Harrison RV, Stanton SG, Ibrahim D, Nagasawa A, Mount RJ. Neonatal cochlear hearing loss results in developmental abnormalities of the central auditory ...
Acta Otolaryngol (Stockh) 1993; 113: 296-302
Neonatal Cochlear Hearing Loss Results in Developmental Abnormalities of the Central Auditory Pathways R. V. HARRISON,'*2.3S. G. STANTON,* D. IBRAHIM,2 A. NAGASAWA' and R. J. MOUNT' From the 'Auditory Science Laboratory, Department of Otolaryngology, Hospital for Sick Children, Toronto and 'Department of Physiology and Institute of Biomedical Engineering University of Toronto, Canada
Harrison RV, Stanton SG, Ibrahim D, Nagasawa A, Mount RJ. Neonatal cochlear hearing loss results in developmental abnormalities of the central auditory pathways. Acta Otolaryngol (Stockh) 1993; 113: 296-302. We have used animal models of long term neonatal cochlear hearing loss to study developmental plasticity of the central auditory pathways. Newborn chinchilla pups and feline kittens were treated with the ototoxic drug amikacin, so as to induce basal lesions in the cochlea. At maturity these animals were used in single unit electrophysiological mapping studies, in which the cochleotopic organization of primary auditory cortex (of the cat) and the inferior colliculus of the midbrain (in the chinchilla) were mapped. We have observed, both in the midbrain and auditory cortex, massive reorganization of frequency representation. Most striking were the presence of large monotonic regions (i.e. large areas in which all neurons have similar tuning properties). Cochlear lesions which involve inner haircells clearly modify the normal development of cochleotopic representation in the midbrain and cortical regions. We suggest that similar abnormal patterns of frequency representation will exist in human subjects with long term neonatal hearing loss. Key words: Cochlear hearing loss, cochleotopic organization, developmental plasticity, inferior colliculus, primary auditory cortex, electrophysiological studies.
INTRODUCTION Damage to the cochlear sensory epithelium, as occurs in sensorineural hearing loss, can affect the central auditory system in a number of ways. First, haircell damage will change the threshold sensitivity and other coding properties of the cochlea such as frequency selectivity and intensity coding (e.g. 1, 2, 3) and thus interfere with higher auditory function. We also know that inner haircell degeneration will cause retrocochlear degeneration of spiral ganglion cells (4), and ultimately degenerative changes in the auditory brainstem and possibly midbrain (5, 6). The present studies focus on a third influence that the cochlear damage has on the central auditory pathways. During development the functional integrity of the central aud,itory system appears to be critically dependent on normal excitation patterns from the periphery. There have been a number of studies in which the effects of total cochlear ablation on the development of central pathways have been explored (7- 10). Needless to say central auditory pathway development is very abnormal in such cases. Recently, Robertson & Irvine (1 I ) reported changes to frequency mapping in auditory cortex after longterm cochlear lesions. Our own studies (12, 13, 14) have also demonstrated that the cochleotopic or tonotopic organization of the auditory CNS appears to be dependent on normal levels of excitation of the cochlear nerve fibre array. In the present paper we report on experiments in which cochlear lesions were induced in experimental animals (the cat and chinchilla) just after birth, and in which the cochleotopic organization of primary
auditory cortex (cat) and the central nucleus of the inferior colliculus (chinchilla) were mapped in detail using single unit microelectrode recording techniques. We have induced cochlear lesions using the ototoxic drug amikacin. This causes inner and outer haircell lesions mainly in the high frequency, basal regions of the cochlea. We have observed massively abnormal cochleotopic mapping in both the primary auditory cortex and in the auditory midbrain. The most striking change is the presence of large monotonic (iso-frequency) regions in which all neurons have similar frequency tuning properties. We conclude that the normal cochleotopic arrangement of pathways within the central auditory system is not a "hard-wired'' feature of the auditory system, but rather, develops according to the pattern of neural excitation across the cochlear nerve fibre array during early developmental periods. Cochlear lesions which involve inner haircells and therefore cause degeneration of primary cochlear neurons interfere with the normal development of central cochleotopic representation. This will also be the case in human subjects after neonatal or long-term cochlear hearing loss. MATERIALS AND METHODS Briefly, our experimental protocol was as follows. Neonatal kittens and chinchilla pups were treated with the ototoxic drug amikacin (400 mg/kg/day, S.C. 4 days) to produce bilateral lesions to the cochleas. The functional deficits were confirmed using auditory brainstem evoked responses to tone-pip stimuli (ABR audiograms).
Acta Otolaryngol (Stockh) I13
At maturity the animals were used in single unit electrophysiological recording experiments in which the primary auditory cortex (in the cat), or the central nucleus of the inferior colliculus (in the chinchilla) were systematically mapped. Multiple microelectrode penetrations were made in these areas and the response properties of individual neurons were recorded. In particular we were interested in the best or characteristic frequency (CF) of response of neurons. We define C F as the low threshold tip of the most sharply tuned region of the response area. Acoustic stimuli (70 ms; 10 ms rise/fall) were presented in a pseudo-random fashion to the ear contralateral to the recording site. Our recording experiments were fully automated, with computer control over stimulus parameters and data collection and analysis. (See: (12, 14) for details of techniques.) The present study was based on recordings from 8 treated kittens and 6 treated chinchilla pups. A similar number of untreated animals of each species were used as controls. All procedures involving animals were carefully carried out according to guidelines provided by the Canadian Council on Animal Care. After the electrophysiological recording experiment the cochleas and cochleas of the experimental animals and their brains were fixed for morphological assessment. We used scanning electron microscopy to define the amikacin induced lesions. Detailed results of the morphological studies have been published
Central auditory system; developmental plasticity
297
elsewhere (14) and are therefore not included in this paper. RESULTS Primary auditory cortex. The cochleotopic organization of normal A1 auditory cortex in the cat is illustrated in Fig. 1. Each data point represents the position of neurons (recorded with microelectrodes) of which the C F or best frequency of tuning is indicated. The posterior cortical area contains. neurons tuned to low frequencies. There is a systematic increase in neuron C F in more anterior regions. In Fig. 1, iso-frequency contours have been superimposed at octave intervals; note the relatively regular spacing of these contours in the normal animal. In animals with cochlear basal lesions from birth we see dramatic departures from the normal pattern. Fig. 2A and B show examples of abnormal cochleotopic mapping. The audiograms of these treated animals, as derived using ABR techniques, are shown in the upper panels. The animal of Fig. 2A has a high frequency (neonatal hearing loss with a cut-off between 4-8 kHz. In this subject, note that in the posterior region low frequencies are normally represented, but that a large region of the anterior cortex contains neurons tuned to a limited frequency range between 6.3-8 kHz (cross-hatched area). We term this region an iso-frequency expansion or monotonic
Fig. 1. Cochleotopic organization of the primary auditory cortex in the normal cat. Positions of microelectrode penetrations are indicated together with the CFs of neurons encountered. Iso-frequency contours at octave intervals have been superimposed. The dashed line anteriorly separates primary auditory cortex from the anterior field. (sf: sylvian fissure; pef: posterior ectosylvian fissure; aef: anterior ectosylvian fissure.)
298
R.V. Harrison et al.
A
Acta Otolarvnaol (Stockh) 113
B
ABR audlogram
-
s 'ii -P0" ea
g
-
-
0
0
m
U
20
20
m m
-0
40
40
P
2a
60
;
80 100
ABR audiogram
60
80
100
r '
'
.5
1
'
I
'
2 4 8 16 32 frequency (kHz1
.5
1
2 4 8 16 3 2 frequency (kHz1
....
.
............. ............................. ............................. ..................................................... ............................
JL30CT06
Fig. 2A and B. Tonotopic organization in primary auditory cortex of two cats after long-term neonatal cochlear hearing loss. The upper panels indicate ABR audiograms for each animal. The tonotopic maps of primary auditory cortex are shown in the lower panels. Where possible, iso-frequency contours at octave intervals are shown. The major monotonic areas indicated by the cross-hatching. SF: sylvian fissure; PEF: posterior ectosylvian fissure; AEF: anterior ectosylvian fissure.
area. The frequency of this area appears to coincide with the cut-dff frequency of the audiogram. In Fig. 2B we show the developmental effects of more severe neonatal cochlear damage. The ABR audiogram indicates hearing loss with a high frequency cut-off between 1-2 kHz; thus there are significant threshold elevations for low frequencies as well as the severe high frequency deficits. In this case auditory cortex appears to be dominated by a large iso-frequency region at 6.6 kHz (cross-hatched area), as well as an over-representation of 0.7 kHz posteriorly. In this animal, there appears to be little evidence of a systematic tonotopic organization even for low frequencies. Inferior colliculus of the midbrain. Our findings at the level of auditory cortex lead us to suppose (see discussion) that similar reorganizations may occur at lower levels in the auditory pathways. We therefore chose to carry out an analogous experiment to explore the auditory midbrain (in this case using the chinchilla as our animal model). The normal cochleotopic representation within the inferior colliculus is
shown in Fig. 3. The upper panel shows the electrode tracts through the central nucleus of inferior colliculus. Note that there is a systematic representation of frequency, with neurons tuned to low frequency located dorsal in the nucleus, and neurons tuned to high frequency represented ventrally. In Fig. 4A and B, the frequency representation in two neonatally lesioned animals are depicted. The ABR audiograms in the upper panels show the degree of neonatally induced hearing loss. In Fig. 4A, the high frequency cut-off slope of the ABR audiogram is around 10 kHz. Frequency representation with the inferior colliculus of this animal is shown to the right; electrode tracks through the central nucleus are indicated together with the general frequency organization of the central nucleus. Throughout most of IC there is a relatively normal cochleotopic arrangement. However, in the ventral, high frequency area there is a significant region in which all neurons have similar response characteristics, with a C F at about 10 kHz. This arrangement is also reflected in the lower plot of electrode excursion
Central auditory system; developmental plasticitv
Acta Otolaryngol (Stockh) 113
299
Electrode Tracks
IORSAL
-
iB ANTERIOR
I C *4
ENTRAL
I
I
I
POSTERIOR
I
I
I
I
I
I
I
I
129 NORMAL CHINCHILLA 0 TRACK 1
00, Y
A
TRACK 2
0 TRACK 3
ITRACK 4
I
0.25
I 0.5
I
I
I
I
I
1
2
4
8
16
I 32
Fig. 3. Cochleotopic organization of the inferior colliculus in the normal chinchilla. In the upper panel, the cross sectional area of the central region of inferior colliculus is indicated together with positions of three electrode penetrations from dorsal to ventral. A sample of the recording sites are indicated together with iso-frequency contours at octave intervals. In the lower plot, neuron CF is plotted as a function of dorsoventral excursion of the microelectrode tracks.
CHARACTERISTIC FREQUENCY (kHz) versus neuron CF; pooled data from three electrode tracks is shown in this graph. The chinchilla of Fig. 4B had a more extensive basal cochlear lesion induced neonatally, as reflected in the 5 kHz cut-off slope of the ABR audiogram. In dorsal regions of IC there is a relatively normal tonotopic progression up to 5 kHz. The ventral area, however, consists of a large “expanded” region in which all neurons have similar response areas and with CFs of approximately 5 kHz. DISCUSSION In the main sensory modalities, i.e. visual, somatosensory, auditory systems, there is a systematic representation of the sensory epithelium at all levels in the
central sensory pathways up to cortical regions (e.g. 15- 18). This topographical mapping of the sensory surface had long been regarded as a stable, “mainline” organizational feature of sensory systems. However, a number of studies have now shown that the cortical areas of sensory systems can be radically reorganized, particularly if there is a restricted or otherwise abnormal sensory input present during early stages of development. Thus for example, Hubel & Wiesel showed that the visual cortex was considerably reorganized when kittens were deprived of sensory input (e.g. monocular deprivation) from an early age (e.g. 19). In the somatosensory system the representation of the body surface in somatosensory cortex can be massively distorted by neonatal lesions to the sensory epithelium or partial deafferentation (20, 21, 22).
300
R.V. Harrison et al.
mR
Acta Otolaryngol (Stockh) 113
*BR WDloQUu
*u)IoQwI
i!
B
1
5
60 v)
l &
F
F
100
Mrmw
0.5 I 2 4 6 8 16 32 FREOUENCY (kHz)
I
1
1
I
1
ot-
P
3
V
1
0.5 I 2 4 8 16 32 FREQUENCY (kHz) 1
1
1
1
CHINCHILLA /39
E
;i
E
I 1 0-
I
I
I
1
I
I
I
-
z I-
% P
3
V
x 2-
W
w
-
E
3-
0 0 0 TRACK 2
0 TRACK I
I
0 TRACK 2
W - I W
0 T W K 3
0 TRACK 3 TRACK 4
4c
A
1
CHINCHILLA 137
Y
x 2
w
80 100
2 0*25C$ACTERISTIC
4 8 16 32 FREQUENCY (kHz)
-I
I 2 4 8 1 6 3 2 0‘25CEACTERISTIC FREQUENCY (kHz1
Figs. 44 and B. Frequency representation in inferior colliculi of two chinchillas after long-term neontal cochlear lesions. The ABR audiograms (upper left hand panels) indicate the functional deficits for each animal. Indicated in the upper right hand panels are the electrode penetrations through the central area of inferior colliculus, indicating large monotonic regions in which all neurons appear to have similar frequency response characteristics. The lower plots show the relationship between neuron CF and the mm excursion of the electrode from the dorsal to ventral areas.
In the auditory system, Robertson & Irvine (11) have also reported the development of abnormal iso-frequency regions in primary auditory cortex after partial cochlear lesions in the guinea pig. Our own studies presented here from auditory cortex have extended these findings. In our study we have specifically modelled human sensorineural hearing loss from birth by inducing bilateral basal cochlear lesions using the ototoxic drug amikacin. We believe that human subjects with long term hearing loss from an early age will have a correspondingly abnormal frequency representation at the level of auditory cortex. During our experimental studies at the cortical level we found circumstantial evidence to suppose that the reorganizations seen were occurring at a sub-cortical, more peripheral stage of the auditory pathway. This evidence came, in part, from analysis of abnormal ABR and myogenic potentials evoked from the midbrain region (23) as well as neural tracer mapping of ascending input to auditory cortex. We therefore carried out an analogous experiment to explore the midbrain. We used the chinchilla, again damaging the cochleas bilaterally using the amikacin.
If we compare tonotopic maps from the cortex (Fig. 2) with those from the midbrain (Fig. 4), there are striking similarities. In particular, both exhibit “expanded” areas in which all the neurons share similar tuning properties. These iso-frequency regions correspond to the boundary of the cochlear lesion, and thus to the high frequency cut-off slope of the audiogram. It should also be noted that the responses of neurons in these areas are of elevated threshold and have abnormal tuning characteristics, as is typical of neural responses originating from cochleas with haircell lesions (1, 2, 3). It is evident that the development of frequency representation in central areas of the auditory system including the midbrain is very much dependent upon the pattern of excitation from the periphery during development. When there is loss of inner haircells in sensorineural hearing loss, we know that there is a loss of the primary cochlear afferents corresponding to the area of haircell damage (4). This partial deafferentation from an early age is responsible for the reorganizations of frequency representation seen in this
Acta Otolaryngol (Stockh) I13
study both in the midbrain and at primary auditory cortex. When the deafferentation is restricted to a small (high) frequency region the areas of cortex which would normally be tuned to that frequency region appear to be “utilized” by the frequency area immediately next to it. Thus, for example, in Fig. 2A the large shaded area which would normally contain neurons tuned t o high frequencies now becomes innervated from the lower (6-8 kHz) frequency region. We might hypothesize that the utilization of neural space during development is made on a competitive basis. Thus ascending neurons will make connections with higher order neurons (e.g. inferior colicullus; auditory cortex) in competition with other active ascending neurons according to their levels of activity. When activity of the whole neural array is equally weighted there will be a tendency for parallel innervation patterns throughout the auditory pathways. If, on the other hand, one frequency region has reduced activity then adjacent neurons may “win out” in competing for target cells. This simple mechanism can result in our observed high frequency monotonic regions as seen in both Figs. 2 and 4. These experimental results raise a number of applied questions, perhaps the most important being whether such reorganization is reversible. Thus for example, if after many developing years with an unaided hearing loss, a hearing prosthesis is provided to an individual such that more areas of the cochlea are activated, can the frequency representation in the midbrain and cortex change as a result of the partially restored input? It may well be that there is a critical period during which reversal is possible, but after which the plasticity of the system is lost. Another unanswered clinical issue concerns the developmental effects of chronic or reccurring conductive hearing loss in neonates and infants. In our studies we induced cochlear damage in which inner haircells and therefore primary cochlear afferent neurons are inactivated. The question arises: How does a (mild, moderate or severe) conductive hearing loss affect the development of central auditory areas? Whilst it is a common belief that the majority of neurons in the cochlear nerve are spontaneously active and therefore not influenced by a conductive loss, we really have no experimental evidence that spontaneous activity alone can provide normal development of central pathways. We will be addressing these and other applied questions in our future studies.
ACKNOWLEDGMENTS This research was supported by the Medical Research Council of Canada and the Masonic Foundation of On-
Central auditory system; developmental plasticity
301
tario. Ms. S. Stanton was supported by an Ontario Ministry of Health Scholarship. The authors thank Ms. Carol Morgan for her assistance in preparing this manuscript. REFERENCES 1 . Kiang N Y S , Moxon EC, Levine RA. Auditory nerve activity in cats with normal and abnormal cochleas. In
Wolstenholme GEW, Knight K eds. Sensorineural hearing loss. CIBA Found Symp., London: Churchill, 1970: 241-68. 2. Evans EF. The sharpening of cochlear frequency selectivity in the normal and abnormal cochlea. Audiology 1975; 14: 419-42. 3. Harrison RV. Rate versus intensity functions and related AP responses in normal and pathological guinea pig and human cochleas. J Acoust SOCAm 1981; 70: 1036-44. 4. Spoendlin H. Retrograde degeneration of the cochlear nerve. Acta Otolaryngol (Stockh) 1975; 79: 266-75. 5. Jean-Baptiste J, Morest DK. Transneuronal changes of synaptic endings and nuclear chromatin in the trapezoid body following cochlear ablation in cats. J Comp Neurol 1975; 162: 111-33. 6. Killackey RP, Ryugo KD. Effects of peripheral auditory system damage on the structure of the inferior colliculus of the rat. Anat Rec 1977; 187: 624-5. 7. Jackson JH, Rubel EW. Rapid transneuronal degeneration following cochlea removal in chickens. Anat Rec 1976; 84: 434-5. 8. Parks TN. Afferent influences on the development of the brainstem auditory nuclei of the chicken: otocyst ablation. J Comp Neurol 1979; 183: 665-78. 9. Nordeen KW, Killackey HP, Kitzes LM. Ascending projections to the inferior colliculus following unilateral cochlear ablation in the neonatal gerbil meriones unguiculatus. J Comp Neurol 1983; 214: 144-53. 10. Kitzes LM. Some physiological consequences of neonatal cochlear destruction in the inferior colliculus of the gerbil, meriones unguiculatus. Brain Res 1984; 306: 171-8. 1 1 . Robertson D, Irvine DRF. Plasticity of frequency organization in auditory cortex of guinea pigs with partial unilateral deafness. J Comp Neurol 1989; 282: 456-71. 12. Harrison RV, Nagasawa A, Stanton S, Smith DW, Mount RJ. Extensive reorganization of cat auditory cortex after high frequency cochlear hearing loss. Hear Res 1991; 54: 11-9. 13. Harrison RV, Smith DW, Nagasawa A, Stanton SG, Mount RJ. Developmental plasticity of auditory cortex; psychophysical and physiological correlates. In: Advances in the biosciences, Vol 83. Auditory physiology and perception. London: Pergamon Press, 1992: 625- 33. 14. Mount RJ, Harrison RV, Stanton SG, Nagasawa A. Correlation of cochlear pathology with auditory brainstem and cortical responses in cats with long term high frequency hearing loss. Scanning Microsc 1991; 5: 1105-13. 15 Penfield W. The excitable cortex in conscious man. Liverpool: Liverpool University Press, 1958. 16. Merzenich MM, Knight PL, Roth GL. Representation of cochlea within primary auditory cortex in the cat. J Neurophysiol 1975; 38: 231-49.
302
R.V. Harrison et al.
17. Hubel DH, Wiesel TN. Functional architecture of macaque visual cortex. Proc R SOCLond (Biol) 1977; 198: 1-59. 18. Reale R, Imig T. Tonotopic organization in auditory cortex of the cat. J Comp Neurol 1980; 192: 265-91. 19. Wiesel TN, Hubel DH. Single cell responses in striate cortex of kittens deprived of vision in one eye. J Neurophysiol 1963; 26: 1003-17. 20. Waite PMS, Taylor PK. Removal of whiskers in young rats causes functional changes in cerebral cortex. Nature 1978; 274: 600-2. 21. Merzenich MM, Kaas JH. Re-organization of mammalian somatosensory cortex following peripheral nerve injury. Trends Neurosci 1982; 65: 434-6. 22. Kaas JH, Merzenich MM, Killackey HP. The re-
Acta Otolarvneol (Stockh) 113
