (CNS) and other noise
sources that are downstream of the afferent synapse.
A small population of afferent axons in the mam-
malian cochlea (type II; Figure 1(a)) are
unmyelinated (Liberman, M. C. et al., 1990). Like
the type I afferents, they are labeled by antibodies
against Nav 1.6 and may initiate action potentials
near their distal tips (Hossain, W. A. et al., 2005).
These outer hair cell (OHC) afferents are thought
to monitor the state of the organ of Corti, but not to
contribute directly to the perception of sound, the
role ascribed to the IHCs. Very little is known about
the physiology of type II afferents, and they will not
be considered further in this chapter.
Although the peripheral neural network in the ear
is less elaborate than that of the retina and contains
far fewer cell types, electron micrographs of the
organ of Corti do show several additional types of
synapses that could provide important functional
interactions, including local feedback (Raphael, Y.
and Altschuler, R. A., 2003; Sobkowicz, H. M. et al.,
2003). These include connections between IHCs and
OHCs via afferent and efferent fibers that contact
both hair cell types (Sobkowicz, H. M. et al., 2004).
Nothing is known about the function of these
synapses in auditory signal processing. Efferent
synapses are typically seen on the postsynaptic
 Author's personal copy
368 Afferent Synaptic Mechanisms

(a)
Spiral
lamina
Type l Node of Tectorial membrane
Ranvier
IHC
Type ll
OHCs

Bas
ilar
(b) (c) mem
Endolymph bran
e
Perilymph Tight
SB junction
10 mV
Modiolar Pillar 2 ms

EPSPs

0.1 MET
mV Tone
Afferent
bouton

Figure 1 Auditory afferent synapses. (a) Schematic cross section through the mammalian organ of Corti, showing the
bipolar cell bodies of afferent neurons (left), which each send one axon peripherally toward the hair cells and another axon
centrally to the brainstem. In mammals, most afferent axons (type I) are myelinated (green), unbranched, and contact a single
inner hair cell (IHC). A smaller population (type II) are unmyelinated and branch to contact several outer hair cells. A high density
of voltage-gated sodium channels (Nav 1.6, red) are located at nodes of Ranvier, on the unmyelinated distal axon segments of
all afferent neurons, and at both axon hillocks of type II neurons. (b) Each IHC supplies the only input to tens of type I afferent
neurons at active zones. Large-diameter afferent axons usually contact the pillar face of the IHC, whereas smaller-diameter
axons contact the modiolar side. (c) Electrophysiological recording of afferent synaptic activity in vivo. A 200 Hz sinusoidal
sound delivered to the goldfish sacculus (bottom trace) evoked an extracellular field potential (middle trace), analogous to a
cochlear microphonic potential (see Chapter Hair Cell Transduction and Adaptation: Physiology and Molecular Mechanisms),
that is primarily a measure of mechanoelectrical transduction (MET). The top trace is an intracellular recording from an afferent
axon, in which spikes had been blocked to reveal the EPSPs, which are synchronized to the sound waveform and show strong
synaptic depression. (c) Adapted from Furukawa, T. and Matsuura, S. 1978. Adaptive rundown of excitatory postsynaptic
potentials at synapses between hair cells and eighth nerve fibres in the goldfish. J. Physiol. 276, 193–209, with permission from
Blackwell Publishing.

terminals of mammalian type I afferents, where they
may modulate synaptic transmission from the hair
cell to the afferent axon. Efferent axons also make
inhibitory cholinergic synapses directly onto hair
cells in most vertebrate auditory and vestibular
organs, including mammalian IHCs and OHCs (see
Chapter Efferent System).
3.22.2.2 Auditory Nerve Fibers
The central axons of spiral ganglion neurons project
to the brainstem in the eighth cranial nerve, where
they are called auditory nerve fibers (ANFs). Nearly
all recordings of action potentials in mammalian
ANFs are presumed to be from axons of type I spiral
ganglion neurons because they are larger and much
more numerous than the axons of type II neurons
(Liberman, M. C., 1982).
In vivo recordings from single ANFs show that
nearly all fire action potentials spontaneously (i.e., in
a quiet laboratory environment with no experimen-
tally imposed stimulus). The baseline or spontaneous
firing rate (SR) in some ANFs exceeds 100 spikes s1,
while others fire fewer than 5 spikes s1, with no
apparent correlation between a fiber’s characteristic
frequency (CF) and SR (Kiang, N. Y. S. et al., 1965;
Tsuji, J. and Liberman, M. C., 1997). CF varies with
position along the tonotopic axis whereas SR has been
correlated with both axon diameter and subcellular
synaptic structure in the guinea-pig and cat
(Liberman, M. C., 1980; Merchan-Perez, A. and
Liberman, M. C., 1996; Tsuji, J. and Liberman, M. C.,
1997). In these studies of afferent structure/function,

The Senses: A Comprehensive Reference, vol. 3, pp. 365-395

 Author's personal copy
intracellular recording and labeling of individual
ANFs followed by serial-section electron microscopy
(EM) showed that ANFs of high SR belong to spiral
ganglion neurons that have large diameter peripheral
axons and contact IHCs predominantly on the pillar
face (Figure 1(b)); however, this segregation of axonal
diameters around the hair cell circumference was not
observed in one study of the gerbil (Slepecky, N. B.
et al., 2000). Low- and intermediate-SR fibers of the cat
and guinea-pig contact the modiolar face and have
synapses with larger presynaptic dense bodies (called
synaptic bodies, SBs, or ribbons) and more vesicles
than synapses with high-SR fibers. Because the 5–30
afferent synapses on each IHC display this range of
ultrastructural characteristics and modiolar/pillar
position, and because ANFs of the same CF differ in
SR, it is inferred that each IHC synapses onto both
high-SR fibers and intermediate-/low-SR fibers.
To quantify the strength of a sound, most experi-
menters report the sound pressure level (SPL),
expressed on the logarithmic decibel scale (dB).
Each 20 dB increase in SPL corresponds to a tenfold
multiplication of root mean square (RMS) sound
pressure. The zero point (0 dB) is often defined as
20 mPa, which is approximately the amplitude of a
barely perceptible sound at frequencies where human
hearing is most sensitive (2–5 kHz). Compared to the
standard atmospheric pressure of 100 kPa, 20 mPa
amounts to a change of only about 1 part in 5 billion.
An ANF’s sensitivity to sound (Figure 2) is typi-
cally measured as the SPL needed to increase its
firing rate by a criterion number of spikes per second
above the SR, or alternatively, to produce a criterion
increase in ANF discharge synchrony with the sound
stimulus (Rose, J. E. et al., 1967; 1971; Johnson, D. H.,
1980). This measure, which is often called the fiber’s
threshold, depends on the response criterion, stimu-
lus duration, and other factors that are specific to the
experimental protocol (Heil, P. and Neubauer, H.,
2003). The term threshold is typically used in biology
to mean a stimulus level below which there is no
response (e.g., the threshold for generating an action
potential), and therefore implies the existence of a
strong nonlinearity in the stimulus–response relation
(a cut-off below which there is no response). The use
of the term threshold in auditory physiology is quite
different. It carries no implication of a threshold
nonlinearity, only that the response has fallen below
some criterion level. In fact, the initial stages of
auditory processing are constructed to avoid thresh-
old nonlinearities.
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 369
A threshold nonlinearity in the peripheral audi-
tory system would be detrimental to subsequent
neural processing because it would eliminate infor-
mation that could otherwise be recovered by
temporal averaging (integration) or filtering, which
can in principle pull an arbitrarily small periodic
signal out of aperiodic background noise. This ability
is crucial for systems designed to detect small noisy
signals. For example, a sinusoidal signal that is super-
imposed on additive white noise can be passed
through a narrow-band linear filter that leaves the
signal unchanged while reducing the noise in propor-
tion to the bandwidth of the filter. Narrowing the
filter’s bandwidth can thus increase the signal-to-
noise ratio as much needed to detect an arbitrarily
small periodic signal, albeit at the expense of tem-
poral resolution (Bracewell, R. N., 1999). The
auditory system’s use of temporal integration to
improve signal-to-noise ratio is seen experimentally as
an inverse relation between threshold and sound dura-
tion ( Jeffress, L. A., 1968; Heil, P. and Neubauer, H.,
2001; 2003). The observation that sounds too soft to
be audible can nevertheless influence the audibility
of subsequent sounds (Plack, C. J. et al., 2006) supports
the idea that subthreshold sounds still produce some
response.
The peripheral auditory system eliminates any
threshold cut-off in the afferent response to faint
sounds by eliminating all slack from the system,
ensuring that arbitrarily weak sounds evoke a change
in ANF firing rate that is proportional to the sound
amplitude, at least for low-frequency tones. This
begins with mechanosensory transduction, which
uses feedback to maintain partial activation of the
transduction channels even in complete silence.
The resulting resting transduction current places
the hair cell’s resting membrane potential in a range
where the presynaptic calcium channels are also par-
tially activated (Koschak, A. et al., 2001; Xu, W. and
Lipscombe, D., 2001; Johnson, S. L. et al., 2005). This
steady-state activation of the calcium current (i.e., the
random, thermally driven, independent gating of
individual calcium channels that occurs at constant
membrane potential) may be sufficient to catalyze
baseline transmitter release (Brandt, A. et al., 2005)
that drives the so-called spontaneous spiking in
ANFs (Liberman, M. C. and Kiang, N. Y. S., 1978;
Sewell, W. F., 1984; Glowatzki, E. and Fuchs, P. A.,
2002; Keen, E. C. and Hudspeth, A. J., 2006). Thus,
each stage of processing is maintained in a partially
active state even in complete silence, thereby avoid-
ing a threshold nonlinearity and the resulting loss of
 Author's personal copy
370 Afferent Synaptic Mechanisms

(a) 120 (b)

110 120

100 100
Threshold (dB SPL)

Threshold (dB SPL)

90 80

80 60

70 40

60 20

50 0
0.1 1
Frequency (kHz)
40
0.01 0.1 1
100 (c)
1
90
0.8
80 0.6
Threshold (dB SPL)

Vector strength
70
0.4
60

50
0.2
40

30
0.1
0.01 0.1 1
20 Frequency (kHz)
0.1 1 10
Frequency (kHz)
Figure 2 Tuning curves and phase locking in avian (emu) auditory nerve fibers (ANFs). (a) The sound pressure level (SPL)
required to evoke a criterion response, in this case, a 20 Hz increase in ANF firing rate above the spontaneous rate (SR),
plotted as a function of the frequency of the tonal stimulus. This type of measurement is often called the fiber’s threshold,
although it does not mean that the fiber fails to respond to less intense sounds, only that the response falls below the
experimenter’s criterion. Data from two ANFs are shown. Each panel shows one fiber’s tuning curve as a continuous line and
phase histograms (insets) from the same fiber, each spanning a time frame of two cycles on the x-axis and displaying the
probability of a spike in particular time bins within the cycle. The stimulus for the insets was at the frequency-level
combinations indicated by the associated filled, inverted triangles. The phase relationship of spike occurrence to the period of
the tonal stimulus changes as a function of frequency, but at the characteristic frequency (the low point on the curve) the
phase relationship does not vary with changes in sound intensity (dB SPL). (b) Tuning curves for four ANFs of different
characteristic frequency. (c) When vector strength (a measure of the degree of phase locking, on a scale of 0 to 1, derived
from phase histograms like those shown in (a)) is plotted as a function of stimulus frequency around characteristic frequency
for 73 different afferent fibers, it is apparent that cochlear ANFs in the emu exhibit a sharp decline in phase-locked firing at
frequencies above 1 kHz. Reprinted with permission from Manley, G. A., Köppl, C., and Yates, G. K., J. Acoust. Soc. Am.,
Vol. 101, pages 1560-1573, 1997. Copyright 1997, American Institute of Physics.

information about small signals. Because several steps
in auditory reception involve discrete all-or-none
events (e.g., the opening of individual ion channels,
the fusion of individual synaptic vesicles (SVs) with
the plasmalemma, and the generation of individual
action potentials in ANFs), the presence of thermally
driven noise in the system is crucial to maintaining
high sensitivity to small sounds (Indresano, A. A. et al.,
2004). This phenomenon has been studied in many
physical systems under the odd name of stochastic
resonance.
At frequencies above a few kilohertz, where sound
perception relies on nonlinear transduction at the
foot of the stimulus–response curve to produce the

The Senses: A Comprehensive Reference, vol. 3, pp. 365-395

 Author's personal copy
DC component of the receptor potential (see Section
3.22.6), a nonlinear response to small sounds may be
unavoidable.
Fibers that are the most sensitive to sound also
have the highest SR. The equivalent statement in
terms of threshold is that SR varies inversely with
threshold. In cats, ANFs have been divided into three
classes based on SR: low-SR fibers (<0.5 Hz) have the
highest threshold (20–60 dB higher than high-SR
fibers), high-SR fibers (18 Hz) have the lowest
thresholds, and medium-SR fibers have an inter-
mediate sensitivity (Merchan-Perez, A. and
Liberman, M. C., 1996). Because ANFs with the
same CF can have high-, intermediate-, or low-SR,
it seems likely that a single IHC forms synapses with
representatives of all three functional groups.
Unlike many types of neurons that fire either with
a fairly constant interspike interval or in bursts, the
exponential distribution of interspike intervals for
baseline activity in mammalian ANFs suggests that
their activity in the absence of sound is driven by
independent random events (i.e., it is a Poisson pro-
cess), except during a neural refractory period
following each spike (Johnson, D. H. and Kiang, N.
Y. S., 1976) rather than by an oscillation in the pre- or
postsynaptic membrane potential or by any of the
other well-known mechanisms that underlie bursting
or rhythmic firing in neurons. The independent
events may be the brief openings of individual pre-
synaptic calcium channels, in which case the afferent
synapses onto ANFs with higher SRs could have
more calcium channels or a higher probability of
transmitter release in response to the opening of a
single calcium channel. Another possibility to explain
the correlation between an ANF’s sensitivity and its
SR is that the voltage dependence of the calcium
channels at high-SR synapses is shifted such that
the channel’s resting open probability is increased,
which would also move the resting potential to a
steeper place on the channel’s sigmoidal activation
curve. In all of these cases, high SR would be asso-
ciated with greater sensitivity to sound. Since a single
hair cell can make synapses onto several ANFs that
have widely varying SRs, these differences must be
specific to individual synapses.
If the Poisson events that trigger exocytosis are
the openings of individual calcium channels, then a
simple calculation shows that the probability of exo-
cytosis associated with each channel opening is likely
to be very small. Calcium channels in hair cells have
mean open times of <1 ms and a steady-state open
probability of 10% at the resting potential,
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 371
implying a mean closed time of <10 ms. If calcium
channels in IHCs have properties similar to these,
then each channel is expected to generate >100
openings per second at the resting potential, and the
ensemble of 100 calcium channels at a typical AZ
are expected to generate >104 openings per second.
Given that the highest SRs in ANFs are 100 s1,
fewer than 1% of calcium channel openings are pre-
dicted to trigger exocytosis.
3.22.2.3 Innervation Patterns and the
Function of Auditory Organs
The pattern of afferent innervation onto hair cells
contributes greatly to the response properties
observed in afferent fiber recordings. To illustrate
this, we will consider two very different auditory
organs of the inner ear: the mammalian cochlea and
the frog sacculus.
As mentioned above, sensory information from a
single IHC in the mammalian cochlea diverges onto
multiple ANFs, which can be regarded as a set of
parallel channels that relay information from a single
IHC to the brain. Because each IHC has a slightly
different CF, and a slightly different phase relation-
ship to the sound, restricting each ANF’s input to a
single IHC preserves these two important features of
the cochlear tonotopy. Having multiple outputs from
each IHC with different sensitivities (range fractio-
nation) may be important for maintaining a high
signal-to-noise ratio for faint sounds in some ANFs
while avoiding saturation by loud sounds in other
ANFs. Even if some of the ANFs have identical
response properties, maintaining redundant informa-
tion channels can be important for increasing the
signal-to-noise ratio. Two important sources of
noise are uncertainty in when (or whether) a spike
will be generated in a given ANF and variation in the
conduction time of ANF spikes to the brain. This
noise can, in principle, be reduced by combining
information transmitted over separate channels that
have independent noise sources. Each ANF postsy-
naptic to a given IHC shares a common input signal
(the hair cell’s membrane potential), but the subse-
quent steps of synaptic transmission, spike generation
and conduction to the brain are carried out indepen-
dently in each AFN, and therefore are subject to
independent noise. The innervation pattern of the
mammalian cochlea thus appears to be optimized to
provide the spectral and temporal information
needed for complex computations involved in speech
recognition and generating a map of auditory space.
 Author's personal copy
372 Afferent Synaptic Mechanisms
The frog sacculus has a very different structure
and pattern of innervation (reviewed in Smotherman,
M. S. and Narins, P. M., 2000). It is (in addition to its
function as a vestibular organ) exquisitely sensitive to
both airborne sound and substrateborne vibration
(Koyama, H. et al., 1982). The frog saccular hair
cells reside in a roughly circular patch of sensory
epithelium beneath the otolithic membrane, and all
are attached to it through their kinocilia. There is no
evidence of tonotopic organization. Instead, the hair
cells fall into only a few anatomical classes, suggest-
ing that many may be functionally equivalent to each
other (Edmonds, B. et al., 2000). Unlike the mamma-
lian cochlea, each saccular afferent neuron branches
to contact many hair cells (Lewis, E. R. et al., 1982).
Summation of synaptic inputs from multiple hair
cells is expected to increase the sensitivity and, per-
haps more important, the signal-to-noise ratio of
ANFs from the frog sacculus, which is expected to
contribute to the organ’s extreme sensitivity to sound
and vibration.
3.22.3 Hair Cells
Hair cells are part of the tight epithelium that sepa-
rates endolymph from perilymph. Each hair cell is
sealed to its neighbors by tight junctions that form an
extracellular diffusion barrier across the epithelium
and divides the cell membrane into an apical domain
that is primarily involved in mechanosensory trans-
duction and a basolateral domain where the synapses
reside (Figure 1(b)). Hair cells are both pre- and
postsynaptic, and have other neuronal properties,
including voltage- and calcium-gated ion channels
that, in concert with the transduction current and the
membrane’s electrical capacitance, determine how
the membrane potential changes with time. The
hair cell’s compact shape has important consequences
for both the electrical signal and the ensuing calcium
signal that controls transmitter release at afferent
synapses.
3.22.3.1 Hair Cell Shape in Relation to
Synaptic Signaling
The hair cell body has a diameter-to-length ratio that
is too large to allow significant spatial variation in the
membrane potential, except in OHCs of the mamma-
lian cochlea, where extensive internal membranes
lining the plasmalemma might impede intracellular
current flow (Nakagawa, T. et al., 2006). Other types
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
of hair cells, including IHCs of the mammalian
cochlea, are expected to be nearly isopotential even
to the tips of the stereocilia (Roberts, W. M. et al.,
1990). This means that the electrical signal for synap-
tic transmission is the same at each of the afferent
synaptic sites, and implies that the different properties
of the afferent synapses on a single hair cell must arise
downstream of the electrical signal. Similarly, the
membrane capacitance is expected to attenuate high-
frequency voltage oscillations to the same degree at all
AZs in a given hair cell, leaving only a steady depolar-
ization in response to high-frequency sounds (Section
3.22.6).
The hair cell’s large radius of membrane curva-
ture compared to the diameter of a single synaptic
AZ (several micrometers vs. hundreds of nanometers)
simplifies the understanding and modeling of afferent
synaptic transmission because for some purposes the
membrane can be regarded as a planar boundary
separating two infinite spaces (Neher, E. and
Augustine, G. J., 1992; Roberts, W. M., 1994).
Clearly, the hair cell’s 1 pl volume is less than
infinite and cannot serve as an infinite sink for cal-
cium entering at synapses, but as long as calcium
extrusion can keep pace with calcium entry, the
details of how and where calcium is pumped back
into the extracellular space are expected to be less
important than in the confined space of a typical
presynaptic bouton.
3.22.3.2 The Synaptic Body and Synaptic
Vesicle Pools
Afferent synapses on all hair cells and related elec-
troreceptors are distinguished by a prominent
presynaptic structure, the synaptic body (SB). One
or more SBs are found at each site of synaptic contact,
in close association with the presynaptic plasma
membrane (Figures 3 and 4). Numerous functions
have been proposed for the SB (Sjostrand, F. S.,
1958; Bunt, A. H., 1971; Gray, E. G. and Pease, H.
L., 1971; Osborne, M. P. and Thornhill, R. A., 1972;
Wagner, H.-J., 1978; King, T. S. and Dougherty, W. J.,
1982; Schmitz, F. et al., 2000; Lenzi, D. and von
Gersdorff, H., 2001; Parsons, T. D. and Sterling, P.,
2003), but none proved.
The hair cell contains a large population of
40 nm, clear-core vesicles that are presumed to
contain the neurotransmitter, glutamate. These SVs
have been partitioned into several conceptual pools,
defined either structurally with EM or functionally
 Author's personal copy
Afferent Synaptic Mechanisms 373

(a) (b) (c)

200 nm
(d) (e)

200 nm

Figure 3 Afferent synapse ultrastructure. All panels are electron micrographs of afferent synapses in frog saccular hair
cells. (a) Each site of synaptic contact between a hair cell (top) and afferent nerve ending (bottom) typically contains one
synaptic body, which is darkly stained and surrounded by 40 nm-diameter vesicles. A second synaptic body (SB) is faintly
visible to the left. The region of close apposition between pre- and postsynaptic membranes is marked by a fuzzy appearance
that extends 100–400 nm beyond the edge of the synaptic body, but the presumed region where synaptic vesicles release
their contents (the active zone) is restricted to the area beneath the SB and perhaps 100 nm beyond. The presynaptic
membrane at the active zone contains several densities that anchor the SB. (b) The SB and its attached vesicles remain
anchored to the presynaptic active zone after the cell is broken open in extracellular saline (contains 1.8 mM calcium) and the
cytoplasm washed away prior to fixation. (c) A freeze-fracture electron micrograph shows rows of particles in the presynaptic
membrane that may align with the presynaptic densities and correspond to ion channels. The membrane dimples around the
periphery of the active zone are most likely the openings of membrane tubules rather than sites of vesicle fusion, which are
expected to occur closer to the ion channels at the active zone. (d, e) Filamentous links tether the SB to the cytomatrix at the
active zone (short arrows) and synaptic vesicles to the SB (long arrows). Synapses that had been inhibited before and during
fixation (d) had a high density of docked vesicles in the space between the SB and the plasmalemma (arrowhead points to
one). Synapses that had been stimulated before and during fixation (e) had few docked vesicles and numerous cisternal
structures in the cytoplasm. (c) Reproduced with permission from Roberts W. M. 1994. Localization of calcium signals by a
mobile calcium buffer in frog saccular hair cells. J. Neurosci. 14, 3246–3262. Copyright by the Society for Neuroscience. (d,e)
Reprinted from Lenzi, D., Crum, J., Ellisman, M .H., and Roberts, W. M. 2002. Depolarization redistributes synaptic membrane
and creates a gradient of vesicles on the synaptic body at a ribbon synapse. Neuron 36, 649–659, Copyright (2002) with
permission from Elsevier.

with electrophysiological assays of SV fusion and
neurotransmitter release.
Experiments on a wide variety of chemical
synapses in many animals have shown that SVs can
generally be assigned to one of three groupings: the
readily releasable pool (RRP), the recycling pool, and
the reserve pool (reviewed in Rizzoli, S. O. and Betz,
W. J., 2005). Parallel definitions of these pools at hair
cell afferent synapses are useful to facilitate compar-
isons with more conventional chemical synapses,
although this is hampered somewhat by their
different anatomical and functional characteristics.
At both types, maximal activation of the presynaptic
calcium current exhausts a small SV pool, defined as
the RRP, within a few milliseconds. At conventional
synapses, the RRP comprises 1–2% of the total vesi-
cle population, but considering the hair cell
cytoplasm as an unrestricted volume for SV move-
ment containing hundreds of thousands of SVs
(Lenzi, D. et al., 1999; Holt, M. et al., 2004), the RRP
in hair cells comprises a miniscule fraction of the
total (Moser, T. and Beutner, D., 2000; Schnee M.
E. et al., 2005; Rutherford, M. A. and Roberts, W. M.,
2006).

The Senses: A Comprehensive Reference, vol. 3, pp. 365-395

 Author's personal copy
374 Afferent Synaptic Mechanisms
(a) Outlying vesicles ( ) concentrate at synapses
Synaptic Nonsynaptic
5.2% of dense-packed 0.4% of dense-packed
(b) (c)
(d) (e)
Figure 4 Tomographic reconstructions of afferent
synapses in frog saccular hair cells. (a) Tethered vesicles
(yellow) surround the SB (blue), which lies close to the
plasmalemma (red) at the Active zone (AZ). Other (outlying)
synaptic vesicles (green) are also highly concentrated within
a few hundred nanometers of the SB, and populate the rest
of the cytoplasm at much lower density. (b) An inhibited
synapse shows mostly spherical vesicles around the SB. (c)
A stimulated synapses has numerous irregular membrane
structures near the SB that could be the result of compound
vesicle fusion and/or membrane retrieval. (d, e)
Tomographic reconstructions give a clear view of the more
extensive cisternal structures (magenta) at stimulated
synapses (e) compared to inhibited synapses (d). Coated
vesicles (orange) and mitochondria (dark blue) are also
shown. Other colors are the same as in (a). (a) Reproduced
with permission from Lenzi, D., Runyeon, J. W., Crum, J.,
Ellisman, M. H., and Roberts, W. M. 1999. Synaptic vesicle
populations in saccular hair cells reconstructed by electron
tomography J. Neurosci. 19, 119–132; Copyright by the
Society for Neuroscience. (d,e) Reprinted from Lenzi, D.,
Crum, J., Ellisman, M. H., and Roberts, W. M. 2002
Depolarization redistributes synaptic membrane and
creates a gradient of vesicles on the synaptic body at a
ribbon synapse. Neuron 36, 649–659; Copyright (2002) with
permission from Elsevier.
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
At conventional synapses, the RRP is generally
thought to correspond to the docked vesicles, those
seen in EM to be in contact with the plasmalemma at
the AZ. The corresponding pool of morphologically
docked vesicles at ribbon synapses consists mostly of
vesicles that occupy the space between the SB and
the plasmalemma at each AZ (Figures 3 and 4).
Differences in the precise criteria used for anatomical
classification of vesicles can cause estimates of the
size of the docked vesicle pool to differ by a factor of
two, even in careful quantitative studies.
There is ambiguity about how to classify vesicles
that are tethered to the SB but do not contact the
plasma membrane. Direct evidence that these vesi-
cles even participate in transmitter release under
normal conditions is scant (Siegel, J. H. and
Brownell, W. E., 1986), but assuming that they do,
are they part of the RRP, the recycling pool, or some
other pool? One reason to think that even though
they reside far from the plasmalemma they might still
be in the RRP is that they may be able to fuse with
each other, forming a chain that connects to the cell
surface (reviewed in Parsons, T. D. and Sterling, P.,
2003; Edmonds, B. W. et al., 2004). Figure 4(c) shows
what may be a chain of fused vesicles. If the morpho-
logical criterion that has been used to determine
whether a vesicle is docked to the plasmalemma is
applied to vesicles on the SB, then many of these
vesicles are morphologically docked to each other,
and thus could conceivably be ready to fuse together,
although the calcium signal is expected to be highly
attenuated more than a few vesicle diameters from the
calcium channels on the cell surface (Roberts, W. M.,
1994). Alternatively, the type of structures shown in
Figure 4(c), which were observed after extended
depolarization in high potassium (Lenzi, D. et al.,
2002), could reflect membrane retrieval.
The available physiological evidence indicates
that the RRP in the frog sacculus and mouse cochlea
is much smaller than the total pool of SB-associated
vesicles, and is close to the number of vesicles docked
at the plasmalemma (Khimich, D. et al., 2005;
Rutherford, M. A. and Roberts, W. M., 2006), which
supports the more conventional hypothesis that only
vesicles docked at the plasmalemma are part of the
RRP, and that other SB-associated vesicles are part of
the recycling vesicle pool that replenishes the RRP.
Physiological correlates of these anatomical defini-
tions are discussed in more detail in the following
sections of this chapter.
Tens of thousands of clear-core vesicles are found
in the hair cell cytoplasm within several hundred
 Author's personal copy
nanometers of the cell surface, and hundreds of thou-
sands more reside in the vast cell interior. They
occur at a low average density in frog saccular hair
cells (0.4% of dense packed ¼ 100 vesicles per
mm3), but are much more concentrated within
700 nm of each synapse (5–10% of dense packed ¼
1200–2400 vesicles per mm3; Lenzi, D. et al., 1999;
2002). The contribution of these vesicles to synaptic
transmission under normal physiological conditions
is unclear. At conventional synapses, there is a phy-
siologically distinct (but sometimes not anatomically
distinct; Rizzoli, S. O. and Betz, W. J., 2005) recycling
pool of SVs that are reused preferentially by local
endocytosis (Pyle, J. L. et al., 2000; Stevens, C. F. and
Williams, J. H., 2000), whereas vesicles in the reserve
pool enter the exocytic pathway only under strong
stimulus conditions that also cause bulk endocytosis
and vesicle reconstitution through an endosomal
compartment. It remains unknown whether hair cell
afferent synapses have distinct recycling and reserve
SV pools.
The AZ at most brain synapses contains only a few
docked vesicles, and a single presynaptic action poten-
tial seldom evokes exocytosis of neurotransmitter
from more than one of them, with frequent failures
(Dobrunz, L. E. and Stevens, C. F., 1997; for a review
see Sudhof, T. C., 2000). In contrast, the docked vesi-
cle pool is considerably larger at AZs in hair cells of
the turtle papilla (mean of 30; Schnee, M. E. et al.,
2005), mouse cochlea (16–30, IHC; Khimich, D. et al.,
2005), and the frog sacculus (mean of 43; Lenzi, D.
et al., 2002). Furthermore, both pre- and postsynaptic
assays of SV exocytosis (see Sections 3.22.5–3.22.7)
have indicated that more than one SV can fuse in
response to a single, brief presynaptic stimulus
(reviewed in Fuchs, P. A., 2005). The stimulus to
open voltage-gated calcium channels (VGCCs) in
hair cells is not an all-or-none action potential that
lasts on the order of 1 ms, rather, they respond to
depolarizations of varying duration and intensity
with graded release of neurotransmitter. Thus, while
some generalizations can be made across all synapse
types, there may be important differences between SV
pools at conventional synapses and hair cell afferent
synapses.
3.22.3.3 Calcium Buffering in Hair Cells
Exogenous calcium buffer can be introduced into a hair
cell through a micropipette recording electrode, and
has been shown to influence synaptic transmission.
Interestingly, a high concentration of calcium buffer
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 375
with fast binding kinetics is required to interfere with
exocytosis from the RRP, whereas refilling of the RRP
can be inhibited by low concentrations of slower buffer
(Moser, T. and Beutner, D., 2000). It is concluded that
although both processes are regulated by calcium,
refilling of the RRP involves signaling over a longer
distance, and is therefore more susceptible to disrup-
tion by calcium buffers. To understand the interaction
between calcium ions and calcium buffers that can
diffuse in cytoplasm one must think about the rates of
binding and diffusion, not just equilibrium binding of
calcium to the buffer. A brief introduction to the essen-
tial concepts needed to understand the nonequilibrium
interaction between a mobile calcium buffer and the
calcium ions that enter through a calcium channel and
diffuse into a large cytoplasmic volume is given below.
High concentrations of calcium-binding proteins
can limit the spread of calcium signals within a cell
(Roberts, W. M., 1994). This may be particularly
important in hair cells, because they use calcium sig-
nals for several distinct cellular processes, including
feedback inhibition of the transduction current, which
involves calcium entry and action at the stereociliary
tips (see Chapter Hair Cell Transduction and
Adaptation: Physiology and Molecular Mechanisms),
and efferent synaptic transmission see Chapter Efferent
System at postsynaptic sites, which intermingle with
afferent presynaptic sites on the hair cell’s basolateral
surface. In addition to localized transmembrane cal-
cium signals, hair cells also exhibit localized calcium-
induced calcium release from intracellular stores near
sites of efferent contact (Lioudyno, M. et al., 2004).
The equilibrium interaction between calcium and a
molecule that has a single calcium-binding site is char-
acterized by a single number, the dissociation constant,
KD (units: molar, M), or its reciprocal, KA (M1):
½Ca½B
KD ¼
½CaB
where [Ca] and [B] are the concentrations (M) of free
Ca and free buffer, and [CaB] is the concentration of
calcium bound to the buffer. Free means that they are
able to interact with each other (i.e., not already
bound to something else, hidden inside an organelle,
etc.); whether or not they are free to diffuse in the
cytoplasm is a separate issue that is also important.
Time does not appear in the units of these variables.
It is assumed that enough time has elapsed for equi-
librium to be attained.
The equilibrium equation does not apply near an
open calcium channel because equilibrium is never
 Author's personal copy
376 Afferent Synaptic Mechanisms
reached while the calcium current persists (which
requires an extrusion mechanism to keep the cell
from filling up with calcium). The existence of a
calcium current is prima facie evidence of disequili-
brium, at least locally around the mouth of the
channel. We will not specifically consider the cal-
cium pump here, except to note that the maximum
flux of calcium ions through a single pump
(103 s1) is roughly one ten thousandth the flux
through an open calcium channel (107 s1), so that
the local calcium gradient around a single pump is
miniscule compared to the gradient near the mouth
of a single calcium channel. Therefore, it makes sense
to consider the calcium extrusion process to be dis-
tributed over a large region of cell surface, and ignore
its effect on the localized region of extremely high
free calcium near an open calcium channel.
Two parameters, kon (M1 s1) and koff (s1), are
needed to describe the kinetics of calcium binding to
a single-site buffer. kon[B] and koff are, respectively,
the binding kinetics are important and the buffer’s
effects can be complex. The diffusion coefficient for
the buffer in cytoplasm is also important because
slow diffusion reduces the rate at which calcium-
free buffer is replenished near the open channel,
which reduces the local [B], and in turn locally
increases both  and . At hair cell afferent synapses,
exocytosis of RRP vesicles falls into the first cate-
gory, where cytoplasmic buffers are irrelevant, but
some other important calcium-dependent steps in the
SV cycle appear to fall into the complex category,
where buffer concentration, binding kinetics, and
diffusion coefficient all come into play.
To assign approximate values to  and  one
needs to know the concentration and kon for the
predominant diffusionally mobile calcium buffer(s)
in the cell. Many types of hair cells are intensely
labeled by fluorescent antibodies against one or
more of the diffusible calcium-binding proteins
calbindin-D28k, calretinin, parvalbumin-, and par-
the rate constants for calcium binding to and unbind-
ing from the buffer. The three parameters, kon, koff,
and KD are related by KD ¼ (koff/kon), so any two
specify the third.
A calcium ion is initially free as it enters the cell
valbumin- (also known as oncomodulin or PV-3).
For example, some hair cells in the frog sacculus
contain calretinin at a concentration above 1 mM,
which provides more than 5 mM calcium-binding
sites (Edmonds, B. et al., 2000). Other frog saccular
through a calcium channel. The mean time required
to encounter and bind to a buffer molecule is
1/(kon[B]), which for the buffering conditions
encountered in many hair cells ([B] >> KD) is close
to the equilibration time constant () in response to a
step change in free calcium:
hair cells contain predominantly parvalbumin , at
the extraordinary concentration of 4 mM (Heller, S.
et al., 2002), which amounts to 8 mM calcium-binding
sites. OHCs in the mammalian cochlea contain different
calcium-binding proteins than are found in the IHCs,
and appear to express them at a tenfold higher concen-
1 to 5 mM calcium-binding sites; Hackney, C. M. et al.,
¼
kon ð½B þ KD Þ
During this time, a free calcium ion diffuses a
characteristic distance of
pffiffiffiffiffiffiffiffiffiffi
¼ DCa
where DCa is the diffusion coefficient for free calcium
in cytoplasm (Neher, E. and Augustine, J. G, 1992;
Roberts, W. M., 1993). The parameters  and  are,
respectively, the time and space constants for calcium
binding to the buffer in the vicinity of an open
calcium channel.
If the distance between the presynaptic calcium
channels and the calcium sensors that trigger exocy-
tosis is substantially less than  then the buffer has
little effect. If the distance is substantially greater
than  then the buffer has time to come into equili-
brium with global free calcium and binding kinetics
are not important. If the distance is similar to  then
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
tration (calbindin-D28k and parvalbumin- equivalent
2005). These extreme concentrations of calcium buf-
fer are certain to have important consequences for
many calcium signals in the cell. It is also worth
remembering that some calcium-binding proteins
may serve both as buffers and as calcium sensors
that influence exocytosis or participate in other
steps in the SV cycle.
For typical values of KD ¼ 1 mM and
kon ¼ 107 M1 s1, the buffer space constant and
time constant in a cell that has a moderate calcium
buffer concentration of [B] ¼ 100 mM are  ¼ 450 nm
and  ¼ 1 ms. Increasing the buffer concentration
100-fold to 10 mM decreases  to 44 nm and  to
10 ms. From this calculation, one can see that an
extremely high concentration of a calcium buffer
can reduce the buffering space constant from
approximately the diameter of the entire AZ
(Figure 3(c)) to approximately the diameter of a
single SV, thereby minimizing the overlap of calcium
 Author's personal copy
domains around each open calcium channel, particu-
larly at barely perceptible sound intensities where
only a few calcium channels are open at each afferent
synapse. The close association between calcium
channels and AZs (Lewis, R. S. and Hudspeth, A. J.,
1983; Art, J. J. and Fettiplace, R., 1987; Roberts, W. M.
et al., 1990; Issa, N. P. and Hudspeth, A. J., 1994;
Martinez-Dunst, C. et al., 1997; Rodriguez-
Contreras, A. and Yamoah, E. N., 2001; Zenisek, D.
et al., 2003; Brandt, A. et al., 2005) means that steep
spatial calcium gradients and the nonequilibrium
calcium binding to buffer are important at afferent
synapses.
At distances on the order of  from an open
channel, the buffer serves as a calcium sink, and can
greatly reduce the free calcium concentration
(Roberts, W. M., 1994). At larger distances, where
free calcium falls below the buffer’s KD, the buffer
becomes a calcium source as it dumps its bound
calcium, although the large volume available to
receive this calcium means that the calcium rise
associated with unbinding from the buffer is orders
of magnitude smaller than the calcium concentration
near the source. Because the buffer molecules are
large compared to a calcium ion, bound calcium
diffuses more slowly than free calcium. The net
result is that the buffer can greatly increase total
calcium near an open calcium channel, and can
slightly increase free calcium at certain distances
from the channel, but its main effect is to greatly
reduce the extremely high concentration of free cal-
cium that accumulates near the channel (i.e., at
distances on the order of ).
More detailed discussions of these and other prop-
erties of mobile calcium buffers are in the literature
(e.g., Stern, M. D., 1992; Roberts, W. M., 1993; 1994;
Dargan, S. L. and Parker, I., 2003). One interesting
result is that the average distance that a calcium ion
travels while remaining bound to the buffer is >1 mm,
which is large compared to the AZ diameter. Thus,
the mobile buffer gives a calcium ion only one brief
chance to exert its effect. Either it binds to its target,
such as the calcium sensor that triggers exocytosis, or
it is captured by the mobile buffer and is carried
away. If the buffer has reasonably fast binding
kinetics (kon ¼ 107 M1 s1) and provides 5–10 mM
of mobile binding sites for Ca (Edmonds, B. et al.,
2000; Heller, S. et al., 2002), then the average time
available for a calcium ion to encounter and bind to
its target (i.e., the binding time constant, ) is only
10–20 ms.
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 377
3.22.4 Ribbon Synapses – A Small
but Diverse Class
The SBs of hair cells are homologous to the presy-
naptic ribbons found in rods, cones, and bipolar cells
in the retina, and pinealocytes in the pineal body.
Like hair cells, these cells usually do not make action
potentials. Because of their anatomical, physiological,
and molecular similarities, we include the afferent
synapses of hair cells in the class of ribbon synapses,
but one should also keep in mind the differences,
including, for example, the slow-time scale of the
visual system compared to the auditory system.
Ribbon synapses are understood to be specialized
for sustained transmitter release. For example, verte-
brate photoreceptors release neurotransmitter
continuously in the dark (Rea, R. et al., 2004) and
the hair cell afferent synapse is active in the absence
of an applied stimulus (Siegel, J. H., 1992). However,
the hair cell is also capable of releasing transmitter
phasically (e.g., EPSPs that are correlated with the
periodicity of the stimulus for afferent fibers at fre-
quencies up to a few kilohertz; Figures 1(c) and 2(c),
see also Section 3.22.6). The details of synaptic phy-
siology that vary among ribbon synapses from
different tissues and cell types are likely to be essen-
tial for their function (for a review, see Sterling, P.
and Matthews, G., 2005). Nevertheless, the first
synapse in the pathways for vision, hearing, balance,
and lateral line senses in all vertebrate animals is a
ribbon synapse.
3.22.4.1 The Synaptic Body (Ribbon) and
Its Vesicles
The term ribbon originates from the flat, almost 2D
morphology of the presynaptic structures observed in
transmission EMs of retinal photoreceptors and bipo-
lar cells. Serial reconstructions of gluteraldehyde-
fixed hair cells from many preparations, including
mammalian cochlea, have shown that the SB in hair
cells has a more spherical or ellipsoid shape. IHCs
rely exclusively on ribbon synapses for afferent trans-
mission. The role of SBs in OHCs is more uncertain
as they (the SBs) are far fewer and often seen only
during development (Smith, C. A. and Sjostrand, F. S.,
1961; Siegel J. H. and Brownell W. E., 1981;
Sobkowicz, H. M. et al., 1986).
Although SBs are heavily stained by osmium,
which typically labels lipid membranes, they are
thought to be composed primarily of protein
 Author's personal copy
378 Afferent Synaptic Mechanisms

(Krstic, R., 1976). The osmiophilic nature of the SB is
well-known because OsO4 is a common EM fixative
and is electron-dense. Perhaps more relevant to func-
tion, the ribbon has been reported to bind both
calcium and phosphate (Schmitz, F. and
Drenckhahn, D., 1991), although there is no evidence
that it binds or releases these ions in the process of
synaptic transmission.
SBs at AZs are always found in association with
40 nm-diameter clear-cored vesicles, which are
attached to it by thin filaments (Usukura, J. and
Yamada, E., 1987) and often cover its surface nearly
completely (Figures 3 and 4; Lenzi, D. et al., 2002).
SBs can sometimes be found in the cell interior,
either singly or in clusters, where they are also cov-
ered in vesicles (Miller, M. R. and Beck, J., 1987). The
molecular identity of the filaments is not known,
except that they are thought not to contain synapsins
I or II, which form filaments linking SVs to cytoske-
leton at other synapses (Hirokawa, N. et al., 1989), but
are absent from ribbon synapses (von Kriegstein, K.
et al., 1999). Based on antibody labeling, the filaments
could contain the KIF3A subunit of the kinesin 2
motor complex (Muresan, V. et al., 1999), which
forms 50 nm-long rods in vitro (Yamazaki, H. et al.,
1995), or possibly one or more of the many other
KIF3 isoforms that may cross react with the antibody.
SBs and ribbons can be regarded as elaborations of
the cytomatrix at the active zone (CAZ) seen at all
synapses (Zhai, R. G. and Bellen, H. J., 2004), but in
EM the SB appears as a separate structure, connected
to the CAZ by filamentous links (Usukura, J. and
Yamada, E., 1987; Lenzi, D. et al., 2002). The tight
association between the SVs and the SB, and between
the SB and the presynaptic plasma membrane is
evident in cells that were broken open prior to fixa-
tion (Figure 3(b)). If the binding were weak or
disrupted by exposure to the high-calcium concen-
tration of extracellular saline, these structures would
not have remained behind while the rest of the cyto-
plasm was washed away.
3.22.4.2 Molecular Specializations of
Ribbon Synapses
Understanding of the molecular composition of hair
cell afferent synapses has begun (Figure 5), but is far
from complete (Wagner, H.-J., 1997). A protein, Ribeye

(a) (b) Ribbon

RIBEYE/CtBP2
Hair cell CtBP1
KIF3A
Piccolo
RIM1

CAZ
Ribbon Bassoon
RIM2
VGCCs
Munc13-1
Glast BKs
CAST1

Observed
in hair cells
AMPARs
Support Inferred from
cell Afferent bouton retinal ribbons

Figure 5 Molecular specializations of the hair cell ribbon synapse. (a) Localization of proteins associated with the synaptic
ribbon and plasma membrane density of the cytomatrix at the active zone (CAZ), labeled in (b). Voltage-gated calcium channels
(VGCCs) are clustered underneath the ribbon and juxtaposed to postsynaptic -amino-3-hydroxy-5-methylisoxazole-4-
propionic acid (AMPA)-type glutamate receptors. In some hair cells, presynaptic calcium-sensitive voltage-gated (BK)
potassium channels intermingle with the VGCCs. Glutamate transporters in supporting cells clear neurotransmitter from the
cleft. Bassoon anchors the ribbon to the Cytomatrix at the active zone (CAZ) plasma membrane density via its interaction with
the RIBEYE. Bassoon, RIBEYE/CtBP2, CtBP1, and Piccolo have been observed at hair cell afferent synapses. The existence of
the other proteins at the hair cell’s ribbon synapse is by inference from observations at retinal ribbon synapses (tom Dieck, S.
et al., 2005). Reproduced from The Journal of Cell Biology, 2005, 168, 825–836 by Copyright permission of The Rockefeller
University Press; and Reprinted from Curr. Opin. Neurobiol. Vol. 13, Fuchs, P. A., Glowatzki, E., and Moser, T., The afferent
synapse and cochlear hair cells, 452–458, Copyright 2003, with permission from Elsevier.

The Senses: A Comprehensive Reference, vol. 3, pp. 365-395

 Author's personal copy
(Schmitz, F. et al., 2000), has been identified that
appears to be a component of all synaptic ribbons,
and is expressed nowhere else. This finding justifies
exclusion of T-bar synapses of arthropods (e.g.,
Drosophila) from the class of ribbon synapses because
the Drosophila genome contains no DNA sequences that
code for a homolog of the unique exon 1 of the Ribeye
protein (Bradford, Y. and Roberts, W. M., unpublished
observations). Ribeye is one of two known protein
products of the CtBP2 gene. The other product,
CtBP2, is a transcription regulator (reviewed in
Chinnadurai, G., 2002). A recent report suggests that
the CtBP1 protein, which is the product of a closely
related gene and also a known transcription regulator,
is a presynaptic component of both hair cell (Moser, T.,
personal communication) and retinal (tom Dieck, S.
et al., 2005) ribbon synapses, as well as conventional
synapses, raising the possibility that the ancestral CtBP
protein had a presynaptic function in addition to being
a transcriptional corepressor, and conceivably could
link synaptic activity to gene expression. The
CAZ-associated protein Bassoon was recently shown
to anchor the ribbon at the photoreceptor synapse via a
physical interaction with Ribeye (tom Dieck, S. et al.,
2005). Bassoon and the related protein Piccolo, are also
present at hair cell afferent synapses (Khimich, D. et al.,
2005). Bassoon appears to interact with both Ribeye
and CtBP1, which are both localized to ribbons even in
the absence of functional Bassoon protein.
Drosophila has a single CtBP gene that produces a
transcription regulator essential for embryonic
development (Nibu, Y. et al., 1998; Poortinga, G.
et al., 1998), but it is not known whether CtBP is
present at synapses in Drosophila. Similarly, the
Ciona genome (a urochordate) has a single CtBP
gene that lacks the coding sequence for Ribeye
(Bradford, Y. and Roberts, W. M., unpublished obser-
vations). The simplest evolutionary scenario
consistent with the data is that a gene duplication
event early in the vertebrate lineage, but after the
divergence of urochordates from vertebrates, gave
rise to paralogous genes, CtBP1 and CtBP2. The
CtBP2 gene then acquired an alternate first exon,
which became involved in aggregation of CAZ com-
ponents (Schmitz, F. et al., 2000) into large structures
(ribbons), with important, but largely unknown con-
sequences for afferent synaptic transmission in the
eye and ear. Ribeye expression became limited to the
small class of ribbon synapses, while CtBP1 continues
to be expressed at other synapses. An additional gene
duplication gave rise to two CtBP2/ribeye genes in
teleost fishes (Wan, L. et al., 2005).
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 379
Besides Ribeye, only one other protein, the B16
antigen, has been proposed to be a unique component
of all ribbon synapses. This protein is known only as an
antigenic epitope (amino sequence: DTYQHPPKDS;
Balkema, G. W., 1991; Nguyen, T. H. and Balkema, G.
W., 1999). A search of the human genome and
expressed sequence tag (EST) databases shows that
only one gene, aconitase, contains a nucleotide
sequence that codes for this exact amino acid sequence.
Aconitase is a transcription regulator and a regulator of
iron metabolism, but reportedly is not present in
synaptic ribbons (Nguyen, T. H. and Balkema, G. W.,
1999). Human CtBP2 and Ribeye contain a sequence
(TYxxPP) near the C-terminus that is part of the B16
antigen and thus might bind the B16 antibody, but
synaptic ribbons in fish reportedly also bind the B16
antibody (Balkema, G. W., 1991) even though this
sequence is entirely absent from their CtBP2 and
Ribeye proteins. Thus, the component of the ribbon
that binds antibodies raised against the
DTYQHPPKDS peptide remains unknown.
There are other proteins that may be unique to
ribbon synapses but are not expressed at all ribbon
synapses. The 1D calcium channel is found only in
hair cells and is essential for normal hair cell Ca2þ
current, neurotransmitter release, and for hearing
(reviewed in Moser, T. et al., 2006a). Most synapses
express N, P/Q, or R-type calcium channels, but hair
cells express L-type calcium channels (Cav 1.3) formed
by this unique splice variant. This channel opens at
unusually hyperpolarized voltages (Koschak, A. et al.,
2001; Brandt, A. et al., 2003; Rodriguez-Contreras, A.
and Yamoah, E. N., 2003; Schnee, M. E. and Ricci, A. J.,
2003), which may cause the transmitter release respon-
sible for the baseline action potential rate (i.e., SR) in
afferent neurons. Retinal rods express a different L-type
calcium channel at ribbon synapses, Cav 1.4 (1F;
Morgans, C. W., 2001).
Another protein that may be unique to ribbon
synapses is complexin IV (Pabst, S. et al., 2000;
Reim, K. et al., 2006). Based on similarity of com-
plexin I and II, and the ability to rescue synaptic
transmission in complexin I/II knockout mice, both
complexin III and complexin IV are potential cal-
cium-dependent regulators of synaptic transmission.
Complexins I and II are known to bind to the soluble
NSF (N-ethylmaleimide-sensitive factor) attachment
receptor (SNARE) complex that forms the core of
the vesicle fusion machinery at many synapses
(McMahon, H. T. et al., 1995; Reim, K. et al., 2001;
Chen, X. et al., 2002; for a review, see Sudhof, T. C.,
2004). Complexins III and IV have both been
 Author's personal copy
380 Afferent Synaptic Mechanisms
described at ribbon synapses in the retina, but only
complexin IV was found to be unique to retinal
ribbon synapses. It was not reported whether com-
plexins III or IV are present in the ear.
Otoferlin is another protein that may be involved
in synaptic transmission by hair cells. This protein
was discovered by genetic mapping of familial deaf-
ness (Yasunaga, S. et al., 1999), and is essential for
hearing. The otoferlin gene is large and complex,
having 48 exons and many splice variants (Yasunaga,
S. et al., 2000). It is similar to the FER-1 protein of
Caenorhabditis elegans, which has been implicated in
vesicle membrane fusion (Achanzar, W. E. and
Ward, S., 1997). Otoferlin is a candidate to substitute
for synaptotagmin as the calcium sensor for transmit-
ter release (Yasunaga, S. et al., 2000).
Several other presynaptic proteins, including
synapsins and synaptotagmins, are notable by their
absence from ribbon synapses (Safieddine S.
and Wenthold R. J., 1999; von Kriegstein, K. et al.,
1999). Syntaxin 1, a component of the core SNARE
complex is also missing, replaced by syntaxin 3
(Morgans, C. W. et al., 1996).
3.22.4.3 Anatomy of the Synaptic Body:
Heterogeneity of Structure
SBs come in a bewildering variety of shapes and sizes.
Most synapses have one SB at the synaptic complex,
others have two or three (Liberman, M. C. et al., 1990).
In addition to the typically spherical or ellipsoidal
shapes of SBs in hair cells from mice, turtles, frogs,
and chicks (Sneary, M. G., 1988; Martinez-Dunst, C.
et al., 1997; Lenzi, D. et al., 1999; Khimich, D. et al.,
2005), SBs of different geometries and densities have
been observed in EM (Giannessi, F. and Ruffoli, R.,
1996). In many types of hair cells, the SB appears as a
uniformly dense blob, with no internal structure but
it also can have a laminar appearance, as in some
cochlear IHCs. In old world monkeys the SBs of
IHCs nearly always appear ring shaped (Bodian, D.,
1980). It is unclear how much of this variability is an
artifact of fixation and staining. For example, in the
frog retina, synaptic ribbons of a lamellar nature were
observed with conventional transmission electron
microscope (TEM) thin sections, but were not so
apparent when freeze-etching and freeze-substitu-
tion methods were employed for tissue preparation
(Usukura, J. and Yamada, E., 1987).
SBs vary in size and number among organs, within
and between species, as well as among cells within a
particular organ. Furthermore, SBs vary in size and
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
shape within individual cells. The dimensions of the
SB are typically submicron, ranging from 200 nm
for the shortest principle axis of the ellipsoid SBs in
mouse IHCs (Khimich, D. et al., 2005) to 400 nm in
hair cells of frog and goldfish sacculus (Gleisner, L.
et al., 1973; Hama, K. and Saito, K., 1977; Jacobs, R. A.
and Hudspeth, A. J., 1990; Lenzi, D. et al., 1999).
Estimates of the number of ribbon-containing
synapses per cell ranged from 10 in mouse cochlear
IHCs (Khimich, D. et al., 2005) to 85 in the high-
frequency hair cells of the turtle papilla (Sneary, M. G.,
1988).
Several studies have provided clues regarding the
function of SBs in hair cells by demonstrating that the
size and number of synaptic ribbons, as well as the
numbers of associated calcium channels and SVs, is
correlated with the position of the hair cell on the
tonotopic axis of the organ. In the turtle papilla, high-
frequency cells have three times more SBs, higher
vesicle densities near SBs, more calcium channels,
and larger calcium currents than do lower-frequency
cells (Sneary, M. G., 1988; Art, J. J. et al., 1995; Schnee,
M. E. et al., 2005). Upon depolarization, the high-
frequency cells released more vesicles and with faster
kinetics. A model of SV trafficking that incorporated
all of these data predicted that these differences
result in transmission being optimized to the CF of
the hair cell (Schnee, M. E. et al., 2005). In the papilla
of the chick, the diameter of the SB ranged from 120
to 210 nm with increasing CF from apical to more
basal positions along the tonotopic axis (Martinez-
Dunst, C. et al., 1997). These two studies used differ-
ent experimental approaches to reach a similar
conclusion; that the morphology of the SB has influ-
ence over the numbers of SVs that are maintained at
well-defined positions with respect to the source of
calcium influx. SVs close to the center of the AZ may
be more responsive to calcium influx if the channels
are clustered there (Furukawa, T. and Matsuura, S.,
1978; Furukawa, T. et al., 1982; reviewed for the case
of CNS synapses in Schneggenburger, R. and Neher,
E., 2005).
A study in the mammalian cochlea has reported
presynaptic morphological correlates of ANF SR and
threshold (Merchan-Perez, A. and Liberman, M. C.,
1996). Analysis of ultrastructure at the EM level
showed that the size and complexity of SBs tended
to increase as SR decreased, accompanied by an
increase in the total number of SVs clustered at the
AZ. These data suggest that the size of the SB has
something to do with the regulation of vesicle release
that determines the baseline afferent discharge rate
 Author's personal copy
and sensitivity to faint sounds, but the relationship is
opposite what is expected based on the prevailing
view that the ribbon collects vesicles and conveys
them to the release sites. Instead, these data seem to
suggest that a larger SB may somehow inhibit SV
fusion for weak sounds. Clearly, more work needs to
be done to understand the relationship between SB
morphology and postsynaptic properties.
3.22.5 The Synaptic Vesicle Cycle in
Hair Cells
The basic features of the SV cycle are similar
at conventional chemical synapses and ribbon
synapses. Both baseline and evoked EPSCs at ribbon
synapses require presynaptic calcium influx through
channels (Robertson, D. and Paki, B., 2002; Keen, E.
C. and Hudspeth, A. J., 2006) and are completely
and reversibly blocked by antagonists of the -ami-
no-3-hydroxy-5-methylisoxazole-4-propionic acid
(AMPA)-type glutamate receptor (Glowatzki, E.
and Fuchs, P. A., 2002; Keen, E. C. and Hudspeth,
A. J., 2006; Holt, J. C. et al., 2006). The calcium-
controlled release of glutamate into the synaptic
cleft is quantal, involving exocytosis from SVs that
fuse with the plasmalemma and are later retrieved for
reuse.
3.22.5.1 Endocytosis
Various routes for vesicle recycling at chemical
synapses have been proposed, including two contro-
versial types of fast, local reuse in which the
neurotransmitter leaves the SV through a fusion
pore without full collapse of the SV membrane into
the plasmalemma. In one of these hypothetical fast
recycling routes, named kiss-and-stay, the vesicle
remains docked to the AZ where it is presumably
refilled with glutamate after discharging its contents.
In the other, named kiss-and-run, the vesicle undocks
and enters a local recycling pool. Slower recycling
pathways involve full fusion of the SV with the
plasmalemma, followed by clathrin-dependent endo-
cytosis of vesicles or bulk membrane retrieval and
reprocessing of SVs either directly or via an endoso-
mal route (reviewed in Sudhof, T. C., 2004). The
extent to which each of these processes contributes
to the SV cycle under natural physiological condi-
tions remains unclear.
Recent work in the cochlea showed that IHCs
may be able to draw upon the vast supply of
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 381
preformed vesicles present throughout the cyto-
plasm, and thus may not have to rely on endocytic
routes for replenishment of releasable SVs for even
the longest stimuli used in many experiments
(Griesinger, C. B. et al., 2005). For example, a hair
cell that contains 1 million vesicles could, in princi-
ple, sustain the exocytosis of 10 vesicles per
millisecond for 100 s, although is hard to imagine
how the hair cell could accommodate the resultant
threefold increase in surface area. Clearly, in the long
run endocytosis still must keep pace with exocytosis.
3.22.5.2 Vesicle Mobility
The movement of vesicular membrane through the
cytoplasm is an essential step in the SV cycle, except
for the case of kiss-and-stay. At conventional
synapses, the vesicle-associated phosphoproteins,
called synapsins, play essential roles in regulating
the movement and clustering of SVs (Rosahl, T.
et al., 1995). Studies on hippocampal synapses in
synapsin knockout mice have shown that synapsins
contribute to the SV cycle by maintaining the reserve
pool and the RRP at glutamatergic and GABAergic
terminals, respectively (Gitler, D. et al., 2004). The
synapsins tether SVs to actin filaments in a phosphor-
ylation-dependent manner (for a review, see Hilfiker, S.
et al., 1999).
Synapsins are absent at ribbon synapses, supporting
the idea that vesicle clustering and mobilization are
different in these cells (von Kriegstein, K. et al., 1999).
Recent work in retinal bipolar cells utilized EM, con-
focal microscopy, and total internal reflection
microscopy to demonstrate that cytoplasmic vesicle
movement is rapid, random, and not altered by cal-
cium influx or cytoskeletal interactions (Holt, M. et al.,
2004, reviewed in Sterling, P. and Matthews, G., 2005).
The authors concluded that SVs 2 are highly mobile in
bipolar cells (D ¼ 1.5  102 mm s1), and that random
vesicle movement (i.e., diffusion) is sufficient to
account for the translocation of vesicles from the
large cytoplasmic pool to release sites.
If baseline spike rates in ANFs are due to fusion of
SVs, then the diffusion coefficient reported for SVs in
bipolar terminals is far too low to support continuous
resupply of release sites at mammalian hair cell
synapses that sustain SRs of >100 Hz unless some
additional mechanism exists to concentrate vesicles
in the vicinity of the synapse. Even assuming that
diffusion is used only to transport vesicles to the SB,
whereupon they are instantly transported to release
sites by some other means, and ignoring the
 Author's personal copy
382 Afferent Synaptic Mechanisms
plasmalemma and other barriers that restrict diffu-
sional access to the SB, the maximum steady-state
rate at which vesicles can reach the SB by diffusion is
F ¼ 4DRC
where C is the global concentration of vesicles in
the volume. Substituting D ¼ 1.5  102 mm2 s1,
F ¼ 100 s1, and R ¼ 100 nm, and solving for C yields
C ¼ 5000 vesicles mm3, which is more than 50 times
the density reported in the nonsynaptic regions of
hair cells. This calculation shows that the average
concentration of SVs throughout the cytoplasm is
too low to continuously supply the SB with vesicles
by diffusion unless some additional process operates
to maintain the locally high concentration of SVs
within a few hundred nanometers of the SB (Lenzi,
D. et al., 1999; 2002).
Experiments like these, combined with the
observed presence of synaptic ribbons in cell types
that support continuous exocytosis have led to the
view that the structural and molecular specializations
of ribbon synapses assist high rates of sustained vesicle
supply to AZs, but crucial questions remain unan-
swered. For example, data for the refilling of the RRP
by vesicles on the SB are indirect; it is still unknown
whether or not vesicles actually move on the ribbon.
3.22.5.3 Docking and Priming
In addition to translocation to the AZ, other things
must occur before a vesicle is ready for exocytosis.
Two important but relatively poorly understood steps
are called docking and priming, each of which likely
embodies a series of associated events. Docking refers
to the placement of a SV in contact with the inner
leaflet of the plasma membrane whereas priming
refers to the molecular reactions, some ATP-depen-
dent, required to make a vesicle immediately ready for
fusion upon calcium influx. It is unclear whether a
vesicle must first be docked in order to be primed.
For clues to the relationship between fast exocyto-
sis and docking/priming at ribbon synapses, we look to
experimental evidence from retinal bipolar cells. Here
the ribbon is plate-like and the anatomical picture of
rows of vesicles tethered to both sides of the ribbon led
to the model in which the vesicles move along the
ribbon to docking sites. One hypothesis is that during a
100 ms strong depolarization all vesicles on the ribbon
are depleted via exocytosis. Assuming that this exocy-
tosis requires movement of vesicles along the ribbon to
the plasmalemma, several lines of evidence argue
against the involvement of ATP-dependent,
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
kinesin-mediated translocation of vesicles to the plas-
malemma (Heidelberger, R. et al., 2002). This
observation led to the alternative hypothesis that SVs
on the ribbon can fuse with each other (reviewed in
Parsons, T. D. and Sterling, P., 2003), creating chains
that extend to the plasmalemma. This mechanism can
account for the large EPSCs recorded postsynaptically
at hair cell synapses (see Section 3.22.6.3), and implies
that vesicles on the ribbon can be primed and fuse with
each other. However the protein Munc13-1, thought
to be necessary for priming (Rosenmund, C. et al.,
2003), was found to cluster with calcium channels at
the plasma membrane and not on the ribbon (tom
Dieck, S. et al., 2005), implying that the final priming
steps must take place near the membrane.
3.22.5.4 Transmitter Release and
Clearance
Glutamate released into the synaptic cleft binds to
postsynaptic AMPA receptors (Ruel, J. et al., 2000;
Glowatzki, E. and Fuchs, P. A., 2002; Keen, E. C. and
Hudspeth, A. J., 2006), causing a depolarization of the
postsynaptic membrane and, if the depolarization is
sufficient, one or more action potentials. N-Methyl-
D-aspartate (NMDA) receptor antagonist had no
effect on ANF activity, suggesting that they are
functionally absent. The pattern of spikes depends
on the temporal pattern of neurotransmitter release,
the kinetics of AMPA receptor activation, stochastic
spike generation, and neuronal refractoriness. Most
in vivo studies have focused on the firing properties of
ANFs in response to pure tones of varying intensities.
If quantal release of neurotransmitter is to accurately
reflect the temporal pattern of the sound stimulus, then
the chemical signal in the synaptic cleft must rapidly
be terminated, which involves uptake of glutamate
from the synaptic cleft by excitatory amino acid
transporters (e.g., GLAST) in surrounding support-
ing cells and afferents (Furness, D. N. and Lehre, K.
P., 1997; Furness, D. N. and Lawton, D. M., 2003).
Voltage-clamp recordings from inner phalangeal
cells in the cochlea of normal rats and GLAST
knockout mice showed the presence and absence,
respectively, of currents produced by these transpor-
ters (Hakuba, N. et al., 2000). Normal glutamate
reuptake from the synaptic cleft is also thought to
be important for the prevention of noise-induced
damage to afferent terminals, which can be mimicked
by GLAST blockers (Hakuba, N. et al., 2000;
Rebillard, G. et al., 2003) and prevented by AMPA
receptor blockers.
 Author's personal copy
Afferent Synaptic Mechanisms 383

3.22.5.5 Relating Anatomical and
Functional Vesicle Pools
The hair cell’s compact shape has proved fortuitous for
experimental measurements of synaptic exocytosis by
observing the increase in electrical capacitance asso-
ciated with the increased cell surface that occurs when
SVs fuse with the plasmalemma (Figures 6(a)–6(d);
Lindau, M. and Neher, E., 1988; reviewed in
Nouvian, R. et al., 2006). Because the surface area of
a single SV can be estimated, capacitance changes
can be translated into numbers of SVs exocytosed,
although given the uncertainties in SV diameter and
specific membrane capacitance, such calculations are
probably only accurate to within a factor of two.
These functionally derived SV counts can be com-
pared to numbers obtained with anatomical studies at
the EM level in order to test hypotheses about the
relationships between the structure and the function
of SV pools. The reliability of this technique in
accurately monitoring the release of neurotransmit-
ter has been verified by combining capacitance

(a) (c) (d)

20 10

Pool size (fF)

8
15
ΔCm (fF)

10 fF 6 τf = 140 ms
10
τs = 3 s
4
5
2
40 pA
0 2 4 6 8 10 32 34 0
200 ms
0 10 20 30 40 50 Time after pool depletion (s)
(b) Duration of stimulation (ms) (e) 450

300 10 ms

Discharge rate (spikes/s)

300 0.1 mM EGTA
5 mM EGTA 150
250 5 mM BAPTA
ΔCm (fF)

200 Endogenous buffer 0

150 450

100 300 160 ms

50
150
0
0
500 1000
0 300 600
Duration of stimulation (ms)
Time (ms)

Figure 6 Exocytosis from inner hair cells, recovery from synaptic vesicle pool depletion, and adaptation/disadaptation of
auditory nerve fiber responses. (a) Plasma membrane capacitance changes (upper) caused by depolarization-induced
synaptic vesicle exocytosis that depends on calcium influx (lower). Two 15 ms depolarizations to the peak of the IV curve for
calcium (15 mV) were separated by 100 ms. Although the calcium currents evoked by the paired pulses were the same, the
exocytic response to the first pulse was greater than the response to the second pulse (secretory depression). (b, c) The
amplitude of stimulus-evoked capacitance changes is plotted as a function of the duration of the depolarization for different
intracellular calcium-buffering conditions. (c) The first 50 ms of (b) at higher temporal resolution. With the cell’s endogenous
buffer present, the readily releasable pool (RRP) is depleted within the first 10 ms, after which exocytosis continues at a slower
rate that is sustained for at least 1 s. The slow calcium buffer (EGTA) of 5 mM abolishes the sustained release while leaving
exocytosis of the RRP intact. The fast calcium buffer (BAPTA) of 5 mM inhibits both fast exocytosis of the RRP and the slower
component of exocytosis. (d) Strong depolarizations sufficient to deplete the RRP were followed by a second identical
stimulus with varying interstimulus intervals as indicated on the x-axis. Recovery of the RRP as a function of time after pool
depletion has both rapid and slow components. (e) Auditory nerve fiber spike rate in vivo as a function of time after sound
onset. The sound is interrupted by 10 ms (upper) or 160 ms (lower). During a sound presentation, the spike rate is initially very
high followed by two phases of fast reduction in firing rate called rapid and short-term adaptation, with time courses of about
10 and 100 ms, respectively. The 10 ms interruption (upper) is not enough for recovery from rapid adaptation of the spike rate
whereas the 160 ms interruption (lower) is sufficient for nearly complete recovery of the initial spike rate. Note the depression
of baseline rate and its recovery after sound offset. Panels (a)–(d) are for the mouse: Reproduced with permission from,
Copyright (2000) National Academy of Sciences, U. S. A., Moser, T. and Beutner, D., Kinetics of exocytosis and endocytosis
at the cochlear inner hair cell afferent synapsis of the mouse. Proc. Natl. Acad. Sci. U S A 97, 883–888. Panel (e) is from the
chick: JARO, Vol. 5, 2004, Evidence that rapid vesicle replenishment of the synaptic ribbon mediates recovery from short-
term adaptation at the hair cell afferent synapse, Spassova, M. A., Avissar, M., Furman, A. C., Crumling M. A., Saunders, J. C.,
and Parsons, T. D., Copyright Springer, with kind permission of Springer Science and Buisness Media.

The Senses: A Comprehensive Reference, vol. 3, pp. 365-395

 Author's personal copy
384 Afferent Synaptic Mechanisms
measurements with simultaneous detection of gluta-
mate release in bipolar cells of the vertebrate retina
(von Gersdorff, H. et al., 1998).
To estimate the size of the RRP in hair cells, synap-
tic transmission is first inhibited by hyperpolarization
(to fill vesicle pools), then a depolarizing voltage stimu-
lus that maximally stimulates exocytosis is delivered,
and the ensuing capacitance increase is observed
(Figures 6(b) and 6(c); reviewed in Fuchs, P. A. et al.,
2003). Pairs of brief strong depolarizations that evoke
maximal calcium currents typically cause paired-
pulse depression (Figures 6(a) and 6(d)), implying
that the first stimulus partially or totally depletes
the RRP, which is refilled more slowly. The size of
the RRP can also be measured by applying strong
continuous depolarizations of various durations,
which evoke continuous calcium currents from the
L-type calcium channels at ribbon synapses.
Exocytosis of the RRP is seen as an initial rapid
increase in capacitance that lasts only a few millise-
conds before giving way to a sustained secretory
component (SSC) that lasts for seconds (Figure 6(b);
Moser, T. and Beutner, D., 2000; Spassova, M. A.
et al., 2004; Rutherford, M. A. and Roberts, W. M.,
2006). Generally, the size of the RRP measured by
change in capacitance is in good agreement with the
size of the docked SV pool, as measured by EM, and
is smaller than the pool of vesicles on the SB. The
presence of a SB and the size of the RRP are both
important for normal ANF responses (Khimich, D.
et al., 2005; see also Section 3.22.7.2).
The density of vesicles on the SB and in the
cytoplasm surrounding the AZ is sensitive to the
physiological state of the hair cell before and during
fixation (Lenzi, D. et al., 2002). In frog saccular hair
cells, prolonged depolarization by high extracellular
potassium reduces the density of vesicles on the SB
from 79% to 42% of 3D dense-packed, with the
greatest reduction occurring in the number of docked
SVs in the narrow space between the SB and the
plasmalemma. Fixation under resting conditions
yielded intermediate values (Lenzi, D. et al., 2002).
Whole-cell measurements of SV exocytosis support
the notion that all of the docked, SB-associated vesi-
cles can fuse with the plasmalemma within 10 ms
(Rutherford, M. A. and Roberts, W. M., 2006).
Vesicles in the cytoplasm that are not attached to
the SB are also massively depleted by prolonged
depolarization (from 10% to 2% of 3D dense-packed
in reconstructed volumes within 800 nm perpendicu-
lar to the plasmalemma; Figure 4).
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Exocytosis of the SSC is slow compared to the
RRP (but still a fast resupply rate of hundreds of
vesicles per second per AZ), and may indicate the
rate at which the RRP is replenished from a recycling
pool in the presence of a continuing calcium influx.
Dialyzing the cell with the calcium buffer EGTA to
inhibit the spread of free calcium away from the AZ
(Figures 6(b) and 6(c)) blocked the SSC but not the
initial exocytic burst from the RRP (Spassova, M. A.
et al., 2004; Moser, T. and Beutner, D., 2000), which
suggests that refilling of the RRP is stimulated by
long-distance calcium signaling that can be blocked
by a slow calcium buffer. Exocytosis of the RRP was
inhibited only with the fast calcium buffer BAPTA,
as is expected for the case of tight spatial coupling
between the release-ready vesicles and the calcium
channels (i.e., the RRP docked at the AZ). At present,
it is unknown whether the SSC involves vesicles on
the SB, other cytoplasmic vesicles that move to the
AZ, or possibly even vesicles that are not located at
AZs. Experiments that combine capacitance mea-
surements with other techniques to monitor the
flow of membrane through the SV cycle will be
required to answer such questions.
3.22.6 Physiology of the Hair Cell
Afferent Synapse
Precise encoding at the first synapse in the auditory
system is absolutely critical for the complete range of
behavioral interactions with the acoustic world. To
produce a neural representation of the acoustic
environment in the CNS, the mammalian cochlea
first decomposes complex stimuli into separate fre-
quency components, each of which has an amplitude
and a phase. The simplest statement that can be made
about sensory encoding in the mammalian cochlea is
that the amplitude of the sound at a given frequency
is encoded by the spike rate in the ANFs of corre-
sponding CF, while phase and other temporal
patterns (e.g., amplitude modulations) are specified
by the timing of ANF spikes. Decades of work have
produced a wealth of detail about sensory encoding
in the auditory system, but much remains to be dis-
covered. The following will barely scratch the
surface of what is known. For additional information,
the mechanisms underlying the temporal precision of
sound coding at the IHC ribbon synapse have been
recently reviewed (Moser, T. et al., 2006b).
 Author's personal copy
Afferent Synaptic Mechanisms 385

3.22.6.1 Encoding of Phase frequency can be attributed to the low-pass filtering

For low-frequency sinusoidal (tonal) stimuli, action
potentials in ANFs occur within a narrow time window
relative to each cycle of the sinusoid (Anderson, D. J.
et al., 1970). The discharge of the ANF is said to be
phase-locked to the stimulus (Rose, J. E. et al., 1967).
Phase locking has been observed experimentally in the
auditory pathways of mammals, birds, reptiles, amphi-
bians, and fish, and is thought to be an important
mechanism for the temporal encoding of low-frequency
sound. For example, phase differences between the two
ears provide important binaural cues for localizing
sounds horizontally (Knudsen, E. I. and Konishi, M.,
1978; see Chapter Encoding of Interaural Timing for
Binaural Hearing).
Phase locking declines as frequency increases until
action potentials in the auditory nerve become nearly
random with respect to the stimulus period. The
frequency where this transition occurs in the cochlear
nerve is generally between 1 and 4 kHz, but it varies
among species and even within an individual cochlea.
In part, the decline of phase locking with increasing
of the AC component of the receptor potential by the
hair cell membrane (Figure 7). When intracellular
receptor potentials were recorded in vivo from a gui-
nea-pig cochlear IHC, the receptor potential was
quasisinusoidal with the periodicity of the stimulus
for low-frequency tones (Palmer, A. R. and Russell, I.
J., 1986; Cheatham, M. A. and Dallos, P., 1993). The
AC (periodic) component of the receptor potential
began to decline at 600 Hz and was not detectable
above 3.5 kHz, leaving only a DC (steady) depolar-
ization (Figures 7(a) and 7(b)). The diminution of the
presynaptic AC component was mirrored by a
decline in postsynaptic phase locking (Figure 7(c)).
Recordings of spike trains in the saccular branch of
the eighth nerve in frogs have shown that vibrations at
frequencies 50 Hz evoke a single, phase-locked spike
for each cycle of the stimulus (Koyama, H. et al., 1982;
Christensen-Dalsgaard, J. and Narins P. M., 1993).
When expressed as spikes per second, the response
thus decreases in proportion to frequency below
50 Hz. Because for low frequencies the depolarizing
phase of the hair cell receptor potential may be

(a) (b) 10.0

AC component / DC component

100
3.0
300
1.0

500
25 mV 0.3

700
0.1
0.03 0.1 0.3 1.0 3.0
900 Frequency (kHz)
(c)
1.2
1000
Synchronization index

1.0
AC
D.C. 2000 0.8
3000 0.6
4000
0.4
5000
0.2
0 10 20 30 40 50 60 70 ms 0.0
0.1 1.0 10.0
Frequency (kHz)

Figure 7 The relationship between receptor potentials in inner hair cells (IHCs) and phase locking in auditory nerve fibers.
(a) Intracellular recordings of receptor potentials in one IHC of the guinea-pig cochlea in response to loud (80 dB sound
pressure level) tones at the frequency (Hz) indicated by the right side of each trace. The receptor potential has an oscillatory
(AC) component and a steady-state (DC) component. The AC component dominates at low frequencies but is almost absent
above 3000 Hz. (b) The ratio of AC to DC component across frequency for the receptor potentials from nine different IHCs.
(c) The synchronization index (a measure of phase locking to the tonal stimulus) plotted as a function of tone frequency.
(Reprinted from Palmer, A. R. and Russell, I. J. 1986. Phase-locking in the cochlear nerve of the guinea-pig and its relation to
the receptor potential of inner hair-cells. Hear Res. 24, 1–15, with permission from Elsevier.)

The Senses: A Comprehensive Reference, vol. 3, pp. 365-395

 Author's personal copy
386 Afferent Synaptic Mechanisms
relatively long (e.g., 5 ms for 100 Hz) compared to the
duration required for fusion of a single vesicle, the
synapse could generate a single spike per cycle if
synaptic exocytosis occurred primarily within a nar-
row time window in response to the onset of each
depolarizing phase of a periodic stimulus, and was
subsequently suppressed until the next depolariza-
tion. This possibility was tested using capacitance
measurements as the assay of SV exocytosis
(Rutherford, M. A. and Roberts, W. M., 2006). A
30 ms depolarization to 55 mV was interrupted
with a return to 80 mV, thus creating a pair of
10 ms depolarizations separated by 10 ms. If the
synapse responds primarily to stimulus onset, then
the response to the interrupted depolarization is
expected be twice the response to a maintained
30 ms depolarization, even though the interposed
hyperpolarization reduces the total calcium influx.
The total capacitance change in response to the
interrupted stimulus was nearly twice as large as
the response to the sustained depolarization. The
result may reflect the mechanism by which hair cell
limits the ANF discharge to a single phase-locked
spike per stimulus cycle at low frequencies.
3.22.6.2 Encoding of Frequency
In general, all ANFs phase lock to low frequencies,
regardless of their CF. Therefore, the timing of ANF
spikes can provide information about the frequency of a
low-frequency sound independent of the fiber’s CF
(Figure 2). For frequencies above 3–4 kHz, the alter-
native to temporal coding is place-rate coding (Javel, E.
and Mott, J. B., 1988), which does not rely upon the
temporal (fine timing) pattern of spikes to communicate
frequency. In the case of the mammalian cochlea, which
performs exquisite mechanical tuning, frequency can be
determined from peaks in action potential rates in
ANFs that innervate specific CF regions (labeled
lines) in the organ of Corti. Therefore, an increase in
the mean firing rate of an ANF that originates from a
particular place in the organ of Corti is a mechanism for
specifying frequency to the brain. Auditory nuclei in
the brain stem may use different decoding algorithms
for different frequency ranges in order to interpret
firing patterns of the auditory nerve.
3.22.6.3 Baseline and Evoked Synaptic
Activity
Whole-cell recordings from postsynaptic afferent
terminals demonstrated EPSCs (Figure 8) resulting
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
from baseline and evoked SV exocytosis from rat
IHCs (Glowatzki, E. and Fuchs, P. A., 2002) and
from hair cells of the frog amphibian papilla (Keen,
E. C. and Hudspeth, A. J., 2006). In both cases, EPSCs
were fast (1–2 ms for the rat, 2–3 ms for the frog) with
very rapid onset and exponential decay. Sensitivity to
AMPA receptor antagonists provided a functional
confirmation of previous in vivo and EM studies
(Matsubara, A. et al., 1996), showing that the gluta-
matergic receptor subunits GluR2/3 and 4 are
present.
The amplitude distribution of EPSCs recorded
under baseline conditions (i.e., extracellular saline
with normal concentration of potassium) showed
20-fold variation in the rat and 10-fold variation
in the frog with mean sizes of 190 and 100 pA, respec-
tively (Figure 8). This is remarkable considering that
spontaneous glutamatergic EPSCs recorded elsewhere
in the CNS are distributed normally with much smal-
ler variation around a quantal size of 20–40 pA.
Presynaptic depolarization via paired whole-cell
recordings in the frog, and by perfusion with high-
potassium saline in the rat, greatly increased the
EPSC rate (Figure 8(a)) while the amplitude distri-
bution remained relatively unchanged (Figures 8(d)
and 8(e)). In both preparations, individual small
events (miniature EPSCs) were observed, but the
amplitude distributions were skewed to larger events
for both baseline and evoked EPSCs. For example, in
the rat the peak in the amplitude distribution chan-
ged from about 18 pA in 5.8 mM potassium (baseline,
Figure 8(d), left) to about 36 pA in 40 mM potassium
(evoked, Figure 8(e), left), as the mean rate changed
from about 1.5 to 27 events per second. Interestingly,
the mean EPSC size was slightly less for evoked
EPSCs, perhaps due to SV depletion with the sharp
increase in EPSC frequency. Combined with the
positive shift in the peak of the amplitude distribu-
tion, this results in a tightening of the EPSC
amplitude distribution (e.g., smaller interquartile
range), although the total range remains unchanged
(cf. Figures 8(d) and 8(e), left). In both frog and rat
hair cells, the larger EPSCs were sometimes multi-
phasic, but the durations of the larger monophasic
events were not appreciably longer than the smaller
monophasic events (Figure 8(b), frog; Figures 8(c)–8(e),
rat), which suggests that if the large EPSCs are due to
the fusion of several SVs, then these fusion events are
highly synchronized (reviewed in Parsons, T. D. and
Sterling, P., 2003; Singer, J. H. et al., 2004).
Two possible mechanisms might explain such multi-
quantal events. One possibility is that the opening of
 Author's personal copy
Afferent Synaptic Mechanisms 387

(a) –45 mV (b) (c)

300 300

EPSC amplitude (pA)

–40 mV 200 200
200
100 100 pA

0 0
–35 mV 0 0.4 0.8 0 0.4 0.8 1 ms
Rise time (ms) Decay time (ms)

–30 mV (d) 800 AF #4–7

+
5.8 mM K

EPSC amplitude (pA)

20 AF #4–7
n = 698
+ 600
5.8 mM K

Number of events
15 190 ± 149 pA
–25 mV n = 1003
400
10
200
5
–20 mV
0 0
0 200 400 600 800 0.0 0.5 1.0 1.5 2.0
EPSC amplitude (pA) 10–90% rise time (ms)
800
–15 mV (e)

EPSC amplitude (pA)

AF #1
30 AF #1 +
+ 600 40 mM K
40 mM K
n = 340
25 133 ± 126 pA
Number of events

n = 802
–10 mV 20 400
15
10 200
5
20 ms 100 pA 0 0
0 200 400 600 800
0 0.5 1.0 1.5 2.0
EPSC amplitude (pA) 10–90% rise time (ms)

Figure 8 Excitatory postsynaptic currents (EPSCs) in afferent fibers indicate the size and frequency of glutamate release
from hair cells. (a) EPSCs recorded in afferent fibers of the frog amphibian papilla while the presynaptic hair cell was
depolarized to the voltages indicated at the right of each trace. The predominant effect of depolarization was an increase in
EPSC frequency, with little change in EPSC amplitudes. (b) The rise and decay times for baseline EPSCs (frog) are tightly
distributed despite a wide range of amplitudes. (c) EPSCs evoked from rat cochlear afferent fibers in response to inner hair
cell depolarization by 40 mM extracellular potassium, aligned in time, exhibit a large range in the amplitude of monophasic
events. (d, e) Left: EPSC amplitude distributions from rat afferent fibers show the number of events observed in each
amplitude bin along the x-axis. Baseline EPSC amplitudes from four afferents are shown in (d), evoked EPSC amplitudes from
one afferent in (e). The distributions are positively skewed, suggestive of multivesicular release events from individual hair cell
active zones. The range of baseline EPSC amplitudes observed in a normal concentration of extracellular potassium (5.8 mM)
spans the entire range of EPSC amplitudes evoked by depolarization in high potassium (see text). Mean amplitudes were
190 pA (baseline) and 133 pA (evoked). Right: The relationship between EPSC amplitude and 10–90% rise time is very similar
for baseline (d) and evoked (e) events. (a, b) Reproduced from Keen, E. C. and Hudspeth, A. J. 2006. Transfer characteristics
of the hair cell’s afferent synapse. Proc. Natl. Acad. Sci. U. S. A. 103, 5537–5542; Copyright (2006) National Academy of
Sciences, U. S. A. (c–e) Glowatzki, E. and Fuchs, P. A., 2002. Transmitter release at the hair cell ribbon synapse. Nat.
Neurosci. 5, 147–154; Copyright (2002) Nature Publishing Group.

a single presynaptic calcium channel, or a few chan- but the slope of the initial rise can be used instead
nels that happen to be close together, triggers the to measure the EPSP size. In vivo recordings from
synchronized fusion of several nearby vesicles. afferent units in the guinea-pig cochlea demonstrated
Another is that several vesicles undergo nonsynchro- a range of EPSP slopes (Siegel, J. H., 1992). Similar to
nous fusion with each other, but their contents are the EPSCs recorded in vitro and discussed above, the
not released until one of them fuses with the distribution of EPSP slopes was positively skewed,
plasmalemma. suggesting the existence of synchronous multiquantal
When spikes are present in intracellular record- release events. Interestingly, nearly all EPSPs that
ings from ANFs, the peak EPSP amplitude is did not occur during neural refractory periods trig-
obliterated by the action potential that it triggers, gered action potentials.

The Senses: A Comprehensive Reference, vol. 3, pp. 365-395

 Author's personal copy
388 Afferent Synaptic Mechanisms
What then is the consequence of multivesicular
release to synaptic function? One possibility is that
large EPSCs trigger action potentials with less tem-
poral jitter (Anderson, D. J. et al., 1970; reviewed in
Fuchs, P. A., 2005), but this leaves open the question
of whether there is some value in having a large
variance in EPSP amplitude rather than all being
equally large. An intriguing possibility has to do
with the role of efferent innervation on the input
resistance of afferent terminals. The terminal bouton
of a type I afferent is typically postsynaptic to both an
IHC and an efferent axon (see Chapter Efferent
System). Spikes in the efferent axon release transmit-
ter that opens postsynaptic channels, thereby
reducing the input resistance of the afferent terminal.
The effect of efferent activity may therefore be to
reduce the size of the EPSP produced by a given
EPSC. Since efferent activity increases in loud envir-
onments, the effect may be to render small EPSCs
from the hair cell unable to trigger ANF spikes in
loud environments. In this case, the wide distribution
of EPSC amplitudes associated with multiquantal
events may be important for efferent control of
ANF firing rate. Similarly, efferent innervation
directly onto hair cells is proposed to evoke inhibi-
tory conductances in hair cells during strong stimuli.
The expected effect would be a decrease in hair cell
depolarization and calcium influx. Since efferent
modulation of synaptic activity is expected to be
largely absent from in vitro preparations, the effects
of such inhibition on the EPSC amplitude and fre-
quency in response to varying presynaptic stimulus
levels are unknown.
If large EPSCs are due to the fusion of vesicles
with each other on the SB before one of them fuses
with the plasmalemma, this could be a mechanism to
integrate sound amplitude over many cycles while
preserving phase locking of glutamate release to the
sound waveform. For this to work, the fusion of SVs
with each other on the SB would have to be triggered
by the relatively small changes in free calcium that
are predicted to occur at distances of tens to hun-
dreds of nanometers from the plasmalemma, whereas
the fusion of the resulting multivesicular structure
with the cell surface membrane would be phase
locked to the large, rapid calcium changes that
occur close to the presynaptic calcium channels. In
favor of this hypothesis, large irregularly shaped
structures that appear to be attached to the SB are
seen much more frequently at stimulated synapses
than inhibited or resting synapses (Lenzi, D. et al.,
1999; 2002, Figure 4). Opposed to this hypothesis, the
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
mean EPSC amplitude did not increase following
stimulation (Glowatzki, E. and Fuchs, P. A., 2002;
Keen, E. C. and Hudspeth, A. J., 2006).
3.22.6.4 Tuning Mechanisms in Hair Cells
Auditory organs of fish, amphibians, reptiles, and
birds all show frequency selectivity resulting from
the properties of individual hair cells. These mechan-
isms are different from the frequency tuning seen in
the mammalian cochlea, which is the result of the
gradient in stiffness of the basilar membrane (acellu-
lar) and active mechanical feedback from hair bundle
motion and OHC motility (cellular). Our focus will
be on tuning in the other four classes of vertebrate
animals.
3.22.6.4.1 Frequency-selective
augmentation of membrane potential
oscillation
In the turtle auditory papilla, there is no apparent
frequency tuning extrinsic to the hair cell (i.e., trans-
duction currents are equal over a broad frequency
range). Instead, each cell is intrinsically tuned to a
different frequency band by a process that has been
called electrical resonance. The resonant frequency
is determined primarily by the number and kinetics
of the cell’s calcium-activated potassium channels,
voltage-activated potassium and calcium channels,
and electrical capacitance (Ricci, A. J. et al., 2000;
and reviewed in Fettiplace, R. and Fuchs, P. A.,
1999). Each hair cell preferentially amplifies the
receptor potential in a frequency band, with peak
amplification near the resonant frequency. The tono-
topic axis in this organ is generated by the systematic
variation of the hair cells’ preferred frequencies along
the length of the sensory epithelium. Low-frequency
cells have low channel density and slower potassium
channel kinetics than high-frequency cells. These
variations in channel kinetics are caused by differen-
tial expression of potassium channel genes and
differential splicing of the mRNAs ( Jones, E. M. C.
et al., 1999). The result is a continuum of resonant
frequencies along the papilla. Similar electrical tuning
has been studied in the frog sacculus (Armstrong, C. E.
and Roberts, W. M., 1998; 2001) and chick cochlea
(Correia, M. J., 1992), but there is no evidence of
electrical tuning in the mammalian cochlea.
3.22.6.4.2 Synaptic frequency selectivity
Experiments employing capacitance measurements
in hair cells typically use very strong depolarizations
 Author's personal copy
that evoke maximal calcium currents in order to
measure vesicle pool sizes by depletion. This
approach has been very fruitful for understanding
vesicle pools, but the stimuli used for these experi-
ments are much larger than in vivo receptor
potentials, drawing into question the physiological
relevance of the evoked exocytosis. A recent study
of frog saccular hair cells (Rutherford, M. A. and
Roberts, W. M., 2006) used smaller, sinusoidal fluc-
tuations in membrane potential as the voltage-clamp
stimulus to mimic the size and frequency of in vivo
receptor potentials and looked for frequency selec-
tivity downstream of the known mechanical and
electrical tuning mechanisms in these cells. Changes
in membrane capacitance were measured in response
to small sinusoidal voltage changes (5 mV centered
at 60 mV) delivered at 5, 50, and 200 Hz to test the
hypothesis that exocytosis would be largest at 50 Hz,
the frequency where the sacculus is most sensitive to
airborne and groundborne vibrations. Consistent
with this hypothesis, exocytosis was twice as large
at 50 Hz than at 5 or 200 Hz, even though the total
calcium influx was the same at the three frequencies.
This preference for 50 Hz disappeared when stronger
sinusoidal stimuli (10 mV centered at 55 mV)
were applied, which is reminiscent of other tuning
mechanisms that are saturated by large stimuli (e.g.,
de Boer, E. and Nuttall, A. L., 2000). It is important to
note that the 5 mV sinusoidal stimulus, while small
compared to the stimuli used in most voltage-clamp
studies, is well above the amplitude of oscillations
that occur at sensory threshold (Dallos, P., 1985). The
exocytosis in response to the small stimulus at 50 Hz
corresponded to the net fusion of 1.7 SV per AZ per
stimulus cycle. This amounts to twice the number of
vesicles in the docked pool fusing during the 1 s
stimulus, and suggests that new vesicles can dock
and be primed for release during the stimulus. This
synaptic tuning, operating in series with mechanical
and electrical tuning, is expected to contribute to the
overall tuning of the sacculus, which has been
measured by action potential firing rates and sensi-
tivity in afferent axons of the eighth cranial nerve
(Koyama, H. et al., 1982; Lewis, E. R. and Narins, P. M.,
1985; Yu, X. et al., 1991; Christensen-Dalsgaard, J. and
Narins, P. M., 1993). Similar synaptic tuning could
contribute to frequency selectivity at other ribbon
synapses, particularly in auditory and vestibular
organs where synaptic transmission results in
phased-locked action potentials in the postsynaptic
neuron.
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 389
3.22.7 Auditory Nerve Fiber Firing
Patterns Mirror Presynaptic Vesicle
Pool Dynamics
Until recently, physiological studies of cochlear out-
put have focused on in vivo recordings of spike trains
in ANFs in response to sound stimuli (e.g., pure tones
of varying SPL). Advances in patch-clamp techni-
ques in vitro, such as capacitance measurements as a
presynaptic assay of SV exocytosis, have allowed
more direct comparisons between presynaptic activ-
ity and postsynaptic spike trains.
3.22.7.1 Adaptation of the Auditory Nerve
Response
Adaptation of ANF responses is thought to be neces-
sary for the peripheral processing of speech
(Delgutte, B., 1980). Two types of fast adaptation
have been described to explain the reduction in
ANF discharge rate following the onset of a sustained
acoustic stimulation (Figure 6(e)). These have been
called rapid adaptation and short-term adaptation,
with durations of approximately 10 and 20–90 ms,
respectively. Although the precise time courses can
vary by species, SR of the fiber, and sound intensity;
both processes are thought to arise from mechanisms
downstream of IHC receptor potential in vivo, which
is not reduced in amplitude during a sustained sti-
mulus (Russell, I. J. and Sellick, P. M., 1978; Holton,
T. and Weiss, T. F., 1983). Fast adaptation is thought
to involve presynaptic depression (as described
below), but not via inactivation of the presynaptic
calcium current which shows no significant decline
over this time course under voltage clamp (Parsons,
T. D. et al., 1994; Moser, T and Beutner, D., 2000;
Schnee, M. E. and Ricci, A. J., 2003; Eisen, M. D. et al.,
2004).
Fast adaptation of ANF firing and its subsequent
recovery seem to be predominantly due to depletion
and refilling of the RRP. Recovery from adaptation
takes hundreds of milliseconds (Smith, R. L., 1977;
Harris, D. M. and Dallos, P., 1979; Chimento, T. C.
and Schreiner, C. E., 1991), and this time course is
thought to reflect refilling of the RRP. In the chick,
using a combination of in vivo single unit recordings
of ANF activity and in vitro patch-clamp recordings
of SV fusion, it was found that sound-evoked ANF
activity adapted and recovered with similar time
courses as RRP depletion and recovery (Figure 6;
Spassova, M. A. et al., 2004). After RRP exhaustion,
 Author's personal copy
390 Afferent Synaptic Mechanisms
its recovery was 95% complete after a 200 ms rest
period, similar to the time course of recovery from
adaptation. In comparison, analogous recovery in
retinal bipolar neurons takes about 100 times longer
(20 s; von Gersdorff, H. and Matthews, G., 1997). In
the mouse IHC, RRP refilling is biphasic, possibly
reflecting the refilling of two functionally distinct
types of AZs (Moser, T. and Beutner, D., 2000).
Differences among AZs in a single IHC could pro-
vide the basis for the variable properties of ANFs
(Merchan-Perez, A. and Liberman, M. C., 1996).
Indeed high- and low-SR fibers exhibit different
times needed for recovery from prior stimuli
(Relkin, E. M. and Doucet, J. R., 1991) that are a
close match to the biphasic refilling of the RRP
(Figure 6(d)).
3.22.7.2 The Role of the Synaptic Body in
Auditory Nerve Fiber Spike Timing
Computational attempts at predicting psychophysi-
cal performance on auditory tasks based on the
activity of ANFs have generally concluded that the
probabilistic nature of action potentials (the variabil-
ity of spike timing during repeated presentations of
the same stimulus) limits the precision of acoustic
information sent to the brain (Mountain, D. C. and
Hubbard, A. E., 1996). Upon repeated presentation of
a pure tone, an ANF will fire a first spike with
variable latency. The variance of first spike latency
in mice with a targeted deletion of the gene for the
cytomatrix AZ protein Bassoon was increased by a
factor of 12 compared to controls (Buran, B. N. et al.,
2006). Bassoon is thought to anchor SBs at AZs by
interacting with Ribeye (Khimich, D. et al., 2005; tom
Dieck, S. et al., 2005); ultrastructural inspection
demonstrated a 90% reduction in the number of
SBs anchored at AZs. This work suggests that the
presence of SBs at release sites helps to reduce the
temporal jitter of ANF discharge, and hence improve
phase locking.
In the absence of an applied stimulus, the tem-
poral firing pattern of ANFs in vivo is often
represented as a poison process, but on a short time-
scale it is more accurately described as a renewal
process that is a function of neurotransmitter con-
centration in the cleft (presynaptic effect) and neural
refractoriness (postsynaptic effect), the combination
of which causes the firing probability to be depressed
for more than 20 ms following a spike (Gaumond, R.
P. et al., 1983). Furukawa and co-workers recorded
synaptic potentials from afferent units in the saccular
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
branch of the eighth nerve of the goldfish and
observed a 20 ms delay in firing upon a decrement
in the intensity of tone. They attributed this delay to
the time required for SVs to refill the depleted low-
threshold release sites between the SB and the cal-
cium channels at the plasmalemma. Furthermore,
quantal analysis of the size of EPSPs showed a nega-
tive correlation between the amplitude of two
consecutive events (i.e., large EPSPs tend to be fol-
lowed by smaller ones and vice versa). These data
support the idea that adaptive rundown of EPSPs is
the result of SV depletion (Furukawa, T. and
Matsuura, S., 1978; Furukawa, T. et al., 1978).
Additional evidence for the involvement of the SB
in maintenance of the functional RRP comes from
the Bassoon mutant mice, in which the first phase of
exocytosis and the calcium current are both reduced
(Khimich, D. et al., 2005). The large reduction of SB-
associated docked SVs at AZs in the mutant provides
additional support for the contribution of these
docked vesicles to the functional RRP measured in
capacitance experiments. The reduced calcium cur-
rent supports previous evidence that calcium channel
expression is coregulated with and restricted spa-
tially to SV release sites adjacent to SBs (Martinez-
Dunst, C. et al., 1997; Zenisek, D. et al., 2003).
Intriguingly, the sustained component of the capaci-
tance rise was relatively intact in the mutant mice
compared to the 50% reduction in ANF discharge
rate after short-term adaptation to tonal stimuli
(Buran, B. N. et al., 2006). One interpretation is that
the sustained component in the mutant mice is rela-
tively asynchronous or not properly localized to
postsynaptic fibers. In contrast, the sustained exocy-
tosis in wild-type mice is likely to reflect synaptic
release as indicated by the massive sustained release
measure by EPSCs in high potassium and higher
levels of sustained ANF discharge rate in vivo. The
sound-evoked onset firing rate (before adaptation)
and steady-state firing rate (after fast adaptation)
were decreased and threshold was increased, in
accordance with the reduction of the functional
RRP measured presynaptically. In spite of this, the
shapes of rate-level functions were normal in the
mutants. Indeed a recent modeling study showed
that the effect of increasing the maximum number
of transmitter quanta held in the immediate store
(analogous to the concept of increasing the size of
the RRP) is to scale the release rate linearly across
the entire dynamic range (Sumner, C. J. et al., 2002).
Work with the Bassoon mutant mouse has shown
that the localization of synaptic ribbons to hair cell
 Author's personal copy
AZs is required for transmitter release to be precisely
synchronized to the onset of sound at each AZ, which
is required for synchronous auditory signaling car-
ried to the brain stem (Khimich, D. et al., 2005; Buran,
B. N. et al., 2006). If simultaneously activated fibers
provide input to a single neuron in the cochlear
nucleus, then converging synaptic inputs arriving
closely in time could be the physiological basis for a
coincidence detector to differentiate between base-
line activity and stimulus-evoked (synchronous) hair
cell afferent synaptic activity. It seems probable that
the precise configurations of SBs and their associated
cytomatrices at AZs are important for establishing
both the sensitivity and the phase relationship of
ANF spikes to the sound.
3.22.8 Synapse to Psychophysics
One of the goals of auditory research is to provide an
explanation of the behavior of the auditory system
based upon the functions of its cellular components
and the circuits they make. For example, numerous
experiments have examined the ability of the audi-
tory system to detect weak sounds. These
experiments have been carried at the level of ANFs,
cortical neurons, and behavior, in a variety of mam-
mals, including humans. Recently it has been
proposed that the inverse relationship between the
duration and the SPL needed to attain a threshold
criterion for detection seen at all levels of auditory
processing can be explained by the properties of hair
cell afferent synapses. One interpretation of these
results is that the statistics of individual events,
which may correspond to the exocytosis of individual
SVs at the afferent synapse, underlies the trade-off
between strength and duration for barely detectable
sounds (Heil, P. and Neubauer, H., 2003) that is
measured in psychophysical experiments of sound
perception.
References
Achanzar, W. E. and Ward, S. 1997. A nematode gene required
for sperm vesicle fusion. J. Cell Sci. 110, 1073–1081.
Anderson, D. J., Rose, J. E., Hind, J. E., and Brugge, J. F. 1970.
Temporal position of discharges in single auditory nerve
fibers within the cycle of a sine-wave stimulus: frequency
and intensity effects. J. Acoust. Soc. Am. 49, 1131–1139.
Armstrong, C. E. and Roberts, W. M. 1998. Electrical properties
of frog saccular hair cells: distortion by enzymatic
dissociation. J. Neurosci. 18, 2962–2973.
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 391
Armstrong, C. E. and Roberts, W. M. 2001. Rapidly inactivating
and non-inactivating calcium-activated potassium currents
in frog saccular hair cells. J. Physiol. 536, 49–65.
Art, J. J. and Fettiplace, R. 1987. Variation of membrane
properties in hair cells isolated from the turtle cochlea. J.
Physiol. 385, 207–242.
Art, J. J., Wu, Y. C., and Fettiplace, R. 1995. The calcium-
activated potassium channels of turtle hair cells. J. Gen.
Physiol. 105, 49–72.
Balkema, G. W. 1991. A synaptic antigen (B16) is localized in
retinal synaptic ribbons. J. Comp. Neurol. 312, 573–583.
Bodian, D. 1980. Presynaptic bodies of auditory hair cells in old
world monkeys. Anat. Rec. 197, 379–386.
Bracewell, R. N. 1999. The Fourier Transform and Its
Applications. McGraw-Hill Co.
Brandt, A., Khimich, D., and Moser, T. 2005. Few Cav1.3
channels regulate the exocytosis of a synaptic vesicle at the
hair cell ribbon synapse. J. Neurosci. 25, 11577–11585.
Brandt, A., Striessng, J., and Moser, T. 2003. CaV1.3 channels
are essential for development and presynaptic activity of
cochlear inner hair cells. J. Neurosci. 23, 10832–10840.
Bunt, A. H. 1971. Enzymatic digestion of synaptic ribbons in
amphibian retinal photoreceptors. Brain Res. 25, 571–577.
Burack, M. A., Silverman, M. A., and Banker, G. 2000. The role
of selective transport in neuronal protein sorting. Neuron
26, 465–474.
Buran, B. N., Moser, T., Gundelfinger, E. D., and Liberman, M.
C. 2006. Role of synaptic ribbons in auditory nerve
responses in mice lacking the synaptic scaffolding protein
Bassoon. Assoc. Res. Otolaryngol (ARO) Poster, 2006,
Baltimore.
Cheatham, M. A. and Dallos, P. 1993. Longitudinal comparisons
of IHC ac and dc receptor potentials recorded from the
guinea pig cochlea. Hear Res. 68, 107–114.
Chen, X., Tomchick, D. R., Kovrigin, E., Arac, D., Machius, M.,
Sudhof, T. C., and Rizo, J. 2002. Three-dimensional
structure of the complexin/SNARE complex. Neuron
33, 397–409.
Chimento, T. C. and Schreiner, C. E. 1991. Adaptation and
recovery from adaptation in single fiber responses of the cat
auditory nerve. J. Acoust. Soc. Am. 90, 263–273.
Chinnadurai, G. 2002. CtBP family proteins: more than
transcriptional corepressors. Bioessays 25, 9–12.
Christensen-Dalsgaard, J. and Narins, P. M. 1993. Sound and
vibration sensitivity of VIIIth nerve fibers in the frogs
Leptodactylus albilabris and Rana pipiens pipiens. J. Comp.
Physiol. A 172, 653–662.
Correia, M. J. 1992. Filtering properties of hair cells. Ann. N. Y.
Acad. Sci. 656, 49–57.
Dallos, P. 1985. Response characteristics of mammalian
cochlear hair cells. J. Neurosci. 5, 1591–1608.
Dambly-Chaudiere, C., Sapede, D., Soubiran, F., Decorde, K.,
Gompel, N., and Ghysen, A. 2003. The lateral line of
zebrafish: a model system for the analysis of morphogenesis
and neural development in vertebrates. Biol. Cell
95, 579–587.
Dargan, S. L. and Parker, I. 2003. Buffer kinetics shape the
spatiotemporal patterns of IP3 evoked Ca2þ signals. J.
Physiol. 553, 775–788.
de Boer, E. and Nuttall, A. L. 2000. The mechanical waveform of
the basilar membrane. III. Intensity effects. J. Acoust. Soc.
Am. 107, 1497–1507.
Delcomyn, F. 1998. Foundations of Neurobiology. W. H.
Freeman.
Delgutte, B. 1980. Representation of speech-like sounds in the
discharge patterns of auditory-verve fibers. J. Acoust. Soc.
Am. 68, 843–857.
tom Dieck, S., Altrock, W. D., Kessels, M. M., Qualmann, B.,
Regus, H., Brauner, D., Fejtova, A., Bracko, O.,
 Author's personal copy
392 Afferent Synaptic Mechanisms
Gundelfinger, E. D., and Brandstatter, J. H. 2005.
Molecular dissection of the photoreceptor ribbon synapse:
physical interaction of Bassoon and Ribeye is essential for
the assembly of the ribbon complex. J. Cell Biol.
168, 825–836.
Dobrunz, L. E. and Stevens, C. F. 1997. Heterogeneity of
release probability, facilitation, and depletion at central
synapses. Neuron 18, 995–1008.
Edmonds, B. W., Gregory, F. D., and Schweizer, F. E. 2004.
Evidence that fast exocytosis can be predominantly
mediated by vesicles not docked at active zones in frog
saccular hair cells. J. Physiol. 560, 439–450.
Edmonds, B., Reyes, R., Schwaller, B., and Roberts, W. M.
2000. Calretinin modifies presynaptic calcium signaling in
frog saccular hair cells. Nat. Neurosci. 3, 786–790.
Eisen, M. D., Spassova, M., and Parsons, T. D. 2004. Large
releasable pool of synaptic vesicles in chick cochlear hair
cells. J. Neurophysiol. 91, 2422–2428.
Fettiplace, R. and Fuchs, P. A. 1999. Mechanisms of hair cell
tuning. Annu. Rev. Physiol. 61, 809–834.
Fuchs, P. A. 2005. Time and intensity coding at the hair cell’s
ribbon synapse. J. Physiol. 566, 7–12.
Fuchs, P. A., Glowatzki, E., and Moser, T. 2003. The afferent
synapse of cochlear hair cells. Curr. Opin. Neurobiol.
13, 452–458.
Furness, D. N. and Lawton, D. M. 2003. Comparative
distribution of glutamate transporters and receptors in
relation to afferent innervation density in the mammalian
cochlea. J. Neurosci. 23, 11296–11304.
Furness, D. N. and Lehre, K. P. 1997. Immunocytochemical
localization of a high-affinity glutamate-aspartate
transporter, GLAST, in the rat and guinea-pig cochlea. Eur. J.
Neurosci. 9, 1961–1969.
Furukawa, T. and Matsuura, S. 1978. Adaptive rundown of
excitatory post-synaptic potentials at synapses between
hair cells and eighth nerve fibers in the goldfish. J. Physiol.
276, 193–209.
Furukawa, T., Hayashida, Y., and Matsuura, S. 1978. Quantal
analysis of the size of excitatory post-synaptic potentials at
synapses between hair cells and afferent nerve fibers in the
goldfish. J. Physiol. 276, 211–226.
Furukawa, T., Kuno, M., and Matsuura, S. 1982. Quantal
analysis of a decremental response at hair cell-afferent fibre
synapses in the goldfish sacculus. J. Physiol. 322, 181–195.
Gaumond, R. P., Kim, D. O., and Molnar, C. E. 1983. Response of
cochlear nerve fibers to brief acoustic stimuli: role of
discharge-history effects. J. Acoust. Soc. Am. 74, 1392–1398.
von Gersdorff, H. and Matthews, G. 1997. Depletion and
replenishment of vesicle pools at a ribbon-type synaptic
terminal. J. Neurosci. 17, 1919–1927.
von Gersdorff, H., Sakaba, T., Berglund, K., and Tachibana, M.
1998. Submillisecond kinetics of glutamate release from a
sensory synapse. Neuron 21, 1177–1188.
Giannessi, F. and Ruffoli, R. 1996. Fine structure of the afferent
synapses in the paratympanic organ of the chicken, with
special reference to the synaptic bodies. Ann. Anat.
178, 127–131.
Gitler, D., Takagishi, Y., Feng, J., Ren, Y., Rodriguiz, R. M.,
Wetsel, W. C., Greengard, P., and Augustine, G. J. 2004.
Different presynaptic roles of synapsins at excitatory and
inhibitory synapses. J. Neurosci. 24, 11368–11380.
Gleisner, L., Flock, A., and Wersall, J. 1973. The ultrastructure of
the afferent synapse on hair cells in the frog labyrinth. Acta
Otolaryngol. 76, 199–207.
Glowatzki, E. and Fuchs, P. A. 2002. Transmitter release at the
hair cell ribbon synapse. Nat. Neurosci. 5, 147–154.
Goldberg, J. M. 1996. Theoretical analysis of intercellular
communication between the vestibular type I hair cell and its
calyx ending. J. Neurophysiol. 76, 1942–1957.
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Gray, E. G. and Pease, H. L. 1971. On understanding the origins
of the retinal receptor synapses. Brain Res. 35, 1–15.
Griesinger, C. B., Richards, C. D., and Ashmore, J. F. 2005. Fast
vesicle replenishment allows indefatigable signaling at the
first auditory synapse. Nature 435, 212–215.
Hackney, C. M., Mahendrasingam, S., Penn, A., and
Fettiplace, R. 2005. The concentrations of calcium buffering
proteins in mammalian cochlear hair cells. J. Neurosci.
25, 7867–7886.
Hakuba, N., Koga, K., Gyo, K., Usami, S., and Tanaka, K. 2000.
Exacerbation of noise-induced hearing loss in mice lacking
the glutamate transporter GLAST. J. Neurosci.
20, 8750–8753.
Hama, K. and Saito, K. 1977. Fine structure of the afferent
synapse of the hair cells in the saccular macula of the
goldfish, with special reference to the anastomosing tubules.
J. Neurocytol. 6, 361–373.
Harris, D. M. and Dallos, P. 1979. Forward masking of auditory
nerve fiber responses. J. Neurophysiol. 42, 1083–1107.
Hashimoto, S., Kimura, R. S., and Takasaka, T. 1990.
Computer-aided 3-dimensional reconstruction of the inner
hair-cells and their nerve-endings in the guinea-pig. Acta
Otolaryngol. 109, 228–234.
Heidelberger, R., Sterling, P., and Matthews, G. 2002. Roles of
ATP in depletion and replenishment of the releasable pool of
synaptic vesicles. J. Neurophysiol. 88, 98–106.
Heil, P. and Neubauer, H. 2001. Temporal integration of sound
pressure determines thresholds of auditory nerve fibers. J.
Neurosci. 21, 7404–7415.
Heil, P. and Neubauer, H. 2003. A unifying basis of auditory
thresholds based on temporal summation. Proc. Natl. Acad.
Sci. U. S. A. 100, 6151–6156.
Heller, S., Bell, A. M., Charlotte, S. D., Choe, Y., and
Hudspeth, A. J. 2002. Parvalbumin 3 is an abundant Ca2þ
buffer in hair cells. JARO 3, 488–498.
Hilfiker, S., Pieribone, V. A., Czernik, A. J., Kao, H. T.,
Augustine, G. J., and Greengard, P. 1999. Synapsins as
regulators of neurotransmitter release. Phil. Trans. R Soc.
Lond. B 354, 269–279.
Hille, B. 2001. Ionic Channels of Excitable Membranes. Sinauer
Associates.
Hirokawa, N., Sobue, K., Kanda, K., Harada, A., and Yorifuji, H.
1989. The cytoskeletal architecture of the presynaptic
terminal and molecular structure of synapsin 1. J. Cell Biol.
108, 111–126.
Holt, M., Cooke, A., Neef, A., and Lagnado, L. 2004. High
mobility of vesicles supports continuous exocytosis at a
ribbon synapse. Curr. Biol. 14, 173–183.
Holt, J. C., Xue, J. T., Brichta, A. M., and Goldberg, J. M.
2006. Transmission between type II hair cells and bouton
afferents in the turtle posterior crista. J. Neurophysiol.
95, 428–452.
Holton, T. and Weiss, T. F. 1983. Receptor potentials of lizard
cochlear hair cells with free-standing stereocilia in response
to tones. J. Physiol. 345, 205–240.
Hossain, W. A., Antic, S. D., Yang, Y., Rasband, M. N., and
Morest, D. K. 2005. Where is the spike generator of the
cochlear nerve? J. Neurosci. 25, 6857–6868.
Indresano, A. A., Frank, J. E., Middleton, P., and Jaramillo, F.
2004. Mechanical noise enhances signal transmission in the
bullfrog sacculus. JARO 4, 363–370.
Issa, N. P. and Hudspeth, A. J. 1994. Clustering of Ca2þ
channels and Ca2þ-activated Kþ channels at fluorescently
labeled presynaptic active zones of hair cells. Proc. Natl.
Acad. Sci. U. S. A. 91, 7578–7582.
Jacobs, R. A. and Hudspeth, A. J. 1990. Ultrastructural
correlates of mechanoelectrical transduction in hair-cells of
the bullfrogs internal ear. Cold Spring Harb. Symp. Quant.
Biol. 55, 547–561.
 Author's personal copy
Javel, E. and Mott, J. B. 1988. Physiological and
psychophysical correlates of temporal processes in hearing.
Hear Res. 34, 275–294.
Jeffress, L. A. 1968. Mathematical and electrical models of
auditory detection. J. Acoust. Soc. Am. 44, 187–203.
Johnson, D. H. 1980. The relationship between spike rate and
synchrony in responses of auditory-nerve fibers to single
tones. J. Acoust. Soc. Am. 68, 1115–1122.
Johnson, D. H. and Kiang, N. Y. S. 1976. Analysis of discharges
recorded simultaneously from pairs of auditory nerve fibers.
Biophys. J. 16, 719–734.
Johnson, S. L., Marcotti, W., and Kros, C. J. 2005. Increase in
efficiency and reduction in Ca2þ dependence of exocytosis
during development of mouse inner hair cells. J. Physiol.
563, 177–191.
Jones, E. M. C., Gray-Keller, M., and Fettiplace, R. 1999. The
role of Ca2þ-activated Kþ channel spliced variants in the
tonotopic organization of the turtle cochlea. J. Physiol.
518, 653–665.
Kandel, E. R., Schwartz, J. H., and Jessell, T. M. 2000.
Principles of Neural Science. McGraw-Hill.
Keen, E. C. and Hudspeth, A. J. 2006. Transfer characteristics
of the hair cell’s afferent synapse. Proc. Natl. Acad. Sci. U. S.
A. 103, 5537–5542.
Khimich, D., Nouvian, R., Pujol, R., tom Dieck, S., Egner, A.,
Gundelfinger, E. D., and Moser, T. 2005. Hair cell synaptic
ribbons are essential for synchronous auditory signalling.
Nature 434, 889–894.
Kiang, N. Y. S., Rho, J. M., Northrop, C. C., Liberman, M. C.,
and Ryugo, D. K. 1982. Hair-cell innervation by spiral
ganglion cells in adult cats. Science 217, 175–177.
Kiang, N. Y. S., Watanabe, T., Thomas, E. C., and Clark, L. F.
1965. Discharge Pattern of Single Fibers in the Cat’s
Auditory Nerve. MIT Press.
King, T. S. and Dougherty, W. J. 1982. Effect of denervation on
‘‘synaptic’’ ribbon population in the rat pineal gland. J.
Neurocytol. 11, 19–28.
Knudsen, E. I. and Konishi, M. 1978. A neural map of auditory
space in the owl. Science 200, 795–797.
Koschak, A., Reimer, D., Huber, I., Grabner, M., Glossmann, H.,
Engel, J., and Striessnig, J. 2001. 1D (Cav1.3) subunits can
form L-type Ca2þ channels activating at negative voltages. J.
Biol. Chem. 276, 22100–22106.
Koyama, H., Lewis, E. R., Leverenz, E. L., and Baird, R. A. 1982.
Acute seismic sensitivity in the bullfrog ear. Brain Res.
250, 168–172.
von Kriegstein, K., Schmitz, F., Link, E., and Sudhof, T. C.
1999. Distribution of synaptic vesicle proteins in the
mammalian retina identifies obligatory and facultative
components of ribbon synapses. Eur. J. Neurosci.
11, 1335–1348.
Krstic, R. 1976. Ultracytochemistry of the synaptic ribbons in
the rat pineal organ. Cell Tissue Res. 166, 135–143.
Lenzi, D. and von Gersdorff, H. 2001. Structure suggests
function: the case for synaptic ribbons as exocytotic
nanomachines. Bioessays 23, 831–840.
Lenzi, D., Crum, J., Ellisman, M. H., and Roberts, W. M. 2002.
Depolarization redistributes synaptic membrane and creates
a gradient of vesicles on the synaptic body at a ribbon
synapse. Neuron 36, 649–659.
Lenzi, D., Runyeon, J. W., Crum, J., Ellisman, M. H., and
Roberts, W. M. 1999. Synaptic vesicle populations in
saccular hair cells reconstructed by electron tomography. J.
Neurosci. 19, 119–132.
Lewis, E. R. and Narins, P. M. 1985. Do frogs communicate with
seismic signals? Science 227, 187–189.
Lewis, R. S. and Hudspeth, A. J. 1983. Voltage- and ion-
dependent conductances in solitary vertebrate hair cells.
Nature 304, 538–541.
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 393
Lewis, E. R., Baird, R. A., Leverenz, E. L., and Koyama, H. 1982.
Inner ear: dye injection reveals peripheral origins of specific
sensitivities. Science 215, 1641–1643.
Liberman, M. C. 1980. Morphological differences among radial
afferent fibers in the cat cochlea: an electron-microscopic
study of serial sections. Hear Res. 3, 45–63.
Liberman, M. C. 1982. Single-neuron labeling of the cat auditory
nerve. Science 216, 1239–1241.
Liberman, M. C. and Kiang, N. Y. S. 1978. Acoustic trauma in
cats. Cochlear pathology and auditory-nerve activity. Acta
Otolaryngol. Suppl. 358, 1–63.
Liberman, M. C., Dodds, L. W., and Pierce, S. 1990. Afferent
and efferent innervation of the cat cochlea: quantitative
analysis with light and electron microscopy. J. Comp.
Neurol. 301, 443–460.
Lindau, M. and Neher, E. 1988. Patch-clamp techniques for
time-resolved capacitance measurements in single cells.
Pflugers Arch. 411, 137–146.
Lioudyno, M., Hiel, H., Kong, J. H., Katz, E., Waldman, E.,
Parameshwaran-Iyer, S., Glowatzki, E., and Fuchs, P. A.
2004. A ‘‘synaptoplasmic cistern’’ mediates rapid inhibition
of cochlear hair cells. J. Neurosci. 24, 11160–11164.
Manley, G. A., Köppl, C., and Yates, G. K. 1997. Activity of
primary auditory neurons in the cochlear ganglion of the emu
Dromaius novaehollandiae: spontaneous discharge,
frequency tuning, and phase locking. J. Acoust. Soc. Am.
101, 1560–1573.
Martinez-Dunst, C., Michaels, R. L., and Fuchs, P. A. 1997.
Release sites and calcium channels in hair cells of the
chick’s cochlea. J. Neurosci. 17, 9133–9144.
Matsubara, A., Laake, J. H., Davanger, S., Usami, S., and
Ottersen, O. P. 1996. Organization of AMPA receptor
subunits at a glutamate synapse: a quantitative immunogold
analysis of hair cell synapses in the rat organ of Corti. J.
Neurosci. 16, 4457–4467.
McMahon, H. T., Missler, M., Li, C., and Sudhof, T. C. 1995.
Complexins: cytosolic proteins that regulate SNAP receptor
function. Cell 83, 111–119.
Merchan-Perez, A. and Liberman, M. C. 1996. Ultrastructural
differences among afferent synapses on cochlear hair cells:
correlations with spontaneous discharge rate. J. Comp.
Neurol. 371, 208–221.
Miller, M. R. and Beck, J. 1987. Heterotopic synaptic bodies in
the auditory hair cells of adult lizards. Anat. Rec. 218, 338–344.
Morgans, C. W. 2001. Localization of the 1F calcium channel
subunit in the rat retina. Invest. Ophthalmol. Vis. Sci.
42, 2414–2418.
Morgans, C. W., Brandstatter, J. H., Kellerman, J., Betz, H., and
Wassle, H. 1996. SNARE complex containing syntaxin 3 is
present in ribbon synapses of the retina. J. Neurosci.
16, 6713–6721.
Moser, T. and Beutner, D. 2000. Kinetics of exocytosis and
endocytosis at the cochlear inner hair cell afferent
synapse of the mouse. Proc. Natl. Acad. Sci. U. S. A.
97, 883–888.
Moser, T., Brandt, A., and Lysakowski, A. 2006. Hair cell ribbon
synapses. Cell Tissue Res. 326, 347–359.
Moser, T., Neef, A., and Khimich, D. 2006. Mechanisms
underlying the temporal precision of sound coding at the
inner hair cell ribbon synapse. J. Physiol. 576, 55–62.
Mountain, D. C. and Hubbard, A. E. 1996. Computational
Analysis of Hair Cell and Auditory Nerve Processes.
In: Auditory Computation (eds. H. L. Hawkins, T. A. McMullen,
A. N. Popper, and R. R. Fay), p. 121. Springer.
Muresan, V., Lyass, A., and Schnapp, B. J. 1999. The kinesin
motor KIF3A is a component of the presynaptic ribbon in
vertebrate photoreceptors. J. Neurosci. 19, 1027–1037.
Nakagawa, T., Oghalai, J. S., Saggau, P., Rabbitt, R. D., and
Brownell, W. E. 2006. Photometric recording of
 Author's personal copy
394 Afferent Synaptic Mechanisms
transmembrane potential in outer hair cells. J. Neural. Eng.
3, 79–86.
Neher, E. and Augustine, G. J. 1992. Calcium gradients
and uffers in bovine chromaffin cells. J. Physiol.
450, 273–301.
Nguyen, T. H. and Balkema, G. W. 1999. Antigenic epitopes of
the photoreceptor synaptic ribbon. J. Comp. Neurol.
413, 209–218.
Nibu, Y., Zhang, H., Bajor, E., Barolo, S., Small, S., and
Levine, M. 1998. dCtBP mediates transcriptional repression
by Knirps, Krueppel and Snail in the Drosophila embryo.
EMBO J. 17, 7009–7020.
Nicholls, J. G., Martin, A. R., Wallace, B. G., and Fuchs, P. A.
2001. From Neuron to Brain. Sinauer Associates.
Nouvian, R., Beutner, D., Parsons, T. D., and Moser, T. 2006.
Structure and function of the hair cell ribbon synapse. J.
Memb. Biol. 209, 1–13.
Osborne, M. P. and Thornhill, R. A. 1972. The effect of
monoamine depleting drugs upon the synaptic bars in the
inner ear of the bullfrog. Z. Zellforsch. Mikrosk. Anat.
127, 347–355.
Pabst, S., Hazzard, J. W., Antonin, W., Sudhof, T. C., Jahn, R.,
Rizo, J., and Fasshauer, D. 2000. Selective interaction of
complexin with the neuronal SNARE complex. J. Biol. Chem.
275, 19808–19818.
Palmer, A. R. and Russell, I. J. 1986. Phase-locking in the
cochlear nerve of the guinea-pig and its relation to the
receptor potential of inner hair-cells. Hear Res. 24, 1–15.
Parsons, T. D. and Sterling, P. 2003. Synaptic ribbon: conveyor
belt or safety belt? Neuron 37, 379–82.
Parsons, T. D., Lenzi, D., Almers, W., and Roberts, W. M. 1994.
Calcium-triggered exocytosis and endocytosis in an isolated
presynaptic cell: capacitance measurements in saccular hair
cells. Neuron 13, 875–883.
Plack, C. J., Oxenham, A. J., and Drga, V. 2006. Masking by
inaudible sounds and the linearity of temporal summation. J.
Neurosci. 26, 8767–8773.
Poortinga, G., Watanabe, M., and Parkhurst, S. M. 1998.
Drosophila CtBP: a hairy-interacting protein required for
embryonic segmentation and hairy-mediated transcriptional
repression. EMBO J. 17, 2067–2078.
Pyle, J. L., Kavalali, E. T., Piedras-Renteria, E. S., and
Tsien, R. W. 2000. Rapid reuse of readily releasable pool
vesicles at hippocampal synapses. Neuron 28, 221–231.
Raphael, Y. and Altschuler, R. A. 2003. Structure and
innervation of the cochlea. Brain Res. Bull. 60, 397–422.
Raymond, J., Puel, J. L., and Devau, G. 2003. Glutamate
transporters in the guinea-pig cochlea: partial mRNA
sequences, cellular expression and functional implications.
Eur. J. Neurosci. 17, 83–92.
Rea, R., Li, J., Dharia, A., Levitan, E. S., Sterling, P., and
Kramer, R. H. 2004. Streamlined synaptic vesicle cycle in
cone photoreceptor terminals. Neuron 41, 755–766.
Rebillard, G., Ruel, J., Nouvian, R., Saleh, H., Pujol, R.,
Dehnes, Y., Raymond, J., Puel, J. L., and Devau, G. 2003.
Glutamate transporters in the guinea-pig cochlea: partial
mRNA sequences, cellular expression and functional
implications. Eur J. Neurosci. 17, 83–92.
Reim, K., Mansour, M., Varoqueaux, F., McMahon, H. T.,
Sudhof, T. C., Brose, N., and Rosenmund, C. 2001.
Complexins regulate a late step in Ca2þ-dependent
neurotransmitter release. Cell 104, 71–81.
Reim, K., Wegmeyer, H., Brandstatter, J. H., Xue, M.,
Rosenmund, C., Dresbach, T., Hofmann, K., and Brose, N.
2006. Structurally and functionally unique complexins at
retinal ribbon synapses. J. Cell Biol. 169, 669–680.
Relkin, E. M. and Doucet, J. R. 1991. Recovery from prior
stimulation. I. Relationship to spontaneous firing rates of
primary auditory neurons. Hear Res. 55, 215–222.
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Ricci, A. J., Gray-Keller, M., and Fettiplace, R. 2000. Tonotopic
variations of calcium signalling in turtle auditory hair cells. J.
Physiol. 524, 423–436.
Rizzoli, S. O. and Betz, W. J. 2005. Synaptic vesicle pools. Nat.
Rev. Neurosci. 6, 57–69.
Roberts, W. M. 1993. Spatial calcium buffering in saccular hair
cells. Nature 363, 74–76.
Roberts, W. M. 1994. Localization of calcium signals by a
mobile calcium buffer in frog saccular hair cells. J. Neurosci.
14, 3246–3262.
Roberts, W. M., Jacobs, R. A., and Hudspeth, A. J. 1990.
Colocalization of ion channels involved in frequency
selectivity and synaptic transmission at presynaptic active
zone of hair cells. J. Neurosci. 10, 3664–3684.
Robertson, D. 1984. Horseradish peroxidase injection of
physiologically characterized afferent and efferent neurons
in the guinea pig spiral ganglion. Hear Res. 15, 113–121.
Robertson, D. and Paki, B. 2002. A role for purinergic receptors
at the inner hair cell-afferent synapse? Audiol. Neurootol.
7, 62–67.
Rodriguez-Contreras, A. and Yamoah, E. N. 2001. Direct
measurement of single-channel Ca2þ currents in bullfrog hair
cells reveals two distinct channel subtypes. J. Physiol.
534, 669–689.
Rodriguez-Contreras, A. and Yamoah, E. N. 2003. Effects of
permeant ion concentrations on the gating of L-type Ca2þ
channels in hair cells. Biophys. J. 84, 3457–3469.
Rosahl, T., Spillane, D., Missler, M., Herz, J., Selig, D.,
Wolff, J. R., Hammer, R. E., Malenka, R. C., and
Sudhof, T. C. 1995. Essential functions of synapsins I and II
in synaptic vesicle regulation. Nature 375, 488–493.
Rose, J. E., Brugge, J. F., Anderson, D. J., and Hind, J. E. 1967.
Phase-locked response to low-frequency tones in single
auditory nerve fibers of the squirrel monkey. J. Neurophysiol.
30, 769–793.
Rose, J. E., Hind, J. E., Anderson, D. J., and Brugge, J. F. 1971.
Some effects of stimulus intensity on response of auditory-
nerve fibers in the squirrel monkey. J. Neurophysiol.
34, 685–699.
Rosenmund, C., Rettig, J., and Brose, N. 2003. Molecular
mechanisms of active zone function. Curr. Opin. Neurobiol.
13, 509–519.
Ruel, J., Bobbin, R. P., Vidal, D., Pujol, R., and Puel, J. L. 2000.
The selective AMPA receptor antagonist GYKI 53784 blocks
action potential generation and excitotoxicity in the guinea
pig cochlea. Neuropharmacology 39, 1959–1973.
Russell, I. J. and Sellick, P. M. 1978. Intracellular studies of hair
cells in the mammalian cochlea. J. Physiol. 284, 261–290.
Rutherford, M. A. and Roberts, W. M. 2006. Frequency
selectivity of exocytosis in frog saccular hair cells. Proc. Natl.
Acad. Sci. U. S. A. 103, 2898–2903.
Safieddine, S. and Wenthold, R. J. 1999. SNARE complex at the
ribbon synapses of cochlear hair cells: analysis of synaptic
vesicle- and synaptic membrane-associated proteins. Eur. J.
Neurosci. 11, 803–812.
Schmitz, F. and Drenckhahn, D. 1991. Influence of Ca2þ on
synaptic morphology of fish cone photoreceptors. Brain Res.
546, 341–344.
Schmitz, F., Konigstorfer, A., and Sudhof, T. C. 2000. RIBEYE, a
component of synaptic ribbons: a protein’s journey through
evolution provides insight into synaptic ribbon function.
Neuron 28, 857–872.
Schnee, M. E. and Ricci, A. J. 2003. Biophysical and
pharmacological characterization of voltage-gated calcium
currents in turtle auditory hair cells. J. Physiol. 549, 697–717.
Schnee, M. E., Lawton, D. M., Furness, D. N., Benke, T. A., and
Ricci, A. J. 2005. Auditory hair cell-afferent fiber synapses
are specialized to operate at their best frequencies. Neuron
47, 243–254.
 Author's personal copy
Schneggenburger, R. and Neher, E. 2005. Presynaptic calcium
and control of vesicle fusion. Curr. Opin. Neurobiol.
15, 266–274.
Sewell, W. F. 1984. The relation between the endocochlear
potential and spontaneous activity in auditory nerve fibres of
the cat. J. Physiol. 347, 685–696.
Siegel, J. H. 1992. Spontaneous synaptic potentials from
afferent terminals in the guinea pig cochlea. Hear Res.
59, 85–92.
Siegel, J. H. and Brownell, W. E. 1981. Synaptic bodies in outer
hair cells of the chinchilla organ of Corti. Brain Res.
220, 188–193.
Siegel, J. H. and Brownell, W. E. 1986. Synaptic and golgi
membrane recycling in cochlear hair cells. J. Neurocytol.
15, 311–328.
Singer, J. H., Lassova, L., Vardi, N., and Diamond, J. S. 2004.
Coordinated multivesicular release at a mammalian ribbon
synapse. Nat. Neurosci. 7, 826–833.
Sjöstrand, F. S. 1958. Ultrastructure of retinal rod synapses of
the guinea pig eye as revealed by three-dimensional
reconstructions from serial sections. J. Ultrastruct. Res.
5, 13–17.
Slepecky, N. B., Galsky, M. D., Swartzentruber-Martin, H., and
Savage, J. 2000. Study of afferent nerve terminals and fibers in
the gerbil cochlea: distribution by size. Hear Res. 144, 124–134.
Smith, R. L. 1977. Short-term adaptation in single auditory
nerve fibers: some poststimulatory effects. J. Neurophysiol.
40, 1098–1111.
Smith, C. A. and Sjostrand, F. S. 1961. Structure of the nerve
endings on the external hair cells of the guinea pig cochlea
as studied by serial sections. J. Ultrastruct. Res. 5, 523556.
Smotherman, M. S. and Narins, P. M. 2000. Hair cells, hearing
and hopping: a field guide to hair cell physiology in the frog.
J. Exp. Biol. 203, 2237–2246.
Sneary, M. G. 1988. Auditory receptor of the red-eared turtle: II.
Afferent and efferent synapses and innervation patterns. J.
Comp. Neurol. 276, 588–606.
Sobkowicz, H. M., August, B. K., and Slapnick, S. M. 2004.
Synaptic arrangements between inner hair cells and tunnel
fibers in the mouse cochlea. Synapse 52, 299–315.
Sobkowicz, H. M., Rose, J. E., Scott, G. L., and Levenick, C. V.
1986. Distribution of synaptic ribbons in the developing
organ of Corti. J. Neurocytol. 15, 693–714.
Sobkowicz, H. M., Slapnick, S. M., and August, B. K. 2003.
Reciprocal synapses between inner hair cell spines and
afferent dendrites in the organ of corti of the mouse. Synapse
50, 53–66.
Spassova, M. A., Avissar, M., Furman, A. C., Crumling, M. A.,
Saunders, J. C., and Parsons, T. D. 2004. Evidence that
rapid vesicle replenishment of the synaptic ribbon mediates
recovery from short-term adaptation at the hair cell afferent
synapse. JARO 5, 376390.
Spoendlin, H. 1969. Innervation patterns in the organ of Corti of
the cat. Acta Otolaryngol. 67, 239–254.
Spoendlin, H. 1985. Anatomy of cochlear innervation. Am. J.
Otolaryngol. 6, 453–467.
Sterling, P. and Matthews, G. 2005. Structure and function of
ribbon synapses. Trends Neurosci. 28, 20–29.
Stern, M. D. 1992. Buffering of calcium in the vicinity of a
channel pore. Cell Calcium 13, 183–192.
The Senses: A Comprehensive Reference, vol. 3, pp. 365-395
Afferent Synaptic Mechanisms 395
Stevens, C. F. and Williams, J. H. 2000. ‘‘Kiss and run’’
exocytosis at hippocampal synapses. Proc. Natl. Acad. Sci.
U. S. A. 97, 12828–12833.
Sudhof, T. C. 2000. The synaptic vesicle cycle revisited. Neuron
28, 317–320.
Sudhof, T. C. 2004. The synaptic vesicle cycle. Annu. Rev.
Neurosci. 27, 509–547.
Sumner, C. J., Lopez-Poveda, E. A., O’Mard, L. P., and
Meddis, R. 2002. A revised model of the inner-hair cell and
auditory nerve complex. J. Acoust. Soc. Am. 111, 2178–2188.
Tsuji, J. and Liberman, M. C. 1997. Intracellular labeling of
auditory nerve fibers in the guinea pig: central and peripheral
projections. J. Comp. Neurol. 381, 188–202.
Usukura, J. and Yamada, E. 1987. Ultrastructure of the synaptic
ribbons in photoreceptor cells of Rana catesbeiana revealed
by freeze-etching and freeze-substitution. Cell Tissue Res.
247, 483–488.
Wagner, H.-J. 1978. Cell types and connectivity patterns in
mosaic retinas. Adv. Anat. Embryol. Cell Biol. 55, 1–81.
Wagner, H.-J. 1997. Presynaptic bodies (‘‘ribbons’’): from
ultrastructural observations to molecular perspectives. Cell
Tissue Res. 287, 435–446.
Wan, L., Almers, W., and Chen, W. 2005. Two ribeye genes in
teleosts: the role of Ribeye in ribbon formation and bipolar
cell development. J. Neurosci. 25, 941–994.
Xu, W. and Lipscombe, D. 2001. Neuronal Cav1.31 L-type
channels activate at relatively hyperpolarized membrane
potentials and are incompletely inhibited by
dihydropyridines. J. Neurosci. 21, 5944–5951.
Yamazaki, H., Nakata, T., Okada, Y., and Hirokawa, N. 1995.
KIF3A/B: a heterodimeric kinesin superfamily protein that
works as a microtubule plus end-directed motor for
membrane organelle transport. J. Cell Biol. 130, 1387–1399.
Yasunaga, S., Grati, M., Chardenoux, S., Smith, T. N.,
Friedman, T. B., Lalwani, A. K., Wilcox, E. R., and Petit, C.
2000. OTOF encodes multiple long and short isoforms:
genetic evidence that the long ones underlie recessive
deafness DFNB9. Am. J. Hum. Genet. 67, 591–600.
Yasunaga, S., Grati, M., Cohen-Salmon, M., El-Amraoui, A.,
Mustapha, M., Salem, N., El-Zir, E., Loiselet, J., and Petit, C.
1999. A mutation in OTOF, encoding Otoferlin, a FER-1-like
protein, causes DFNB9, a nonsyndromic form of deafness.
Nat. Gen. 21, 363–369.
Yu, X., Lewis, E. R., and Feld, D. 1991. Seismic and auditory
tuning curves from bullfrog saccular and amphibian papillar
axons. J. Comp. Physiol. A 169, 241–248.
Zenisek, D., Davila, V., Wan, L., and Almers, W. 2003. Imaging
calcium entry sites and ribbon structures in two presynaptic
cells. J. Neurosci. 23, 2538–2548.
Zhai, R. G. and Bellen, H. J. 2004. The architecture of the active
zone in the presynaptic nerve terminal. Physiology 19, 262–270.
Further Reading
Rizzoli, S. O. and Betz, W. J. 2004. The structural organization
of the readily releasable pool of synaptic vesicles. Science
303, 2037–2039.