Representing the Acoustic World within the Brain: Normal and Abnormal Development of Frequency Maps in the Auditory System

Full text

Conference Keynote Address
Representing the Acoustic World within the Brain:
Normal and Abnormal Development of
Frequency Maps in the Auditory System
Robert V. Harrison
Introduction
This discourse is concerned with a fundamental
way in which acoustical aspects of the outside world
are represented within the brain. More specifically we
consider the way in which spectral components of
acoustic signals are represented by neural activity in
tonotopically organized arrays.
The review will start with general notions about
how a version of the world (outside of our heads) is
recreated in the brain, using (limited) information
from our various sensory organs. The process involves
both the coding of stimuli at the level of the sensory
epithelia, and the faithful transmission of informa-
tion from the periphery to the cerebral cortex. We will
look briefly at these concepts in visual and somato-
sensory systems before considering the equivalent
organization in the auditory system, namely cochle-
otopic or tonotopic mapping.
We will discuss the normal development of cen-
tral tonotopic maps and then the abnormal develop-
ment of these representations. We will consider the
plasticity of the developing system, as revealed by
studies that show a reorganization of central tono-
topic maps after experimental manipulations of the
sensory (auditory) input. We will explore some differ-
ences between the tonotopic map plasticity of the
developing auditory system and that of the adult
animal. Finally we will discuss some clinical implica-
tions which arise from this and related work. Be it
known here that the references provided are but
a small sample of representative work; no attempt
is made to provide a exhaustive list. For further,
Address correspondence to: Dr. Robert V. Harrison, Department of
Otolaryngology, The Hospital for Sick Children, 555 University
Avenue, Toronto, Ontario, Canada M5G 1X8
original sources one might consult the reference lists
from those papers cited.
Representation of the External World
in the Brain
One of the earliest theories that a version of the
external world is reproduced within our brain is to be
found in the work of Rene Descartes. This concept is
beautifully illustrated in figure 1 taken from the work
of the 17th century philosopher. The world in which
we live is really an illusion. We believe that there are
objects and sounds around us, but in reality these
percepts are all within our brain. What we see and
hear and feel are actually portrayals of the outside
world within the cerebral cortex. These representa-
tions are only as good as the ability of our sensory
organs to transduce the external stimuli and the abil-
ity to transfer this sensory information to cerebral
cortex. We only “see” an image that impinges on the
retina, produces a pattern of neural activity at the
level of the photo-receptors, and is faithfully trans-
mitted to visual cortex. Similarly, we only “hear”
those acoustic signals that activate the sensory epi-
thelium of the cochlea, and which produce a pattern
of neural activity that is faithfully transmitted to cen-
tral auditory areas.
It is clear that the basic organization of the audi-
tory system is very similar to other sensory systems.
As shown in figures 2 and 3, images coded on the
retina are faithfully transmitted through the thal-
amus to primary visual cortex in a systematic way
such that we talk about retinotopic projections. This
is essentially a spatially organized or topographic pro-
jection system. Thus as shown in figure 3 we see how
CHAPTER ONE
3

Figure 1. Representation of the outside world, as visualised by
Rene Descartes.
Figure 2. Ordered projections from the retina to cortex in the visual system.
Figure 3. The retinotopic organisation of primary visual
cortex.
4 a A Sound Foundation Through Early Amplification

Figure 4. The sensory homunculus; how somatosensory cortex
proportionally represents the sensory epithelium of the body
surface.
Figure 5. Original illustration of the sensory homunculus by Wilder Penfield.
the representation of the retinal
activity (including that in foveal
and extra-foveal areas) is system-
atically re-represented at the level
of primary visual cortex. Similarly
for the somatosensory system, as
shown in figure 4, the body surface
(more specifically activity from the
sensory epithelium of the skin) is
re-represented in somatosensory
cortex. A more familiar represen-
tation of the sensory homunculus
is illustrated in figure 5 (Penfield
1958).
Sound Frequency Maps
in the Auditory System
In the auditory system we
have exactly the same type of
“main-line” organization as in
other sensory systems. Neural
activity patterns generated along
the sensory epithelium of the coch-
lea (figure 6) are re-represented at
all levels within the auditory
pathway up to and including pri-
mary auditory cortex and areas
Figure 6. Sensory epithelium of the cochlea.
Representing the Acoustic World within the Brain a 5

Figure 7. Central projections in auditory system as illustrated by Netter.
6 a A Sound Foundation Through Early Amplification

beyond. This organization has been beautifully illus-
trated by Netter, as shown in figure 7 (a black and
white and therefore somewhat degraded adaptation
of his color original). This figure also serves to remind
us that the sensory epithelium of the cochlea is
performing a place-coding of sound frequency. Strictly
Figure 8. Multiple cochleotopic maps in auditory cortex of cat.
Figure 9. Cochleotopic or tonotopic organisation in primary auditory cortex of
chinchilla.
speaking, we should say that the auditory pathways
have cochleotopic projections (in analogy to retino-
topic projections in the visual system), but because
the cochlea performs a place coding of sound fre-
quency we can also use the terms tonotopic projection
or frequency map. However, it is always very useful to
remember that neural activity in
any area of a “frequency map” rep-
resents sound frequency only in a
very indirect sense. Strictly speak-
ing, it is a representation of activ-
ity generated at a particular pos-
ition along the cochlear length
that happens to be selectively
detecting a particular frequency of
acoustic signal. This then is pri-
marily the way in which the audi-
tory system re-represents the
exterior acoustic world within
cerebral cortex. For the more
sophisticated reader, yes, this is a
“stripped-down” version of the sys-
tem; we ignore for the moment
(and for the rest of this review)
temporal coding and binaural
mechanisms including sound
localization.
In figure 8, for the cat (based
on Woolsey and Walzl 1942) we see
that primary auditory cortex has a
cochleotopic or tonotopic represen-
tation and that many other sec-
ondary auditory areas also have
tonotopic maps.
Figures 9 and 10 illustrate
two experimental methods that
can reveal the tonotopic maps in
auditory cortex. Figure 9 (from
Harrison, Kakigi, Hirakawa,
Harel and Mount 1996b) shows a
map based on recording the best
frequency of response (character-
istic frequency) of individual
neurons at many different sites in
temporal cortex (of chinchilla).
Figure 10 (adapted from Harel,
Mori, Sawada, Mount and Har-
rison 2000) shows a tonotopic map
derived by optical imaging of the
blood flow changes in temporal
cortex that result from pure tone
Representing the Acoustic World within the Brain a 7

Figure 10. Chinchilla cortical tonotopic map revealed by optical imaging of hemo-
dynamic changes that result from acoustically driven neural activity.
Figure 11. Simple conceptualisation of central auditory projections during early
development.
stimulation at different frequen-
cies (as shown by the gray scale
key).
Development and
Plasticity of
Tonotopic Maps
Central (cortical) tonotopic
maps are almost certainly not
present at birth or during early
stages of development. Whilst the
human cochlea is completely func-
tional at birth (as well as the mid-
dle ear, give or take a drop or two
of fluid), the central auditory brain
is very immature. During infancy
and adolescence there is a continu-
ing maturation of the central audi-
tory pathways as revealed, for
example, by the change in proper-
ties of the auditory evoked poten-
tials (e.g. Eggermont 1985,1988).
Auditory brain stem evoked
responses (ABRs) take up to five
years to “mature”. Middle latency
responses (MLR) and other poten-
tials from auditory cortex are not
adult-like” for 12–15 years. Much
of this maturation is paralleled by
anatomical developments of cere-
bral cortex.
Figure 11 illustrates schemat-
ically the projections from the
cochlea to auditory cortex. During
early development it is known
that the projections from one level
to the next are not direct point-to-
point connections, but rather there
is considerable divergence of the
connections at all levels. Under
these conditions then, patterns of
neural activity that represent
sounds at the level of the cochlea
will not be perfectly transmitted
from the sensory epithelium to the
cortex.
However, in the adult animal
as modeled in figure 12, we see
that the divergent connections
8 a A Sound Foundation Through Early Amplification

have become much more direct. Now, with a precise
point-to-point projection system, neural activity pat-
terns that represent sound frequencies at the level
of the cochlea are faithfully transmitted to the cortex.
There are various mechanisms whereby the
organization of the pathways becomes more ordered,
i.e. changes from that in figure 11 to that of figure 12.
For example, some lateral (divergent) connections die
out, by mechanisms not clearly understood. More dir-
ect connections remain or are made because they “win
out” in the competition to make synapses on the tar-
get cell soma. The rules of the “competition” are likely
to be very complex, based, for example, on physical
space available at target sites, as well as complex
chemical signaling mechanisms for axonal and syn-
aptic growth. The synapses which become part of the
most direct pathway are strengthened compared with
those in less direct connections, perhaps because of
Hebbian type processes which favor connections
between cells which have highly correlated activity
(Hebb 1949). In respect to Hebbian strengthening it is
worth mentioning that acoustically driven activity
will be a more influential stimulation than (randomly
distributed) spontaneous activity. Various mechan-
isms for synaptic strengthening including, for
example “long term potentiation” have been explored
at the membrane level including the now well known
activity-driven modifications to NMDA post synaptic
receptor channels. Space does not permit any detailed
review, but suffice to say that a host of mechanisms
Figure 12. Cochlea and cortex connected up with good point-to-point linkages.
probably contribute to the early
development of point-to-point con-
nectivity, and to the strengthening
of active pathways.
Experiments Exploring
the Developmental
Plasticity of Tontopic
Maps
Let us now turn our attention
to ways in which these tonotopic
maps can become reorganized, or
fail to organize correctly. Under
discussion here is the plasticity of
the auditory system, i.e. the in-
herent ability of the auditory sys-
tem to modify or reorganize. There
are many ways in which the plas-
ticity of the auditory system has
been investigated. Some of the very earliest experi-
ments were anatomical studies (e.g. Levi-Montalcini
1949; Webster and Webster 1979; Parks 1979; Rubel
1985) in which the developing auditory system was
lesioned and the anatomical changes to neural path-
ways were investigated (e.g. making cell counts or
looking at how axonal pathways are changed). There
have been physiological studies in which the devel-
opment of the auditory system has been monitored
using auditory evoked potentials (e.g. Eggermont
1985, 1988). Another way of investigating the plas-
ticity of the auditory system is to carry out behavioral
studies. Anyone who does a psychophysical study
(including clinicians measuring an audiogram) and
repeats that study over a number of days or weeks
and sees changing results may be revealing the plas-
ticity of the auditory system.
Here, I want to focus on data from single unit
electrophysiological mapping studies that show some
aspects of the developmental plasticity of the auditory
system, particularly in tonotopic map development.
Essentially these studies have measured tonotopic
maps at various levels in the auditory system after
changing the input to the system by making lesions to
the cochlea or by inducing other abnormal patterns of
activity at the level of the cochlea.
It is important here to draw your attention to the
difference between developmental plasticity, and
adult plasticity. There are a number of experimental
plasticity studies that have been carried out in the
Representing the Acoustic World within the Brain a 9

mature subject. In fact, the first
studies to look at tonotopic map
changes in auditory cortex result-
ing from lesions to the cochlea
were done in the adult animal
(Robertson and Irvine 1989;
Rajan, Irvine, Wise and Heil 1993;
Rajan and Irvine 1996, 1998b;
Kakigi, Hirakawa, Harel, Mount
and Harrison 2000). However, the
data that I show here is related to
developmental plasticity, where
changes to the input to the system
(e.g. by lesioning the cochlea), are
made during early development.
In this case, the animal experi-
ences an abnormal input to the
system during important early
developmental periods (e.g. Willott
1984; Harrison, Nagasawa, Smith,
Stanton and Mount 1991; Har-
rison, Smith, Nagasawa, Stanton
and Mount 1992; Harrison, Ibra-
him, Stanton and Mount 1996a;
Stanton and Harrison 1996, 2000).
Figure 13 shows the general
protocol for a typical develop-
mental plasticity experiment. It
depicts the time course of the sub-
ject from birth to maturity. In
these studies we are changing the
condition of the cochlea at birth
using amikacin injections to cause
cochlear hair cell loss. Amikacin is
an aminoglycoside antibiotic that,
in high concentration, is ototoxic.
It primarily causes damage to the
basal (high-frequency) region of
the cochlea. After making a coch-
lear lesion, the animal is allowed
to mature and then we carry out
various functional and anatomi-
cal studies.
The first experiments that I
review here were made in the cat.
If we do nothing to the newborn
kitten, i.e. if it matures normally,
then we will record a normal fre-
quency map in primary auditory
cortex as is illustrated in figure 14.
We can suppose that in this
Figure 13. General protocol for experiments on developmental plasticity.
Figure 14. Tonotopic map in normal cat cortex based on single unit electro-
physiological responses.
10 a A Sound Foundation Through Early Amplification

Figure 15. Tonotopic map in cat cortex after basal cochlear lesion induced during an
early neonatal period.
subject there has been a good
development of point to point
projections such that in auditory
cortex there is a very accurate rep-
resentation of what is happening
at the level of the cochlea (as de-
picted in the “model” of figure 12).
Figure 15 shows the results
from an experimental animal in
which we induced a basal cochlear
lesion within a few days of birth.
This cochlear damage results in a
high frequency cochlear hearing
loss as shown in the audiogram
(derived from ABR thresholds to
tone pips). We allow the cat to
mature (for 6–9 months) and then
make our cortical mapping study.
We find a very abnormal tonotopic
map in primary auditory cortex.
First, note that the (low frequency)
area of this tonotopic map, which
corresponds to the normal (apical)
region of the cochlea, is normal.
For example this region has rela-
tively normal separation of octave
spaced iso-frequency contours. We
can suppose that where there is a
normal level of (driven) activity in
cochlear afferent neurons there
will be a normal cortical map. The
big change in the frequency map of
this experimental animal is in the
de-afferented cortical region.
Where normally there would be
input from high-frequency regions
of the cochlea, all of the neurons
respond best to sounds around 6–8
kHz. In other words we have a
very large cortical area which
appears to connect up to one par-
ticular region of the cochlea. This
cochlear location approximates to
where there are surviving hair-
cells on the border of the cochlear
lesion.
A simple interpretation of
these experimental findings is
depicted in figure 16. We have
damaged the base of the coch-
lea, and the associated cochlear
Representing the Acoustic World within the Brain a 11

afferent neurons have degenerated. At the next level
in the system (cochlear nucleus), the divergence
which we find during normal early development is
maintained because there is no competition from
adjacent neurons to make or strengthen synaptic con-
tacts. We could imagine this process occurring at all
levels in the auditory system such that we end up
with a large iso-frequency region of auditory cortex
where all the neurons are essentially connected up
from one point along the cochlear length, at the bor-
der of the experimental lesion.
Figure 17 shows data from another experiment.
In this case, the lesion that we have made to the coch-
lea during early development is much more extensive,
as reflected in the sloping ABR audiogram. Histo-
logical studies of this cochlea showed various degrees
of haircell degeneration all the way up to the apex. In
the cortical frequency map of this animal we note a
large iso-frequency region in which all the neurons
respond best to sound frequencies near to 6.6 kHz. We
also see that there is another iso-frequency region
where neurons respond best to about 0.6–0.7 kHz. In
between these regions we see a very abnormal
tonotopic map; the spacing of iso-frequency contours
and their shape are clearly unusual. We can conclude
from these data that the establishment of tonotopic
maps in auditory cortex depends, in large part, on the
pattern of activity coming from the periphery. If that
Figure 16. Simple conceptual model of central projections after basal cochlear dam-
age induced during an early post-natal period.
pattern of activity is abnormal,
then an abnormal central repre-
sentation develops.
The studies described so far
have investigated the effects
removing or reducing the input
that comes from the cochlea dur-
ing early development. What
might happen to central frequency
maps if we have an increased level
of activity in certain cochlear
regions? To address this question
we have carried out a study in
which, during an early postnatal
period, an acoustic stimulus was
used to constantly activate a
particular region of the cochlea
(Stanton and Harrison 1996). We
reared kittens in an acoustic
environment in which there was a
constant 8 kHz tone which was
modulated ±1 kHz, at modulation
rate of 1 Hz. We thus had a
stimulus that was frequency specific but also con-
stantly changing so as to avoid both adaptation
effects and acoustic trauma to the cochlea. This
stimulus was presented such that at the kittens’
ears it was approximately 60 dB SPL. Kittens were
born into this acoustic environment and remained
in it for 3 months. Six to nine months later,
the mature cats were used in cortical mapping
experiments.
Figure 18 shows the tonotopic map in cortex of a
normal control animal (upper panel) compared with
an experimental “augmented” animal (lower panel).
On these maps, iso-frequency contours are spaced at
one octave intervals. Note in particular the spacing of
the 8–16 kHz region (cross-hatched area). The map
from the subject which was reared in the environ-
ment containing that 8 kHz signal shows a significant
increase in the amount of cortex which is coding, or
representing frequencies at 8 kHz and above. Figure
19 shows pooled results from 3 experimental animals,
versus 3 age-matched controls. The graph shows the
percentage of auditory cortex that is devoted to par-
ticular sound stimulus frequencies. In the animals
that were reared in an environment where there was
the constant 8 kHz signal, we see a peak in the cor-
tical representation for frequencies from 8–12 kHz.
The over-representation extends above 8 kHz because
of the suprathreshold level of the 8 kHz conditioning
12 a A Sound Foundation Through Early Amplification

Figure 17. Effects of neonatal cochlear hearing loss on frequency mapping in cat auditory cortex.
Representing the Acoustic World within the Brain a 13

Figure 18. Lower panel: over representation of the 8–16 kHz region of cortical frequency map (cross-hatched) after rearing kitten
in presence of a constant 8 kHz acoustic signal. Upper panel shows normal control data.
14 a A Sound Foundation Through Early Amplification

Figure 19. Amount of auditory cortex devoted to different sound frequencies in kit-
tens (n = 3) reared in presence of a constant 8 kHz acoustic stimulus (continuous line).
Dotted line shows data from three normal controls.
Figure 20. Simple schematic model suggesting the pattern of cochleotopic projections
in kitten reared with an over activation of one cochlear region.
signal which produces a basal
(higher frequency) spread of the
activated cochlear region.
Our “model” to explain what
might be happening is shown in
figure 20. Because there is an
over-stimulation of a certain
region of the cochlea, we suppose
that these more active neurons
win in the competition to form
synapses on target cells. The
synapses associated with the
active neurons will be more
strengthened than less active sur-
rounding neurons. This strength-
ening will involve Hebbian type
mechanisms, and well as those
known to mediate various types of
long-term potentiation. Ultimately
one can see how this could result
in an over-representation at the
level of cortex of one particular
frequency region of the cochlea.
These data are further evidence
that central tonotopic maps do not
result from some hard wiring pro-
cess governed by a genetic blue-
print, but rather they form, in
large part, as a consequence of the
pattern of activity that comes from
the level of the cochlea during
early development.
Plasticity of
Sub-Cortical
Frequency Maps
We next asked the question,
are these reorganizations of cor-
tical frequency maps just limited
to cortex, or are there similar
changes at lower levels in the
auditory pathway? If the “models”
of figures 16 or 20 are half accur-
ate, one might predict sub-cortical
changes. In the following set of
experiments we have looked at
changes to tonotopic maps at the
level of the mid-brain, i.e. the cen-
tral nucleus of inferior colliculus.
Representing the Acoustic World within the Brain a 15

Figure 21. Normal frequency mapping in the inferior colliculus of the chinchilla.
Figure 22. Tonotopic representation in auditory mid-brain of chinchilla after neonatal
cochlear lesion.

The experimental protocol is similar to that outlined
in figure 13. The animal (in this case the chinchilla) is
born and then we immediately induce a lesion to the
cochlea taking advantage of the ototoxic effects of the
aminoglycoside amikacin. We allow the animal to
mature (3 months) and then we investigate tonotopic
mapping in inferior colliculus.
Figure 21 shows the tonotopic map in the inferior
colliculus (central nucleus) of a normal chinchilla. As
depicted in the lower left panel, an electrode intro-
duced into the dorsal region will record from neurons
responding best to low frequencies. As the electrode
progresses ventrally, we encounter cells responding
best to higher frequencies. The progression of best (or
characteristic) frequency as a function of electrode
excursion is clearly shown in the upper left plot. Note
that in the normal animal, there is a
very clear linear relationship between
electrode position and characteristic
frequency on a logarithmic frequency
scale. As can be noted in the lower left
panel, the octave interval iso-frequency
contours are relatively evenly spaced.
Results from one experimental
animal are shown in figure 22 (adapted
from Harrison, Ibrahim and Mount
1998). In this chinchilla, as a neonate,
we made a widespread cochlear lesion
which started in the basal cochlear
turn and extended apically. The func-
tional consequences of this cochlear
damage are reflected in the ABR
audiogram. In this animal we can note
(upper right panel) an iso-frequency
region in the mid-brain tonotopic map
at around 2.7 kHz. This matches up
with the mid-point of the cut-off slope
of the ABR audiogram, which in turn
corresponds to the main border area of
damage along the cochlear length. In
many ways, the mid-brain tonotopic
map changes are rather similar to
what we found at the level of auditory
cortex in our other experiments. In this
chinchilla we also note that the tono-
topic map for lower frequency regions
is abnormal (compare with control
animal of figure 21). We suppose that
this is related to abnormal activity pat-
terns arising from the damaged sens-
ory epithelium of apical regions of the
cochlea. This again reflects what we have previously
observed at the level of cortex in other experiments.
Therefore, in a developmental model, the
reorganization of tonotopic maps is not exclusively
cortical but is present at lower levels in the system.
Indeed one might speculate that if there is abnormal
sensory input during early development, some degree
of reorganization is likely to be present wherever
there are synapses to be modified.
As already pointed out, after the cochlea is
lesioned in the mature animal one can see evidence of
reorganization to cortical tonotopic maps (e.g. Robert-
son and Irvine 1989; Rajan et al. 1993; Kakigi et al.
2000). However at the mid brain level we have a
different picture. Figure 23 shows results from two
tonotopic mapping experiments in inferior colliculus
Figure 23. Tonotopic mapping at level of inferior colliculus in chinchilla after basal
cochlear lesion induced in the neonatal animal (left-hand panels) versus in the
adult subject (right-hand panels).
Representing the Acoustic World within the Brain a 17

in which the cochlear lesion was
made neonatally (left) compared
with in the adult animal (right).
After a neonatal cochlea lesion, we
see evidence of map reorganiza-
tion (e.g. the large iso-frequency
region at 10 kHz). In contrast, for
the subject lesioned as an adult
there is no iso-frequency region or
other evidence of map reorganiza-
tion. To reiterate, at the level of the
inferior colliculus, the reorganiza-
tion of tonotopic maps shown in a
developmental model (i.e. where
the auditory system is interfered
with during an early postnatal
period) is not found in an adult
plasticity model. A similar lack of
plasticity in the adult animal has
been previously reported at the
level of the cochlear nucleus
(Kaltenbach, Meleca and Falzar-
ano 1996; Rajan and Irvine 1998a).
One conclusion from these
studies is that whilst the highest
levels of the auditory system (cor-
tex) can remain plastic in the
adult animal, lower levels of the
auditory system may only be plas-
tic during certain early stages in
development. In a sense we are
saying that the plasticity of lower
levels of the auditory system is age
related and may even have some
critical period of plasticity after
which there is little possibility for
reorganization.
What can Interfere
with the Central
Representation of
Complex Signals?
In most of the studies reported
here, there has been a radical
change to the sensory input. In the cochlear lesion
experiments the animals have the equivalent of a
moderate to severe sensorineural hearing loss from
birth, that clearly translates to very abnormal activ-
ity patterns arising from the cochlea. However, I want
to suggest that much more subtle changes, which
could occur not just at the level of the cochlea but at
higher levels in the system, might potentially inter-
fere with the development (and maintenance) of
point-to-point projections.
Figure 24. A complex acoustic signal such as this speech utterance can be broken
down, as shown here, into its spectral components.
Figure 25. Part of the bank of narrow, overlapping band pass filters that constitute
the peripheral hearing mechanism.
18 a A Sound Foundation Through Early Amplification

Figure 24 serves to illustrate
the task of the peripheral auditory
system. This spectrogram shows
how a complex speech signal (“we
play ball on summer days”) can be
analyzed into its frequency com-
ponents, as a function of time. The
cochlea carries out a similar oper-
ation. This image could easily
represent the pattern of neural
activity generated in the cochlear
nerve in response to this speech
signal. Imagine that the ordinate
is not a frequency scale but rather
the neural array from the apex to
the base of the cochlea. (This
transformation has been made in
figures 26–29.) This pattern of
neural activity is achieved by the
narrow band-pass filtering ele-
ments of the cochlea. Figure 25
serves to remind us about the
characteristics of the filtering
elements at the level of the cochlea
which are performing this fre-
quency analysis. This diagram
shows the frequency threshold
curves or tuning curves from
normal cochlear nerve fibers
(recorded from the guinea pig).
The process of the initial coch-
lear frequency analysis is very
important. We are all well aware
that there are many pathological
cochlear conditions that can de-
grade this analysis. For example
all types of sensorineural hearing
loss involving outer hair cell dam-
age will reduce not just the sensi-
tivity of the system, but also its ability to separate out
frequency components (e.g. Kiang, Moxon and Levine
1970; Evans and Harrison 1976; Dallos and Harris
1977; Harrison and Evans 1977, 1979). However, the
initial frequency analysis is only part of the whole
process of getting patterns of neural activity faithfully
transmitted to cerebral cortex. The point-to-point
transmission system can be potentially degraded in
many ways.
Figures 26–29 show, diagrammatically, some of
the possible ways in which the projection system from
cochlea to cortex might be degraded. The lower panel
of each figure represents the neural array of activity
at the cochlear nerve level evoked by the speech
signal “we play ball on summer days” (adapted from
figure 24). The upper panels show (two) stages of pro-
gressive degradation of the signal. For illustrative
purposes the clarity of the auditory percept is repre-
sented in each panel by the text. This is a sort of
visual depiction of the degradation of the sound
percept!
For figure 26 the small degradation is a “blur” in
the clarity of representation of frequency across the
neural array. This degradation in spectral frequency
Figure 26. Schematic diagrams of the spatio-temporal patterns of neural activity
representing a speech signal at three levels within the auditory system. Cumulative
effect of a poor maintenance of spatial activity pattern across neural array (spatial
“blur”).
Representing the Acoustic World within the Brain a 19

Figure 27. Schematic diagrams of the spatio-temporal patterns of neural activity
representing a speech signal at three levels within the auditory system. Effect of
abnormal noise levels.
Figure 28. Schematic diagrams of the spatio-temporal patterns of neural activity
representing a speech signal at three levels within the auditory system. Effect of
progressive reduction in number of information channels.

representation could initially be due to
poor cochlear filtering as found in sen-
sorineural hearing loss. It could also
result from poor lateral inhibition in
neurons throughout the central auditory
system, or perhaps a de-synchronization
of neurons due to myelination disorders.
More importantly, it could result from a
failure to establish good point-to-point
projections during early development. In
other words it could represent a failure of
the projection system to change from its
early divergent innervation pattern
depicted in figure 11 to the more direct
pattern shown in figure 12.
Another way in which we can degrade
information reaching cortex from the
cochlea is by adding noise, as illustrated
in figure 27. Physiologically, such noise
may result from a lack of inhibition in
the neural systems involved. There may
be pathological conditions that produce
noise, e.g. too many “leaky” synapses, neu-
ronal injury discharges. Perhaps there is
a link here to tinnitus.
The next example of a potential
source of degradation is depicted in figure
28, and relates to reducing the number of
channels of information, which can carry
neural representations of sound to cortex.
This is the situation, for example, in sub-
jects having a cochlear implant device in
which the channel numbers are consider-
ably reduced (down to the number of
implanted electrodes). Auditory neur-
opathy might also result, in many cases,
from a significant reduction in the num-
ber of channels bringing information from
the cochlea to the cortex (see Harrison 1998 for
further arguments).
The final example of information degradation is
perhaps the one most likely to have a real biological
basis. In figure 29, the “blur” occurs in the time
domain. Anything that interferes with the transmis-
sion of action potentials or electrotonic conduction
(e.g. various postsynaptic potentials) could smear the
neural representation of a signal. One major design
weakness of the cochlea-to-cortex projection system is
that it involves multiple stages, each interrupted by
synapses. There is always some timing uncertainty
at synapses. There are many pathological conditions
in which synaptic function might be impaired, leading
to increased time jitter. More importantly a delay or
failure to produce synaptic strengthening during
early development might lead to a degraded system
in the mature subject. Given that lower levels of the
auditory pathways may only be “plastic” for a certain
early time period, the adults system might be per-
manently impaired.
Figure 29. Schematic diagrams of the spatio-temporal patterns of neural activ-
ity representing a speech signal at three levels within the auditory system.
Effect of increasing timing uncertainty (“temporal blur”).
Representing the Acoustic World within the Brain a 21

Clinical Implications
One of the main “take home messages” from the
experimental animal studies is that during an early
developmental period, activity patterns arising at the
cochlear level are instrumental in establishing cen-
tral tonotopic maps (as well as other central sound
coding mechanisms). From a clinical perspective, any
early conditions that might change cochlear acti-
vation patterns will possibly disrupt normal central
development. These conditions could for example
include chronic conductive hearing loss caused by
repeated middle ear infection.
Another important message from the studies is
that the plasticity of the auditory system appears to
be much greater in the developing animal compared
to the adult. Thus our studies (Harrison 2001) have
indicated that at the level of the mid-brain, there
is little evidence of plasticity in the adult animal com-
pared with that found during early development.
Therefore, when we consider the plasticity of the
auditory system we have to bear in mind that differ-
ent levels in the system have different degrees of plas-
ticity. We have evidence that at the level of primary
auditory cortex (and beyond) plasticity is present
perhaps for life. This does not appear to be the case
for the lower level “gateway”. At lower levels there is
“age related plasticity” and perhaps some well-
defined “critical periods”.
Clinically, if we are trying to ameliorate or com-
pensate for congenital auditory deficits, for example
with a hearing aid or a cochlear implant, the best
results are going to be obtained when we do this in
young infants. Indeed if they are not young enough at
time of intervention, it may be too late. This impera-
tive is, of course, driving the establishment of
universal infant hearing screening programs. Prac-
tically, the early hearing screening is feasible, but
then the greater challenge is to provide useful inter-
vention very soon after. That issue was the main topic
of the conference reported in this volume.
Important Summary Points
•
Speech and other complex sounds are repre-
sented in the brain by activity patterns within
large neural arrays.
•
These activity patterns are established along the
sensory epithelium of the cochlea by mechanisms
that can separate out closely spaced frequency
components, and represent them at different
places along the cochlea.
•
These tonotopically mapped patterns have to be
faithfully re-represented in auditory cortex, for
further processing.
•
Both the cochlear mechanisms and the transmis-
sion pathways to cortex can be disrupted such
that central representations of sounds are
abnormal.
•
The central pathways develop, in part, as a result
activity patterns from the cochlea, and can
develop abnormally if there is cochlear dysfunc-
tion. Even very small changes to transmission in
neural signals will result in poor representation
of sound at cortex
•
There appears to be considerable plasticity in
whole pathway during early development, but in
the adult subject plasticity at lower level (e.g.,
brainstem, midbrain) is considerably reduced,
perhaps lost.
•
Clinically these data indicate that the early post
natal period is very important for the establish-
ment of auditory pathways that can accurately
represent complex sounds at the cortical level.
•
There appears to be a critical period of plasticity,
at least for the lower “gateway” pathways of the
auditory system. Thus, every effort should be
made to detect hearing problems early (e.g., with
neonatal hearing screening programs).
•
If infants need intervention (e.g., amplification,
cochlear implant, auditory verbal therapy), it
should be provided as soon as possible.
References
Dallos, P., and Harris, D. 1977. Properties of auditory
nerve responses in the absence of outer haircells.
Journal of Neurophysiology 41: 365–383.
Eggermont, J.J. 1985. Physiology of the developing audi-
tory system. In S.E. Trehub and B. Schneider (eds.),
Auditory development in infancy (pp. 21–45). New
York: Plenum Press.
Eggermont, J.J. 1988. On the rate of maturation of
sensory evoked potentials. Electroencephalography
and Clinical Neurophysiology 70: 293–305.
Evans, E.F., and Harrison, R.V. 1976. Correlation between
outer hair cell damage and deterioration of cochlear
nerve tuning properties in the guinea pig. Journal of
Physiology 256:43–44.
22 a A Sound Foundation Through Early Amplification

Harel, N., Mori, N., Sawada, S., Mount, R. J., and
Harrison, R.V. 2000. Three distinct auditory areas
of cortex (AI, AII, AAF) defined by optical imaging of
intrinsic signals. NeuroImage 11: 302–312.
Harrison, R.V. 1998. An animal model of auditory neur-
opathy. Ear and Hearing 19: 355–361.
Harrison, R.V. 2001. Age-related tonotopic map plasticity
in the central auditory pathways. Scandinavian
Audiology 30: 8–14.
Harrison, R.V., and Evans, E.F. 1977. The effects of hair
cell loss (restricted to outer hair cells) on the thresh-
old and tuning properties of cochlear fibres in the
guinea pig. In M. Portmann and J.M. Aran (eds.),
Inner ear biology, Colloquium INSERM 68 (pp. 105–
124). Paris: INSERM.
Harrison, R.V., and Evans, E.F. 1979. Cochlear fiber
responses in guinea pigs with well defined cochlear
lesions. Scandinavian Audiology Suppl. 9: 83–92.
Harrison, R.V., Ibrahim, D., and Mount, R.J. 1998.
Reorganization of frequency representation in the
inferior colliculus after neonatal cochlear lesions in
the chinchilla. Experimental Brain Research 123:
449–460.
Harrison, R.V., Ibrahim, D., Stanton, S.G., and Mount,
R.J. 1996a. Reorganization of frequency maps in
chinchilla: Auditory midbrain after long-term basal
cochlear lesions induced at birth. In R.J. Salvi, D.
Henderson, F. Fiorino, and V. Colletti (eds.), Auditory
system plasticity and regeneration (pp. 238–255). New
York: George Thieme.
Harrison, R.V., Kakigi, A., Hirakawa, H., Harel, N., and
Mount, R.J. 1996b. Tonotopic mapping in auditory
cortex of the chinchilla. Hearing Research 100: 157–
163.
Harrison, R.V., Nagasawa, A., Smith, D.W., Stanton,
S.G., and Mount, R.J. 1991. Reorganization of audi-
tory cortex after neonatal high frequency cochlear
hearing loss. Hearing Research 54: 11–19.
Harrison, R.V., Smith, D.W., Nagasawa, A., Stanton, S.G.,
and Mount, R.J. 1992. Developmental plasticity of
auditory cortex in cochlear hearing loss: Physiological
and psychophysical findings. Advances in the Bio-
sciences 83: 625–633.
Hebb, D.H. 1949 The organization of behavior. New York:
John Wiley.
Kakigi, A., Hirakawa, H., Harel, N., Mount, R.J., and
Harrison, R.V. 2000. Tonotopic mapping in auditory
cortex of the adult chinchilla with amikacin-induced
cochlear lesions. Audiology 39:153–160.
Kaltenbach, J.A., Meleca, R.J., and Falzarano, P.R. 1996.
Alterations in the tonotopic map of the cochlear
nucleus following cochlear damage. In R.J. Salvi, D.
Henderson, F. Fiorino, and V. Colletti (eds.), Auditory
system plasticity and regeneration (pp. 317–332). New
York: George Thieme.
Kiang, N.Y., Moxon, E.C., and Levine, R.A. 1970. Auditory
nerve activity in cats with normal and abnormal
cochleas. In G.E. Wolstenholme, and K. Knight
(eds.), Sensorineural hearing loss (pp. 241–268).
CIBA Foundation Symposium. London: Churchill
Press.
Levi-Montalcini, R. 1949. Development of the acoustico-
vestibular centers in the chick embryo in the absence
of the afferent root fibers and of descending fiber
tracts. Journal of Comparative Neurology 91: 209–
242.
Parks, T.N. 1979. Afferent influences on the development
of the brainstem auditory nuclei of the chicken:
Otocyst ablation. Journal of Comparative Neurology
183: 665–678.
Penfield, W. 1958. The excitable cortex in conscious man.
Liverpool UK: Liverpool University Press.
Rajan, R., Irvine, D.R.F., Wise, L.Z., and Heil, P. 1993.
Effect of unilateral partial cochlear lesions in adult
cats on the representation of lesioned and unlesioned
cochleas in primary auditory cortex. Journal of
Comparative Neurology 338: 17–49.
Rajan, R., and Irvine, D.R.F. 1996. Features of, and
boundary conditions for, lesion induced reorganiza-
tion in adult cortical maps. In R.J. Salvi, D. Hender-
son, F. Fiorino, and V. Colletti (eds.), Auditory system
plasticity and regeneration (pp. 224–237). New York:
George Thieme.
Rajan, R., and Irvine, D.R.F. 1998a. Absence of plasticity
of the frequency map in dorsal cochlear lesions. Jour-
nal of Comparative Neurology 339: 35–46.
Rajan, R., and Irvine, D.R.F. 1998b. Neuronal responses
across cortical field AI in plasticity induced by per-
ipheral auditory organ damage. Audiology Neuro-
otology 3: 123–144.
Rubel, E.W. 1985. Strategies and problems for future
studies of auditory development. Acta Otolaryn-
gologica (Stockholm) Suppl. 421: 114–128.
Robertson, D., and Irvine, D.R.F. 1989. Plasticity of fre-
quency organization in auditory cortex of guinea pigs
with partial unilateral deafness. Journal of Com-
parative Neurology 282: 456–471.
Stanton, S.G., and Harrison, R.V. 1996. Neonatal audi-
tory augmentation modifies cochleotopic mapping in
primary auditory cortex of the cat. Auditory Neuro-
science 2: 97–107.
Stanton, S.G., and Harrison, R.V. 2000. Projections from
Representing the Acoustic World within the Brain a 23

the medial geniculate body to primary auditory
cortex in neonatally deafened cats. Journal of
Comparative Neurology 426: 117–129.
Webster, D.B., and Webster, M. 1979. Effects of neonatal
conductive hearing loss on brainstem auditory nuclei.
Annals of Otology Rhinology Laryngology 88: 684–
688.
Willott, J.F. 1984. Changes in frequency representation in
the auditory system of mice with age-related hearing
impairment. Brain Research 309: 159–162.
Woolsey, C.N, and Walzl, E.M. 1942. Topical projection of
nerve fibers from local regions of the cochlea to the
cerebral cortex. Bulletin of Johns Hopkins Hospital
71: 315–344.
24 a A Sound Foundation Through Early Amplification