Monday 14 November 2011

Dead-spots revisited....

Ted Venema's excellent article on Cochlear dead-spots: of particular interest is the reverse sloping SN loss as in figure 5, which I've run into quite recently. It doesn't 'fit' into the normal models we are used to for gain - too much LF just masks out the better high frequencies, for what seems like not a lot of return in spatial awareness/loudness etc.

http://www.hearingreview.com/issues/articles/2005-03_06.asp


Identifying Cochlear Dead Spots

by Ted H. Venema, PhD
A primer on cochlear function as it relates to cochlear dead regions
How cochlear dead regions can be identified, what kinds of hearing losses are often associated with them, and why
Editor’s Note: This article1 and the interview that follows2 were originally published in the July/August 2003 (Vol 52, No 4) and March/April 2004 (vol 53, No 2) editions of The Hearing Professional, the official journal of The International Hearing Society (IHS). The articles are adapted and reprinted here with permission.
The cochlea is the “retina of the ear.” It changes sound into electrical impulses, and those impulses are the language the brain understands. Just as dead areas of the retina can create holes in one’s field of vision, dead hair cell areas of the cochlea can produce audiometrically useless frequencies. At these frequencies, hearing aid amplification does little or no good.
Brian Moore, PhD, a Cambridge University researcher in areas of psychoacoustics, has developed a protocol, called the Threshold Equalizing Noise (TEN) test, to clinically identify cochlear dead spots. The most interesting thing about this test is not its clinical utility and how well and consistently it identifies cochlear dead spots. Instead, the best thing about this test is that, in order to understand its rationale and how it works, one is forced to understand how the cochlea works.
Outer and Inner Hair Cell BasicsBy way of a brief overview, each cochlea contains one row of approximately 3,000 inner hair cells and 3-5 rows of about 12,000 outer hair cells (Figure 1).

Figure 1. The hairs or stereocilia of the outer hair cells are jammed into the underside of the tectorial membrane, while those of the inner hair cells are not. When soft sounds enter the cochlea, the test-tube shaped outer hair cells shrink, thus pulling the tectorial membrane down, so the stereocilia of the jug-shaped inner hair cells can be bent or sheared.
The jug-shaped inner hair cells send all sound information to the brain; without them we are totally deaf. These hair cells have one fundamental limitation, however: they cannot sense sounds softer than conversational speech.3 More specifically, the inner hair cells cannot sense signals below about 50 dB SPL for the low frequencies and below about 65 dB SPL for the high frequencies.4
The outer hair cells work in the opposite direction; that is, they receive messages from the brain and from within the cochlea telling them to rapidly stretch or shrink. These test-tube shaped outer hair cells are the active mechanism of the cochlea—the moving parts. Their movements help the inner hair cells sense soft sounds.
Sound hitting the eardrum results in a traveling wave of fluid motion inside the cochlea, thus causing a ripple along the floor upon which the hair cells stand (known as the basilar membrane). The stereocilia of the inner hair cells bend or become sheared where the wave peaks. This is what stimulates the hair cells at the cochlea’s wide base (high frequencies) or narrow apex (low frequencies) or at some unique place in between. In short, the wave grows (and slows) as it goes up the spiral-shaped cochlea until it reaches peak amplitude and stops. By the way, the main reason that the wave actually gets a peak in the first place is because it meets impedance along its travels up the spiral. As it is forced to slow down along its spiral route, its energy has to go somewhere; hence, its peak of “vertical” amplitude.
The wave’s peak is even further defined as a result of the action of the outer hair cells. The stretching/shrinking action of the outer hair cells temporarily alters the basilar membrane on either side of the peak. This mechanically forces the peak into a sharper point that, in turn, increases our ability to distinguish between frequencies that are close together. In someone with outer hair cell damage, the traveling wave peak is dull and rounded, and their ability to distinguish frequencies that are close together is diminished (Figure 2). Is it any wonder that those with sensorineural hearing loss (SNHL) and damaged outer hair cells have difficulty separating speech from background noise?

Figure 2. Without the action of the outer hair cells, the traveling wave has a dull and rounded peak. This passive traveling wave stimulates many adjacent frequencies simultaneously. The sharpening of the peak is accomplished with the action of the outer hair cells, and this increases the ability to distinguish between frequencies that are close together.
Why WDRC is Often RecommendedIt can safely be said, therefore, that the most common type of damage to the ear is damage to the outer hair cells. This results in the most common type of hearing loss: a moderate SNHL, where soft sounds below conversational speech (50–65 dB HL) are inaudible, yet 90-100 dB HL sounds are perceived as loud as they would be to someone with normal hearing. For this person, hearing aids should amplify the soft sounds significantly and amplify louder sounds by progressively smaller and smaller increments. Wide Dynamic Range Compression (WDRC) hearing aids that accomplish this are specifically intended to imitate what the outer hair cells once did. Outer hair cells begin their work for sounds below 50–65 dB SPL; hence, the knee-point of WDRC is most often found at input levels of around 50 dB, as well.

Figure 3. These idealized, schematic shapes represent three traveling wave envelopes. The top shows a normal traveling wave envelope, resulting from stimulation of two tones different in frequency. The middle shows a traveling wave envelope that is reduced in amplitude. Note also that the peaks are rounded, due to outer hair cell damage. The bottom shows what would happen with amplification. The original traveling wave size or amplitude is restored, but the peaks are still rounded. In other words, the ability to separate speech from background noise has not been restored.
It must be emphasized, however, that no matter how good a hearing aid is, it cannot restore a normal-functioning cochlear traveling wave (Figure 3). When the sharpened peak of a traveling wave becomes dull, it is dull for good. Hearing aids can only amplify and, by so doing, can only enlarge a diminished traveling wave. They cannot restore one’s original sharp frequency resolution or the ability to separate frequencies that are close together. Amplification only increases audibility of sounds, but does not come close to the majesty and wonder of the healthy cochlea.
Essential Cochlear ConceptsCochlear dead spots occur where there is complete destruction to both the inner and outer hair cells. As mentioned earlier, the gain provided by the outer hair cells to very soft input levels is about 50 dB for the low frequencies and about 65 dB for the high frequencies. Moore says that additional inner hair cell damage can only result in another 25-30 dB of hearing loss beyond 50 dB in the lows and 65 dB in the highs.4 This would make the maximum hearing loss possible from only hair cell damage about 75-80 dB in the lows and 95 dB in the highs.
An important fact to keep in mind is that the traveling wave is asymmetrical in shape (Figures 2-4). This concept is essential in understanding Moore’s test for cochlear dead spots. The traveling wave has a long tail towards the cochlea’s wide base (high-frequency region) and a steep front that is facing the cochlea’s low-frequency apex. This is our hearing physiology and explains “the upward spread of masking,” or that low frequencies mask high frequencies better than vice versa (Figure 4).

Figure 4. The traveling wave is asymmetrical in shape. Soft, high-frequency stimulation results in a small traveling wave at the base of the cochlea (right), which would easily be overcome or masked by the wave resulting from intense low-frequency stimulation at the apex (left). The reverse would not be true. Intense, high-frequency stimulation results in a traveling wave confined to the base of the cochlea (right) and, thus, it would not interfere with the wave resulting from soft low-frequency stimulation (left).
TEN Test ProceduresMoore’s TEN test for cochlear dead spots is available on a CD that can be played over a two-channel audiometer (http://hearing.psychol.cam.ac.uk/dead/ dead.html). The CD plays puretones, as well as a single broadband masking noise (noise that includes all audiometric frequencies). This broad-band noise is quite different from the narrow bands of noise used in our audiometers. The puretones and the masking noise have to be directed toward the same ear, and this can only be done with a two-channel audiometer. One can separately adjust the intensity of the tones and the masking noise by way of the intensity controls on the audiometer and send both to either the right or left ear. You begin by testing for thresholds of the puretones from the CD, and then testing the same ear for thresholds while the masking noise is presented into that ear (ie, ipsilateral masking).
When this article was first published in The Hearing Professional, the TEN test puretones and broadband masking noise were all calibrated in dB SPL, not HL. This is important to note when using the CD. If a client has normal hearing, the thresholds on the typical audiogram will look a bit like a barn roof (Figure 5), with best thresholds showing for the mid frequencies and borderline-to-mild hearing loss appearing for the low frequencies and high frequencies. The reason for this audiogram shape is that normal-hearing ears are most sensitive to frequencies between 1-4 kHz. Incidentally, this is why equalizer buttons on some stereo systems are shaped like a smile; we need the artificial boost for the lows and highs in order to hear all of the frequencies at equal loudness levels. To be sure, there are some complicated calibration issues that would need to be addressed in order to accurately translate the TEN test results to the typical audiogram with which we are all familiar. These, however, are beyond the scope of this introductory article. As of late last year, the TEN test5 is available in calibration of dB HL. This makes it easier to relate test results directly to the audiogram.
The procedure for Moore’s TEN is to first test for hearing thresholds for the puretones from the TEN CD in quiet, and then to retest for the same thresholds in the presence of ipsilateral masking. When using the test, be sure to go to a level whereby the better thresholds of the person are affected (ie, made worse by the masking). Compare the unmasked thresholds to the masked thresholds. The TEN should affect the better thresholds because it is audible to the person at these frequencies. The worst thresholds, however, should not be affected because the TEN should not be audible to the person at these frequencies. If they are, then suspect cochlear dead spots at these frequencies.

Figure 5. Moore’s TEN test on someone with normal hearing. Note that the thresholds from the puretones of the TEN CD produce a convex, “barn-roof” shape. This is due to the calibration of the CD tones in dB SPL, whereas the audiogram is measured in terms of dB hearing loss. Note also, however, that the thresholds masked by 30 dB TEN are only those thresholds that can hear it and not, for example, the worst threshold at 8000 Hz.
For a normal-hearing person, for example, 30 dB of the TEN from the CD should affect most thresholds (Figure 5). If the decibels on the audiogram were in dB SPL, all thresholds would be elevated or shifted to show a flat 30 dB hearing loss. On the typical audiogram (where dB HL rather than dB SPL is used), the barn-roof shaped thresholds for the normal-hearing person are still affected by the TEN. Figure 5 shows that, for any frequency where the broadband TEN is audible, thresholds within the TEN are shifted to at least the intensity of the TEN, so they are simply “pushed lower” down on the audiogram.
For someone with hearing loss, the main idea is to provide enough TEN masking so that the better thresholds are shifted, and determine if the worse thresholds are affected. For example, consider someone with a mild hearing loss for the low frequencies and a moderate hearing loss for the highs. The puretones played from the CD will show a similar trend; namely, better hearing for the low frequencies than for the high frequencies (Figure 6). Note that in the presence of ipsilaterally presented TEN, at 30 dB the thresholds for the puretones from the CD are tested again. A shift for the low-frequency thresholds appears, but this does not occur for the high-frequency thresholds. This only makes sense, because the person was not even able to hear the broadband TEN in the high frequencies.

Figure 6. The ipsilateral masking with 30 dB TEN affects the better low-to-mid frequency thresholds of the sloping SNHL, because the TEN is audible to the person at these frequencies. The TEN does not, however, affect the high-frequency thresholds because the TEN is not audible to the person at these frequencies. This would indicate a typical high-frequency SNHL that is due to damaged hair cells at these frequencies, but not due to high-frequency cochlear dead spots.
Similarly, according to Moore,4 if you masked the worst thresholds by their own minimum masking levels with the TEN, these thresholds should theoretically only be shifted to the level of the TEN used to mask them. Consider now that these worst thresholds are caused by cochlear dead spots: in this case, the minimum TEN level would actually shift the worst thresholds at least 10 dB beyond the decibel levels of the TEN itself. This is because these “worst” audiometric thresholds are not real; they are caused by cochlear dead spots and, thus, are actually far worse than the audiogram would suggest!
More “Suspicious” CasesOne type of SNHL that should give rise to suspicion of cochlear dead spots is a moderate degree of reverse hearing loss; another type is a severe degree of precipitous high-frequency hearing loss. For either type, amplification for the worst thresholds might not be the best course of action. For example, excess high-frequency gain can result in feedback for the person with precipitous hearing loss.
Moderate Reverse-Sloping SNHL. Be suspicious of reverse-sloping SNHL, as it could be indicative of low-frequency dead regions. It is very possible that the person could be completely deaf in the low frequencies; however, due to the asymmetric shape of the traveling wave, only a moderate reverse hearing loss may be revealed.
Consider someone who has completely dead inner and outer hair cells for the frequencies below 1000 Hz. In this case, intense low-frequency stimulation results in a traveling wave with a peak at the apical (low-frequency) hair cell region of the cochlea. The long tail of the traveling wave, however, may still extend into the healthy mid-frequency regions (Figure 7). Even though these low-frequency hair cells might be dead, a moderate amount of low-frequency stimulation might still excite living mid-frequency hair cells, thus causing the person to raise a hand, indicating he/she indeed heard a tone. In this case, the person might be “hearing” these low frequencies with their healthy mid-frequency hair cells, and not by means of their dead low-frequency hair cells!

Figure 7. Low-frequency dead spots may reveal only a moderate, low frequency SNHL with a reverse audiogram. Due to the long tail of the traveling wave, intense, low-frequency stimulation may “excite” the healthy mid-frequency hair cell regions (gray area of traveling wave diagram at top of figure). In this case, the person will indicate a response, but it will not truly arise from hearing in the low-frequency hair cell regions.
Severe Precipitous High-Frequency SNHL. Severe precipitous high-frequency SNHL can indicate high-frequency cochlear dead regions. Here, it is possible that the high-frequency thresholds do not truly arise from damaged high-frequency hair cells. On the contrary, these thresholds might result from indirect stimulation of low-frequency hair cells.
High-frequency stimulation would have to be quite intense to enable the steep front of the traveling wave to extend into the living, healthy mid-frequency hair cell regions. The steep slope of the precipitous high-frequency hearing loss thus might reflect the steep front of the traveling wave as it occurs in the cochlea(Figure 8). In this case, even though the high-frequency hair cells might be totally dead, an intense high-frequency tone might stimulate mid-frequency hair cells, causing the person to raise a hand indicating he/she heard something. The high-frequency thresholds are not truly indicative of high-frequency sensitivity; rather, they are a result of indirect stimulation of remote living hair cell regions.

Figure 8. High-frequency dead spots may reveal an audiogram showing a precipitous, pronounced degree of high-frequency SNHL. Due to the steep front of the traveling wave, intense, high-frequency stimulation may “excite” the healthy mid-frequency hair cell regions (gray area of traveling wave diagram at top of figure). In this case, the person will indicate a response, but it will not truly arise from high-frequency hair cell regions.
Reverse and Precipitous High-Frequency SNHLWith cochlear dead regions of hair cells, one actually hears by means of remote, living hair cells. This is called “off-frequency hearing.” A small amount of ipsilaterally presented broad-band TEN masking noise would elevate the normal (or better) thresholds in the living hair cell regions. In the reverse loss, it would make the good mid-to-high-frequency thresholds worse; in the precipitous high-frequency SNHL, it would make the good low-to-mid frequency thresholds worse. If the reverse or precipitous high-frequency hearing loss were due to cochlear dead spots, the TEN would, however, also elevate the thresholds for the worst thresholds, even though the TEN would theoretically be inaudible to the person at these frequencies!
Specifically, in the case of reverse SNHL, the ipsilaterally presented TEN would shift the low-frequency thresholds, even if these thresholds were greater than the intensity of the masking noise. In the case of the precipitous high-frequency SNHL, the TEN would also make the high-frequency thresholds worse, even if they were greater than the intensity of the masking noise.
Using conventional thinking, we would consider this impossible because, at these thresholds, the listener should not even be able to hear the masking noise. The reason why these thresholds are affected, however, is that, when one has dead hair cell regions at any frequency, one hears tones in these dead areas by means of a small piece of the traveling wave that extends into living hair cell regions. If the TEN masking noise does shift the worst thresholds by 10 dB or more, according to Moore, these thresholds are spurious, and do not actually arise from stimulation of damaged hair cells at these frequencies.4 Instead, in these cases, the worst thresholds arise from indirect stimulation of remote, living hair cells at other frequency regions. Therein lies the rub of Moore’s TEN test!
Of course, if the hair cells in question are only damaged and not truly dead, the same ipsilateral TEN masking noise would not shift these worst thresholds. In the case of the reverse SNHL, a small amount of TEN that was enough to shift the better mid-to-high-frequency thresholds would not affect the poorest thresholds at the low frequencies. In the case of the precipitous high-frequency SNHL, a small amount of TEN might indeed shift the good low-frequency thresholds, but would not affect the worst high-frequency thresholds by anywhere near 10 dB. These hearing losses would then, respectively, be a true reverse hearing loss and a true precipitous high-frequency hearing loss.
Figure 9 shows the thresholds for puretones from the TEN CD, followed by the thresholds found in the presence of 50 dB TEN from the CD. The ipsilateral TEN stimulation of 50 dB SPL should not have any effect on the high-frequency thresholds. In fact, the high-frequency thresholds are indeed affected because they show a shift of at least 10 dB. This finding, says Moore, would indicate cochlear dead spots for the high frequencies.4

Figure 9. The ipsilateral masking with 50 dB TEN affects the low-to-mid frequency thresholds of the sloping SNHL because the TEN is audible to the person at these frequencies. However, the TEN also shifts the high-frequency thresholds by at least 10 dB—even though it is not audible to the person at these frequencies. This would indicate a high-frequency SNHL that is due to high-frequency cochlear dead spots. The high-frequency thresholds thus do not truly arise from damaged high-frequency hair cells; rather, they are a result of stimulation of remote hair cells at the low-to-mid frequencies that are responding to intense high-frequency stimulation (ie, what Moore4 refers to as “off-frequency hearing”).
The implications for amplification are important here; namely, don’t focus on amplification in these extreme high-frequency thresholds. It might be best, in this case, to amplify the low-to-mid frequencies as well as the transition of the audiogram where the thresholds drop.
Summing Things UpIn ears with cochlear dead spots, tones are processed in the dead areas by means of living hair cells located on surrounding regions of the basilar membrane—what Moore refers to as “off-frequency hearing.” Some patients report that these tones do not sound natural or tonal in quality, or that puretone stimulation in these regions gives them the perception of a scratch or a tickle. These subjective reports, however, are not always consistent from person to person, even though dead areas might be indicated.4
The TEN test on the CD is not, in my opinion, a required part of any new test battery for our patients. On the other hand, the presence of reverse hearing loss or precipitous high-frequency hearing loss should make dispensing professionals suspicious that cochlear dead spots might exist. Furthermore, it is a good idea to ask these patients what their perceptions of audible tones presented to their worst thresholds are like. These two items—initial suspicions and secondary questions—might really help in our consideration of how much amplification to provide for a patient’s poorest thresholds. Should we provide low-frequency gain and output for reverse SNHL? For precipitous high-frequency SNHL, should we focus on amplifying the worst high-frequency thresholds or should we concentrate on amplifying the transition or steep slope itself?
Not to be forgotten, of course, is the education of ourselves; that is, understanding the TEN test for cochlear dead spots requires an appreciation for the fascinating way in which our cochleae function.

No comments:

Post a Comment