I. INTRODUCTION

Top Previous Next References

The role of the auditory system in mammals is to inform the subject about the acoustical environment and about the presence of species-specific acoustical communication signals. In humans, the latter function is dominant since one of the specific features of humans is the ability to learn and use language. It is the loss of auditory receptors in childhood, either inherited or acquired, that precludes the development of normal speech in a child, leaving the child deaf and dumb. Yet contemporary rehabilitation techniques, based on electronic devices such as hearing aids with robust amplification or, particularly, cochlear implants, are able to substitute for the missing receptors and, when initiated in early childhood, to guarantee quasi-natural language development. The rehabilitation of hearing function with cochlear implants is also possible in adults, especially when the period between the loss of hearing and the introduction of the cochlear implant is not too long. A limited ability to differentiate sounds with the aid of cochlear implants is also acquired in adults who were born deaf. These examples show that not only during childhood, but also in adulthood, the auditory system is plastic and may recover some of its basic functions.

Many plastic changes in the brain and in the receptor organs occur only in developing organisms as a continuation of embryological development, i.e., after birth (285). In humans, the activation of the auditory system, as indicated by the occurrence of acoustic brain stem reflexes, starts in the 26-28th gestational week, and the development of hearing function continues for several years. Similar precocial development of hearing function occurs in experimental animal models used in auditory research such as guinea pigs or chinchillas, whereas altricial development of hearing, i.e., with the start of hearing after birth, is known to be present in the mouse, rat, ferret, gerbil, and cat. In mammals, the loss of hearing receptors at an early postnatal age results in plastic changes in the central auditory system similarly like in the visual and somatosensory systems (118, 129, 273). In mammals, in contrast to other vertebrates, the hair cells in the cochlea do not regenerate and the plastic processes in the central auditory system may be different from other vertebrates; therefore, this review is limited to the plasticity occurring in the auditory system of mammals and does not cover the problems of plasticity in birds and other submammalian orders.

The physiology of hearing has been positively influenced by a related scientific discipline, psychoacoustics. Psychoacoustics has been defined as the study of the relationship between the physical properties of sound stimuli and the behavior this stimulation evokes (401). Detection, discrimination, identification, and scaling have been the primary measures of behavior studied by psychoacousticians. However, psychoacoustics by definition presupposes that the behavior is more or less steady and therefore does not expect that essential changes occur in the auditory system during the investigation. Yet several recent studies have demonstrated that if the stimulation has some informational meaning for the animal, the properties of receptive fields and maps in the central auditory system change, mainly at the cortical and thalamic level (366). Cortical and thalamic neurons may change their characteristic features in the process of learning and memory. Such plastic changes, together with plastic changes after partial and total hearing loss, serve as evidence for the dynamism and plasticity of the auditory system and against the traditional view of the auditory system as a hard-wired system where change is impossible.

The loss of hearing function may not necessarily be complete, but even a partial decline in the number of inner ear receptors results in a limitation of hearing function in humans, especially in the intelligibility of speech. The deterioration of hearing function, particularly at high frequencies, is a process that inevitably accompanies aging in humans as well as in other mammals (379). The dominant component of presbycusis is the gradual degeneration of hair cells mainly in the basal part of the cochlea; however, its central component is usually based on sclerotic processes in the brain. It is the aim of contemporary research to recognize the compensatory processes that exist in the adult auditory system with the aim of utilizing them for the rehabilitation of hearing function.

The auditory system, unlike other sensory systems, is in modern humans exposed to ever-increasing overstimulation. Noise exposure induces degenerative changes of the sensory receptors in the inner ear, with an obvious destruction of the outer hair cells at first. With continuing noise stimulation or with higher intensity noise, the loss of outer hair cells is followed by the loss of the inner hair cells. Ultimately, the postexposure changes in the inner ear are similar to those following mechanical lesions of the receptor organ, lesions induced by ototoxic drugs, such as aminoglycoside antibiotics, or degenerative changes occurring during aging (328). The changes in the central auditory system after any of these pathological states are to some extent similar; however, in the case of noise exposure, the plastic changes resulting from the loss of function of peripheral receptors are preceded by specific changes in the function of auditory neurons, elicited by overstimulation. Understanding the mechanisms of the changes in the auditory system occurring during and after noise exposure may be helpful in the prevention of this pathology and in its rational cure.

Many aspects of plasticity in the auditory system, both in the developmental period and during adulthood, are similar to the plasticity in other sensory systems and share similar mechanisms. It was the intention of the reviewer not to extend the scope of the review in this direction, since many excellent reviews exist covering the problems of plasticity in other sensory systems and the problem of plasticity in sensory systems in general, independent of the individual systems (75, 129, 273). For more detailed information about individual aspects of plastic changes in the auditory system, other recent reviews can be consulted (11, 113, 214, 240, 261, 274, 292, 328, 329, 379).

	II. PLASTIC CHANGES IN THE AUDITORY SYSTEM AFTER TOTAL OR PARTIAL LESIONS OF THE INNER EAR

Top Previous Next References

A.  Effects of Complete Removal of the Inner Ear

1.  Neonatal ablations

A common feature of sensory systems is a topographical projection from the sensory surface to various levels within the central nervous system. Thus retinotopic and somatotopic maps are internal representations of the retina and the body surface, respectively. In the auditory system, the sensory epithelium is rerepresented in all successive nuclei within the auditory pathway up to and including the auditory cortex. Because sound frequency is space mapped in the cochlea, the cochleotopic representation is usually named tonotopic representation. Numerous studies in recent years have shown that the central sensory maps are subject to reorganization as a result of manipulation of the peripheral sensory input particularly during early development. One of the first studies exploring the phenomenon of plasticity in the sensory systems was the pioneering studies by Hubel and Wiesel (108, 375, 376), who demonstrated changes to ocular dominant columns of striate cortex in monocularly deprived kittens. In the somatosensory system, a restricted peripheral deafferentation in developing animals results in the reorganization of subcortical and cortical somatotopic maps such that areas that would normally receive input from the denervated regions become used for the representation of the adjacent skin surface (130). In the auditory system, many of the studies of developmental neuroplasticity have focused on the effect of deafening on the immature auditory system. This work has shown that surgical removal of one cochlea in neonatal mammals results in a loss and shrinkage of neurons in the cochlear nucleus (92, 212, 346) and superior olivary complex (213, 247), in the formation of new connections between the cochlear nucleus on the intact side and various target structures in the brain stem (144, 286) and midbrain (234, 218, 215), and an increase in the responsiveness of inferior colliculus (142, 145, 217) and primary auditory cortex (276) neurons to acoustic stimulation of the ipsilateral, intact ear. In all of these experiments, altricial animal species were used, in which the development of hearing as well as other senses continues for some time after birth (cat, ferret, and gerbil).

In initial experiments with adult cats (276), it was shown that in cats reared with a neonatal cochlear ablation, the primary auditory cortex (AI) contralateral to the operated ear showed a normal tonotopic map derived from single neurons and neuronal clusters. The thresholds at many recording sites were as low as those observed in the AI ipsilateral to the operated ear. This was in marked contrast to the results of recordings in normal adult cats, in which only ~65% of AI neurons were excited by sound delivered to the ipsilateral ear, and thresholds to ipsilateral ear stimulation were significantly higher than thresholds to contralateral ear stimulation. In support of this finding, a significant increase in dendritic length and a change in their orientation in the auditory cortex contralateral to the ablated ear was observed in rabbits with neonatal unilateral ablation (192). Earlier studies in the cat of the anatomic consequences of cochlear ablation were restricted to adult animals. In these studies, lesions resulted in the shrinkage of cells in the deafferented cochlear nuclei (259) and contralateral medial nucleus of the trapezoid body (121). Neonatal cochlear ablation in gerbil (145, 234), however, has been reported to produce far greater anatomic changes in the central auditory pathways. The cochlear nuclei on the nonoperated side contained a greater number of cells projecting to the ipsilateral inferior colliculus and a reduced number of cells projecting to the contralateral inferior colliculus compared with controls (218, 234).

The increased input to the colliculus ipsilateral to the intact ear in gerbils after unilateral cochlear ablation in the early postnatal period was accompanied by signs of increased excitatory influence from the ipsilateral ear. Multiple unit activity (234) as well as single unit activity (145) dramatically increased in adult age in these animals, with more vigorous responses, broader dynamic ranges, and lower thresholds than in adult intact animals. The responses of inferior colliculus (IC) neurons to ipsilateral stimulation closely resembled those found in response to contralateral stimulation in the binaurally intact animal. Interestingly, the functional reorganization in pathways projecting to the IC was not accompanied by changes in synaptic density within the IC (83), since adult cats deafened neonatally in one ear did not exhibit changes in synaptic density with respect to normal hearing controls. The synaptic density in the IC was, however, significantly lower in cats deafened bilaterally (83). In later studies by Kitzes et al. (144), detailed structural changes were reported in the superior olivary complex of gerbils after unilateral neonatal ablation of the cochlea. Projections from the ventral cochlear nucleus to higher brainstem nuclei, which normally were directed to the contralateral side, instead were directed to individual nuclei of the superior olivary complex (medial and lateral superior olive, medial nucleus of the trapezoid body) on both sides after cochlea ablation. Such induced projections first appeared within 24 h of cochlear ablation and continued to develop over at least 11 subsequent days. Thus, before the day when the cochlea in the gerbil becomes functional (postnatal day 12 or 13), the ventral cochlear nucleus (VCN) has established specific ectopic projections to loci normally innervated by the VCN on the ablated side. All induced ectopic projections observed in neonatal animals were also present in neonatally ablated adult animals. The fact that elements of the induced projections appear within 24 h of the cochlear ablation raises questions about the mechanism of induction. It was suggested (143) that neurons from the medial nucleus of the trapezoid body (MNTB) ipsilateral to the lesioned ear might release a chemotropic molecule after the growth cones of the ipsilateral calyciferous axons have already traversed the nucleus and are beyond the effective range of the molecule. With the loss of the normal source of calyciferous axons from the VCN on the operated side, the chemotropic molecule would then be available to influence proximal axons from the ipsilateral side. Neonatal destruction of auditory receptors is accompanied in experimental animals by a loss of spiral ganglion cells (SGC), by a loss of nerve cells in the cochlear nuclei, and by a reduction in the size of some nerve cells in the cochlear nuclei. Because these changes are partially reversible by electrical stimulation of the inner ear, comprehensive discussion of this topic is presented in section VI.

Although little is known at the present time about the processes subserving the formation of new connections after neonatal cochlear ablation, substantial data about changes in the brain stem auditory nuclei are available with respect to the effects of cochlear ablation in adult animals. Manipulations of the cochlea in adult gerbils (248) have been shown to produce transneuronal changes in the cochlear nucleus. The cross-sectional area of large spherical cells in the VCN was reduced by ~25% within 48 h of removing input to the VCN by ablation of the cochlea or by administration of tetrodotoxin to the cochlea. The authors concluded that the electrical activity of the auditory nerve fibers or sequelae of the electrical activity is a major factor in transneuronal regulation of cell size. Functional deafferentation in another study (27) was performed in adult guinea pigs by intracochlear infusion of the glutamate agonist -amino-3-hydroxy-5- methyl-4-isoxazole-propionic acid (AMPA), which is known to transiently disconnect inner hair cells from primary auditory dendrites (263, 264).

Measurements of immunostaining intensities demonstrated rapid and reversible changes in calcium-binding protein levels. Twenty-four hours after AMPA treatment, calretinin immunostaining was reduced in deafferented neurons of the cochlear nuclei and their axons of the superior olivary complex. In contrast, calbindin D28k immunoreactivity levels by this time were higher in deafferented neurons of the medial nucleus of the trapezoid body and their axons in the lateral superior olive. Parvalbumin immunostaining was also generally increased in deafferented neurons, but the changes were less evident and more complex. Normal levels of all three calcium binding proteins were reached within 5 days after AMPA treatment, when afferent activity had almost completely recovered. Apparently the increased buffering capacity in deafferented neurons, as indicated by rises in calbindin D28k and parvalbumin immunostaining, may be involved in the mechanisms promoting neuronal survival after loss of sensory input. Increases in calbindin D28k expression in brain stem auditory nuclei were also observed in mature rats after permanent unilateral cochlear ablation (67).

Calbindin is characteristically expressed in many auditory brain stem nuclei during ontogeny and disappears from many of these nuclei during maturation. Growth-associated protein (GAP)-43 shows a similar time-limited expression during ontogeny (106). Extensive data link the presence of GAP-43 to axonal elongation and synapse formation during development. While it is expressed at high concentrations in cochlear nuclei and the superior olivary complex on the first postnatal day in the rat, by the 16th postnatal day the immunoreactivity is almost lost, with only discrete levels persisting to adulthood. However, GAP-43 was expressed de novo in fibers and boutons of the ipsilateral cochlear nucleus after unilateral cochlear ablation in adult rats (110). The occurrence of GAP-43 after cochlear ablation in adult animals suggests that sudden loss of spiral ganglion cells leads to a reactive synaptogenesis in several auditory brain stem nuclei. Recent biochemical studies (320, 321) also suggest that after unilateral cochlear ablation in young adult guinea pigs, changes develop in glycinergic and GABAergic inhibitory transmission in the brain stem nuclei. Glycinergic transmission is persistently weakened ipsilaterally in the ventral cochlear nucleus and lateral superior olivary nucleus (LSO) and bilaterally in the dorsal cochlear nucleus (DCN).

Initial experimental results, in which changes in the function of the central auditory system after cochlear ablation in adult animals were studied, did not show any significant effect on the neuronal responses in the IC. Unilateral cochlear ablation in adult gerbils did not affect the proportion of IC neurons excited by the ipsilateral ear (234). Large lesions of the cochlear nucleus in the adult gerbil had no apparent effect on the responsiveness of single neurons in the contralateral IC to stimulation of the intact ear (219). However, experiments in the ferret (220) found that even after unilateral cochlear ablation in adult animal, the proportion of neurons excited from the intact ear in the ipsilateral IC was unusually high. In subsequent experiments (216), performed in the ferret auditory cortex, the authors confirmed that the adult ferret auditory system responded rapidly and dramatically to cochlear ablation. The topographic representation of the ipsilateral ear in the primary auditory cortex expanded after ablation of the contralateral cochlea. Also, the proportion of units in the AI driven by stimulation of the intact ear substantially increased. While in control animals the ipsilateral ear provided excitatory drive to only ~50% of the recording sites that were excited by contralateral stimulation, after acute cochlear ablation a much more widespread distribution of ipsilateral excitation was found. "Acute" cochlear ablation in the ferret (i.e., within 24 h before recording) produced a quantitative reduction in mean unit thresholds and an increase in the spontaneous discharge rate of cortical neurons on the side of the intact ear relative to normal, binaurally intact animals. Although these experiments have shown that changes in the excitability of the central auditory system after cochlear ablation occur very early, it has been demonstrated (216) that the number of loci in the ferret IC excited from the ipsilateral ear increased with the time of survival after cochlear ablation. Similarly, the thresholds of neuronal responses to ipsilateral stimulation decreased with survival time after ablation. These results contradict a simple unmasking hypothesis (35) of the changes in excitability, which would predict that all changes occur suddenly. The term unmasking implies immediate release from suppression after ablation of an inhibitory input. This unmasking may be contrasted with longer term changes that include the modification of existing synapses, synaptogenesis, and the formation of new connections (43, 44).

Results similar to those in the ferret were obtained in experiments in adult guinea pigs with quasi-complete hair cell destruction induced by injection of an ototoxic antibiotic into one ear (254). The thresholds of evoked potentials in the auditory cortex ipsilateral to the nondamaged ear decreased, and the amplitudes increased over the course of several days or weeks after hair cell destruction. These results also do not support the unmasking hypothesis, the slow progress of change during several days or weeks after cochlear lesion speaks in favor of less dramatic and more structural changes.

B.  Effects of Partial Cochlear Lesions

1.  Lesions induced at birth

Initial experiments with partial lesions of the cochlea in newborn animals were performed in cats where changes were found in the primary auditory cortex (87). A neonatal high-frequency loss was induced by injections of the ototoxic antibiotic amikacin in newborn kittens, and tonotopic maps were constructed in adult animals after 12-mo survival. As in experiments by Robertson and Irvine (280) with adult guinea pigs, the AI in the cat was extensively reorganized with regions that in normal nonlesioned controls were excited by high frequencies being taken over by lower frequencies, primarily those associated with the border of the cochlear lesion (see below for details). However, retrograde tracer injections into different regions of the AI produced a normal pattern of labeling in the ventral division of the medial geniculate body (MGB) (318). The deafened cats did not develop more divergent thalamocortical projections compared with normal control animals, indicating that an abnormal spread of thalamocortical afferents across the frequency domain of the AI is not responsible for the altered cochleotopic map in these neonatally deafened animals.

The effects of cochlear lesions produced by injections of amikacin in newborn animals were studied further in the inferior colliculus in chinchillas (86), i.e., in a precocious animal. As a result of the cochlear lesion, the tonotopic maps in the IC of adult chinchillas were substantially reorganized, usually with a large "monotonic" or "isofrequency" region in the ventral part of the central nucleus of the IC, i.e., in the area where normally high frequencies are represented. The tuning curves of neurons in this expanded region were pathological and their tuning, which was almost identical, corresponded with the tuning of the boundary areas of the cochlear lesion.

In contrast to the plasticity exhibited by the developing nervous system, it was for many years generally assumed that representational maps in adult animals were stable and not susceptible to change. Studies over the last 20 years have shown that this is not the case and that sensory projection areas of the adult brain are subject to reorganization after peripheral receptor lesions. In the auditory system, a pioneering study was published in 1989 by Robertson and Irvine (280), who induced restricted mechanical lesions of one cochlea in mature guinea pigs and assessed the extent of the resultant hearing loss by recording the compound action potential audiogram and by histological examination of the cochlea. The lesion produced a relatively sharp notch of decreased peripheral sensitivity over a range from ~10 to 20 kHz, associated with a discrete region of abnormal inner and outer hair cells at the corresponding cochlear location (Fig. 1). The frequency organization of the auditory cortex in the hemisphere contralateral to the lesioned cochlea was investigated with a microelectrode mapping technique 35-81 days after cochlear lesioning. The region of cortex that in normal animal would be occupied by an orderly and continuous representation of frequencies from 10 to 20 kHz was found to be almost entirely occupied by enlarged representations of limited frequency ranges corresponding to the borders of the cochlear lesions (9.9-10.8 kHz and 20.6-22.0 kHz). The responses of neuronal clusters in the reorganized regions of the cortex were strong, and their thresholds at the newly acquired characteristic frequencies were comparable to those observed in normal control animals.

View larger version (34K): [in this window] [in a new window]   Fig. 1. Two examples of "notch"-type lesions in the cochlea with corresponding cortical frequency maps. Top panels: compound action potential (CAP) audiograms. The postrecovery audiogram for each animal (open circles, solid lines) is compared with the mean audiogram for 10 normal animals (solid circles, broken line). Shaded areas around the normal mean audiogram show 1 SD. Bottom panels: frequency organization of the auditory cortex after lesion. Each dot represents the location on the cortical surface of an electrode penetration. The number beside each dot is the characteristic frequency (CF) in kHz of clusters at that point. Points marked X were unresponsive to acoustic stimulation. Points marked A were acoustically responsive, but a clear CF could not be assigned. Lines are approximate isofrequency contours drawn by eye through loci of the same CF as indicated by the numeral at the end of each line. Broken lines delineate areas of expanded frequency representations with sensitive drive. D, dorsal; R, rostral. [From Robertson and Irvine (280). Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.]

The changes in the representation of the lesioned cochlea in the contralateral hemisphere were assessed using acoustical stimulation of the lesioned ear. However, as discussed above, the unilateral ablation of the cochlea produces changes in the representation of the unlesioned ear in the inferior colliculus and auditory cortex. Therefore, Rajan et al. (268) investigated the effects of unilateral restricted cochlear lesions in adult cats on the representation of the lesioned and normal cochleas in the primary auditory cortex (area AI) contralateral to the lesioned ear. High-frequency loss, with edge frequencies in the range of 18-24 kHz, was produced by mechanical lesions of the cochlea. In confirmation of the study in adult guinea pigs (280), in adult cats 2-11 mo after a unilateral cochlear lesion, the map of the lesioned cochlea in the contralateral AI was also altered so that the AI region in which the frequencies with lesion-induced elevations in cochlear neural sensitivity would have been represented was occupied by an enlarged representation of lesion-edge frequencies. There was no topographic order within this enlarged representation. In contrast to the change in the map of the lesioned contralateral cochlea, the map of the unlesioned ipsilateral cochlea did not differ from those in normal animals. This finding is in sharp contrast to the results of experiments with complete unilateral ablation of the cochlea in which the excitatory influence of the unlesioned ear on the central auditory system was strengthened. A comparison of the tuning curves of cortical neurons in the reorganized area with those in normal control animals did not disclose any essential differences. The tuning curves were slightly, but not significantly, less sharply tuned, and in some animals the latencies of responses were significantly shorter than normal. Further evidence bearing on the effects of experimentally produced cochlear lesions in adults on the frequency organization of the auditory cortex was obtained in adult macaques who were treated by kanamycin and furosemide to produce a high-frequency hearing loss (303). When the cortex was mapped 80-90 days after ototoxic treatment, the high-frequency region of the AI was found to be occupied by an enlarged representation of lesion-edge frequencies.

The lack of alignment of ipsilateral and contralateral projections in the AI of the adult cat after unilateral restricted lesions led to the conclusion that the plastic changes in the contralateral AI are a product of changes at the subcortical level. When partial cochlear lesions were performed in the same manner as in previous experiments looking at changes in the auditory cortex (268), no plasticity of the frequency maps was found in the dorsal cochlear nucleus (267) and only patchy signs of plastic changes were observed in the central nucleus of the inferior colliculus. Similarly, no evidence of reorganization was reported in the dorsal cochlear nucleus of adult hamsters in which restricted cochlear lesions were produced by intense noise exposure (132). These data suggest that the pathway from the dorsal cochlear nucleus to the central nucleus of the inferior colliculus is not involved in the lesion-induced reorganization in the colliculus.

One possible explanation of the mechanisms of cortical map reorganization is based on the loss of cortical inhibition, specifically surround inhibition. Receptor organ damage leads to an anatomically defined loss of inhibitory inputs in the somatosensory and visual cortices (e.g., Refs. 72, 99, 360, 372) and a physiologicaly defined unmasking of normally silent inputs in the deprived sensory cortex (29). Surround inhibition is inhibition deriving from the receptor surface regions adjacent to those providing excitation, and it has been shown to restrict physiological response spread within cortical neurons. However, in the auditory system, receptor organ damage produces a loss of surround inhibition in the auditory cortex without topographic map plasticity (266). The level of inhibition was examined in cats with restricted unilateral cochlear damage. The threshold shifts assessed by measurement of the compound action potential were ~25 dB or less. For assessing the inhibitory parts of the neuronal tuning curve, a conditioning (forward masking) experimental design was used involving the presentation of two successive stimuli. The second of these tones was fixed at the cortical neuron's characteristic frequency (CF; the frequency to which the neuron is most sensitive) and the first was varied in frequency and intensity outside the excitatory response area of the cortical neuron. This procedure revealed that surround inhibition was almost absent in cortical neurons with CFs within the cochlear hearing loss range (Fig. 2). This was in sharp contrast to neurons with a CF outside the area of hearing loss and with neurons from normal control animals. In the cortical area corresponding to the lesioned cochlear region, further signs of decreased inhibition were also present: nonmonotonic spike-intensity functions were less numerous, the responses to white noise stimulation exhibited less inhibitory influences, and some neurons expressed dual response lobes in the excitatory response areas. However, the loss of surround inhibition was not accompanied by a reorganization of the topographic CF map. It is therefore probable that the loss of surround inhibition is not responsible for the topographic reorganization of the auditory cortex after restricted peripheral lesions. The results, however, demonstrate clearly that hearing loss, and in this case hearing loss induced by noise in most animals, results in a profound decrease of inhibition in the auditory cortex and possibly also in some subcortical structures. This is in agreement with data obtained in experiments where the effects of noise exposure on the auditory cortex function were assessed (see sect. IV).

View larger version (17K): [in this window] [in a new window]   Fig. 2. Restricted cochlear hearing losses result in the near absence of cortical surround inhibition, which alters neuronal response patterns shaped by inhibition. A: examples of cochlear hearing losses in two cats. Cochlear hearing sensitivity, measured as frequency-specific thresholds for the auditory nerve CAP, is expressed as threshold loss relative to mean normative thresholds, which are plotted as zero. The hatched area represents the range within 1.64 SD of the mean normative threshold. B: incidence of surround inhibition in primary auditory cortex (AI) neurons of normal animals (open bars) and in neurons of test animals grouped according to whether the neuron's CF was at a frequency within the range with CAP threshold losses (test neurons, solid bars) or at a frequency outside this range (control neurons, hatched bars). C: patterns of responses to pure tone stimuli at CF: incidence of AI neurons with monotonic or nonmonotonic intensity functions. AI neurons were grouped as detailed for B. [From Rajan (266)].

	III. CHANGES RESULTING FROM AGING

Top Previous Next References

A common denominator for the cause of plastic changes in both total ablations or partial lesions of the cochlea is the loss of receptors, the hair cells. Apparently it is not relevant whether the loss is induced by mechanical intervention or by the destruction of hair cells with ototoxic drugs. In addition, several other causes of hair cell loss exist, which may be accompanied by plastic changes in the central auditory system. Damage of hair cells may result from overexposure of the auditory system to noise or as a consequence of aging. The latter case is usually characterized as presbycusis. Presbycusis in humans refers to a constellation of age-related auditory deficits that include a loss of hearing sensitivity and a decreased ability to understand speech, particularly in the presence of background noise. The hearing loss tends to accelerate with age, with high-frequency losses exceeding low-frequency losses at all ages. Data from large populations screened for noise exposure and otologic disease (for review see Ref. 379) show a progressive increase of hearing loss amounting to ~20 dB at frequencies below 1 kHz, when hearing thresholds at ages 30 and 70 years are compared, and increasing to 60 dB difference at 8 kHz (337). Age-related hearing deficits in humans have usually been attributed to changes in the cochlea, including a loss of sensory cells, atrophy of stria vascularis, and a loss of spiral ganglion cells. However, an important part of presbycusis consists of changes in the central auditory system. The changes were demonstrated in many recent animal experiments, particularly in mammalian species with a relatively short life span, such as mouse, rat, or gerbil. The results of these experiments will be briefly reviewed further.

A.  Animal Models of Presbycusis

1.  Animal strains used in the study of presbycusis

Animals suffer a similar deterioration in hearing function with aging as do humans. However, the extent of the deterioration is different not only among animal species but also within a species. Particularly in mice, there are many inbred strains that exhibit rapid progress in presbycusis as well as many mutants that succumb to progressive hearing loss with aging.

One of the best-characterized models of human presbycusis is the Mongolian gerbil (2, 20, 21, 81, 97, 206, 208, 300, 301, 341). The gerbil has an audibility curve similar to that of humans, with greatest sensitivity around 4 kHz. It is an outbred strain, and it is therefore genetically diverse, like humans. As a result, variability in the magnitude of hearing loss that gerbils develop over their life span is extremely high. Another model of human presbycusis is represented by the C57BL/6J (C57) strain of mouse (101, 109, 188, 189, 200, 358, 377, 379-383, 385, 386, 388-390). In contrast to gerbils, all individuals of the C57 strain are genetically identical. The disadvantage of this model of human presbycusis is that the audibility curve in the C57 mouse is shifted upward by several octaves, with greatest sensitivity around 16 kHz (102, 166, 188, 200). The C57 mouse strain, which demonstrates a progressive auditory decline with onset at an early age, is usually compared with another model, the CBA/J (CBA) mouse strain, which displays a moderate auditory impairment with onset late in life (100, 102, 109, 136, 165-167, 245, 307, 310, 317, 373, 379, 384, 387-390). Another model, used particularly in studies of the biochemical aspects of presbycusis, is represented by the Fischer 344 rat or Fischer 344 × Norway F1 strain of rats (32, 64, 201-203, 241, 242, 309, 396). Compared with human, the audibility curve in rat is (as in the mouse) shifted upward to higher frequencies with the best sensitivity around 16 kHz (334).

The three mentioned animal models now dominate presbycusis research; limited data are also available on hearing loss connected with aging in other species. Changes in the cochlea and in the central auditory system with aging were previously described in guinea pig (39, 50, 51, 112), chinchilla (23, 187), cat (85), and monkey (45, 93). The main obstacle for detailed analysis of presbycusis in these species relates to their long life spans, which in rhesus monkey exceeds 30 years. All animal models are of limited value in explaining the sequelae of presbycusis in humans, since in humans the main problem of presbycusis is the deterioration in speech comprehension (see sect. IIIB).

A characteristic feature that dominates cochlear pathology in presbycusis in animal models is a loss of hair cells in the inner ear. For example, in the C57 mouse strain, the genetically programmed pattern of sensorineural loss starts in early adulthood, ~2 mo of age, and progresses to near total deafness by the second year of life (166, 378). The cochlea undergoes progressive hair cell loss, starting in the basal coils (high-frequency region) and progressing to involve increasingly lower frequencies along the cochlear membrane (102, 317). A loss of both outer hair cells (OHC) and inner hair cells (IHC) is present, although the percentage of lost OHC is always higher. In contrast, the CBA mouse shows little evidence of hair cell damage until late in life, i.e., until 18 mo of age. By 26 mo of age, the OHC losses in the apex (low-frequency region) and base of the cochlea in the CBA mouse equal approximately one-half of the total number. At the same stage, one-quarter of IHC are lost. The hair cell loss is, in the C57 mouse, accompanied by a loss of spiral ganglion cells with nearly complete loss in the basal cochlea during the second year of life (379). The measurements of hearing threshold shifts with evoked potentials during the life span of C57 mice reflect hair cell loss in the inner ear (378). Threshold shifts are mostly expressed at high frequencies, later progressively spreading to low frequencies.

From species with life spans longer than mice or rats, more extensive data about the anatomic and functional consequences of aging are available in guinea pigs and chinchillas.

In guinea pigs, cochlear pathology has been studied in animals up to an age of 5 years. No significant loss of hair cells was seen in the basal or middle turn of the cochlea of aged animals (111). In the apical (low-frequency) turn, there was a significant loss of hair cells in all rows of the outer hair cells (up to ~20%), most severe in the third row. There was no loss of apical inner hair cells in the aged animals. The apical loss of hair cells corresponded with the originally described decrease in the numbers of spiral ganglion neurons, which also occurred near the apex (39). Recordings of auditory brain stem responses (ABR) (51, 112) demonstrated a marked elevation of hearing thresholds in this species; however, the reason for this was considered to be conductive hearing loss (112). There were no significant age-related changes in the interpeak intervals of the ABR. Detailed histopathology in chinchillas, at ages up to 19 yr, demonstrated small but progressive losses of OHC and IHC (with the outer row of OHC most severely affected), more expressed at the apex and base. In many cases IHC losses were accompanied by damage to spiral ganglion cells (23). A later study by McFadden et al. (187) confirmed the losses of OHC and IHC in the basal and apical regions of the cochlea in aged chinchillas. The losses were paralleled by decreases in distortion product otoacoustic emissions and small but significant declines in auditory sensitivity, indicated by measurements of evoked potentials with high-frequency losses exceeding low-frequency losses.

Although most changes in hearing function connected with aging may be considered as a reflection of cochlear pathology, the results of several studies suggest that changes in the function of the central auditory system might appear as well. Willott (377, 378) compared the activities of neurons in the inferior colliculus of C57 and CBA mice. The results showed several changes in the function of neurons in the fast-aging C57 mouse, which were not present in the CBA mouse. As high-frequency sensitivity in C57 mice declined as a result of cochlear pathology, neurons in the ventral IC that normally have low thresholds for high-frequency tones could no longer respond to them. The best frequencies (the frequency for which a neuron has the lowest threshold) of ventral IC neurons shifted from high to middle frequencies so that the best frequencies tended to be similar throughout the IC. That is, tonotopic organization was severely disrupted in middle-aged C57 mice. The "shifted" best frequencies had thresholds that were lower than the thresholds of the same frequencies in the tuning curve "tails" of neurons in the same location in young mice. As hearing loss progressed even more, thresholds for all frequencies were eventually elevated. Willott and colleagues (382, 390) later completed the study by investigating single and multiple unit responses in the cochlear nuclei and auditory cortex of the C57 mouse. In principle, changes in the auditory cortex were similar to those observed in the IC, i.e., with aging virtually the entire auditory cortex contained neurons with best frequencies in the middle-frequency range. In contrast, neurons in the ventral nucleus of the C57 mouse did not show the age-related sensitization of tuning curve tails seen in the IC and auditory cortex (AC) neurons from high-frequency tonotopic regions. Altogether the results demonstrated that the plastic changes in the thresholds of tuning curves, observed in the aging C57 mice, are limited to regions of the central auditory system higher than the ventral cochlear nuclei.

Many studies of plasticity mechanisms have suggested that a diminished inhibition could be involved in changes observed during brain aging. In the central auditory system such investigations were mostly performed in the IC of the Fischer 344 rat. The IC has been shown to display age-related changes including decreased numbers of GABA immunoreactive neurons, decreased basal levels (concentrations) of GABA, decreased GABA release, decreased glutamate decarboxylase levels, decreased GABA_B receptor binding, decreased numbers of presynaptic terminals, and subtle GABA_A receptor binding changes (33, 34, 201). Age-related changes of the GABA_A receptor in the IC of young-adult, middle-aged, and aged Fischer 344 and Fischer 344/Brown-Norway F1 hybrid rats were related to receptor subunit composition and receptor function (32). In both strains, the aged group exhibited significant increases in ₁-subunit protein and a decrease in ₁-subunit protein. These results suggested an age-related change in GABA_A receptor composition. Age-related changes have also been noted in calbindin D-28k and calretinin immunoreactivity in the IC of young and old C57 and CBA mice (403). As mentioned above, the CBA/CaJ mouse maintains good hearing until very late in life, whereas the C57Bl/6 strain exhibits severe sensorineural hearing loss at an early age. Cell counts revealed a 22.3% decrease in the number of CB+ cells in old CBA mice and a 25.1% decrease in old C57 mice. Calretinin immunoreactivity was high in the pericentral regions of the IC, but the central nucleus was devoid of CR+ cells. The pericentral parts of the IC showed increases in the number of CR+ cells, but only in the old CBA mice. No significant change was observed in the old C57 mice. The increase in CR+ cells was interpreted as a possible compensatory adaptation to the decrease in CB+ cells. Changes with aging were observed also in the structure of the IC in the C57 mice (135). A significant loss of the axosomatic synapses on the principal neurons in the central nucleus of the IC started at 6 mo of age (early middle age) and became progressively severe into old age (24 mo). In contrast, in the CBA mouse, aging was not accompanied by changes in the number or type of synapses, the length of synaptic apposition, or the size of the synaptic terminal area. The preservation of synapses on principal neurons in this strain suggests that synaptic loss is not an inevitable event in aging but may be related to the preservation of peripheral auditory function and input to the neurons (136).

The changes accompanying aging were demonstrated in the IC of Fischer 344 rats also with electrophysiological methods (241). A comparison of responses of single units to acoustical stimuli in aged and young rats revealed that although the average threshold increased from 25.4 dB sound pressure level (SPL) in young rats to 56.1 dB SPL in aged rats, no differences were noted in spontaneous activity, first spike latency, dynamic range, percentage of units with nonmonotonic contralateral CF tone rate-intensity functions (RIFs), or percentage of units sensitive to change in CF tone presentation rate. In aged rats, a higher percentage of units were poorly responsive to auditory stimulation than in young animals; there was a reduction in the maximum discharge rate, an increase in the percentage of units classified as onset in their temporal response pattern, and an increase in the breadth of the isointensity functions at 30 dB above threshold. Age-related changes in the central nucleus of the IC (CIC) frequently differed from those observed in the external cortex of the IC (ECIC). For example, the percentage of units classified as having nonmonotonic contralateral tone RIFs decreased with age in the CIC but increased with age in the ECIC. Altogether these results support the hypothesis that there is an age-related shift to higher intensities in the working range of most CIC units along with a small, selective deficit in inhibitory processing. However, despite peripheral deficits that lead to reduced input to the IC and neurochemical changes affecting neurotransmitter levels, IC neurons in aged rats were able to respond to simple auditory stimuli in a fashion quite similar to that observed in young rats. Even though a decrease in GABA and GABA receptor levels in the old Fischer 344 rat IC was present, the binaural interaction as measured by responses of single units in the IC to binaural stimuli essentially did not differ between young and old Fischer 344 rats (242). Similar results were observed in the superior olivary complex (64), which together with results from the IC demonstrate that the mechanisms of binaural interaction might not be significantly influenced by aging.

Changes in the levels of mediators and their receptors accompanying aging were demonstrated in the Fischer 344 rat not only in the IC but also in other parts of the central auditory system. In the cochlear nucleus, a significant age-related decrease in [³H]strychnine binding was observed, suggesting a decrease in strychnine-sensitive glycine receptors (203). Similarly in the neocortex of Brown Norway × Fischer 344 F₁ rats, an age-related depletion of presynaptic boutons as indicated by a decrease in synaptophysin immunostained punctae (396) was found.

Interesting aspects of the auditory system's plasticity were shown by the results of studies in which animals were exposed to an augmented acoustical environment, i.e., their auditory system was continuously stimulated by broad-band noise for several months (348, 391, 392). Such procedures were able to ameliorate aging-elicited changes in the auditory function in strains of mice with progressive hearing loss. Mice of strains that exhibit high-frequency hearing loss during young adulthood [the C57BL/6J strain (C57)] or around the time of weaning/adolescence [the DBA/2J strain (DBA)] were exposed, starting at age 25 days, to a 70 dB SPL broad-band noise for 12 h every night. The augmented acoustical environment was maintained until age 14 mo in C57 mice and 9 mo in DBA mice. Control mice were age matched and maintained under normal vivarium acoustic conditions. The ABR, acoustic startle response amplitude, and prepulse inhibition (PPI) were used to assess the function of the auditory system. Exposure to the augmented acoustical environment resulted in improved auditory performance in both strains (better PPI, lower ABR thresholds, larger startle amplitudes). Similar results were obtained in mouse strains that exhibit various degrees and time courses of progressive hearing loss (BXD-22, BXD-12, BXD-16, BXD-14, BALB/cJ) and in normal-hearing CBA/CaJ mice. The augmented acoustical environment was maintained for at least 30 days in each strain. Exposure to the augmented environment resulted in improved auditory performance (better PPI, lower ABR thresholds) when hearing impairment was present, but not when hearing was normal. The ameliorative effects occurred irrespective of the mouse's age at the onset of hearing loss, as long as the initiation of treatment with the augmented environment preceded the occurrence of severe hearing loss. If the treatment was delayed beyond such a point, loss of threshold sensitivity progressed as usual, although PPI could still benefit. The results have shown that augmented environment treatment can slow, but not prevent, the occurrence of severe genetically determined hearing loss. The effects of the exposure to an augmented acoustical environment may be compared to some extent with the effects of similar experiments in which animals were exposed to a so-called enriched environment (for review see Ref. 260).

The signs of presbycusis in humans are similar to those observed in experimental animals (316, 379). Loss of IHC is typically present in the basal turn of the cochlea, whereas the loss of OHC is more scattered, spreading up to the apical part. The loss of OHC is always more pronounced than the loss of IHC. Less severe signs of degeneration in the cochlea comprise the occurrence of different disarrays of stereocilia of hair cells, including the occurrence of giant stereocilia. Concomitantly with the loss of hair cells, presbycusis is accompanied by a loss of spiral ganglion cells, which usually comprises the loss of peripheral dendrites as well as central projections, i.e., auditory nerve fibers. The efferent system seems to be more robust and does not degenerate to the same extent as the afferent system. In the central auditory system a marked reduction in the cell size of neurons in the ventral cochlear nucleus was found in subjects with profound hearing loss (221). The size of neurons was larger in subjects with greater survival of cochlear ganglion cells.

The results of electrophysiological investigations of the aging auditory system correspond with the described morphological changes. The amplitude of the acoustic reflexes decreases with age similarly to the amplitude of auditory brain stem potentials. The latencies of ABR waves may be slightly prolonged. In some cases increases in the amplitude of middle latency responses (MLR) at higher intensities of sound stimulation were found, probably as a sign of a decreased amount of inhibition in the central auditory system (397). Changes with age in the hearing thresholds for pure tones have been well documented in many audiometric studies (e.g., Ref. 337) (Fig. 3). In listeners screened for the absence of subjective indications of hearing loss of different origin than aging, the thresholds for low-frequency tones change little with age, while in sharp contrast, high-frequency sensitivity is greatly diminished (for review see Ref. 379). The decline in the hearing sensitivity is different for men and women; in men it declines twice as fast as in women at most ages and frequencies (249, 337). Longitudinal declines in hearing sensitivity are detectable among men by age 30, but the age of the onset of the decline is later in women at most frequencies. Women have more sensitive hearing than men at frequencies >1,000 Hz, but men have more sensitive hearing than women at lower frequencies (246, 249).

View larger version (20K): [in this window] [in a new window]   Fig. 3. Effect of age on hearing threshold levels in decibels at 1, 2, 4, and 8 kHz. Solid circles represent male, and open circles represent female. [From Takeda et al. (337), copyright 1992 Springer-Verlag.]

In addition to increases in hearing thresholds, elderly listeners manifest deficiencies in their ability to process information about frequency, intensity, time, and space. In most cases the presbycusis is based on cochlear impairment, a typical sign of which is the occurrence of a phenomenon called loudness recruitment. In this case relatively small increments in sound intensity are perceived as disproportionate increases in loudness. Loudness recruitment is present in 20-50% of older listeners (379). Frequency resolution declines with aging as hearing loss increases (177). When hearing level is accounted for, there is only a minor decline in age-dependent frequency resolution. Temporal resolution, which is usually measured as gap detection thresholds, declines with sensorineural hearing loss. However, in older adults with good hearing, significantly higher thresholds in gap detection studies were also found. Older adults who were matched to a group of younger listeners with respect to their audiometric thresholds had gap detection thresholds that were ~30% higher than those of the young listeners (302, 311, 319). Spatial hearing based on binaural cues also deteriorates with aging. Higher thresholds for interaural time differences were found in elderly adults with clinically normal hearing than in matched young listeners (319), and similarly, the results of a test of binaural hearing based on the precedence effect were found to be affected by age (40). Because in both studies only subjects without peripheral hearing loss were included, the results speak in favor of specific changes in the central auditory system with aging.

By far the most important function of the auditory system in humans is speech recognition. Speech understanding deteriorates with aging, especially in situations in which the voice of the speaker is accompanied by a noisy background or by other voices. The details of this topic are, however, out of the scope of the present review. Improving speech recognition is an important goal in the treatment of hearing loss with hearing aids. There is only limited information available about improvements in hearing function with long-term wearing of hearing aids (240). For example, it has been shown that in bilateral, hearing-impaired subjects using a single hearing aid, speech identification for the unaided ear declined over a long period of time relative to the performance of the normally aided ear after unilateral amplification (308). In other experiments (73) with long-term users of a single hearing aid, the normally aided ear performed better for speech identification in the presence of background noise than the unaided ear at high SPL (>75 dB). At low SPL (<75 dB), the situation was reversed, i.e., the unaided ear showed better performance than the normally aided ear. The advantage of the normally aided ear only became apparent after listeners had used their hearing aid for 4-6 wk (74). The advantage at high presentation levels is probably the result of increased sensitivity to intensity changes at levels to which the ear is normally exposed. As the tested listeners had sloping sensorineural hearing impairment that was bilateral and symmetric, their hearing aid provided gain at the impaired, higher frequencies, and little to no gain at frequencies that were close to normal hearing levels. This was confirmed in intensity discrimination tests in which a greater sensitivity to changes in intensity was observed in a normally aided ear tested without the aid than in the unaided ear at sound pressure levels greater than 85 dB (281). Plastic changes in normally aided ears definitely do occur; research in this topic is just beginning. For example, because hearing aids are almost always used only unilaterally, very little is known about benefits of bilateral hearing aids.

In summary, presbycusis is a process that is present in all mammals. However, in humans it comprises not only the loss of auditory receptors and the resulting changes in the central auditory system, but also inevitable impairment of speech comprehension. The practical solution of the problems associated with presbycusis in humans, such as auditory threshold shifts, intensity recruitment, and loss of frequency and temporal resolution, lies in refinements of prosthetic devices such as hearing aids. A knowledge of the plastic changes in the central auditory system occurring in presbycusis may open new vistas in the pharmacological treatment of the impairment. For example, it will be necessary to understand the influence of decreased inhibition in the central auditory system on the manifestation of presbycusis symptoms with the aim of starting rational treatment. Practical outcomes may also be expected from further studies of the long-lasting effects of wearing hearing aids. An understanding of the plastic changes in the auditory system during aging should serve to ameliorate the negative sequelae of this inevitable process or at least move the problems to an older age.

	IV. NOISE-INDUCED CHANGES IN THE AUDITORY SYSTEM

Top Previous Next References

Noise exposure produces various effects in the auditory system, the detailed description of which is out of the scope of this review. For more information several other reviews can be consulted (11, 42, 292, 328).

The usual characterization of noise-induced changes in the auditory system is based on changes in the hearing threshold. Short-term exposure to noise with relatively low sound intensities results in a temporary hearing loss with a limited threshold shift and recovery of hearing sensitivity after some time (temporary threshold shift; TTS). Exposure to high intensities of noise even for a short period produces damage in the cochlea without recovery in hearing sensitivity; the threshold shifts are therefore permanent (permanent threshold shifts; PTS).

While the primary reason for PTS is the loss of OHC and IHC in the cochlea, the mechanisms of TTS still remain to be elucidated. The primary source of mechanical input to the cochlear hair cells during acoustic stimulation is the relative shear between the tectorial membrane and the hair cells' apical parts produced by basilar membrane motion. The displacement of stereocilia modulates their transducer conductances. Stimulation by a loud sound may cause changes in the mechanical properties and consequently in the conductances of hair cells. Indeed, several studies have indicated that intense acoustical stimulation changes the micromechanical properties of the sensory hair bundle at the apical pole of the hair cells (297, 298). Damage to the stereocilia correlates well with alterations in the tuning curves of auditory nerve fibers (169).

Calcium ions may play a potential role in the pathological processes following acoustical overstimulation, as significant elevations of Ca²⁺ were found in the cytoplasm of OHC after acoustic overstimulation (68). Similarly, the increased permeability of the endolymph-perilymph barrier to potassium and sodium ions after noise exposure (126, 147, 148, 290) could be involved in the mechanisms of TTS.

Another mechanism for the TTS has been proposed in connection with changes that occur in the synapses between IHC and peripheral dendrites of the primary auditory neurons, i.e., spiral ganglion cells. Acoustic trauma causes a disruption of dendrite endings below the IHCs, leading to synaptic uncoupling. The dendrite damage might be due to an excessive release of the neurotransmitter glutamate from the IHCs, which is toxic to the structure and function of the primary auditory neuron (262-264). Excess glutamate in the extracellular space below the IHCs could result in permeability changes in the postsynaptic membrane of the dendrite. The resulting osmotic imbalance would cause water to move into the dendrites, resulting in dendrite swelling and ultimately membrane disruption. Local application of the glutamate agonist AMPA induced a destruction of primary auditory dendrites similar to that observed after acoustic trauma, while the application of the glutamate antagonist kynurenate protected dendrites of the primary auditory neurons from noise-induced damage. The time course of the synaptic repair after acoustic trauma corresponded with the recovery of the auditory threshold (262, 264).

Threshold shifts may be modulated in several ways. Simultaneous acoustical stimulation with a tone of the same frequency delivered to the other ear may decrease the value of monaural TTS (36, 269, 271, 272). The TTS in guinea pigs evoked by a 10-kHz pure tone at 107 dB SPL for 1 min decreased by ~30% by simultaneous exposure of the other ear to a 10-kHz pure tone at 85 dB SPL. When the other ear is stimulated with a different frequency, the reduction in TTS does not occur. The reduction in TTS is mediated by the activation of the descending olivocochlear bundle and usually is demonstrated by a suppression of otoacoustic emissions elicited by stimulation of the contralateral ear (258, for review, see Ref. 174). Another type of modulation or suppression of threshold shifts is represented by the conditioning or toughening procedure. The procedure is based on experimental results (269, 270) showing that an initial monaural low-level sound exposure could significantly reduce the threshold shift caused by a subsequent high-level loud sound exposure to the same cochlea, even when there was no residual TTS to the priming sound. Later it was shown (31) that protection against noise trauma can occur when guinea pigs are sound conditioned to a continuous low-level, long-term, nondamaging stimulus before the traumatizing exposure. Sound conditioning provides significant protection against the traumatizing stimulus compared with a control, unconditioned group. Another way of providing protection against noise trauma is by using interrupted repetitive stimulation.

Mechanisms put forth to explain the conditioning phenomenon include a decrease in calbindin in the cochlea (30), increases in heat shock proteins in the cochlea (6), changes in the cochlear antioxidant activity (117), and the effects of cochlear efferents on the hair cells. For example, in chinchillas exposed to a conditioning paradigm, protection of the hair cells in the organ of Corti from noise-induced damage was associated with increased levels of antioxidants in the stria vascularis (117). The role of cochlear efferents in acquired resistance to noise-induced hearing loss is less clear. Complete deefferentation of the ear, i.e., transection of both crossed and uncrossed fibers of the olivocochlear bundle, in chinchillas exposed to noise after a conditioning paradigm revealed substantially more TTS, greater PTS, and larger cochlear lesions in the OHC in the deefferented ears (404). The results of experiments in conditioned guinea pigs (154) confirmed a larger PTS in animals after deefferentation; however, the authors found that animals undergoing sham surgery (brain stem cuts that failed to transect the crossed olivocochlear bundle) appeared protected whether or not they received the conditioning noise exposure. Conditioning-related protection may arise from a generalized stress response, rather than from the function of the olivocochlear bundle (154). The protection afforded by conditioning may not be completely explained by amplification of the olivocochlear reflex but rather by a general enhancement of cochlear sensitivity (155).

Hair cell loss in the cochlea after auditory overstimulation is connected with the degeneration of auditory nerve fibers. Patterns of axonal degeneration were found in the brain stem of adult chinchillas (140, 222). The pattern of terminal degeneration in the ventral cochlear nucleus correlates with the position of IHC loss in the cochlea, i.e., it is cochleotopic. This process is consistent with the dystrophy resulting from the disappearence of IHC and auditory nerve fibers. However, in addition to the dystrophic process (present mostly in the ventral cochlear nucleus), fine fiber degeneration occurs without expressed cochleotopic correlation in the dorsal cochlear nucleus and trans-synaptically in the superior olive and IC. It is possible that an excitotoxic process, similar to that described at the synapses between IHC and auditory nerve fibers, plays a role in this type of terminal degeneration.

The effects of noise exposure on the central auditory system depend on the intensity, frequency content, and duration of the exposure. Differences may also be present immediately after exposure or at later time periods. The characteristic consequences of peripheral sensorineural hearing loss resulting from noise exposure are threshold shifts, abnormally rapid loudness growth, poor speech discrimination, and a reduction in the temporal summation of acoustic energy. Threshold shifts correspond with the degree of hair cell dysfunction in the inner ear (either temporary or permanent) and may be detected by any type of gross auditory evoked potentials (usually brain stem-evoked potentials or middle latency responses) as a shift in the potential threshold. Similarly, threshold shifts are reflected in shifts of the tips of tuning curves to higher intensities (22, 98, 170, 256, 359). Neuronal threshold shifts have been reported to be different at different levels of the auditory pathway with higher shifts present in neurons of the IC than in neurons of the cochlear nucleus (CN) in the same animal (293, 331). These findings did not receive further experimental support, but significant differences between auditory nuclei were found in response to suprathreshold sounds. In contrast to the situation in the auditory nerve and cochlear nucleus, the amplitudes of evoked responses in the IC (19, 294) and in the auditory cortex (257, 334, 335) were found to be enhanced compared with the preexposure situation (Fig. 4). The enhancement was observed immediately after noise exposure and lasted for many hours or days. The intensity of sound exposure must be >105 dB SPL. The recovery of the enhanced amplitudes to normal values exhibited a different time course than the recovery of threshold shifts (334, 335). Greater susceptibility to noise exposure, expressed as threshold shifts and enhanced amplitudes of cortical evoked potentials, was found in rat pups during the first 6 wk after birth (287).

View larger version (29K): [in this window] [in a new window]   Fig. 4. Average amplitude-intensity functions for the auditory nerve action potential (CAP), evoked responses in the inferior colliculus (IC-ER), and evoked responses in the auditory cortex (AC-ER) measured before noise exposure and at different times after exposure. The ordinate expresses the potential amplitude in percentage of the amplitude measured at 90 dB SPL before exposure (A). Inset: principle of amplitude measurement. [From Popelá et al. (257).]

It is possible to explain the enhanced amplitude of evoked responses in the IC and AC by the lack of inhibition in these structures after exposure. This idea was supported by the results of experiments in which neurons were recorded in the CN (22) and in the IC (359) before and after exposure to high-intensity pure tones, the frequency of which was just above the neuron's CF. If a unit had an inhibitory response area at frequencies above the CF and if the traumatizing tone reduced the magnitude of the inhibitory response, then the neuron's discharge rate to suprathreshold tones at the CF increased (Fig. 5). Similar to the case of exposure to wide band noise (331), the exposure to a pure tone at a frequency just above the CF did not change the shape of the poststimulus time histogram, i.e., the pattern of the unit response (22, 292, 295, 296, 359). The discharge rate increased particularly in those neurons that had nonmonotonic rate-level functions, i.e., in neurons where the inhibitory influence was large at high sound intensities. Decreased inhibition after noise exposure was also found in other studies. For example, a decreased influence of the GABA antagonist bicuculline on one part of the IC evoked response was noticed after noise exposure (336). Decreased levels of glutamate decarboxylase (GAD) isoforms were found in the rat IC 2-30 days after exposure producing 20-30 dB TTS and some limited PTS (1). Surprisingly, immediately after exposure, the levels of GAD were elevated compared with IC levels in normal animals. It has to be mentioned at this point that the decrease of inhibition in the cortex of experimental animals is a general phenomenon accompanying a loss of receptors in the periphery, not only in the auditory system but also in other sensory systems (see sect. IIB2).

View larger version (26K): [in this window] [in a new window]   Fig. 5. Preexposure and postexposure tuning curves (A) and discharge rate level functions at the CF (B) of a neuron from the central nucleus of the inferior colliculus. CF, CF threshold and exposure conditions shown at top of panel. [From Wang et al. (359).]

One of the changes that can be observed after noise exposure is a change in the frequency tuning of neurons. When exposure was to a tone above the CF and the neuron had inhibitory sidebands flanking the excitatory area (292, 296, 359), the low frequency area around the tail of the tuning curve expanded and in some cases the high-frequency excitatory area expanded as well. Changes in the frequncy tuning were not observed in the IC neurons with tuning curves that had extended low-frequency tails, a lack of inhibitory side bands, and monotonic spike-intensity functions.

The spontaneous activity of auditory nerve fibers and neurons in the central auditory system decreases after noise exposure as a rule (137, 291). An exception was an increase observed in units of the cochlear nucleus in the rhesus monkey after exposure to short-lasting tones of moderate intensity (173). Specific changes are present in the auditory cortex of the cat after 30-min exposure to a loud tone (93-123 dB); spontaneous activity increases significantly in the primary auditory cortex (AI), does not change in the anterior auditory field (AAF), and decreases in the secondary auditory cortex (AII) (141). Also, other parameters of neuronal function are affected to a different extent in different cortical fields: the mean threshold at the CF increases significantly in the AI and AAF but not in the AII, the mean frequency-tuning curve bandwidth decreases significantly in the AII, whereas an enhancement of frequency response areas around the CF is present most frequently in the AAF. These results, along with the results of others (28, 256, 257, 331, 334, 335), strongly suggest that the function of the auditory cortex is seriously disturbed after noise exposure.

Psychoacoustical experiments and behavioral experiments in animals have shown that acoustic overstimulation leads to a breakdown in temporal integration so that hearing-impaired listeners show relatively little improvement in threshold with increasing stimulus duration (315). In normal subjects, the threshold to a long-duration tone is ~12 dB lower than to a short-duration tone. In noise-exposed subjects, the extent of improvement with increased duration declines (122, 207, 399). No change was found in temporal integration in guinea pigs exposed to noise-producing asymptotic threshold shifts (330); however, the threshold shifts were relatively mild. Another type of function, frequently studied in experiments with temporal integration, the gap function (i.e., recognition of a silent period in a long duration acoustical stimulus), was also found to deteriorate after noise exposure. Gap thresholds increased in experiments with noise-exposed chinchillas only when the asymptotic threshold shifts exceeded 40 dB (77). In patients with sensorineural hearing loss, the gap resolution was found to be significantly poorer than in normal listeners (65, 78). Larger gap thresholds tended to be associated with slow rates of recovery from forward masking. Similarly in experimental animals, increased time constants fitting forward masking data were found (9).

Up to 15% of the world's population frequently or permanently suffers from subjective tinnitus, which is a phantom auditory perception without a counterpart in the outside world (105). A frequent cause of tinnitus is noise exposure and presbycusis, but tinnitus may appear as a side effect of treatment with ototoxic drugs (e.g., salicylates or chinin). There are several other mechanisms of tinnitus generation including involvement of the middle ear muscles, inner ear hydrops, etc. In many cases, however, tinnitus originates in the central auditory system or in brain structures related to the central auditory system. This statement is supported by the fact that tinnitus persists in patients with acoustic neurinoma after transection of the auditory nerve (107, 185). Recently, accumulated evidence from magnetoencephalography and positron emission tomography (PET) studies speaks in favor of a central origin for several types of tinnitus. In patients who could alter tinnitus loudness by performing voluntary oral facial movements, these loudness changes affected cerebral blood flow in the auditory cortex contralateral to the ear in which tinnitus was perceived, whereas unilateral cochlear stimulation caused bilateral effects (171, 172). Patients, compared with controls, showed evidence for more widespread activation by tones at aberrant links between the limbic and auditory systems. In a similar study (76), patients were used with another rare form of tinnitus elicited by eye movements. In this case the phantom auditory sensation increased regional blood flow bilaterally in temporoparietal association auditory areas but not in the primary auditory cortex. A marked shift of the cortical representation of the tinnitus frequency into an area adjacent to the expected tonotopic localization was observed with magnetoencephalography. The subjective strength of the tinnitus was found in this case to correlate strongly with the amount of cortical reorganization (224). The data obtained in patients with tinnitus thus confirmed the results of experiments in animals with partial cochlear lesion (see sect. II) and suggested that tinnitus in many cases is related to plastic changes in the auditory cortex and may be compared with phantom limb pain occurring after amputation. In patients with tinnitus, a slight shortening of the latency of the wave V in brain stem-evoked audiometry was also observed (209), as well as changes in N1-P2 latency of late auditory evoked responses (235) and changes in the modified contingent-negative variation paradigm (105).

Several other facts, known from clinics as well as from laboratory experiments, support the hypothesis of a central origin of tinnitus. Tinnitus can be suppressed in a large proportion of patients by electrical stimulation of the inner ear by temporary adjusted electrodes (94, 156) or in patients with cochlear implants (for review, see Refs. 349, 350). Some methods, borrowed from clinical and experimental psychology, such as cognitive behavior modification (326) or habituation (119), are successfully used