Reproduced with permission from
Mechanisms of Tinnitus, Chapter 11
Jack A. Vernon, Aage R. Møller editors
Allyn and Bacon, 1995; ISBN 0-205-14083-1

A MODEL FOR COCHLEAR ORIGIN OF
SUBJECTIVE TINNITUS: EXCITATORY DRIFT IN
OPERATING POINT OF INNER HAIR CELLS

 

Eric L. LePage, Ph.D.
Hearing Conservation Research Unit
National Acoustic Laboratories
126 Greville Street, Chatswood, N.S.W. 2067
AUSTRALIA

INTRODUCTION

 

There are many purported sites for the origin of tinnitus. However, for the most part, the incidence of tinnitus is so strongly connected with the onset of sensorineural hearing loss (whether it be permanent, or temporary due to acoustic overexposure) as to implicate the cochlea as the key site. The objects of this chapter include

1. to set down an itemised list of salient properties of tinnitus of cochlear origin as reported in the literature;
2. to set down a further itemised characterisation of single relevant case study, like that of Wegel (1931) the only case of which the author has intimate experience;
3. to produce a descriptive treatment of the physiological regulation of tectorial membrane position along the lines suggested by Davis (1954);
4. to show that this model constitutes a testable hypothesis for explaining the origin of the listed items in terms of a phantom input to the cochlea (Jastreboff, 1990);
5. to observe that the model explains the difficulty of characterising tinnitus in terms of equivalent external sounds;
6. the issue of severity is strongly dependent not just on masking influences but also on selective attention and this accounts for the heavy psychological involvement and difficulty of treatment in severe cases; and
7. much of the difficulty in obtaining a conceptual purchase on the causes of tinnitus stems from the long held but questionable assumption that hearing sensitivity provides a sensitive measure of the extent of cochlear damage.

RELEVANT CHARACTERISTICS PREVIOUSLY REPORTED

Repeated studies have provided an invaluable characterisation of the facets of tinnitus and these relate to basic issues in cochlear physiology and noise-induced hearing loss. A non- exhaustive review of relevant literature provides some key characteristics:

Nature of sensations:

1. The phenomenon can be categorised as continuous or discontinuous (transient).
2. Continuous tinnitus (CT), at least in cases of high frequency hearing loss, or Continuous High-Frequency Tinnitus (CHFT), is of a predominantly high frequency nature (80% over 3kHz, 55% over 5kHz, 42% over 7kHz, Meikle and Taylor-Walsh, 1984), but can be of a very low frequency (<100Hz) (CLFT) (Wilson, 1979).
3. CHFT is heard as narrow bands of noise, likened to the sounds of cicadas; the most debilitating seem to be like externally applied intense sinusoids (Reed, 1960).
4. CT can also be like “water rushing”, “sand blasting” (wide band) or in the case of Menière’s syndrome “roaring” and of low frequency (Nodar and Graham, 1965).
5. CT can be uniform or pulsating (Harris, et al., 1979). In cases of otosclerosis it is frequently dynamic in character, like bell pealing, hooting noises causing great difficulty obtaining an auditory assessment.
6. Transient “spontaneous” tinnitus (TST) is invariably unilateral and frequently preceded by a transient non-audible pressure-like sensation within the affected ear accompanied by dulled hearing acuity (Vernon et al., 1978).
7. Subjective judgements of tinnitus pitch are highly variable (Harris, et al., 1979; Penner, 1983; Burns, 1984).

Loudness and severity:

1. The loudness of tinnitus is variable and also depends on ambient noise-level and becomes louder with decrease in ambient noise-level (Axelsson and Ringdahl, 1989).
2. There exists an inherent problem (in common with pain) of assessing severity (Meikle and Taylor-Walsh, 1984; Tyler et al., 1992; Burns, 1984);
3. While on a scale of 1 to 10, most patients rate the severity of their tinnitus between 5 and 8, 76% percent of patients match the loudness of their tinnitus at <6dB SL (Meikle, 1991c).
4. Severe cases require higher masking levels; it often requires sound levels approaching 100 dB to mask the tinnitus (Vernon, 1991).
5. Patients “fix” upon their tinnitus (Hazell, 1979a); i.e. selective attention or focussing is evidently an important component in assessing disability,

Physical/Physiological attributes:

1. The production of beats between an external tone and CT is very rare, even when the tinnitus is tonal (Vernon, 1991)
2. Fowler (1944) describes tinnitus as objective or subjective (“nonvibratory”).
3. Subjective tinnitus most often has no associated spontaneous otoacoustic emission (Penner, 1992; Shulman et al., 1992).
4. Ears with tinnitus have much larger cochlear microphonics than unaffected ears (Hazell, 1979c)
5. There is a distinction to be made between masking and suppression (Hazell) which may be electrically induced (Hazell, 1979c).
6. Electrical suppression of tinnitus is possible with positive but not negative current produced auditory sensations (Grapengiesser, 1801 see Vernon, 1991; Portmann et al., 1979; Aran, 1981).

Possible causative factors:

1. Purported factors in etiology of tinnitus include: medications including aspirin (Jastreboff et al., 1988; Day et al., 1989) oral contraceptives, changes in dietary (caffeine and cholesterol), noise-exposure (Reed, 1960; Man and Naggan, 1981; Coles et al., 1981; Chung et al., 1984; Axelsson and Barrenäs, 1992), diabetes and cardiovascular status (Meikle, 1991).
2. May be due to steady mechanical pressure producing a slight but permanent displacement of the tectorial membrane in relation to the hair cells (Davis, 1954) and there is a fairly strong connection with this type of origin in Menière’s syndrome (Nodar and Graham, 1965) since in Menière attacks a pressure build up is well documented histologically.
3. Barotrauma can produce a change in tinnitus (Farmer, 1977;Tonkin and Fagan, 1975).
4. Tinnitus can be manipulated by a wide variety of pharmacological agents known to have a direct effect on the cochlea, e.g. glycerol, ethacrynic acid etc.

Relation to hearing loss, incidence:

1. Like presbycusis, tinnitus cannot be regarded as an independent disease but a manifestation of accumulative damage or aging, predominantly affecting the 40 to 70 year age group (Meikle et al., 1991; Coles et al., 1981).
2. Sex distribution for tinnitus is about the same (Reed, 1960; Bailey, 1979); by contrast a higher incidence is seen in males, 71 percent, females 29 percent (Meikle, 1991a).
3. Laterality: It appears more commonly in the left ear (46%) than the right (29%) (Hazell, 1981). “Head” tinnitus apparently constitutes the most severe form.
4. Although CHFT is strongly associated with the onset of hearing loss (Loeb and Smith, 1967; Coles et al., 1981), particularly noise-induced hearing loss (NIHL), there is a poor correlation with the extent of the hearing loss.
   1. There may be no associated hearing loss; Hazell et al., (1979b) report a high incidence of tinnitus in people with normal or near normal hearing and,
   2. It is frequently associated with mild to moderate hearing losses
   3. It can remain following sectioning the auditory nerve to eliminate it, analogous to phantom pain from a severed limb.
5. CT seldom impairs hearing threshold by more than 10dB (Vernon et al., 1980).
6. CT is regarded like a rise in noise level (Tonndorf, 1981) by up to 55dB. Patients frequently complain that tinnitus masks their normal hearing; if they could be rid of the tinnitus they feel they could hear normally ( Meikle and Taylor-Walsh, 1984; Meikle, 1991b; Meikle et al., 1991).

Masking of tinnitus:

1. CT can be masked as can external sounds (Hazell, 1979c; Vernon et al., 1978, 1980).
2. Masking conditions are not predictable; masking is not always possible (ibid);
3. Masking of CHFT is most successful when the hearing loss is minimal (ibid).
4. It is easier to mask a relatively quiet CT (ibid).
5. Tinnitus falls into three categories according to the ease with which it may be masked, 18% easy to mask with tones; 52% can be masked at low sensation levels but only by tones near in frequency to the tinnitus; 22% where excessive levels are needed, and 8% other (ibid).
6. Broadband noise provides effective masking for just under half of those with CHFT (Meikle, 1991).
7. Tinnitus is slightly better masked when in the right ear (Hazell, Williams and Sheldrake, 1979).
8. Masking of tinnitus differs from masking external sounds (Vernon et al., 1980):
   1. an external noise can more easily mask an internal noise-type tinnitus than it can mask another external sound;
   2. the usual upward spread of masking for external sounds below 11kHz is not the case for tinnitus;
   3. The sensation level of a tone required to mask tinnitus is not necessarily related to the perceived loudness of the tinnitus;
   4. unlike external sounds which are re-instated once the masking noise is removed, release of the masking sound is often followed by temporary cessation or reduction of the tinnitus (residual inhibition, RI).
9. More likely to get residual inhibition with CHFT (Vernon et al., 1978)

 

SOME NEW OBSERVATIONS — ONE CASE HISTORY

In an area of sensation, like pain, which tends to defy objective quantification, we tend to find patients’ descriptions highly non-specific and limited to comparison with everyday sounds even though the description may alter substantially when presented with actual sounds which they give as examples. In the author’s experience it is quite rare to find a tinnitus patient who can express the sensory experience of their tinnitus in more precise terms by using the languages of science, engineering or music. To try to bridge this shortfall, the following are a series of personal observations by a single tinnitus sufferer (ELP), a male aged 47, with a background in relevant disciplines such as mathematics, physics, engineering plus 20 years researching auditory physiology. More relevant here is his educated ear for analysing musical sounds, his early training as a concert pianist and organist and his abiding interest in musical acoustics (which includes 32 years’ experience tuning pianos). He is experienced at a variety of psychophysical discrimination tasks such as distinguishing the difference between place pitch and repetition pitch, and estimating relative levels of noise and coherence in various frequency bands. He has contributed to experiments on “absolute pitch” (Siegel, 1976; i.e. he can effortlessly name a note, plus octave, over the musical pitch range without external reference), and hence can determine an equivalent frequency (±0.05 octaves) for external and phantom sounds. Figure 1 shows for each ear his pure tone airconduction audiograms plus click-evoked otoacoustic emission spectra obtained using the Otodynamics ILO88 Analyser used in standard screening mode. The tinnitus of this observer has obviously not been as debilitating as that of some clients, but it can often be very troublesome, either in terms of its loudness at low-level ambient noise conditions, or its interference with perception of low level sounds.

FIGURE 1 Audiometric and objective measures of ear function of tinnitus subject ELP. The top two panels show pure tone audiograms for right and left ears respectively. The lower two panels show the corresponding click-evoked otoacoustic emission spectra obtained using the Otodynamics ILO88 analyser in standard screening mode (80 µs, 80dB SPL clicks, average of 260 presentations of the nonlinear click sequence of 4 reversing clicks). The black regions indicate the cross power spectrum for coherent part of the response while the grey regions are for the incoherent part of the response (a measure of external and internal noise sources). The subjective measures indicate mild low and high frequency hearing losses, in contrast to the inferred substantial scattered OHC damage indicated by the objective results.

The following list itemises the prime features of his tinnitus, in order of appearance summarised in Figure 2A.

Transient Spontaneous Tinnitus (TST) – referred to elsewhere as “Ping”

 

1. TST has occurred since early teens; long before any manifestation of continuous tinnitus, it occurred only with a single pitch corresponding to a frequency in the range never below 1.0 kHz and seldom above 4 kHz.
2. The incidence has been episodic, maybe 20 times per week during periods of frequent music exposure or great tiredness, to less than once per week at other times; total number of occasions estimated to be many thousands.
3. Incidence of TST is estimated to be approximately equally divided between ears, never both ears simultaneously.
4. TST was mostly preceded (and often followed) by feelings of dullness (reduction in auditory sensitivity) and fullness (sensation of pressure) in the same ear. This lasted no more than a few seconds. This sensation is vaguely similar to the change in pressure equalization during ascent or descent of an airplane. It is followed by the TST rising to a peak of around 15 to 20 dB SL, typically in 1 to 2 s.
5. The incidence of fluctuant sensations of pressure is far more common than the incidence of TST itself.
6. TST decreased steadily in loudness lasting various lengths of time ranging from 5s for minor disturbances, to 45s for loud cases occurring in very quiet ambient levels suggesting that the decay of the excitation may be exponential-like in nature. This possibility is depicted in Figure 3a showing sensation level (dB) as a function of time (s). The abscissa is the level above which the disturbance is audible.
7. TST may possess harmonic distortion; the percept is that of an external pure tone with a second harmonic discernible at low background noise levels.
8. Only rarely is TST associated with a specific incidence of overexposure, but high level TST (>80dB SL) can be evoked by a brief, high level impulse and is accompanied by an equally strong pressure sensation or even pain.
9. TST is not easily masked by whistling even though a pitch match is readily generated.

Continuous Low-Frequency Tinnitus (CLFT)

1. CLFT first presented after age 40 only in the right ear and is ever present in conditions of low ambient noise, such as in the quiet of the night.
2. It normally has a sensation level of 5 to 10 dB or even up to 20 dB in an anechoic chamber.
3. The pitch is scarcely discernible; like a broadband rumble centred around 50Hz. If several pitches are present covering the band, there are no obvious beats. There are no upper harmonics or partials.
4. Masking is readily and rapidly produced providing the masker has a frequency less than 300 Hz, i.e. there is a downward spread of masking for low frequencies. Lower frequencies are more effective maskers, e.g. humming. The duration of residual inhibition is a strong function of frequency, extending up to 5 seconds for very low frequencies. A sound of similar sensation level to the tinnitus can readily mask the CLFT. The sensation often has a throbbing quality if the ambient level varies around the masking level, e.g. a car passing down the street outside; even continuous head movement can suppress the CLFT.
5. The period of residual inhibition (RI) depends directly on the strength of the masker in a way which suggests that the perturbation recovery function is exponential-like. This possibility is depicted in Figure 2D showing sensation level (dB) as a function of time (s). Again, the abscissa is the level above which the disturbance is audible, so that the CLFT is always audible except when masked, after which recovery occurs.
6. Experiments have failed to reveal any objective counterpart to the CLFT, i.e. a spontaneous emission at very low frequencies which can be masked as above.
7. Sudden counter-clockwise rotation of the head accentuates the CLFT; clockwise rotation masks the CLFT very briefly; this phenomenon occurs irrespective the orientation of the head.

Continuous High-Frequency Tinnitus (CHFT)

1. CHFT occurs at frequencies above 8 kHz in both ears simultaneously; there is no musical note value; the pitch is too high; the percept is central. This sounds like a thin tone or set of tones like once was heard emanating from the 10 kHz line-oscillator on a television set. However, unlike external sounds of that frequency which are highly directional, the location of CHFT within the head cannot be controlled or modulated by rotating the head.
2. The sensation level is highly dependent on:
   1. The ambient background noise level, and
   2. Attention paid to it. In low-noise conditions especially, the sensation level can be made to increase substantially over a period of seconds to a minute or more, simply by focussing on the high frequency “mixture”, whereupon a single frequency emerges as dominant. This, however, tends to become more unilateral in origin.
3. Focussing on this single frequency sensation for minutes can cause the sensation level to rise still further with a marked increase in the coherence of the percept.
4. Relief is achieved more by diverting attention away from the sensation more so than by actively trying to ignore it.

Transient High-Frequency Spontaneous Tinnitus (THFST).

1. Most recent in appearance is TST at discrete frequencies above 6 kHz at relatively low sensation levels. No harmonics are discernible at these frequencies.

FIGURE 2 Relevant features of the subjective tinnitus of ELP. This panel and the left below show two diverse representations of the pitch of the tinnitus of the continuous high frequency tinnitus CHFT, continuous low frequency tinnitus (CLFT); the mid-range transient spontaneous tinnitus (TST)also known as “Ping” and the transient high frequency spontaneous tinnitus (THFST). The frequency representation in Panel A shows the order of appearance; the ordinate indicates presence or absence. TST present since early teens, develops at a single frequency mostly in the mid frequency range, 1 to 4 kHz and possesses even harmonics. CLFT developed around age 40 in the right ear only. It has non discernible pitch around 50 H and has no upper harmonics or timbre. The CHFT which has developed since is mostly perceived as tight clusters of pseudo-sinusoids above 8 kHz (as distinct from the percept of broadband noise). The coherence of the combination is strongly dependent on selectively attending to it.

A musical representation is provided to illustrate firstly that place pitch may extend to much higher frequencies than does repetition (musical) pitch; secondly that CHFTmay generally be thought of as being above the musical pitch range and thirdly, that it may be thought of as clusters of pseudo-tones. Once CHFT extends downward into the musical pitch range as in the case of Smetana’s “high E” (2637 H, “S” arrow) it may be considerably more debilitating because it is closer to the centre of the tonotopy. The indeterminate pitch of the CLFT in the right ear suggests an apical segment of the cochlear partition has collapsed or sagged perhaps due to separation of the outer spiral ligament from the bone. The tectorial membrane is now normally set in the position where inner hair cells (IHC) are excited. The central hypothesis is that this is because the normal setpoint for OHC length has permanently drifted (see text). The right panel shows the perceived sensation level of transient spontaneous tinnitus (TST). Prior to some transient disturbance such as a sudden burst of efferent activity or the death of an OHC there is no sensation because the IHCs are biassed off. In the model the sudden mismatch of opposing forces holding the tectorial membrane in place leads to a excitatory deflection until the previous equilibrium state can be re-established by the remaining OHC which now must generate the same net force as before the sick cell stopped contributing its steady tonic force. The presence of second harmonic distortion (Property 44) suggests that TST does involve an oscillation with some mechanical limiting or saturation which may be attributable to the nonlinear electro-mechanical properties of the OHC. The instability arises when the cochlear partition is displaced from its normal position and because the damping is reduced during such displacement. Panel D shows that CLFT may arise for very similar reasons as TST despite having very different origins and percepts. The explanation derives form the same baseline regulation model, but now the OHC are regulating at a different operating point where the tectorial membrane has drifted in the excitatory direction of the IHC. This leads to the percept of a low frequency hum which can be suppressed if there are sufficient viable surrounding OHC excited to generate a suppressive displacement. The fact that the sensation has no harmonics suggests the generator is not an oscillation.

A PHYSIOLOGICAL MODEL FOR NORMAL OUTER HAIR CELL ACTION

Electro-mechanical regulation of inner hair cell operating point by the outer hair cells.
For many years theories of cochlear function have been based solely on the finding that a travelling wave, like a water wave, propagates along the basilar membrane producing vibrational peaks at places corresponding to frequency components in the sound (von Békésy, 1960) which are detected as peaks by the hair cells. Since von Békésy’s innovative experiments we have learned that the outer hair cells (OHC) are not just sensory cells; they are also motor cells which presumably were not displaying motor activity during his post-mortem experiments. It is now largely accepted that his passive travelling wave picture is incomplete. The OHC activity has fast and slow aspects but the roles of each have yet to be integrated into any comprehensive theory. The fast activity of the OHC (Ashmore, 1987; Zenner et al, 1987; Kachar et al., 1991) is thought to boost the vibration to be large enough to be detectable by the inner hair cells (IHC) even at high frequencies. To model this phenomenon, it was necessary to invoke an amplifier (Neely and Kim, 1983; Davis, 1983) to counteract the viscous effects of the cochlear fluids damping out the sound-produced vibration. Indeed, there are some well advanced ideas about the origins of tinnitus and most deal with the assumption that, like spontaneous otoacoustic emissions, tinnitus has its origins in an anomalous boosted mechanical oscillation (e.g. Plinkert et al., 1990). However, the incidence of spontaneous emissions is not strongly tied to the incidence of subjective tinnitus (Penner, 1992; Shulman et al., 1992).

What still needs to be considered in depth is the possible role(s) of the slow motility of the OHC — the length changes. These cells have the capacity to change their length reversibly by 10 percent of their length (i.e. up to 10 µm) (Zenner et al., 1985). The OHC are like microscopic sausage balloons with a raised internal pressure (Brownell and Shehata, 1990). They also have a very specialised cell border which appears to act like a girdle (Brownell and Shehata, 1990; Holley and Ashmore, 1990). Cell swelling may be due to the cell taking in water and increasing its internal pressure. The girdle appears to have limits to its elasticity and cannot take too much of this. The cell will start to assume a more spherical shape, the swelling starting from its base resulting in shortening of the cell. Eventually the cell can explode and even eject the nucleus (Evans, 1990). When this happens the cell dies, but in the process of doing so the passive spring in the cell wall returns it towards its original cylindrical shape. Also over a period of time, the cell membrane can degrade leading to increased compliance (LePage et al, 1992b) suggesting that the cell has become slack.

Isolated outer hair cells display sinusoidal variations in length in response to low frequency sinusoidal electric stimulation (Brownell et al, 1985; Zenner et al., 1985). The direction of the length change is a systematic function of frequency and place of origin of the OHC (Canlon et al., 1988). By contrast, OHC in vitro, in an intact explanted segment of the basilar membrane, organ of Corti and tectorial membrane, display a steady length change superimposed on the oscillating movement (LePage et al., 1993a). This summating behaviour ceases when the stimulus is turned off. Direct in vivo measurements in the cochleas of guinea pigs have also shown substantial summating displacements of basilar membrane (LePage, 1989) and also in the organ of Corti (Brundin et al., 1991). A recent study has failed to replicate these measurements (Cooper and Rhode, 1992), highlighting the need for further investigation (LePage, 1993). The fact remains that under certain conditions, baseline summating movements of the basilar membrane are observable (LePage, 1981; 1987) and mirror the frequency dependence of the summating potential measured simultaneously from an electrode on the round window (LePage, 1987). Similar results have recently been produced by Brundin et al., 1992. The slow summating movements are prima facie evidence that the OHC regulate the mean position of the basilar membrane thereby stabilising the operating point of the inner hair cells (IHC) against global fluctuations in the pressure in the endolymph relative to perilymph pressure.

FIGURE 3 Integral action of the basilar membrane, organ of Corti and tectorial membrane. Panel A (top) contains two views of the cross-section of scala media bordered below by scala tympani (ST) and above by scala vestibuli (SV). The basilar membrane (BM) is divided into two radial segments, the inner arctuate zone (AZ) and the outer pectinate zone (PZ) which is joined to the wall via the spiral ligament (SLg). The PZ is passive in its properties, whereas the AZ embodies the organ of Corti and pillars which make up the arch which is hinged at the spiral lamina (SL). The two images are redrawn from histological sections of the cochlea to illustrate that slow pressure variations in the endolymph relative to pressures in ST and SV cause the basilar membrane and Reissner’s membrane to flex inward and outward. Typically the basilar membrane is 150 µm wide in the basal turn and under experimental conditions may flex up and down by up to 10 µm. Under normal conditions the OHC deliver a tonic force to regulate the position of the basilar membrane to compensate for such pressure fluctuations and stabilize hearing sensitivity.

Panel B (redrawn from Lim, 1986) shows the key mechanical elements involved in OHC control and IHC stimulation. The OHC are suspended from the reticular membrane in which are embedded their cuticular plates (black) containing roots of the OHC stereocilia. The reticular membrane itself is supported by the arch made up of the inner and outer pillars (IPi,OPi). The bases of the OHC sit in sockets of the Deiter cells (DC) which are attached to the reticular membrane by the diagonal processes. Attached to the outer rim of the reticular membrane are the Hensen cells (HC). The OHCSC are embedded in the underside of the tectorial membrane (TM), glued by proteoglycand molecules. The IHC stereocilia are not embedded in the TM but brush on its underside. On the upper outer edge of the tectorial membrane is the marginal net (MN).

Figure 3A depicts two images of the cross-section of scala media, redrawn from post mortem sections of hydropic cochleas, illustrating that the basilar membrane (lower membrane) and Reissner’s membrane (upper membrane) may undergo substantial baseline variations due to fluctuations in endolymph pressure in scalar media (SM). In addition, variations in BM bias position produced by high-level low- frequency tones have been shown to have a marked influence on cochlear sensitivity (Zwicker, 1977, LePage, 1981; Patuzzi et al., 1984). Low-level, low-frequency tones have no biasing effect because the baseline regulation process actively compensates for the biasing influence. This same process appears to be intimately linked to the normal mode of stimulation of the IHC. In terms of the model the position of both BMand TM is either steady or continually moving to some new set point, depending upon the history of the stimulus (LePage, 1989, 1992a) and descending neural influences via the medial efferent pathway (see Warr, 1992 for a recent review). In the presence of a fixed pure tone, the basilar membrane apparently adopts a “biphasic” or “triphasic” pattern of dc-displacement which is excitatory at the characteristic frequency (CF) of the place under measurement. Displacements toward ST (strongly suppressive) have been observed for frequencies either side of CF (LePage, 1989; 1992a).

Assumptions of the baseline regulation model

1. The OHC detect high frequency energy (Brundin et al., 1989; Iwasa et al., 1991) and adapt to it by change in their length (LePage, et al., 1993).
2. The length change of the OHC is delivered as a change in displacement of the tectorial membrane and to a lesser extent the basilar membrane in opposite phase.
3. The IHC are simple displacement detectors with stereocilia which brush the underside of the tectorial membrane and are deflected by changes in its position (Figure 3B).
4. The IHC are non-adapting receptors, i.e. a constant deflection of the IHC stereocilia (IHCSC) brings about a constant level of depolarisation of the IHC and a constant afferent firing rate (Kiang, et al.,1965).
5. The IHC have a sensory bandwidth which does not exceed the bandwidth fixed by maximal firing rates of nerve fibres (speculation).
6. The bandwidth of the primary force generation mechanism by the outer hair cells need not be much greater than that of the IHC(speculation). Fast length changes have velocities which come within the sensitivity range of sensitive Doppler shift measurement techniques (Sellick et al., 1982), which invalidly assume the motion is pure-sinusoidal (LePage, 1993).
7. In response to a pure tone, the OHC establish a “triphasic” pattern of baseline shift in the position of the tectorial membrane as a function of place, so that the suppression lobes may mechanically bias the IHC off (generate masking) for frequencies either side of CF(speculation).
8. The medial efferent pathway is a negative feedback system which stabilises IHC operating point against long-term degradation of the organ of Corti. There are limits to the effectiveness of the stabilization; it requires some OHC to be present to generate displacements in the suppressive direction (TM moves away from the IHCSC). Total loss of OHC opens the loop. Partial degradation of the active elements can result in “creep” or drift toward the excitatory direction.
9. Regulation of OHC tonic force must effect regulation of this tension in the radial fibres of the tectorial membrane and basilar membrane through the lever action of the arch (LePage, 1990; speculation) which, based on the modified-Helmholtz model (LePage, 1992a), will be highest at the base of the cochlea.

In this model, the primary stimulus for high frequency detection by the IHC is thus assumed to be OHC length change, modifying the transverse position of the tectorial membrane (Figure 3B) by about 10 µm. For sound transduction to occur, the very narrow region of high sensitivity of the IHC (about 0.1 µm) must be physically aligned with the connection to the OHC. (That is, the angle of the tectorial membrane needs to be such that the IHC are minimally able to respond to deflection of their stereocilia).

FIGURE 4 Regulation of tectorial membrane position. In the traditional view of operation of the organ of Corti, vibratory shear between the reticular membrane (RM) and the tectorial membrane (TM) deflects the OHCSC. In the current picture the direction of action is reversed. The three pairs of panels show views of the OHC, RM and TM seen from a radial cross-sectional view seen looking from base toward apex (left panels; redrawn from Lim, 1986) and the corresponding longitudinal section (right panels). The center panels show the OHC at their standard length which gives rise to the normal operating point of the IHC (see Fig 5). At their standard length, the OHC exert a tonic longitudinal tensile force upon the diagonal Deiter cell processes (arrows) which are inextensible, like strings. The regulation process requires that there exist a line of static force running radially through the radial fibers of the TM (TMRF), the arch, the BM and spiral ligament attached to the outer wall of the cochlea, such that the OHCSC act as tiny levers acting to increase the tension in the TM (see arrow). Preparation artefacts readily show that the TM must be under radial tension because the TM retracts when separated from the tips of the OHCSC. The torque thus applied across length of the OHCSC serves to move the arch and the TM up and down, varying the degree of deflection of the IHC stereocilia (IHSC) in the process. In the top panels, elongation of the OHC pushes the reticular membrane upwards causing flexion in the longitudinal direction so the result is a rotatory movement to the cuticular plates which tends to open the gap and suppress IHC response. The tectorial membrane may be raised well clear the IHCSC (the separation is limited by string-like trabeculae) removing all excitatory displacement (Figure 5, operating position A). The lowest two panels show shortened OHC resulting in the generation of shear between the reticular membrane and the tectorial membrane. This shear has both radial and longitudinal components. As shown, clockwise rotation of the arch will close the subtectorial gap, impinging upon the tips of the IHCSC and causing them to slide along the undersurface of the TMknown as Hensen’s stripe. Any drift in the regulating processes tending to close the sub-tectorial gap will result in excitation of the IHC. Various states of IHC excitation are possible (Fig.5:B,C and D) resulting from various degrees of rotation of the arch toward ST.

Mechanical action of the outer hair cells

A hypothesis for the manner of action of the OHC stems from the geometry of the cells which support the OHC, the Deiter cells. These cells have a string-like process which is attached at one end to the cup in which the base of the OHC sits. The other end connects with the reticular membrane one or two cells more apical than their own location (Vold ich and Úlehlová, 1987). Rather than being directly supported by the basilar membrane (BM), the base of OHC is thus supported rigidly from the overlying structure in such a manner that elongation of the OHC results in an upwards force upon the reticular membrane in the direction of scalar vestibuli (SV) and a smaller force downwards on the BM in the direction of ST (see left panels of Fig.4). Conversely, a reduction in length of the OHC will cause the tectorial membrane to move towards ST. The OHC elongation is accompanied by a twisting motion longitudinally along the reticular membrane which simultaneously tends to increase the tension not just in the Deiter processes but the radial fibres of TM as well. (Note that the Deiter cell processes must be diagonal or the OHC would have no freedom to elongate). The three right panels of Fig. 4 show this process in cross-section so that the longitudinal flexing of the reticular membrane can be appreciated. The OHC stereocilia are specialised to withstand this shear force and to maintain tension in the TM and BM (LePage, 1990). Each hair cell has three rows of stereocilia. The smaller length stereocilia are tied together with filaments and butt against the base of the tallest to prevent them deflecting radially toward the modiolus while still allowing them to freely deflect radially outward.

IHC operating point in the normal cochlea

In normal undamaged cochleas the OHC motor unit control system has ample capacity to control the sensitivity of the IHC by setting the operating point on the IHC transfer characteristic to some position which the brain normally interprets as no sound input (Figure 5, inset position A). This condition never actually constitutes zero sound input, but rather a sound level currently deemed as background within the frequency band being coded by that IHC. With significant background noise, the IHC will depolarise to, say, C. In response the motor unit control system will apply negative feedback to try to reset the operating point back towards the same point A but the closest position is B. The dynamic range of the primary neuron is thus extended. Only an increase of sound level in that frequency band will be registered, and by definition, will become a signal of interest (see below).

FIGURE 5 Shortening of the OHC results in excitation of the IHC. The IHC stereocilia (IHCSC) are deflected by various degrees according to the displacement of the TMunder the control of the OHC. They are believed to slide freely under the TM. In position A the IHC is maximally polarised. The TM is not touching the IHCSC. Small movements increase the depolarisation of the IHC until saturation occurs at D, leading to transmitter release and excitation of the afferent terminal (AT).

This scheme constitutes a more comprehensive role for the OHC than pure amplifiers boosting the vibration. It may be a more realistic one because it also provides that the role of the OHC is to be the primary detectors of high frequency energy, delivering their output to the IHCas a low band width signal containing only the envelope of the sound (LePage, 1989). As such it appeals to the principle of Occam’s Razor: Why should the IHC also be envelope detectors, if the OHC already have an adaptive role, changing their length according to amount of the high frequency energy reaching them? That is, the baseline regulation function of the OHC means that the IHC need not also be specialised for detection of high frequency vibration. The transition of OHC function from ac amplifiers to ac detectors and integrators occurs over the frequency range 1 to 4kHz.

In turn this simple scheme does not specifically require the cochlear amplifier be sharply tuned with resonant elements in the model proposed by Gold (1948). Frequency selectivity is modelled in the spatial domain as the triphasic pattern of varying OHC length along the cochlear partition (LePage, 1987; 1989) depicted in Figure 6. The top panel (6A) shows the effect of a variation in the length (greatly exaggerated) of the representative OHC along a 3mm segment of the cochlear partition in response to a pure tone. Shortening of the OHCresults in excitation of the IHC; lengthening results in suppression by virtue of the resulting displacement of the attached TM. Figures 6B and 6C assist with visualising the effect of superimposing a uniform bias component on such a triphasic pattern of excitation. The pattern of Fig. 6A is inverted so that positive-going (downward) displacements of the TM result in excitation (Fig. 6B) (see again Fig. 5). Such a uniform bias could be established by the medial efferent system as a direct function of sound level, higher sound levels resulting in more positive values of dc bias. In this schematic, the size of the triphasic pattern is fixed for simplicity, and only values of TM displacement resulting in a positive deflection of the IHC stereocilia (IHCSC) (Fig. 5, point “A”) are shown. Highly positive peaks (greater than 40 arbitrary units) in the displacement will result in deflections of the IHCSC past set point “D” resulting in saturation of the IHC electric response as shown. For each level of the stimulus, the IHC frequency-response profiles (Fig. 6C) are remarkably similar to the iso-SPLfrequency response characteristics of single neurons in the squirrel monkey (Geisler et al., 1974, Fig.2), leading to the standard representation of tuning at threshold (dark horizontal line representing the spatial extent of the excitation region).

FIGURE 6 Schematic of the baseline regulation model. Panel A shows a representative sample of OHC along a 3mm (ca) length of the cochlear partition under stimulation of a single sinusoidal stimulus. The OHC sense the mean vibratory stimulus and respond by adjusting their lengths accordingly. This is a dynamic, history-dependent process in which the OHC length varies about some mean length which fixes the transverse position of the TM. In one of the several possible mechanical schemes each OHC shortens in order to pull (down) upon the TM, deflecting the IHC stereocilia (IHCSC) and depolarising the IHC (as in Fig 4). Conversely, the OHC elongate to lift the TM away fromIHCSC to produce suppression. The figure utilizes a “triphasic” (here M-shaped) excitation pattern (the amplitude is exaggerated a hundredfold for display) in which an excitatory region (a downward movement of the TM) is bordered by an inhibitory surround (upward movement of the TM). The relative lengths of the excitatory region versus the inhibitory regions depend strongly on the normal operating point (OP) of the OHC relative to the displacement required to just excite the IHC (OP). Clearly if the OHC must deliver a power-stroke in the suppressive direction, the level of suppression at any time depends upon the integrity of the force-generation mechanism leading to an elongation of the OHC. Suppression will be reduced by any agent which tends to disable this force component. The bottom panel shows how shortening of the OHCresults in a downward movement of the TM and excitation of the IHC.

In panel B the ordinate represents the deflection of the IHCSC which results from the M-shaped pattern (Panel A), this time inverted to appear W-shaped so that positive indicates cell excitation. Each of the four curves can be thought of as a stimulus-produced pattern upon which is superimposed a different value of dc-bias (the OHC lengths are changing) due for example to a broadband stimulus, or a small change in the endocochlear potential — or a global change in OHC turgor pressure. The curve in front represents the value of the bias necessary for the underside of the tectorial membrane to just touch the IHC stereocilia for the excitatory peak at the CF. As the bias is raised, due to a rise in OHC dc-operating point, the frequency band for which the TM is in contact with the IHCSC is increased and the IHC are deflected more at the CF. There are still frequencies removed from CF for which the TM does not contact the IHCSC. As the bias is raised further (OHC length increase) the IHCSC register the whole excitatory peak, but not the suppressive displacements. For the rear curve in the top panel the TM is continuous contact with the IHCSC and so the IHC are now excited by the pattern for all frequencies. It is important to realise that the ability of the OHC to generate these biases is large (up to ca. 10µm) relative to the stimulus-produced dc-excitation pattern which, at threshold may be of just- detectable amplitude (ca. 1 nm), yielding a very different picture of cochlear mechanics from the traditional travelling wave envelope.

The bottom panel (C) shows the effect of these displacements on the electrical behaviour of the IHC (see again Fig. 5); the ordinate values are arbitrary but may be helpful in thinking in terms of mV depolarisation. Assuming that the IHC operating point is at A, the IHC will register the smallest deflection as a tiny depolarisation B. Progressively larger deflections will cause more depolarisation C and D, however, because the IHC has a narrow displacement range (ca. 0.1µm) the membrane potential will saturate for larger deflections. The resulting pattern of depolarisation shows very clearly how the bias plays a very important role in the definition of the tuning curve formed by joining the edges of the excitatory region with straight lines in the horizontal plane. Indeed, the series of profiles mimics strongly the patterns of firing of primary afferent neurons. Note that because of the W- shaped pattern, sinusoidal stimuli at other frequencies will generate a negative bias which will be superimposed on the existing stimulus/bias pattern. This will have the effect of driving the W-shaped excitation pattern downward and the operating point of the IHC back toward A, reducing the level of excitation at the CF and masking the response. As is described in the text, this scheme applies equally well for any pseudo-stimulus excitatory pattern such as seen in Fig 8. Provided there are enough OHC in the neighborhood to generate a negative-going displacement (by elongation, drawing the TM away from the IHCSC) the pseudo stimulus may be masked. The greater the loss of OHC the smaller the net force they can deliver on elongation and the more difficult it will be to mask.

The property of this model which is highly relevant to tinnitus is that two-tone suppression can occur by virtue of locating the frequency of the suppressing tone so that one or other of the negative phases is spatially aligned with the excitation peak of the primary tone. The central excitation region may be due to a real input or equally well, a phantom input. In the model the operating point is controlled very precisely so that any drift in that position will have the effect of varying not only the firing rate of the nerve if the operating point is within the displacement-sensing range of the cell but the number of afferent fibres stimulated — the size of the excitatory region. This scheme provides for suppression of the firing rate by turning off the IHC with a negative-going displacement of the cochlear partition, produced by efferent control of OHC length. Clearly the size of the excitatory region (mm) depends upon the value of the bias which is under central control.

There is considerable redundancy in the numbers of OHC. Humans have some 12000 in each ear at birth (see summary in Kim, 1984). Not only are there three rows but each motor unit will consist of the 20 to 50 OHC synapsing with a single medial efferent neurone (see Fig. 7, redrawn from Spoendlin, 1970). The scheme allows that different motor units may give rise to differing local values of bias so that the size of the excitatory region for any sound may be varied with respect to any other sound. The neural innervation pattern of these motor units suggests a likely candidate for the improvement in signal-to-noise ratio (SNR). Here the definition of “noise” is any signal in which one has no interest, whether or not it contains information which has meaning if attention were directed at it.

FIGURE 7 The main neural synapses with the inner and outer hair cells (redrawn from Spoendlin). The left panel shows the form of the cochlear spiral calculated from a Cornu integral. In the right panel most afferent neurones synapse with the single row of inner hair cells. The dotted lines show the efferent or motor neurones which control patches of outer hair cells or “motor units” (LePage, 1989). Three such motor units are shown. The brain would appear to be able to vary the sensitivity of the cochlea for different narrow frequency bands, independently from those either side. This pattern of innervation offers the possibility for a listener to selectively attend to sounds at specific frequencies (presumably either real or phantom) by setting up a comb-shaped filter (LePage, 1989).

PHANTOM ACOUSTIC INPUT : THE OPERATING POINT DRIFT HYPOTHESIS

 

Numerous histological studies (e.g, Engstrom et al., 1970; Bohne et al., 1987) have shown permanent damage patterns ranging from loss of OHC to loss of the supporting cells to loss of the whole organ of Corti. The lesioned areas may be quite irregular giving rise to loss of control of the input to the IHC via the tectorial membrane at these locations. The current state of excitation of the IHC is set by the angle of deflection of the IHC stereocilia (IHCSC) is termed its operating (set) point which varies along the transfer curve (see again Fig. 5). In the model anything which results in steady deflection of the IHC stereocilia in the excitation direction can be interpreted by the brain as a real acoustic excitation pattern. It is clear that any coding system in which the OHC are the fundamental detectors delivering their response to the IHC as a shift in the baseline position of the tectorial membrane, must incorporate precise regulation of baseline position so that the organ may detect very tiny signals buried in ambient noise. Moreover, the scheme should be able to continue to do so even if the normal range of operation of the mechanism is compromised temporarily or permanently (Property 1). The following speculations stem from the baseline regulation model:

Potential conditions which may give rise to drift in IHC operating point:

1. IHC with permanently distorted stereocilia (Saunders et al., 1986; Thorne, et al., 1986; Canlon et al., 1987), if still functional, could be interpreted by the brain as a phantom or virtual acoustic input.
2. Passive shift of the cochlear partition due
   1. to pressure buildup in scalar media (Fig. 3A) (Properties 41,41),
   2. changes to the morphology of the TM (Canlon, 1987),
   3. detachment of the tectorial membrane from the tips of the OHC (von Békésy, 1960),
   4. loss of elasticity of the spiral ligament (Voldrich and Úlehlová, 1982.
3. Active shift due to a drift in the OHC electromechanical set-point e.g. due to:
   1. temporary or permanent change in tonic efferent firing rates (Liberman, 1989) or,
   2. loss of regulation in biochemical gradients due to noise (Konishi and Salt, 1983),
   3. OHC degeneration, e.g. due to:
      1. loss of volume and/or length regulation
      2. loss of force generation per cell
      3. decrease in numbers of cells.

The common factor in each of these potential disturbances of regulation is abnormal deflection of the IHC stereocilia in the excitatory direction (Figure 5) exciting the afferent terminals. In the first case the permanent deflection is independent of OHC response. In cases 2 and 3 recovery depends on the viability of the remaining OHC in the neighborhood to perform some regulation.

In cochleas where there is significant permanent OHC loss, the capacity of the remaining OHC to oppose drifts in the set point will be reduced resulting in a reduced range of suppression or excitation. Damaged OHC may drift in their length, changing the transverse position of the tectorial membrane. The changes in electromechanical set point of the OHC may be expected to be accompanied by changes in the size of the microphonics to external tones (Property 16).

When the OHC drift is sufficient they may no longer respond to the medial efferent influence and the displacement range over which active baseline restoration can take place will be reduced. Presumably at the limit (complete OHC loss) no suppression can occur since no OHCexist in the neighborhood to generate a suppressive displacement (cell elongation). Clearly, no permanent loss of OHC function need occur to produce tinnitus. If the mechanical coupling is damaged at any point in the lever system, e.g. temporary separation of the tectorial membrane from the tips of the OHC stereocilia, force generation by the OHC can no longer maintain that tension and the structure will become slack.

Incessant character of continuous tinnitus

The brain will not be able to distinguish between active deflection by OHC action in response to sound as per the model, or mechanical drift due to OHC inactivation or collapse of the organ of Corti due to acoustic overexposure (Saunders, et al., 1986; Harding et al., 1992). Constant deflection of the IHC stereocilia produces a constant firing rate, consistent with properties 14,15,17-22; 43,51-53. Therefore, in terms of the model, neither CHFT nor CLFT needs a physical oscillation to generate it, accounting for the lack of perceived beats with an external tone (Properties 13,49).

FIGURE 8 Hypothesised origin of subjective tinnitus. The left schematises a short (ca 40µm) segment of degraded OHC surrounded either side by healthy OHC. OHC tend to shorten with degradation and can no longer extend to deliver tonic force to raise the tectorial membrane. There is consequently a short longitudinal segment of tectorial membrane which is depolarising IHCs. The black region constitutes the pseudo-input exciting the IHC (refer again to Fig. 5). The middle panel depicts the hypothesised effect of efferent stimulation of the OHC resulting in shortening of the OHC in the region. A longer segment of the tectorial membrane now lies below the threshold stimulus level of the IHC and so the hypothesised effect of attention is shortening of the OHC leading to a greater number of IHC and afferent fibers being excited. Elongation of the OHC (right panel) due to acoustic or electric stimulation (but not voluntary control via efferent activity) results in suppression of IHC response.

Figure 8A depicts a localised area of collapse in the structure of the organ of Corti due to loss of OHC turgor, leading to an increased deflection of the stereocilia of the IHC mimicking a real input (cf. Fig. 6A). The loudness of the tinnitus will likely depend upon the depth of the deflection. This will be determined by the extent of the OHC lesion and whether all three rows are affected or missing. In turn this will set the number of IHC being stimulated. If the OHC loss is not extensive but scattered, the condition may constitute a mild wideband tinnitus. The adjacent OHC can still exert some regulation function; it simply requires a higher level acoustic stimulus to evoke an increase in OHC length which acts to reduce the amount of deflection of IHC (Fig. 8C). The sound level required to reduce the depolarisation of the IHC to zero is effectively the masking level (Properties 29-37).

Tinnitus pitch
If a short length of the cochlear partition were collapsed or the OHC tonus decreased (Figure 8A), the resulting dynamic excitation pattern is similar to the triphasic pattern for a pure tone (Figure 6). The model therefore suggests that the percept of this condition will be a place-code input, i.e. the percept of a pure sinusoid. To hear harmonics of a fundamental frequency might suggest the existence of a complex waveform propagating along the basilar membrane. Yet for CLFT and CHFT one does not hear anything other than a very `thin’ single pitch even when intense (Properties 13,14; 54) without the harmonic richness (repetition pitch) which accompanies a complex time waveform.

The continuous aspects of the tinnitus are also characterised in musical pitch notation in Figure 2B. The purpose of this representation was to illustrate just how high the pitch of CHFT may be. To the majority of people not generally acquainted with the distinction between place pitch and repetition pitch, describing a sound as “high-pitched” may tend to convey the impression of the sound of a musical instrument in the high register e.g. a trumpet, violin or piccolo. Indeed in Smetana’s string quartet in E major the composer uses a high E from the first violin apparently to record the sound of his CHFT, yet the note chosen is at 2637Hz, just 2.6 octaves above A=440Hz (see “S” arrow Fig. 2B). However, the most common and debilitating kind of tinnitus described by clients of the author, particularly those with industrial hearing loss, is between one and two octaves higher again than this — well above the range of periodicity or musical pitch. It seems very few texts on tinnitus have made much distinction between the two types when the pitch of tinnitus is considered. Confusion about the nature of CHFTmay exist (Property 7) because place pitch is normally implied in pure-tone pitch-matching experiments. If CLFT were due to an oscillation one might expect to hear upper partials like those for the TST but at least in this case none exist (Property 49).

The audio tape produced by the British Tinnitus Association uses a musical synthesizer to give an “artist’s impression” of how different individuals perceive their tinnitus. Interestingly, over half the examples of synthesised tinnitus provided on the tape from the British Tinnitus Association are centred in the low to mid ranges (i.e. <2kHz). Few kinds of tinnitus have any inherently musical quality about them. They generally appear to fall into broadband noise, multiples of tone-like peaks, or the same amplitude modulated at rates ranging from 0.5Hz up to 10 Hz (Property 5). Accordingly one can view broadband noise-induced tinnitus as being due to extensive scattered OHC damage over a length of the cochlear partition, particularly involving the basal end leading to poor regulation of basilar membrane position. The percept will be like many tones which have no harmonic or phase relationship, depending on the number and location of the lesions.

In the case of the hydropic condition is caused by displacement of the basilar membrane position downwards (Fig 3A, right panel and Figure 5, position D). Menière’s syndrome in particular is marked by large, positive pressure fluctuations which are sensed jointly by the vestibular hair cells as vertigo and, according to this model, the IHC as roaring tinnitus. The spectrum of the tinnitus will likely contain a preponderance of low frequencies, because a raised static pressure in scalar media along the cochlear partition will produce greatest displacement where the static compliance is greatest (at the apex). Indeed Nodar and Graham (1965) report mostly frequencies below 1kHz in Menière syndrome attacks.

Onset of tinnitus

As the ear ages, the strength of click-evoked otoacoustic emissions decreases, particularly at high frequencies (LePage and Murray, 1992) suggesting that either the numbers of OHC are decreasing or that their functionality decreases. In the regulation model (Assumption 9) theOHC at the highest frequency end must sustain the highest static force in the fibers of the TM (Fig. 4) and the BM (PZ, Fig. 3). (LePage, 1990). The site of greatest tension, and therefore the site most likely to display damage first, is the high frequency region of the cochlea. Tinnitus tends to range from 4 to 12 kHz judged from pitch-matching experiments in which octave errors frequently occur. Vernon reported one case at 17 kHz which is consistent with our case study here. It is therefore expected that continuous tinnitus will most frequently first present at high frequencies, consistent with Properties 2-4; 23, and 26. However, tinnitus may occur at any pitch depending on the frequency of acute noise or other trauma and the resulting site(s) of cochlear damage (Property 26).

Many of the synthesized examples produced on the British Tinnitus Association tape contain amplitude modulations, some with a modulation frequency about the heart rate (Property 5). Presumably spurious contractions of the tensor tympani and/or stapedius muscles may also be responsible for pulsations. It follows from the mechanical regulation function of the OHC, that deterioration of that function may lead to sloppy regulation; the system will become slack, and be more subject to perturbing influences, such as minor vascular pulsations or jaw movements (Vernon, 1991) which normally do not influence tectorial membrane regulation (Properties 42).

This model does not exclude the possibility that physical oscillation may occur simultaneously (LePage, 1992) and modulate the percept of set point drift (Properties 3,5). The percept of cicadas and crickets by some sufferers suggests that some physical oscillation may be involved, since the percept of low frequency amplitude modulation is unlikely to be the result of neural interaction centrally. In this case the condition may be caused by combination of collapse of the cochlear partition plus a spontaneous mechanical oscillation. Certainly, in the case of TST the presence of harmonics (Property 44) suggests that a mechanical oscillation occurs, in which case such unpredictable phenomena should have an objective counterpart.

Exponential-type behaviour: residual inhibition

The temporal behaviour, particularly the decay of both the TST and CLFT suggest that the time course of the regulation process is exponential in nature (Properties 43, 51). As is shown in Figures 2C and 2D this descriptive model has merit because the same baseline-regulation model can apply to both kinds of tinnitus; the spontaneous and the continuous, despite their diversity. In the figure the ordinate corresponds to sensation level for the tinnitus. However, in the case of the CLFT tinnitus the set point for the IHC has drifted above the audible level where it remains permanently until some transient masking influence occurs. Alternately, in the case of the spontaneous tinnitus the set point is normally just below the sensation level until triggered to go above by some perturbation (Property 45) such as an OHC exploding or a spurious burst of activity in one of the motor units generating a transient bulge in the tectorial membrane. This subsides once the remaining OHC re-establish their estimate of the “normal” operating point of the IHC.

The decrease of TST with time (Property 43) suggests that damping of oscillation on the BM is a function of baseline position. Given that a parasitic oscillation has begun, the baseline regulation process will act so as to return the system to a just-stable condition. Indeed the profile of the OHC length change with place (Fig. 6) could be tested as controlling damping of the cochlear partition.

In the model, the OHC normally undergo length changes while a stimulus is maintained, but will return to their prestimulus length when the stimulus is removed. The time it takes for the decay to occur may depend on several factors. There is a strongly nonlinear relationship between OHC potential and the length change (Evans et al., 1991) so it is to be expected that some kind of electrical exponential decay will take place before any mechanical change occurs. Residual inhibition (RI) (Fig. 2d; Property 36d) could simply be the result of the lag between the electrical driving force and the slow electro-mechanical response.

Poor correlation of tinnitus incidence with hearing loss

Over the history of auditory science, we have become conditioned to think that because the mammalian hearing organ is an exquisitely sensitive mechano-receptor that this is equivalent to the assertion that behavioural thresholds are a sensitive indicator of damage to the organ. An alternative hypothesis is that audiometric thresholds may be a late indicator of damage (LePage, 1991) because damage must be extensive before there is a significant change to the audiogram. The field of otoacoustic emissions has resulted in the development of the concept of cochlear mechanical loss (Kemp, 1989) determined by objective measurement, which is in contrast with the concept of a hearing loss determined psychophysically. Evidence is accumulating that the ear may suffer considerable mechanical damage over a long period of time before it is actually manifested as a hearing loss. Click-evoked otoacoustic emissions suggest that loss of motor function (either due to loss of numbers of OHC or loss of OHC function due for example, to decline in endocochlear potential) may in some cases be determinable many years prior to a loss in hearing sensitivity (Murray and LePage, 1990). In this sense the ear “fails gracefully”.

By contrast, tinnitus can be manifest at a much earlier stage once there are punctate lesions along the cochlear partition involving only small patches of defunct cells or small regions of mechanical disruption which, due to OHC redundancy, has not yet been revealed as a decline in hearing sensitivity (Properties 26a,b,c). Figure 1 illustrates for our case study near-normal thresholds despite the fact that the oto-emissions are very low in amplitude indicating the existence of substantial scattered OHC damage over the length of the cochlea (LePage and Murray, 1993).

In the case where substantial high frequency hearing loss has already occurred, leading to a sloping audiogram, the pitch of the tinnitus is frequently matched with a pure tone of frequency where pure tone thresholds have risen by 25 to 30 dB (Meikle, 1991b) by which stage otoacoustic emissions will have dropped to non-measurable levels (Kemp, 1989). This co- incidence suggests that there exists a broadband tinnitus at all frequencies, but the tinnitus is identified perceptually by an edge effect which normally accompanies band-limited noise. One can speculate that this may be due to the presence of a working region of cells bordered on the basal side by a region of substantial damage.

Pure tone audiometry shows that left ears preferentially have higher permanent losses than right ears and this is registered using otoacoustic emissions as lower emission strengths (LePage and Murray, 1993). This suggests that left ears acquire more permanent damage before right ears (Pirilä et al., 1991) so it is also not surprising that left ears show a higher percentage of tinnitus (Property 25). If tinnitus is generally less well-developed in the right ear, it follows that with less damage it will be easier to find masking conditions for the right ear than the left ear (Property 35). On the basis of acquired hearing loss, therefore, one might expect to find more tinnitus in males than females (Property 24).

Difficulty of obtaining masking conditions

Not all tinnitus can be easily masked (Properties 30-33). The conditions required for masking seem largely unpredictable particularly in respect of the frequency of the masker. The variability of masking conditions may be explained in terms of the baseline regulation model. The more extensive the neighborhood of OHC loss, the higher the level of sound required to mask the tinnitus. For scattered OHC loss, a broadband masker will be more effective. Alternately for a `punctate’ lesion, a pure tone masker may be more effective, its frequency being any frequency over a wide range which will produce a suppressive displacement (refer again to Fig 6A and 6B).

The fact that there is no upward spread of masking for CHFT (Property 36b) stems from the fact that there is no spatial pattern of excitation extending over several millimeters length of the cochlear partition as occurs with an external tone (Fig 6). Lesions causing CHFT can be punctate and limited to just a few OHC extending along the cochlear partition less than 100µm (Fig 8). Indeed there is no reason to suppose any given relationship between the loudness of a tinnitus and that of an external masking tone (Property 36c); the TMdisplacement patterns due to both pseudo- and real-tones will depend on the patterns of OHC damage. The ability to mask with an external sound is strongly linked to the ability of the OHC remaining in the neighborhood to respond to the external sound, to generate a shift of the TM in the suppression or repolarising direction (returning to point A in Fig. 5), rather than in the excitatory direction. In Fig. 8C the normal cells adjacent to the damaged cells will be able to generate the required extension to suppress IHC response. Eventually they too may fail resulting in an increase in size of the lesion, leading to a longer excitatory region and a louder percept (Property 11). If the compromised region is small masking conditions will be easier to obtain (Properties 31-33). As the OHC damage increases the ability for the fewer remaining OHC to raise the TM will be less. In the cases where a whole contiguous patch of OHC is missing, no masking will be possible (Property 30).

Under some circumstances masking could depend on just a couple of remaining OHC and the task is then to set up stimulus conditions to find at what frequency and level they will be caused to extend. This may not be a simple task because the place being stimulated by a particular tone apparently does not depend solely on stimulus frequency. At least at higher sound levels, the place also normally depends on the level of the stimulus due to the motor action of the OHC. Once very high sound levels are necessary for masking (>100dB SPL) the excitation pattern varies with functional remapping of frequency to place (LePage, 1990; 1992a) which will tend to move apically depending on the pattern of the loss (Property 30, 32– 34).

Location of Tinnitus

Finding an adequate explanation for central tinnitus has perhaps represented one of the most challenging aspects to our understanding and management (Møller, 1984). Individuals who have had the eighth nerve sectioned may still experience debilitating tinnitus. While mechanisms proposed for central origin of tinnitus remain equally valid, the baseline regulation model opens the possibility that even central tinnitus, stems originally from cochlear disfunction. The fact that the sensation appears diffusely located within the head may simply reflect that normal binaural processing operations in the brainstem break down when the inputs from both ears cannot be correlated and their relative phases on the basilar membrane be voluntarily manipulated by turning the head. By induction, phantom neural input from just one ear alone following nerve section may be sufficient for maintenance of central tinnitus.

Variability in severity

Inherent in the condition is the fact that there is no adequate measure of its severity (Properties 8-11) which tinnitus has in common to constant pain. Variation in the apparent loudness of constant level sounds is generally appreciated as a basic property of how we hear, e.g. what makes a clock ticking basically inaudible during the day yet `loud’ in the quiet of the night. It is currently being modelled in terms of an internal automatic volume control mechanism (Lyon, 1990) such as is commonly used in hearing aids. [Although the idea has been intuitively obvious for a long time, it was not until OHC length changes were observed both in vitro (Brownell, et al., 1985; Zenner, 1986) and in vivo (LePage, 1981, 1987, 1989) that a physical parameter could be assigned as a likely parameter controlling cochlear sensitivity. This single fact can account for the absence of any absolute sense of loudness in human hearing.] Automatic volume control (AVC) systems are adaptive systems which over a time period adapt the detector to registering small signals above background noise level. Typically the adaptation period may be much longer than the repetition rate of information usable by the detector, e.g. in the visual system dark adaptation typically takes a few minutes.

Tinnitus is paradoxically `loudest’ in low ambient background levels (Properties 8,32; 47,48). Typically a patient complains of trying to get to sleep at night (Axelsson and Ringdahl, 1989), when all is quiet and the daily sounds which normally provide masking are absent. The model provides some insight into why this is so. Because of the partial similarity which exists between real acoustic (Fig. 6) and pseudo inputs (Fig. 8) (Properties 3-5, 28) we can re-utilise the external tone profile in Fig. 6B to represent a pseudo-tonal input. Broadband noise generates masking which will generally have the effect of applying a net suppressive bias to the TM because for any frequency component the excitatory region is much shorter than the suppressive region. Suppose that the kind of pathology depicted in Figure 8 gives rise to an excitation pattern such as is shown in Figure 6B. We can picture the unmasked displacement pattern as the rear profile. The spatial dependence of firing rates of primary fibres (Figure 6C) constitutes a significant input to the brain stem. With increasing background noise we have a situation of increasing broadband suppression and the displacement profile of the tectorial membrane descends in the hyperpolarising direction with the consequent disappearance of the profile below the detection range of the IHC (front profile, Figure 6B). It is clear that the excitatory region is now only represented by the tip of the profile, so that under the higher ambient noise conditions only a small number of afferent fibers carry the phantom input; while under broadband masking the pseudo-input is much less significant (Property 34).

Selective attention and the “masking” of external sounds by tinnitus

.
There are some aspects where tinnitus behaves like external sounds (Properties 29-35). Conversely there are some aspects where it does not (Properties 36a-d). To approach an understanding of the masking of tinnitus by external sounds and vice versa it is first necessary to appreciate the general nature of auditory selection or the voluntary control of improvement of signal-to-noise. All active signal processing systems have a `noise floor’ and it seems that Wegel (1931) recognised that all people have tinnitus to the extent that the internal noise of their ears can be appreciated independently when external noise sources are eliminated, or reduced below the internal noise, for example in an anechoic chamber. Indeed, these are the conditions sought for pure tone audiometry (Murray and LePage, 1991). A pure tone audiogram is in reality the result of an experiment in selective attention. It is a measure of how well the auditory system can focus on a pure tone when it is being masked by internal noise only. In a normal ear, the internal noise level is low. It is still an open question whether the reason auditory thresholds rise is because there is an effective rise internal noise level. However, in terms of the theory there need be no real conceptual need to distinguish between a loss of hearing sensitivity and a rise in internal noise level. This is because the origin of the “pathological noise” need be nothing more than a drift in the set point of the IHC for some length of the cochlear partition, which can no longer be reset by OHC activity of the kind depicted in Figure 6. Both result in a decreased ability to select, or identify, the spatial pattern generated on the basilar membrane due to a pure tone — traditionally designated the travelling wave envelope — which gives rise to cochlear tuning.

By virtue of being able to consider tinnitus as an internal signal adding to external noise, it is no surprise therefore that there should be a heavy overlay of selective attention-related issues in the perception of tinnitus (Properties 12; 55-57) which constitute a major complication in dealing with the severity of tinnitus. Selective attention has been the subject of considerable research in the areas of neuroscience (see Näätänen, 1990 for a recent review) and cognitive psychology (see Johnston & Dark, 1986 and Kinchla, 1992 for recent reviews). Researchers in both areas have argued for many years about whether there are mechanisms located at the earliest stages of sensory processing which play a role in attention. This has become known as the peripheral filtering hypothesis and was first proposed by Hernandez-Peon in 1966. It proposes that exclusion of irrelevant and/or facilitation of relevant information occurs in the sensory pathway, even at the most peripheral levels by processes under the control of higher brain centres via centrifugal or efferent fibres. It is clear that the OHC motor unit (LePage, 1989) with its efferent innervation has the capacity to modify sensory processing in this way. However, data supporting attentional effects at the level of observations of an effect measured at the round window were subsequently criticized for their lack of important controls, but better controlled animal experiments by Oatman and colleagues (Oatman, 1971, 1976; Oatman & Anderson, 1977, 1980) demonstrated suppression of irrelevant auditory information at the level of the cochlea. Repeated attempts to find robust evidence of peripheral effects of attention in humans using brainstem auditory evoked responses however have been generally unsuccessful (see Hirschhorn & Michie, 1991 for a review). This has led some to assert that attention is mediated and implemented by higher level brain mechanisms. To date the earliest replicable effect of attention has been observed on cortical evoked potentials with a latency of approximately 50-60 ms and is presumably mediated by cortical mechanisms (Michie et al., 1990; Näätänen, 1990).

Recently, there has been suggested that OHC-motor units may be implicated selective attention tasks (LePage, 1989). The pattern of efferent innervation (Figure 7; see also Warr, 1992) strongly suggests that the machinery exists for the central nervous system to influence cochlear sensitivity globally or selectively depending upon the variation in tonic firing rates of the medial efferent neurons. For example, a comb-shaped filter could be established to match the spectral peaks of any sound signal deemed important to the listener. The effect of such a selection operation is hypothesized to increase the excitation of the IHC through shortening of the OHC (Fig 8A) so that the size of the excitatory region for a given input is increased still further. Like the brainstem evoked response results, recent experiments show that the effect of attention on otoacoustic emissions, if any, is very small (Puel et al., 1988; Froehlich et al., 1990; Giard et al., 1992). The outcome of these experiments may be highly paradigm dependent and also subject-dependent according to the state of cochlear damage in test subjects not revealed by pure-tone audiometry.

The baseline regulation model also suggests that if the postulated motor-controlled selection process is disabled, the sufferer may not be able to ignore either a virtual input (tinnitus), or indeed a real signal input (hyperacusis?). If the process of masking an input depends upon the ability of OHC to generate a suppressive displacement, decreasing numbers of functional OHC would mean a reduction in ability to select. Depending upon how auditory selection is governed centrally, an increase in firing rate may accordingly be deemed a “signal of interest” even if it is only caused by a drift in the set point which is outside the control of the efferent feedback loop.

Trying to actively ignore tinnitus by a variety of schemes, for example biofeedback (House, 1978) may bring variable results if the efferent system is only programmed to focus by shortening of selected OHC motor units (Figure 8B) (Properties 55-57). That is, if the efferent activity only works in one direction which is not the direction needed to mask the tinnitus (Figure 8C). That is, trying to ignore tinnitus, with effort, may be rather akin to trying to sleep in a hurry. Mental effort plus an absence of distractions thus serves to focus subject attention on the offending input. One suspects it is not even possible to actively focus attention on something else; because the act of focussing establishes an internal program of arbitrary cycle time of attention-switching between the thing being ignored and its replacement.

Likewise in an experiment to test the relative effectiveness of a broadband masker on tinnitus, or on an external tone, the masker might be predicted to have less effect on the external tone, by virtue that it is possible to focus on the tone (or any coherent input) which is inherently more difficult with a broadband tinnitus (Property 36a).

Relaxation techniques (e.g. Tyler et al, 1992) are therefore more likely to be of benefit clinically while methods aimed as masking (acoustic or electrical) depend on there remaining sufficient numbers of OHC in the neighborhood of the lesion capable of raising the TM away from the IHCSC. Pure tone masking should work well for punctate lesions while broadband masking will be predictably better for scattered OHCloss.

In the model, if efferent action is unidirectional, a phantom input cannot actually “mask” an external sound, because masking requires generation of a suppressive displacement (TM towards SV) and such masking is not possible at places where OHC loss is complete. In this sense a pseudo-input interferes with external sound reception only in the sense that the internal noise level has risen, requiring external sounds to be higher in level than the maximum voluntary selective improvement normally achievable (ca. 10dB), accounting for the limit to such “pseudo-masking” (Property 27). Since “true” masking is produced only by OHC elongation in response to external tones, masking of external tones is merely a manifestation of an spastic patch of OHC.

SUMMARY AND CONCLUSIONS

Perhaps the most significant obstacle to progress in tinnitus research has been the lack of a sufficiently tangible and testable model for its origin. The model proposed here for continuous subjective tinnitus is not new but is an elaboration of a “homeostatic” mechanism for cochlear tinnitus proposed by Stephens and Davis (1938:1983), Davis, (1954) and furthered by Tonndorf (1981). Accordingly, this chapter presents the results of an experiment to investigate how many of the salient properties of tinnitus have a potential explanation in the morphology of the organ of Corti and in particular, speculations concerning the function of the slow activity of the OHC. In terms of this theory, phantom acoustic input to the central nervous system is the natural consequence of degradation of OHC activity or alternately disturbance to passive mechanical structure such as the spiral ligament. The form of this degradation is essentially loss of OHC control of regulation resulting in collapse, or static distortion of the structure to be contrasted with the more popular notion of a mechanical oscillation. If the resulting position shift is localised it may resemble the triphasic displacement pattern of IHC excitation in response to a pure tone and thus be regarded by the brain as a real input having the appropriate properties such as no associated spontaneous emissions; no beats with external tones and no perceived harmonic structure. This hypothesis hinges critically upon the primary afferent synapses at the base of the IHC being essentially non-adapting (Kiang et al., 1965, pp73–79.). Apart from refractory effects in neural excitation (Gaumond et al., 1982) adaptation seen in primary afferent fibers (Kiang et al, 1965) is presumed to be due exclusively to OHCelectromechanical adaptation (LePage et al, 1993).

If, in the vicinity of any damaged region, there are still OHC with normal functionality, hearing loss may not be manifest at all and masking of the tinnitus will also be possible. This is due to OHC elongation which raises the tectorial membrane and reduces the deflection of the IHCstereocilia. Hence the model offers the possibility that the incidence of tinnitus and hearing loss not being directly related, while each is directly relatable to the extent of damage to the cochlear partition. When the damage becomes more extensive it will become increasingly difficult to mask the tinnitus because there will be fewer OHC remaining to generate suppressive displacements of the tectorial membrane. Hence within the framework of OHC motility it is possible to describe subjective tinnitus of cochlear origin in fairly general terms using a single parameter — steady deflection of the IHC stereocilia and less generally in terms of tectorial membrane displacement toward scala tympani offering a direct explanation of the roaring tinnitus in Menière’s syndrome. The TST described by the author, by virtue of its harmonic content suggests the existence of a mechanical oscillation which should be detectable as an ear emission if any such event could be captured. More generally, recovery from TST supports the notion that the basilar membrane is maintained in a “conditionally-stable” state and that the OHC length is regulated until any parasitic oscillation is just damped out.

Tinnitus can occur at any pitch corresponding to a site of OHC degradation leading to a shift in the direction of IHC excitation. However, when investigators talk of high pitch and low pitch, what they mean and what any sufferer may mean may be quite different because of the different sense in which the term “pitch” is used. The variety of forms of damage range from punctate lesions leading to percepts mimicking pure tones, to damage over a length of the cochlear partition leading to a wideband hiss. In the case of punctate lesions the “masking” which they perform on real signals will not have any spread. In terms of the model such “pseudo-masking” is not true masking which requires OHC elongation.

Like pain, the severity of tinnitus appears to have a significant attentional component leading to the wide diversity of outcomes of reported investigations. It is observed that the mammalian cochlea receives innervation which theoretically may assist in the selection process at the mechanical level. However, it is known that the process is important in selecting very low level sounds from background noise and that when this selection process becomes disabled subjects have difficulty in listening as distinct from hearing. Tinnitus may be regarded as an increase in the internal noise of the ear which, like auditory deficits, becomes clinically significant when the sufferer can no longer select any external signal. This condition is hypothesised to arise when the range of tectorial membrane baseline regulation decreases to the extent that the OHC cannot generate elongatory displacements to hyperpolarise the IHCs for frequencies not present in the external signal. On the other hand, by attending to the tinnitus, the brain invokes a motor program which is hypothesised increases the contrast in the region of the tinnitus relative to the surrounding region. The motor program is hypothesised to work to shorten the OHC in the region which has the effect of spreading the excitation region and stimulating a larger number of IHCs. If such a program could be demonstrated in physiological experiment it would go a long way to explaining the “uncertainty principle” which tends to interfere with experiments on tinnitus, and even with individual responses to questionnaires. The act of thinking about the symptoms, for example, “How loud is the tinnitus?”, modifies the outcome of the experiment.

A model has been presented which establishes a framework for further physiological investigation and at the same time establishes a basis for establishing a stronger understanding of the origins of hearing loss and tinnitus. Many features of tinnitus have a potential explanation in terms of accumulated damage to the motor cells in the cochlea.

ACKNOWLEDGEMENTS

The author thanks N. Murray, P.T. Michie and J. Macrae for helpful comments on the text and particularly to P.T. Michie for contributions to the section on selective attention.

REFERENCES

Aran, J.-M., (1981). Electrical stimulation of the auditory system and tinnitus control. In A. Shulman (chrmn) Tinnitus. Proceedings of the first international tinnitus seminar. J. Laryngol and Oto., Suppl. 4, 153–163.
Ashmore J.F., (1987). A fast motile response in guinea-pig outer hair cells: The cellular basis of the cochlear amplifier. J. Physiol. 388, 323-347.
Axelsson, A. and Ringdahl, A., (1989). Tinnitus — a study of its prevalence and characteristics. Brit. J. Audiol., 23, 53-62.
Axelsson, A. and Barrenäs, M.-L., (1992). Tinnitus in Noise-Induced Hearing Loss. In: The Effects of Noise on the Auditory System, A. Dancer, D. Henderson, R.J. Salvi, and R.P. Hamernik (Eds.) B.C. Decker, Toronto. Ch. 23, 269–276.
Bailey, Q. (1979). Audiological aspects of tinnitus. Australian J. Audiol. 1, 19–23.
Békésy, G. von. (1960). Experiments in hearing. McGraw-Hill, N.Y.
British Tinnitus Association, 14-18 West Bar Green, Sheffield S1 2DA, England.
Bohne, B.A., Yohman, L. and Gruner, M.M., (1987). Cochlear damage following interrupted exposure to high-frequency noise. Hear. Res., 29, 251–264.
Brownell, W.E., Bader, C.R., Bertrand, D., de Ribaupierre, Y. (1985). Evoked mechanical responses of isolated cochlear hair cells. Science, 227, 194–195.
Brownell, W.E. and Shehata, W.E., (1990). The effect of cytoplasmic turgor pressure on the static and dynamic mechanical properties of outer hair cells. In: The mechanics and biophysics of hearing. Lecture notes in biomathematics. Edited by P. Dallos, C.D. Geisler, J.W. Matthews, M.A. Ruggero and C.R. Steele, Springer-Verlag, New York, 87, 52–59.
Brundin, L. Flock, Å. and Canlon, B. (1989). Sound-induced motility of isolated cochlear outer cell is frequency specific. Nature, 342, 814–816.
Brundin, L., Flock, Å., Khanna, S.M. and Ulfendahl. M., (1991). Frequency-specific position shift in the guinea pig organ of Corti. Neuroscience Letters, 128, 77– 80.
Brundin, L., Flock, B. and Flock, Å. (1992). Sound induced displacement response of the guinea pig hearing organ and its relation to the cochlear potentials. Hear. Res., 58, 175–184.
Burns, E.M., (1984). A comparison of variability amoung measurements of subjective tinnitus and objective stimuli. Audiol., 23, 426–440.
Canlon, B. (1987). Acoustic overstimulation alters the morphology of the tectorial membrane. Hear. Res. 30, 127-134.
Canlon, B., Miller, J., Flock, Å. and Borg, E. (1987). Pure tone overstimulation changes the micromechanical properties of the inner hair cell stereocilia. Hear. Res. 30: 65-72.
Canlon, B., Brundin L., and Flock, Å. (1988). Acoustic stimulation causes tonotopic alterations in the length of isolated outer hair cells from the guinea pig hearing organ. Proc. Natl. Acad. Sci. USA, 85, 7033–7035.
Chung, D.Y., Gannon, R.P. and Mason, K. (1984). Factors affecting the prevalence of tinnitus. Audiol., 23, 441–452.
Coles, R.R.A., David, A.C. and Haggard, M.P. (1981). Epidemiology of tinnitus. Ciba Fdn. Symp. 85, 16–34.
Cooper, N.P. and Rhode, W.S., (1992). Basilar membrane mechanics in the hook region of cat and guinea-pig cochleae: Sharp tuning and nonlinear in the absence of baseline position shifts. Hearing Res., 63, 163–190.
Crane, H.D., (1982). IHC-TM connect-disconnect and efferent control. V. J. Acoust.Soc. Am., 72, 93–101.
Davis, H., (1954). Tinnitus– Physiological Aspects Trans. Amer. Acad. Opthalmol. Otol., 58, 527–528.
Davis, H., (1983). An active process in cochlear mechanics. Hearing Res. 9, 79–90.
Day, R.O., Graham, G.G., Bieri, D., Brown, M., Cairns D., Harris, G., Hounsell, J., Platt- Hepworth, Reeve, R., Sambrook, P.N., and Smith, J. (1989). Concentration-response relationships for salicylate-induced ototoxicity in normal volunteers. Br. J. clin. Pharmac., 28, 695–702.
Engstrom, H., Ades, H.W. and Bredberg, G. (1970). Normal structure of the organ of Corti and the effect of noise-induced cochlear damage. In: Sensorineural hearing loss, edited by G.E.W. Wolstenhome and J. Knight, Williams and Wilkins, Baltimore, 127– 156.
Evans, B.N., (1990). Fatal contractions: Ultrastructural and electromechanical changes in outer hair cells following transmembranous electrical stimulation Hear. Res., 45, 265- 282.
Evans, B.N., Hallworth, R. & Dallos, P. (1991). Outer hair cell electromotility: The sensitivity and vulnerability of the DC component. Hear. Res. 52, 288-304.
Farmer, J.C. (1977). Diving injuries to the inner ear. Ann. Otol. Rhinol. Laryngol. (Suppl.), 86(1 Pt.3), 1–20.
Fowler, E.P.(1944). Head noises in normal and disordered ears. Arch.Otolaryngol.39, 498- 503.
Froehlich, P., Collet, L., Chanal, J-M. and Morgon, A. (1990) Variability of the influence of a visual task on the active micromechanical properties of the cochlea. Brain Res. 508, 286-288.
Gaumond, R.P., Molnar, C.E., and Kim, D.O., (1982). Stimulus and recovery dependence of cat cochlear nerve fiber spike discharge probability. J. Neurophys. 48, 856– 873.
Geisler, C.D., Rhode, W.S. and Kennedy, D.T., (1974). The responses to tonal stimuli of single auditory nerve fibers and their relationships to basilar membrane motion in the squirrel monkey. J. Neurophysiol., 37, 1156–1172.
Giard, M.-H., Bouchet, P., Collet, L., and Pernier, J. (1992). Auditory selective attention at the cochlear level. Presented at the 5th International Conference on Cognitive Neurosciences (ICON V) Jerusalem, Israel, June 14-19, 1992.
Gold, T. (1948). Hearing II. The physical basis of the action of the cochlea. Proc. Royal Soc. Lond. B: 135, 492–498.
Harding, G.W., Baggot, P.J. and Bohne, B.A., (1992). Height changes in the organ of Corti after noise exposure. Hear. Res. 63, 26–36.
Harris, S., Brismar, J. and Cronqvist, S. (1979). Pulsatile tinnitus and therapeutic embolization. Acta Otolaryngol., 88, 220–226.
Hazell, J.W.P. (1979a). Tinnitus. Brit. J. Hosp.Med. 22, 468–471.
Hazell, J.W.P. (1979b). Tinnitus Research. The Current Position. Hearing, 34(1), 10-15 (reprinted as RNID pamphlet).
Hazell, J.W.P. (1979c). A Tinnitus Synthesiser. Physiological Considerations. Proceedings 1st International Tinnitus Seminar, New York, 8 June 1979.
Hazell, J.W.P. (1981). Patterns of tinnitus: Medical audiologic findings. J. Laryngol.Otol., Suppl. 4, 39–47.
Hazell, J.W.P., Williams G.R. and Sheldrake, J.B. (1979). Tinnitus Maskers – Success and Failures. A Report on the State of the Art. Proceedings 1st International Tinnitus Seminar, New York, 8 June 1979.
Helmholtz (1875:1954). On the sensations of tone. Dover, N.Y.
Hernandez-Peon, R. (1966). Physiological mechanisms in attention. In: Frontiers in Physiological Psychology, R.W.Russell editor. Academic Press, New York,121- 147.
Hirschhorn, T.N. and Michie, P.T. (1990). Brainstem auditory evoked potentials (BAEPs) and selective attention revisited. Psychophysiology, 27, 495–512.
Holley M.C. and Ashmore, J.F. (1990). Spectrin, actin and the structure of the cortical lattice in mammalian cochlear outer hair cells. J. Cell Sci., 96, 283–291.
House, J. W. (1978). Treatment of severe tinnitus with biofeedback training. Laryngoscope, 88, 406–412.
Iwasa, K.H., Minxu, Li, M.J. and Kachard, B. (1991). Stretch sensitivity of the lateral wall of the auditory outer hair cell from the guinea pig. Neurosci. Letters, 133, 171– 174.
Jastreboff, P.J. (1990). Phantom auditory perception (tinnitus): Mechanisms of generation and perception. Neurosci. Res., 8, 221–254.
Jastreboff, P.J., Brennan, J.F., Coleman, J.K. and Sasaki, C.T., (1988). Phantom auditory sensation in rats: An animal model for Tinnitus. Behavioral Neurosci, 102, 811- 822.
Johnston, W.A. & Dark, V.J. (1986). Selective attention. Annual Review of Psychology, 37, 43–75.
Kachar, B., Kalinek, F., Iwasa, K., Holley, M. and Lim, D. (1991). A novel membrane-based mechanism for high frequency shape changes in outer hair cells. Presented at the 28th Workshop on Inner Ear Biology, Tübingen, Sept. 8-11.
Kemp. D.T. (1989). Otoacoustic emissions: basic facts and applications. Audiology in practice, 6/3. Excerpta Medica Medical Communications BV Amsterdam.
Kiang, N. Y.-S., Watanabe, T., Thomas, E.C. and Clarke, L.F., (1965). Discharge patterns of single fibres in the cat’s auditory nerve. MIT Press, Cambridge, Massachusetts.
Kim, D.O. (1984). Functional roles of the inner- and outer-hair-cell subsystems in the cochlea and brainstem. In: Hearing Science: Recent Advances, edited by C.I. Berlin, College- Hill Press, San Diego, CA, 241–262.
Kinchla, R.A. (1992). Attention. Annual Review of Psychology, 1992, 43, 711–742.
Konishi, T. and Salt, A.N. (1983). Electrochemical profile for potassium ions across the cochlear hair cell membranes of normal and noise-exposed guinea pigs. Hear.Res. 11, 219–234.
LePage, E.L., (1981). The role of nonlinear mechanical processes in mammalian hearing. Ph.D. Thesis. The University of Western Australia.
LePage, E.L., (1987). Frequency-dependent self-induced bias of the basilar membrane and its potential for controlling sensitivity and tuning in the mammalian cochlea. J.Acoust.Soc.Am.,82,139-154.
LePage, E.L., (1989). Functional role of the olivocochlear bundle: A motor unit control system in the mammalian cochlea. Hear.Res., 38, 177-198.
LePage, E.L., (1990). Helmholtz revisited: direct mechanical data suggest a physical model for dynamic control of mapping frequency to place along the cochlear partition. In: The Mechanics and Biophysics of Hearing, P.Dallos, C.D. Geisler, J.W. Matthews, M.A. Ruggero and C.R. Steele (Eds.). Lecture Notes in Biomathematics, Spinger-Verlag, New York, 87, 278-287.
LePage, E.L., (1991). Hysteresis in cochlear mechanics and a model for variability in noise- induced hearing loss. In: The Effects of Noise on the Auditory System, A. Dancer, D. Henderson, R.J. Salvi, and R.P. Hamernik (Eds.) B.C. Decker, Toronto. Ch. 10, 106- 115.
LePage, E.L. (1992). Models of subjective and obective tinnitus based on outer hair cell turgor. Proc. IVth International Tinnitus Seminar, Edited by J.-M. Aran and R. Dauman, Kugler Press, Amsterdam, 73–78.
LePage, E.L. (1993). In preparation.
LePage, E.L. and Murray, N.M. (1993). In preparation.
LePage, E.L., Reuter, G. and Zenner, H.P. (1993). Summating baseline shifts and mechanical adaptation in a guinea pig cochlear explant shown with two displacement measuring techniques. (Under revision).
LePage, E.L., Reuter, G. and Zenner, H.-P. (1993). Threshold for mechanical response of isolated outer hair cells, (under revision).
Liberman, M.C. (1989). Rapid assessment of sound-evoked olivocochlear feedback: Suppression of compound action potentials by contralateral sound. Hear. Res., 38, 47– 56..
Lim, D.J., (1986). Functional structure of the organ of Corti: A review. Hear. Res., 22, 117– 146
Loeb, M. and Smith, R.P., (1967). Relation of induced tinnitus to physical characteristics of the inducing stimuli. J. Acoust.Soc.Am., 42 453–455.
Lyon, R.F. (1990). Automatic gain control in cochlear mechanics. In: The Mechanics and Biophysics of Hearing., Springer-Verlag, Lecture notes in biomathematics, 87, 395– 402.
Man, A. and Naggan, L. (1981). Characteristics of Tinnitus in Acoustic Trauma. Audiology 20, 72–78.
Meikle, M and Taylor-Walsh, E., (1984). Characteristics of tinnitus and related observation in over 1800 tinnitus patients. J. Laryngol. Otol. Suppl., 9, 17–21.
Meikle, M. (1991a). The onset of tinnitus: causal and precipitating factors. In: Proceedings of the 3rd Bi-Annual Workshop, Tinnitus – Assessment and Re-Habilitation, Melbourne, Australia, 15-17 March 1991, 1-3.
Meikle, M. (1991b). How tinnitus is related to hearing impairment. In: Proceedings of the 3rd Bi-Annual Workshop, Tinnitus – Assessment and Re-Habilitation, Melbourne, Australia, 15-17 March 1991, 4-6.
Meikle, M. (1991c). The issue of severity: Evaluating the severity of Tinnitus. In: Proceedings of the 3rd Bi-Annual Workshop, Tinnitus – Assessment and Re- Habilitation, Melbourne, Australia, 15-17 March 1991, 7-9.
Meikle, M., Griest, S.E., Press, L., and Stewart, B.J. (1991). Relationships between tinnitus and audiometric variables in a large sample of tinnitus clinic patients. Proc. IVth International Tinnitus Seminar, Edited by J.-M. Aran and R. Dauman, Kugler Press, Amsterdam.
Michie, P.T., Crawford, J.M., Bearpark, H.M. and Glue, L.C. (1990). The nature of selective attention effects on auditory event-related potentials. Biological Psychology, 30, 219–250.
Møller, A.R. (1984). Pathophysiology of Tinnitus. Ann. Otol. Rhinol. Laryngol., 93, 39- 44.
Murray, N.M. and LePage, E.L. (1990). Otoacoustic emissions — sounds coming out of the ear. Australian Hearing and Deafness Review, 7, 10–14.
Murray, N.M. and LePage, E.L. (1991). Estimated maximum acceptable background noise levels for audiometric testing with sound-excluding earphone enclosures — an update. Aust. J. Audiol., 13, 47–52.
Näätänen, R. (1990). The role of attention in auditory information processing as revealed by event-related potentials and other brain measures of cognitive function. Behavioral and Brain Sciences, 13, 201–288.
Neely, S.T. and Kim, D.O., (1983). An active cochlear model showing sharp tuning and high sensitivity. Hear.Res., 9, 123–130.
Nodar, R.H. and Graham, J.T. (1965). An investigation of frequency characteristics of tinnitus associated with Meniere’s disease. Arch Otolaryngol., 82, 28-31.
Oatman, L.C. (1971). Role of visual attention on auditory evoked potentials in unanesthetized cats. Experimental Neurology, 32, 341–356.
Oatman, L.C. (1976). Effects of visual attention on the intensity of auditory evoked potentials. Experimental Neurology, 51, 41–53.
Oatman, L.C. and Anderson, B.W. (1977). Effects of visual attention on tone burst evoked auditory potentials. Experimental Neurology, 57, 200-211.
Oatman, L.C. and Anderson, B.W. (1980). Suppression of the auditory frequency following response during visual attention. Electroenceph.Clin. Neurophysiol., 49, 314– 321.
Patuzzi, R., Sellick, P.M. and Johnstone, B.M., (1984). The modulation of the sensitivity of the mammalian cochlea by low frequency tones. I. Primary afferent activity. Hear.Res. 13, 1–9.
Penner, M.J. (1983). Variability in matches to subjective tinnitus. J. Speech and Hearing Res., 26, 263–267.
Penner, M.J. (1992). Linking spontaneous otoacoustic emissions and tinnitus. Brit. J. Audiol., 26, 115-123.
Plinkert, P.K., Gitter, A.H., and Zenner, H.-P. (1990). Tinnitus associated with spontaneous otoacoustic emissions. Active outer hair cell movements as common origin. Acta Otolaryngol. (Stockh), 110, 342–347.
Pirilä, T, Jounio-Ervasti, K. and Sorri, M., (1991). Hearing Asymmetry among left-handed and right-handed persons in a random population. Scand. Audiol. 20, 223–226.
Portmann, M., Cazals, Y., Negrevergne, M., and Aran, J.-M., (1979). Temporary tinnitus suppression in man through electrical stimulation of the cochlea. Acta Otolaryngol., 87, 294–299.
Puel, J.-L., Bonfils, P. and Pujol, R. (1988). Selective attention modifies the active micromechanical properties of the cochlea. Brain Res. 447, 380–383.
Reed, G.F., (1960). An audiometric study of two hundred cases of subjective tinnitus. Arch. Otolaryngol., 71, 95–104.
Saunders, J.C., Canlon, B. and Flock, Å. (1986). Changes in stereocilia micromechanics following overstimulation in metabolically blocked hair cells. Hear.Res. 24: 217- -226.
Sellick, P.M., Patuzzi, R.B. and Johnstone, B.M. (1982). Measurement of basilar membrane motion in the guinea pig using the Mössbauer technique. J.Acoust.Soc. Am. 72, 131– 141.
Shulman, A., Goldstein, B.A. and Bhathal, B., (1992). Spontaneous evoked otoacoustic emissions and tinnitus — its correlation/uncorrelation with specific clinical types of tinnitus. Proc. IVth International Tinnitus Seminar, Edited by J.-M. Aran and R. Dauman, Kugler Press, Amsterdam, 95–102.
Siegel, W., (1976). The paradox of pitch categories. J. Acoust.Soc.Am., 60, S6.
Spoendlin, H., (1970). Structural basis of peripheral frequency analysis. In: Frequency analysis and periodicity detection in hearing, edited by R. Plomp and G. F. Smoorenburg, Sijthoff, Leiden, 2–40.
Stephens, S. S. and Davis, H. (1938:1983). Hearing — its psychology and physiology. Published by the American Institute of Physics, Inc., New York, NY, 351–352.
Thorne, P.R., Duncan, C.E. and Gavin, J.B. (1986). The pathogenesis of stereocilia abnormalities in acoustic trauma. Hear. Res. 21, 41–50.
Tonkin, J.P. and Fagan, P. (1975). Rupture of the round window membrane. J. Laryngol. Otol. 89(7), 733-756.
Tonndorf, J. (1981). Stereociliary Dysfunction, a Cause of Sensory Hearing Loss, Recruitment, Poor Speech Discrimination and Tinnitus. Acta Otolaryngol., 91, 469– 479.
Tyler, R.S., Aran, J-M. and Dauman, R., (1992). Recent advances in Tinnitus. Am.J. Audiol.1/4, November, 36–44.
Vernon, J. (1991). Testing and Evaluation of Tinnitus. In: Proceedings of the 3rd Bi-Annual Workshop, Tinnitus – Assessment and Re-Habilitation, Melbourne, Australia, 15-17 March 1991, 10–15.
Vernon, J., Johnson, R., Schleuning, A. and Mitchell, C. (1980). Masking and tinnitus. Ear and Hearing 1, 5–9.
Vernon, J., Schleuning, A., Odell, L. and Hughes, F. (1978). A Tinnitus clinic. Ear, Nose and Throat J., April, 58–71.
Vernon, J. and Schleuning, A., (1978). Tinnitus: A new management. Laryngoscope, 88, 413– 419.
Voldrich, L. and Úlehlová, L. (1982). The role of the spiral ligament in cochlear mechanics. Acta Otolaryngol., 93, 169–173.
Voldrich, L. and Úlehlová, L. (1987). Cochlear micromechanics. Acta Otolaryngol., 103, 661- -664.
Warr, W. B. (1992). Organisation of olivocochlear efferent systems in mammals. In: Mammalian Auditory Physiology: Neuroanatomy. Springer Handbook of Auditory Research, 410–448.
Wegel, R.L. (1931). A study of tinnitus. Arch. Otolaryngol., 14, 158–165.
Wilson, S. (1979). Mystery of people who hear the hum. New Scientist, 13th December, 868– 870.
Zenner, H.P., (1986). Motile responses in outer hair cells. Hear.Res., 22, 83-90.
Zenner, H.P., Zimmermann, U. and Schmitt, U., (1985). Reversible contraction of isolated mammalian cochlear hair cells. Hear.Res., 18, 127-133.
Zenner, H.P., Zimmermann U. and Gitter, A.H., (1987). Fast motility of isolated mammalian auditory sensory cells. Biochem.Biophys.Res.Comm., 149, 304–308.
Zwicker, E.(1977). Masker period patterns produced by very-low-frequency maskers and their possible relation to basilar membrane displacement. J. Acoust. Soc. Am. 61, 1031– 1040.