Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 Facts, Theories, and Issues ETIOLOGY OF TINNITUS A student of tinnitus has no difficulty finding lists of conditions or mechanisms that cause, or are at least believed to cause, tinnitus. Often much more difficult to uncover is the evidence on which these beliefs are based. Simple intuition seems to be the most common basis. For example, it is intuitive that a localized lesion along the organ of Corti might cause an ~irrita- tion" that would in turn lead to continuous discharge in a small population of primary auditory fibers and, thus, produce a tonal or narrowband tinnitus. Similarly, local- ized damage in a small population of neurons in the coch- lear nucleus or other auditory structure might be expected to produce a perceptual experience of sound where none exists. A "sensory epilepsy" due to an interruption in the afferent-efferent loop sounds like it might be respon- sible for a tinnitus experience in the spectral region of the interruption. And it sounds reasonable that emotional changes might produce changes in the composition of the inner-ear fluids and/or changes in the cochlear blood supply that might in turn produce differential neural activity that would produce auditory sensations. But while these and dozens of other equally attractive proposals may be partially or entirely correct, there is little evidence to support any of them. To be sure, the etiology of a particular tinnitus has occasionally been established--particularly for those having clearly audible vibratory concomitants--but in the vast majority of cases the cause is unknown. There can be no doubt that reducing this ignorance about site of origin is desirable, for, presumably, type of treatment will eventually come to be based on diagnoses of specific causes. However, it is 10
11 doubtful that to date this ignorance has been a major deterrent to effective treatment, for so little is known about etiology and treatment that typically nearly every- thing is eventually tried if a patient persists in his or her complaints. Over the years a truly staggering number of proposals have been made about the etiology of tinnitus in all of its various forms. It would serve little purpose to attempt a comprehensive review of these here; instead, a few selected proposals are discussed. The topic of drugs that cause tinnitus is considered in Chapter 4. Miscellaneous Unequivocal Sources of Tinnitus mere are some unambiguously established causes of tin- nitus. The lack of ambiguity is due to different factors in different cases, but some common factors are (1) sudden occurrence or appearance of both the cause and the tin- nitus, (2) frequent or invariable associations between the cause and tinnitus, and (3) disappearance of the tinnitus following removal or treatment of the cause. As might be expected, many of these unequivocal sources of tinnitus involve gross causes. Severe blows to the head, for example, can produce transient, long-term, or permanent tinnitus of various types (Shucart and Tenner, 1981). Overdoses of various drugs and general anesthetics used during surgery can initiate or exacerbate a tinnitus. Anemia, hypertension, hypothyroidism, and migraine have been linked with tinnitus (CIBA Foundation, 1981:232-236), as has multiple sclerosis (Shucart and Tenner, 1981). Partial or total immobilization of the middle-ear struc- tures, such as that produced by otosclerosis or even im- pacted cerumen (ear wax), can cause tinnitus. The onset of tinnitus sometimes coincides with pregnancy, but so does otosclerosis, so the latter may be the true culprit. Meniere's Disease has a characteristic tinnitus as one of its primary symptoms (see "Meniere's Disease" in this chapter). So-called sensorineural hearing loss such as that induced by chronic or acute exposure to intense noise is frequently accompanied by tinnitus. And, of course, already mentioned are the well-established causes of cer- tain forms of objective tinnitus--anomalies of the vascu- lature or musculature of the head, neck, and jaw. In regard to tinnitus caused by ear wax, Feldmann (CIBA Foundation, 1981:234) believes that simple occlusion of the canal is not the source of the tinnitus but that
12 attachment or contact of wax with the tympanic membrane is required. Coles (CIBA Foundation 1981:235) believes that the wax is serving to simply attenuate external sounds, thereby allowing a preexisting tinnitus to be revealed, an effect that can sometimes also be observed by inserting plugs into the ear canal. It has been asserted that the tinnitus that accompanies otosclerosis is typically present only as long as the hearing loss is less than 50-60 dB (Lempert, 1946; Saltz- man, 1949; J. T. Graham, 1965); beyond this value, it gradually diminishes and then disappears altogether. This assertion deserves confirmation, for, if true, it would have important theoretical implications. Specifi- cally, it appears to be in accord with Kemp's (1981) ideas about the origin of so-called spontaneous otoacoustic emissions (see "The Objective/Subjective Issue" in this chapter). Among the types of tinnitus originating in the muscu- lature of the head is the rapid clicking sound produced by involuntary, rhythmic contractions of the muscles of the soft palate. The condition is known as palatal myo- clonus (see MacKinnon, 1968), and the actual sound source is believed to be the snapping together of the walls of the Eustachian tube. Tumors of the Eighth Nerve Unilateral tinnitus is frequently an early symptom of a tumor that affects the eighth nerve. Other--typically later-developing--symptoms are unilateral high-frequency hearing loss and vertigo. The overwhelming majority of these tumors are not malignant. Most are unilateral and most begin on the vestibular branch of the eighth nerve inside the internal auditory canal. (Because most eighth- nerve tumors are associated with the Schwann cells and are on the vestibular branch of the nerve, the term vestibular schwannoma is preferable to the older term acoustic neu- roma or neurinoma.) The damage done by these tumors is primarily sue to their compressing other tissues as they slowly grow in size. Brackmann (1981a) reported that tinnitus was present in 83 percent of a group of 500 patients suffering from unilateral tumors of the eighth nerve and that tinnitus was the first symptom noticed by about 10 percent of that group. Ronis (1981) agreed with the latter figure but found tinnitus to be an initial symptom in 92 percent of
13 his cases. Postoperatively, "virtually all" of Ronis' patients had some tinnitus; Brackmann felt that the tin- nitus was better postoperatively in about 40 percent of his patients and worse in about 50 percent (House and Brackmann, 1981). No information is offered about simi- larities and differences in tinnitus quality or magnitude, preoperative and postoperative. This lack of information makes it impossible to know if the "old tinnitus" persists or a "new tinnitus" is created by the surgical procedure. A fair summary of this issue appears to be that: (1) unilateral tinnitus and hearing loss can arise in so many ways that they are poor predictors of eighth-nerve tumors, (2) the vast majority of people with an eighth-nerve tumor do report tinnitus either as the initial symptom or in combination with hearing loss and vertigo, and (3) even successful surgery to remove the tumor rarely eliminates the tinnitus. m e clinician should remain alert to the possibility of an eighth-nerve tumor when unilateral tin- nitus and hearing loss combine with vertigo, but it is far from inevitable that a tumor will prove to be the cause. For some time, the prevailing wisdom was that early surgery is in order for all eighth-nerve tumors. Conse- quently, there was (and still is) concern that nonsurgical treatment of the tinnitus per se may work to the detriment of the patient by "covering up" an important symptom of a serious disorder (see "Medical Examination" in Chapter 5). In recent years however, there has been an increasing re- alization that many eighth-nerve tumors grow slowly and cause the patient only minimal discomfort and alteration of lifestyle. Thus, the feeling of urgency about tumor removal has been reduced somewhat; exceptions are tumors in children and young adults, which are still viewed as candidates for early removal. Further, the concern about "covering up" the tinnitus may be wholly unwarranted. It may well be that tinnitus caused by eighth-nerve tumors is refractory to most or all known treatments for tinnitus-- that is, that this important symptom cannot be "covered up." At this point, essentially nothing is known about this possibility, notwithstanding its practical and diagnostic importance. Noise Trauma and Presbycusis Prolonged and repeated exposure to intense sound produces a distinctive pattern of hearing loss not unlike that associated with aging (presbycusis). The loss is typi
14 cally greatest at the highest frequencies and less at the lower frequencies, often with a reasonably sharp transi- tion region between the affected and unaffected regions. Such patterns of hearing loss are very common and they are frequently accompanied by tinnitus. Feldmann (1971) estimated that about one-third of his patients followed this pattern and Schleuning's (1981) estimate was about one-half. Reed (1960) diagnosed 16 percent of his pa- tients as having acoustic trauma and 38 percent as pres- byousic. m e accompanying tinnitus is reported to be of high pitch, and, when it is matched to a tone, the match- ing frequency is often located in the sharp transition between regions of greater and lesser hearing loss (Feld- mann, 1971; Penner et al., 1981). MECHANISMS OF TINNITUS A distinction can be made between presumed or established causes of tinnitus and the actual mechanisms through which the tinnitus is produced. For example, what is the nature (and/or location) of the altered neural activity that pro- duces tinnitus following noise exposure, drug overdose, or a blow to the head? It is intuitive that in many cases of tinnitus, knowledge of the mechanism will eventually prove more valuable for diagnosis and treatment than will knowledge of the cause. Over the years, many proposals about underlying mech- anisms have been offered (a number are mentioned by Durrant, 1981). No comprehensive review is attempted here. Instead, some general comments are made, and then the physiological evidence for two particular mechanisms are examined. It is intuitive (which is not to say correct) that in order for a person to have an experience of a spectrally distinct sound for which there is no acoustical concom- itant, something about the neural activity in that per- son's auditory pathway must be misleading higher auditory centers into the erroneous experience. m at "something n might be a mechanical force--for example, a tumor or a blood clot causing a local compression of neural tissue-- or a biochemical or biomechanical upset affecting a subset of neurons at some level in the auditory nervous system. (It is tempting to think of the resulting "tinnitus sig- nal" as an increase in neural activity, but a marked de- crease might be equally likely to be detected and to pro- duce an erroneous experience.) Note that the upset could
15 logically manifest itself at a number of different places in the auditory system; for example, a set of abnormal fibers in the efferent system might produce aberrant be- haviors in several disparate locations. A popular place to hypothesize tinnitus-producing upsets is the cochlea itself. It is a complex biochemical and biomechanical system, and it is relatively easy to imagine various spe , cific malfunctions that might be associated with tinnitus Although numerous suggestions have been made over the years about possible mechanisms of tinnitus, an accurate characterization about the state of knowledge appears to be that nothing definite is yet known about the roles of biochemical or biomechanical effects operating in the cochlea or beyond to produce tinnitus. Attempts to estab- lish such relations are obviously to be encouraged. One such line of research is discussed in the following section. One point to keep in mind regarding mechanisms is that if the phenomenology of tinnitus is ever an accurate re- flection of its origins, then some of those people with narrowband or tonal tinnitus may have highly localized lesions, while some of those with broadband or complex tinnitus may have numerous or widespread sites of tine; tus origin. The former is more difficult to imagine originating from a general systemic upset than is the latter, and the latter is more difficult to imagine originating from a localized lesion or anomaly. Spontaneous Rates of Primary Fibers A recurring idea in the recent tinnitus literature is that some cases of tinnitus man be traced to an abnor . - , 1- mality in the resting (or "spontaneous") rates of firing of a localized set of primary auditory fibers. (The ab- normality may originate in these primary fibers themselves or in more peripheral cells.) The suggestion has appeared in many forms, but typically the idea is that a small set of fibers is in a state of "irritation" and as a conse- quence is firing more rapidly than normal. In order to examine the evidence pertaining to this proposal, it is necessary to briefly review the facts of spontaneous firings in normal primary fibers. The range of spontaneous firing rates is 0-100 action potentials per second in what are believed to be normal primary auditory fibers, but the distribution of rates is bimodal, not rectangular. About 25 percent of the fibers
16 (in the cat at least) have spontaneous rates below about 15 firings per second, and about 75 percent have spontane- ous rates greater than 15 per second (see, e.g., Kim and Molnar, 1979). In order to test the idea of abnormal spontaneous rates underlying tinnitus, it is necessary to administer to animals stimuli that are known to produce tinnitus in humans and to presume that the same physio- logical changes have been induced, and thus, that the ani- mals would also report tinnitus were they able. Three such animal models of tinnitus have been used in attempts to determine changes in the spontaneous rate--kanamycin- and salicylate-induced hearing loss and noise-induced hearing loss. Kiang et al. (1970) measured the spontaneous rates of primary auditory fibers following administration of kana- mycin and the consequent destruction of much of the organ of Corti in the basal end of the cochlea. In those fibers with high and middle characteristic frequencies (CF) that could still be found, Kiang et al. observed both elevated thresholds of response and nearly total elimination of spontaneous firings. This led Kiang et al. to suggest that the tinnitus heard by patients who have received ototoxic drugs, and perhaps also by those with presbycusic and high-frequency, noise-induced hearing loss, is due not to an elevation in the spontaneous rates of a small set of neurons, but rather, to "the existence of dis- tinctly different distributions of activity in tonotopi cally adjacent elements of the auditory nerve" (p. 264) m at is, the absence of spontaneous activity in those still-functional neurons at the "edge" of a cochlear lesion might lead to a sensation of sound at some higher brain location. There is some psychophysical evidence that tinnitus does occur in the transition region between normal and impaired hearing--that is, at the edge of a pattern of loss (Penner et al., 1981). . Liberman and Kiang (1978) studied the spontaneous fir- ing rates of primary auditory fibers in cats that had been exposed to intense noise bands for an hour or two. The effects differed in the different spectral regions surrounding the exposure, but in those regions where responsivity to sound was greatly reduced, the spontane- ous rates were significantly depressed. Salvi et al. (1978) exposed chinchillas to an octave band of noise of sufficient intensity to produce about 40 dB of asymptotic threshold shift and then recorded from neurons in the cochlear nucleus (not primary fibers). The postexposure spontaneous rates were lower than normal in the frequency
17 regions in which behavioral thresholds were elevated and hair-cell lesions were found. Schmiedt et al. (1980) studied primary fibers in ger- bils after administration of either kanamycin or noise exposure. In accord with Kiang et al. (1970), they re- ported a tendency toward lower-than-normal spontaneous rates in the kanamycin animals, but contrary to Salvi et al. (1978), they found a tendency toward higher-than- normal spontaneous rates in the animals exposed to noise. A curious aspect of this latter outcome is that the cells with increased spontaneous rates had low characteristic frequencies; tinnitus following noise exposure is typi- cally high in frequency. In contrast wicn Kiang et al. (1970), Dallos and Harris (1978) observed no differences in the spontaneous rates of primary fibers in chinchillas following kanamycin administration and consequent hair-cell loss. Evans et al. (1981) recorded from primary fibers in cats before and after administering salicylate and thus were able to report predrug and postdrug measures of spontaneous rates. They found essentially no change in the low spontaneous rate fibers, but the mean rate of fibers with high spontaneous levels increased signifi- cantly. Interpretation of this outcome is difficult, for the implication is that the effect did not apply differ- entially across cell CF, yet the tinnitus reported by salicylate users is typically of high pitch. Without a local, or a differential effect by CF, it is difficult to see how an "edge" could be set up. Thus, while the idea of altered spontaneous rates in primary fibers is a plausible explanation of noise- or drug-induced tinnitus, the neurophysiological evidence is still very mixed. One reason for this may be species differences in the lesions produced by noise and ototoxic drugs (Dallos and Harris, 1978). The example does clearly point out the great need for an acceptable animal model of tinnitus. Decoupling of Stereocilia The tips of the stereocilia of the outer hair cells insert into small invaginations on the underside of the Rectorial membrane; thus, as the traveling wave produced by an acoustic stimulus displaces the basilar membrane, a shear- ing action operates on the stereocilia. One possible con- sequence of noise exposure, and of exposure to certain
18 ototoxic drugs, is that the physical characteristics of the stereocilia may be altered. Tonndorf (1980) has argued that any ciliary change that alters the degree of coupling between the stereocilia and the Rectorial mem- brane will alter the inherent noise level at that inter- face. Specifically, loss of stereociliary stiffness would lead to a partial decoupling at the hair-cell/tectorial membrane interface and a consequent increase in the inher- ent noise level. Such partial decoupling over a rela- tively long section of the cochlea might produce a broad- band tinnitus, while a localized decoupling might produce a narrow-band or tonal tinnitus. Tonndorf suggested that partial decoupling may be the underlying mechanism for tinnitus (and hearing loss) in Meniere's Disease and in acute noise exposure. THE OBJECTI:VE/SUBJECTIVE I SSUE As noted previously, over the years some dramatic in- stances have been reported of loud sounds emanating from people's ears (Glanville et al., 1971; Huizing and Spoor, 1973). Since some of these passed unheard by their owners, they were not truly instances of tinnitus, and the old distinction between objective and subjective tinnitus has been further blurred by recent discoveries. Recall that, until now, the dichotomy had been based simply on the question of whether or not the examining professional could hear the subject's tinnitus, with or without the aid of a stethoscope. Obviously, the uncon- trolled variables in such a situation are many. They in- clude the intensity of the source, the amount of attenua- tion from source to receiver, the examiner's own hearing level in the frequency region of the source, and the ambient noise level in that frequency region. Stimulated by the pioneering work of Kemp (1978, 1979a,b) on what has come to be called the evoked coch lear mechanical response (ECMR) or, more colloquially, the cochlear echo, several investigators have developed techniques for inserting sensitive miniature microphones into the external auditory meatus (Wilson, 1979, 1980a; Wilson and Sutton, 1981; Zurek, 1981). Somewhat to their surprise, they have discovered that it is quite common for normal ears to emit acoustic energy at one or more frequencies. These spontaneous otoacoustic emissions (OAEs) typically are too weak to be heard by an externa 1 observer, even using a stethoscope (sealing the auditory
19 canal provides amplification of these signals that pre- vious observers did not have). In addition, most of these OAEs are also inaudible to the people who have them; thus, in most instances OAEs do not satisfy the technical definition of tinnitus--a conscious experience of a sound that originates in the head. However, in some instances a perceptual experience does correspond to an OAK, and it is those instances that most interest us here. Before discussing details, we should emphasize that examples of all three logical possibilities for tinnitus are now known to exist. First, there are perceptions of sounds that have no (as yet detected) objective counter- part and that are presumably caused by abnormal activity at some place or places in the auditory nervous system (classical subjective tinnitus). Second, there are per- ceptions that do have an objective, vibratory concomitant originating from some structure in the middle or inner ear or elsewhere in the head and neck (classical objec- tive tinnitus). m ird, there are sounds--apparently emitted by the cochlea--that are physically detectable in the ear canal but that are not audible to the people who have them (unheard OAEs). The levels of OAEs (measured in the occluded ear canal) vary from about 0-30 dB SPL, but the measured level is apparently not a good predictor of whether the OAK will be heard by its owner. According to Kemp's (1981) mea- surements, OAEs are not strictly tonal, but are noise bands, about 1.2-4.7 Hz in width (measured with a reso- lution of 0.16 Hz). Zurek (1981) and Wilson and Sutton (1981) have shown that OAEs can be suppressed by tones of other frequencies and, further, that the pattern of sup- pression shows a frequency selectivity that associates the phenomenon with cochlear activity. (Information about the time course of this suppression is provided by Kemp and Chum  and Zurek and Clark .) Thus, the narrowband emissions measured acoustically in the ear canal are generally believed to originate in the cochlea. To date, there have been reports of about a half-dozen cases of people who experience a tonal tinnitus that cor- responds to an OAK detectable in their ear canals (Kemp, 1981; Wilson and Sutton, 1981; Zurek, 1981). In contrast, easily 10 times this number of OAEs have been found that do not have a corresponding tinnitus. (Numerous subjects have had several unheard OAEs, sometimes accompanied by one that is heard.) In the most comprehensive survey to date, Tyler and Conrad-Armes (no date) studied 25 subjects with sensorineural hearing loss and tinnitus and found
20 only one OAE--and that was low frequency and did not match the subject's tinnitus. Hey also studied 20 normal- hearing subjects, and while five OAEs were found, none was audible to its owner. Zurek (1981) tested nearly three dozen people covering a wide range of age and auditory conditions; more than 20 OAEs were found, but none cor- responded to a reported tinnitus. In addition, Zurek (personal communication) studied 16 people who complained of tinnitus of various sorts and found no emissions whatsoever. In contrast to the negative results of these surveys, Wilson and Sutton (1981) did find an OAK corresponding to a reported tinnitus in 4 of 10 people studied; other, unheard OAEs were also present in these ears. Most of these people had apparently responded to a newspaper ad soliciting tinnitus sufferers, but it appears that none was afflicted with a particularly severe form. The 4 People having an OAK corresponding to a conscious exper i _ ence may not have been actual surrerers-- so muon an n"-- urally careful introspectors, and perhaps with practice others might also come to hear their normally unheard OAEs. In those four subjects for whom an OAK did correspond ~ . . , . ~ ~ . ~ ~ _ _ to a conscious experience, Wilson ana Button `'Yo' J c~emon- strated in a number of ways that the two phenomena are related, but not in all of the following ways for all sub- jects. Contralateral pitch matching has shown a good cor- respondence between the tinnitus pitch and the OAK fre- quency, the tinnitus and the OAK have both been cancelled by precise adjustment of the level and phase of an exter- nal tone, and changes in air pressure introduced into the outer-ear canal have shifted the frequency of the OAK up- ward and produced a corresponding increase in the pitch of the tinnitus. Elsewhere, Wilson (1980a) reported that the tinnitus and the OAK have been observed to undergo simultaneous fluctuations in magnitude. Further, heard OAEs are typically not described as tonal, but as rough or noisy (Wilson and Sutton, 1981; Zurek, 1981), in apparent accord with Kemp's (1981) demonstration that physically OAEs are noise bands, 1.2-4.7 Hz in width. Kemp (1981) indicates that stimulation with sounds intense enough to produce temporary threshold shift (TTS) _ ~ _ _ is, can produce diminutions in OAK magnitude that can last for several minutes, but it is not clear that the OAEs studied had a concomitant tinnitus that behaved simi- larly. Apparently not yet studied is whether an OAK and its associated tinnitus both undergo similar patterns of
21 residual inhibition (see Residual Inhibitions" in Chapter 4). Most OAEs detected to date have been in the low- to mid-frequency region, but this is surely due in large part to the high-frequency attenuation characteristic of the middle ear. Any conductive hearing loss would offer a similar impediment to detection of OAEs. Reported fail- ures to detect an OAK in association with an existing tin- nitus (e.g., Tyler and Conrad-Armes, no date) must always be evaluated with this present technical limitation in mind. That is, the dearth of OAEs detected above about 4000 Hz--where much reported tinnitus lies--should not yet be taken as evidence against OAEs being associated with high-frequency tinnitus. To date, few OAEs have been found in nonhumans (see CIBA Foundation, 1981:133; Zurek and Clark, 1981). Decker and Fritsch (1982) described a dog that emitted a continu- ous narrow band of noise centered at about 10.3 kHz; since it was a pet, this animal was not fully studied physio- logically, but the description suggests an OAK. Evans et al. (1981) did study a single guinea pig with a single OAK and found that: (1) the OAK was not affected by paralysis of the middle-ear muscles; (2) it was raised in frequency by both increases and decreases of air pressure in the middle ear; (3) its acoustic level was reduced, but its round window magnitude was unaltered by changes in middle- ear pressure; (4) it could be suppressed by introducing external tones of the appropriate frequency and intensity; (5) it was abolished under hypoxia and returned following restoration of a normal oxygen supply; and (6) it was basically unchanged by sectioning of the tendons of the middle-ear muscles. Of great interest would be animal experiments on OAEs using some of the drugs that have been shown to be effective against tinnitus in humans (see drug m erapy for Tinnitus" in Chapter 4). Also of interest is whether and how OAEs might be produced in animal ears so that experimental study could proceed. Zurek and Clark (1981) were able to induce OAEs in only 2 of 17 chinchillas exposed to intense noise. Much speculation surrounds the relationship between OAEs and the so-called cochlear echoes discovered by Kemp (1978). mese echoes are acoustic energy detectable (using a microphone sealed in the outer-ear canal) several milliseconds following the presentation of a click or brief tone burst. m e two effects may prove to originate from the same cochlear structure(s)--the echo being a normal consequence of cochlear design and the OAK being
22 the result of a local anomaly in the relevant structures. Whatever the eventual relationship of these two phenomena, the newcomer should be aware of their common historical · ~ OrlglnS e Heard and unheard OAEs can be found in pathological ears of various sorts as well as in normal ears. It is as yet unclear whether an OAK of either type can exist in a spectral region showing substantial sensorineural hear- ing loss (an interesting point theoretically), but belief seems to be running against it. Rutten (1980) has shown that cochleae echoes can exist at low frequencies even when there is considerable hearing loss for frequencies above about 2000 Hz. There is a final fact about OAEs that--while yet to be tied to tinnitus directly--is worthy of mention. Several investigators (Kemp, 1979a; Wilson, 1980a) have noted the relationship between OAEs and what has come to be known as the microstructure of the audiogram. Elliott (1958), Comas (1975), Cohen (1982), and Kemp (1979a) have all shown that a person's audiogram can have numerous local inversions that are highly stable across long periods of time and that are predictable from certain acoustical measures of cochlear behavior (Kemp, 1979a). mese peaks and troughs have been discussed as being normal conse- quences of the same mechanisms responsible for OAEs and cochlear echoes. Wilson (1980a) has shown that OAEs (heard and unheard) do not invariably occur at peaks or troughs in the audiogram, but interestingly, Glanville et al. (1971) and Huizing and Spoor (1973) did observe cir- cumscribed regions of hearing loss in the region of tonal emission, and Flottorp (1953) believed that his idiotones always resided in a region of localized hearing loss. Also, Minton (1923) reported diminished sensitivity in the frequency region of a subjective tinnitus. To summarize this topic, many--and perhaps the majority of--normal ears "spontaneously" emit continuous, narrow- band acoustic signals that appear to originate in the cochlea. These signals are low in level, and the vast majority exist unheard by those who emit them. The signals that are heard fit the standard definition of objective tinnitus, but it appears that few, if any, of these cases constitute a problem tinnitus. Several attempts have failed to find concomitant emissions in the ears of people who could truly be regarded as tinnitus sufferers, as opposed to careful introspectors. Further, OAEs do not share certain properties with the common forms of tinnitus (see Some Ways Tinnitus Is Not Like an Ex
23 ternal Sounds in Chapter 3). While it is disappointing to many who had momentarily hoped that some forms of tinnitus would prove to be objective and more easily studied (and perhaps more easily treated) than in the past, it appears that OAEs will not prove to be respon- sible for many instances of severe, problem tinnitus. Students of hearing interested in the general topic of otoacoustic emissions are referred to the visionary paper by Gold (1947/1948). A broader review of otoacoustic emissions can be found in McFadden and Wightman (1983). CAN TINNITUS EXIST IN THE ABSENCE OF HEARING LOSS? It is common to see the assertion that most, but not all, sufferers from tinnitus have some hearing loss (85 per- cent according to Vernon et al., 1980; also see CIBA Foundation, 1981:29). Tinnitus without hearing loss is definitely a possibility--for example, when the cause of the tinnitus is a vascular anomaly--but when considering this issue, it is important to note two points. First, hearing sensitivity is not typically measured above 8 kHz--indeed, on most commercial audiometers it cannot be measured above that frequency. Since hearing loss of various types often proceeds from high frequencies to low, it is very possible that some of those tinnitus sufferers thought to have normal hearing have in fact lost some hearing, but (so far) only in the untested region above 8 kHz. This is obviously consistent with the commonly reported observation that tinnitus is a frequent harbinger of hearing loss. Until it is demon- strated to be incorrect, it appears parsimonious to believe--particularly when the tinnitus is of very high frequency--that there Is an accompanying hearing loss, although it may exist beyond the normal audiometric range. Support for this idea can be found in the data of Jacobson et al. (1969). Second, pure-tone audiometry is typically done only at a set of standardized frequencies spaced at octave inter- vals. m us, localized regions of hearing loss related to the tinnitus--above, below, or at the tinnitus frequency-- might be missed by this sampling procedure (see Kemp, 1979a; Wilson, 1980). m e point is that even modern audiometers are poor research tools, and, when it comes to tinnitus research, a fixed-frequency audiometer is particularly inadequate.
24 TINNITUS IN CHILDREN Two reports exist on the prevalence of tinnitus in chil dren. Nodar (1972) asked a set of tinnitus questions during routine auditory screening of students in grades 5-12 (approximately ages 10-18). Of the approximately 6,000 students passing the screening . about 13 percent reported having tinnitus. - ~ , , ~ ~ _ ~ m is low percentage may re- flect an ambiguity in the relevant question--it may have been unclear whether the question pertained to ever or to _ that moment. Of those not passing the screening, about 59 percent reported tinnitus. ___ Alarms The two groups, the most common characterizations of the tinnitus were "high" and "ringing." ~ : ~ ~ _ group, it is not surprising that the tinnitus-inducing · . . ~ since this was prlmar Ply a normal-nearing . events mentioner were the same as are reported by normal- hea~ing adults--incidents of noise exposure, illness, stress, etc. From such evidence we might tentatively con- clude that young ears are about as prone to short-term tinnitus as older ears, but that children have been less likely to report it. More interesting perhaps are the observations of Graham (1981b), who questioned 158 partially and pro- foundly deaf students (ages 12-18) about tinnitus and found that about half had it on occasion. m is is re- ported as being a great surprise to the hearing profes- sionals who work with these children. Graham estimates that about two-thirds of the tinnitus sufferers had at least one episode a week. A curious finding was that the , tinnitus was much more likely to be localized toward the better-hearing ear. m e discussion following Graham's paper is recommended to those interested in the topic of tinnitus in children (CIBA Foundation, 1981:182-192). POSSIBLE EXPERIMENTAL MODELS OF TINNITUS A well-established research strategy in the medical sci- ences is to develop procedures for inducing the malady of interest in weakened or reversible form in otherwise healthy humans or in species other than man. Therapies that are successful on these "models of the malady are then evaluated on volunteers actually suffering from the condition. The goal of developing experimental models of this sort is often not just the discovery of an effective treatment for ths malady of interest, but also insight into the basic physiological or neurophysiological mech
25 anisms involved. Experimental models of tinnitus appear possible, but little research has yet been done using them. Tinnitus can be reversibly induced in otherwise normal ears in several ways--for example, administration of cer- tain drugs, exposure to intense sounds, and exposure to waveforms with steep spectral skirts. This fact is im- portant for several reasons: (1) it offers at least the possibility of developing procedures to study tinnitus in animals, with all of the usual gains in flexibility over human research, (2) studying (induced) tinnitus in normal human ears would allow certain within-subject controls that may prove valuable and would be otherwise not achiev- able, and (3) there is the potential for insight into the mechanisms underlying some forms of tinnitus. With very few exceptions, however, induced tinnitus in normal ears has yet to be studied as a model of tinnitus in patho- logical ears. Two of the exceptions--Loeb and Smith (1967) and Atherly et al. (1968)--induced a short-term tinnitus through exposure to intense sound. One generalization emerges from these experiments: the induced tinnitus does not occur at the same frequency as the maximal temporary threshold shift (TTS), and where it does occur depends upon the nature of the exposure sound. -- ~ Following exposure to octave-band noise, the tinnitus frequency was well below the maximum TTS frequency; with one-third octave bands, the tinnitus frequency continued to be lower than the maximum TTS frequency, but the difference was smaller than with octave bands; and with tonal exposure stimuli, the tinnitus frequency was higher than the maximum TTS frequency. From personal experience it is clear that monaural exposure to an intense sound can produce tinni- tus in both ears. The qualities are typically different, with the experience in the exposed ear being more broad- band and temporally complex, and the experience in the contralateral ear being more tonal. Thielgaard (1951) and Thompson and Gales (1961) mention this contralateral effect, but to our knowledge it has never been widely known or systematically studied. Its existence might be evidence that the efferent system is involved in tinnitus production. Another procedure for inducing a short-term tinnitus involves the use of noise bands having steep spectral skirts. In our experience, listening for a few minutes to a relatively weak (35-50-dB spectrum level) noise band having skirts of about 400-500 dB per octave can produce
26 a high-pitched tonal tinnitus that will last several min- utes (see McFadden and Plattsmier, 1982a). Lummis and Guttman (1972) used similar waveforms and found that the induced pitch was matched to a frequency about two-thirds octave higher than the edge of a low-pass noise band and about three-fourths octave lower than the edge of a high-pass noise band. Finally, the tinnitus induced by aspirin and other common drugs also offers research opportunities not yet fully explored (see ~Salicylates" in Chapter 4). For example, comparison of the onset and recovery times of the tinnitus and the hearing loss induced by aspirin could be revealing about the mechanics underlying both. Evans et al. (1981) did study the effects of salicylate and lidocaine (see "Lidocaine" in Chapter 4) on the response properties of orimarv auditory neurons i n An _, ~ attempt to find correlates to tinnitus, but ignorance about whether tinnitus was actually present and, if so, about its characteristics makes interpretation of their findings difficult. All of the above procedures for inducing tinnitus could be utilized in connection with the ear-canal monitoring systems (mentioned in "m e Objective/ Subjective Issue" in this chapter) to examine whether the induced subjective experience has an objective counter- part and, if so, whether their spectral, temporal, and other characteristics are in accord. Zurek and Clark (1981) have taken the first step in this direction by inducing otoacoustic emissions (OAEs) in chinchillas by exposing them to noise. Evans et al. (1981) monitored a spontaneous OAK in a guinea pig while performing various psychophysical and physiological manipulations. Kemp (1982) noted some changes in the human cochlear echo (see The Objective/Subjective Issue") immediately following exposure to intense sounds, and he argued that these may be related to the postexposure experience of tinnitus. Three attempts to utilize the brain-stem-evoked re- sponse (BSER) with tinnitus have been reported (Berlin and Shearer, 1981; Dickter et al., 1981; Shulman and Seitz, 1981). Shulman and Seitz believe that people with tinnitus of central origin have BSERs different from those of people with normal hearing, but the effect requires confirmation. Dickter et al. demonstrate that patients with very similar audiometric configurations and tinnitus complaints can yield very different BSER data.
27 MENIERE'S DISEASE The tinnitus associated with Meniere's Disease is being discussed in a separate section because it is more homo- geneous and predictable in its character than is tinnitus associated with other disorders and because much is known about the constellation of physiological changes present in this disorder. Thus, the opportunities for isolating the origin(s) of this tinnitus appear great. True Meniere's Disease consists of three primary symptoms--episodic vertigo, hearing loss, and tinnitus (see Barber et al., 1972). me onset of vertigo is fre- quently sudden and unanticipated. m e episodes may last minutes or hours, but aftereffects often persist for sev- eral days. Severe attacks seem to alternate with milder ones in an unpredictable manner. The involvement is typi- cally unilateral, particularly in the early stages of the disorder; estimates of bilateral Meniere's Disease range from 10 percent to 40 percent, although these may include some instances of other disorders that mimic true Meniere's Disease. W. F. House (1975) indicates that the disorder is more common in whites than in blacks (also see CIBA Foundation, 1981:28), is associated with indus- trialization and urbanization, and is believed to be absent in species other than humans. m e hearing loss experienced during an episode of Meniere's Disease is typically greater, if not exclu- sively, in the low-frequency region. Typically, the loss is unilateral, present in both air-conduction and bone- conduction tests, and accompanied by a feeling of full- ness or pressure in the affected ear. The hearing loss is typically between 15 and 30 dB; it is usually described as fluctuating, since it varies with time and with the episodes of vertigo; and over repeated episodes there is typically a gradual buildup of permanent hearing loss. Recruitment is typically present, as is a diplacusis in the direction of raised pitch in the affected ear. Thus, the general pattern is much like that seen with a cochlear pathology. House (1975) indicates that the typical time course for the various symptoms is as follows: there is first an awareness of fullness or pressure in the ear and a buildup of tinnitus. men, hearing loss develops, and finally, vertigo sets in. Recovery involves first a dimi- nution in vertigo, then reduction in the feeling of full- ness and the tinnitus, and, finally, a subjective return of hearing to normal.
28 In accord with the primary effects on hearing being in the low-frequency region, the tinnitus associated with an episode of Meniere's Disease is also typically low fre- quencY in quality. Common descriptions are "roaring" and _ Objectively, Nodar and Graham (1965) found that all of a sample of 11 patients with confirmed Meniere's Disease matched their tinnitus to tones lower than 1000 Hz; the median was 320 Hz. Note that this is not to say that the tinnitus was tonal in character. Zurek (1981) could find no otoacoustic emission (see "The Objective/Subjective Issue" in this chapter) correspond- ing to the tinnitus heard by the one Meniere's sufferer he tested. According to Goodhill (1979), current belief is that the primary cause of Meniere's Disease is an anatomical/ physiological anomaly. The consequence is a dysfunction _ "buzzing. n The consequence is a of the endolymphatic duct and sac system that somehow produces an excess accumulation (hydrous) of the endo lymph. used synonymously with Meniere's Disease.) ~ _ (Endolymphatic hydrops has in fact come to be In addition, certain other factors appear to be able to potentiate episodes of the disorder--for example, certain vascular conditions, allergies, endocrine deficiencies, and per- sonality variables. The primary histological finding is of a distension of scale media, only occasionally accom- panied by degeneration of the organ of Corti. More recently, anomalies of the vestibular aqueduct have also been reported. Dozens of treatments of Meniere's Disease have been advanced (and criticized) over the years. Surgical treatments have ranged from the extremes of labyrinth- ectomy and vestibular neurectomy to more selective (and less destructive) procedures aimed at the endolymphatic sac. Arenberg and Bayer (1977) have recently argued that in the past many decisions for surgical intervention were . . . . . . . · . . based too heavily on the vertigo symptoms, with too little attention paid to the consequences for hearing. They claim that a large number of initially unilateral cases eventually come to be bilateral, and, thus, that the more destructive surgical procedures should be used sparingly. Arenberg and Bayer believe that recent modifications of the original Portmann procedure for decompressing the endolymphatic sac have yielded good-to-excellent results against vertigo as well as against the symptoms of tin- nitus and pressure, without adverse effects on hearing, either immediately or in the long term. Emus, following Fisch (1976), they suggest that surgery of the endo
29 lymphatic sac be accomplished as early as possible in the course of the disease--during the period when the hearing losses are fluctuating and reversible--and that vestibular neurectomy and excision of Scarpa's ganglion be done in the latter stages of the disease. In this way, existing hearing and long-term potential for hearing are maximally protected in both groups. As an example of the diversity of procedures suggested for treating Meniere's Disease, consider Johnson (1954), who advocated various actions against the sympathetic ner- vous system. He believed in blocking the stellate gang- lion (located along the sympathetic trunk at the level of the seventh cervical vertebrae) with an injection of pro- caine hydrochloride as early in a Meniere's episode as possible, and, if there were several recurrences, he felt that a partial dorsal sympathectomy was advised. He also performed stellate blocks on sufferers from tinnitus of other origin, and when the block was successful, he did dorsal sympathectomies on them as well. (Blocks were apparently ineffective against tinnitus arising from noise exposure.) mese procedures were never widely adopted. The second major form of treatment for Meniere's Disease is drug therapy. Over the decades, this has ranged from simple reduction of sodium in the diet to the use of diuretics, vasodilators, histamines, antihista- mines, tranquilizers, etc. (for a review, see Arenberg and Bayer, 1977). m is research has often been lacking in adequate controls, but some recent controlled research makes betahistine hydrochloride (Serc) appear highly effective against the vertigo, hearing loss, and tinnitus of Meniere's Disease (Frew and Menon, 1976). Glycerin does have the ability to produce dramatic short-term improvements in the hearing of some Meniere's patients and, in the process, to eliminate the tinnitus, but this substance is used primarily in diagnosis, not treatment (see Snyder, 1974; Klockhoff, 1975). Arenberg and Bayer (1977) conclude that no drug has yet proven to be satis- factory in its general ability to alter the natural course of the disease process, and--as previously indicated-- thev favor surgical intervention to accomplish this end. . . . - · . · · · ~ As noted above, histology reveals a distension or scale media due to hydrops of the endolymph in sufferers from Meniere's Disease of long standing. m is distension is primarily at the apical end of the cochlea, where the basilar membrane is most compliant. (Interestingly, it is widely believed that the hydrops appears not to be due
30 to an overproduction of endolymph, but rather to altered absorption characteristics in the strict vascularis that lead to fluid accumulation.) The twin facts that the loss of hearing is greatest at the low frequencies and that the tinnitus is of low frequency are, of course, in accord with this apical locus of disturbance. Exactly which consequence of the fluid accumulation and disten- sion is responsible for the hearing loss and tinnitus produced by Meniere's Disease has been the object of some attention. An excess of fluid pressure could produce changes in the mechanics and micromechanics of the coch- lear duct as well as possibly cause some compression of certain critical structures--such as the Rectorial mem- brane, the hair cells, or the nerve fibers themselves-- thus causing stimulation in the acoustic chain. An argu- ment sometimes raised against the possibility of direct compression/stimulation is that histology reveals little or no hair-cell or other damage at the apical end of the cochlea. This objection may not be cogent, for there is no apparent need for the compression/stimulation to be so great as to damage the cells, just great enough to acti- vate them. Second, the common visual inspection tech- niques may overlook subtle effects. . Tonndorf (1957) has given examples of changes in basilar-membrane mechanics that can occur with an endolymphatic overpressure; many of these are in accord with the symptoms of Meniere's Disease, but unfortunately no direct inference to tinni- tus is possible. Recently, Tonndorf (1980) has argued that the assumption of diminished coupling between the hair cells and the Rectorial membrane, due to stereo- ciliary dysfunction, might account for a number of com- monly observed auditory anomalies, including sensori- neural-type hearing loss and tinnitus. Electrophysio- logical recordings from animals in which endolymphatic hydrops has been experimentally induced (see, for exam- ple, Kimura, 1968) might shed light on the origin(s) of tinnitus in Meniere's Disease and other disorders as well. Vernon et al. (1980) have reported that Meniere's sufferers constitute only 1 percent of all the cases they see at their Oregon tinnitus clinic. There are various possible explanations of this peculiar fact, but the field is narrowed by a second fact. The tinnitus des- cribed by these Meniere's sufferers is not the low- frequency roaring that tvoicallv accompanies acute eoi ~we- ~_ ~_ ,^ ~ ~ ~, ~- ._.., it is middle- or high-pitched and frequently tonal in quality. The implication is that these patients were seeking relief from a tinnitus that
31 was either not, or was only indirectly, related to the Meniere's Disease. As it turns out, the tinnitus com- plained about was easily masked, and a hearing aid or tinnitus masker/instrument typically provided acceptable relief. It is unclear whether the characteristic roaring tinnitus of Meniere's Disease is responsive to masking. me data may not be easy to get, for a person in the throes of a Meniere's episode is typically in great distress and, thus, is not an ideal subject for psycho- physical experimentation.