Hearing Aid Research

When contemplating the purchase of a hearing aid to reduce these hearing problems, however, it is very easy for consumers (and professionals as well) to be heavily influenced by the technical descriptions of modern hearing aids, the more so since these descriptions are based on truly impressive technical accomplishments. For example, it would be difficult indeed to resist the allure of such terms as: "64-band spectral enhancement system"; "laser speech enhancement system"; "dynamic speech recoding"; "speech enhancement algorithm". "multi-channel DSP technology"; "digital loudness control"; "adaptive, intelligent narrow-band noise reduction." These are all terms taken from recent issues of trade journals. As one reads them, and lots more besides, one can't help but respond, "yes"," yes", I want them all! They all sound so good! It is necessary, therefore, to keep the overriding purpose of a hearing aid in mind as one tries to sift through all of these marketing appeals and technical descriptions. What this implies is that any development in hearing aid capabilities must be validated by clinical research studies, to determine how well they actually perform in real life.

It does not follow that an impressive engineering feat, even those modeled on the psychoacoustical function of the normal ear, necessarily improves speech comprehension for people with hearing loss. This has to be tested directly. This is the standard I use when I review the status of hearing aid research. It doesn't matter if some hearing aids can break sound down into a hundred or a thousand channels, performing all kinds of impressive analyses at awesome speeds, if this technical capacity does not result in improved hearing performance. As the old saying has it, "the proof of the pudding is in the eating." Using this standard, I can't recall any specific technical advance in hearing aids that in itself produced a "eureka" moment, some engineering marvel that resulted in a dramatic increase in speech perceptual skills companed to what was possible previously. No hearing aid can completely overcome the inherent limitations of an impaired ear. Even the advent of digital signal processing in wearable hearing aids, as powerful and promising as this is, offers basically incremental increases in listening performance. But this is not to be sneered at or minimized. I am not knocking current hearing aid developments; every little, and not so little, bit helps and they can add up to be quite a lot.

But that's the point. They add up and don't revolutionize. There is an evolutional pattern to the history of hearing aid developments: each new technical advance builds on an earlier foundation. Without careful and continuous hearing aid research, however, it is difficult to know whether a specific technical change is successful and should be further investigated and which should be discarded in future hearing aid design. Taking the long view, hearing aid users are clearly better off than they were 50, or even five years ago. Still, this process is marked by a lot ot trial and error, with two or three steps forward or sideways, and one or two steps back. In this view, we can learn from our failures as well as our successes, as long as both are communicated openly and obiectively.

I've reported on this feature a number of times in the past and it is generally a positive development, though perhaps a bit overstated.

With wide dynamic range compression (WDRC), the hearing aid provides different degrees of amplification depending upon the loudness of the sound. 5ofter sounds are given more amplification than louder ones. This capability has led to the elimination of a volume control in many hearing aids; if the hearing aid adjusts the sound automatically, what need is there for a person to do it? And, indeed. there have been a number of studies that found that at low input levels, speech perception scores with the WDRC were higher than that found with linear amplification systems. (In linear amplification, there is a one-to-one relationship between intensity changes in the input sound signal and resulting changes in the output.

But, as one looks more closely at these comparison studies, the results only apply when the experimenter initially sets the degree of amplification (the "gain") and keeps it at that point throughout the study. Under this circumstance, as the input level drops, the gain of the WDRC is automatically increased. Listeners, therefore, still receive a clearly audible signal, since the hearing aid circuit is essentially increasing the volume. With a linear hearing aid, on the other hand, the gain stays the same no matter how soft or intense the input sound.

In the comparison studies that demonstrated the benefits of a WDRC, users of linear hearing aids were not permitted to adjust the volume control when listening to soft sounds. In ordinary life, however, people who wear hearing aids with volume controls will turn them up when listening to soft sounds and down again when the sounds get loud (or to coin my own acronym, "AFC", automatic finger controI).

In a recent study conducted at the University of Washington, the investigators (Pamela Souza and Virginia Kitch) wondered what would happen if the subjects were permitted to adjust the volume control during a comparison of WDRC, linear, and compression limiting (these are also essentially linear) hearing aids. They tested the subjects at three sound levels: soft, medium, and loud. At each one, they permitted the subjects to adjust the volume control for their preferred listening level, the point at which they achieved "maximum clarity."

Under these conditions, there was no difference in speech audibility between the three processing strategies (WDRC, linear, and compression limiting) in quiet or in noise. What is meant by "speech audibility" is the degree to which amplified speech exceeds a person's impaired auditory thresholds (speech perception scores can be deduced from this measure). In other words, according to this study, there would be no difference in speech comprehension in a real-life situation between WDRC and linear hearing aids.

In real-life, however, the situation is somewhat more complex than these results would indicate. On one hand, as we can see from the results of this study, the presumed advantage of WDRC is minimized if a volume control was included in a linear hearing aid. As it happens, most experienced hearing aid users, even those with hearing aids that contain a WDRC circuit would like to have a VC, perhaps in addition to the automatic operation of the WDRC circuit. As a long-time "fiddler" myself (I keep fiddling with the VC), I like having control over the loudness of the sounds, on a moment-to-moment basis and in different types of acoustical circumstances. The loudness choices made during the initial programming of the hearing aids (to soft. medium, and loud sounds) do not necessarily apply at other times in a person's life. Sometimes people want to listen to the sounds softer or louder than the initial programming decisions. They should be given that option.

But, on the other hand, there are people for whom a WDRC is an unalloyed blessing. They may not have the finger dexterity or desire to "fiddle" with the volume control. They want to insert their hearing aids in the morning and forget about them for the rest of the day. And even for people who use and like volume controls, there are those occasions when the input sounds change so often and so drastically that one simply can't fiddle quickly enough. At these times, a WDRC hearing aid would be advantageous. The final conclusion? Maybe we can have it both ways, that is a hearing aid with a WDRC, but one that includes a VC that can override the loudness programming.

This is the hearing aid that sells for $39 and lasts for 40 days (at an assumed 12 hours per day), after which it is discarded and a new one, an exact replica, is obtained for another 40 days, and so on. At this time, I would like to review a very interesting study that compared the Songbird Disposable Hearing Aid (SDHA) to conventional modern hearing aids and to discuss the implications of this research.

This study was conducted at Walter Reed Army Medical Center by Therese Walden and her colleagues. They compared the SDHA to custom fit hearing aids using three different types of measures: one for speech perception; one for self-assessed hearing aid benefit; and one that examined satisfaction for amplification in daily life. All three tests are well known, standardized audiological tools.

The subjects were fitted with their custom hearing aids in accordance with accepted clinical procedures. These include modifying the electroacoustic response of the hearing aid to conform to pre-determined "targets." As in most modern hearing aids, these custom fit hearing aids are capable of a wide range of modifications, thus ensuring that targets can be closely approximated if not reached. The SDHA, on the other hand, comes with only seven different "formats." These include three varied power hearing aids for people with moderately sloping hearing losses and three for people with more sloping high frequency hearing loss. The seventh format is meant for those with a mild sloping hearing loss. The aid is not designed, therefore, for people with severe and greater hearing losses or odd shaped audiometric configurations.

What the examiners did was select one of these seven formats to match, as closely as possible, the responses obtained with the subject's custom-fit hearing aid. None of the matches were perfect. On average, close matches were achieved only in the mid frequencies. At the higher frequencies, the SDHA provided more output, probably because this hearing aid is insented deep in the ear canal.

After wearing the SDHA for an average of two weeks, the subjects were tested and the results compared to those obtained with their custom fit hearing aids. There proved to be no significant difference between the custom fit hearing aid and the SDHA for all three performance measures (speech perception, HA benefit, and HA satisfaction). In other words, the subjects did as well with the disposable hearing aids as the custom hearing aids in all three measures.

There were some downsides to this generally rosy picture of the SDHA. Retention of the deep insertion proved to be a problem for some people. The aid tended to work its way out after a few hours of use. In addition, feedback proved to be a problem for nine of the twelve subjects and this took a bit of effort to rectify.

The examiners recommend additional design modifications to the ear tip to decrease these problems and to increase the number of people who can successfully be fit. Their experience demonstrates that even though these are "disposable" hearing aids, it still takes care and follow-through to make sure they are fit properly.

The generally positive results obtained with the SDHA do raise some important questions. Here we have a hearing aid that can provide only seven different choices in its operating characteristics, yet seems to perform as well as hearing aids that can deliver a multitude of amplification options. One conclusion of the researchers is that for people with predominately high frequency hearing loss, "most can be fit equally well with a limited number of fixed electroacoustic configurations."

What they are implying is that maybe an impaired ear simply can not benefit from the very precise "fine-tuning" now possible in custom fit hearing aids. But what does this say about the current generation of sophisticated (and expensive) hearing aids? Are they being over-engineered? Are all the bells and whistles necessary, or do they just add additional expense without an accompanying increase in actual performance?

I feel that it is too soon to draw this kind of conclusion for several reasons. First, this is only one study; replication, using other custom-fit hearing aids, would be necessary to verify the original results. Second, these results do not apply to people with more severe hearing losses or to those with different audiometric configurations than the subjects in this study. The SDHA was not designed for these people. Third, and perhaps most importantly, the custom hearing aids used by the subjects, to which the SDHA was compared, did not include such potentially valuable options as directional microphones, noise and feedback reduction circuits, telecoils, and direct audio input capability. For some people, these features may well justify the extra cost of a custom fit hearing aid. (We should recall that over a five-year period, the life span of the average custom fit hearing aid, the cost would be about the same as five years worth of the disposable hearing aids.

Right after I finished writing the above, I learned that songbird has introduced a digital version of its disposable hearing aid. Only this one cost $79 and is supposed to last 70 days when used 10 hours a day. For either model, it the aid were used for a longer period of time each day then this would presumably reduce how long it could be used before it had to be discarded. Other than the manufacturer's statement that this new unit provides "additional clear, crisp digital sound" (and I'm not sure exactly what this means), no evidence was presented to prove that this newer unit will indeed perform better than their analog version.

Since it costs twice as much as the analog version and about three cents more per hour to operate, consumers should know if the extra benefit is worth the extra cost. This should be relatively easy to determine with a study that compares the performance of the analog versus the digital version. The lack of such a study (as far as I know) is an excellent example of the point I was making above regarding the need to validate new hearing development with clinical research.

Nobody who has a hearing loss needs to be reminded of the hearing dificulty they experience under noisy circumstances.

Often, difficulty hearing in noisy surroundings is the main reason why people get hearing aids in the first place. As it happens, the hearing aids we had years ago often seemed to make the situation worse; in noisy conditions, they often added listening problems rather than easing them. The major reason for this was that hearing aids introduced their own internal distortions into the listening situation. It was (and is) quite a challenge for someone with an impaired auditory system (which introduces its own distortions), to understand speech through a distorting "low-fi" hearing aid in a noisy environment. This is the kind of situation where one and one adds up to be more than two. That is, there is a multiple distortion effect in which the total effect is greater than the sum of its individual parts.

More modern hearing aids do not make speech perception poorer because they generally do not introduce additional distortion into the listening process. Indeed, because modern hearing aids can help people perceive previously inaudible sounds they should provide some help regardless of the listening situation.

However, we would like to do more than just make all sounds louder without, at the same time, diminishing speech perception capabilities. The goal is not just the status quo, but to actually improve speech perception in noisy conditions. This is what noise reduction circuits aim to do.

About 10-15 years ago, many hearing aids were introduced which included circuits that automatically reduced the degree of amplification (the "gain") at low frequencies as the sound level increased at these points. The rationale behind this was that ambient noise was composed mostly of low frequency sounds and that by reducing the gain, much of the noise was also eliminated. It was, and is, not a bad idea, but the problem is that this also reduced the speech energy at these same low frequencies. But since the middle and higher frequencies are more important for speech understanding, and these frequencies are less affected by these circuits, this was felt to be a good trade off. One way or another, this type of circuit is still being included in modern hearing aids with varying degrees of success.

Noise reduction circuits took a major leap forward with the re-introduction and perfection of hearing aids incorporating directional microphones. With such an aid, the gain is reduced for sound signals arriving from any other angle but from directly in front. Those sounds arriving from a frontal direction are given full amplification, while all others are suppressed somewhat. Since a person wearing a hearing aid presumably faces the person he or she is talking to, the resulting effect is to increase the level of the desired frontal signal (the speech relative to the level of all other sounds arriving from the rear and the sides). This results in an increase in the speech to noise ratio, still the most important factor underlying speech perception (why a personal FM system can work so well!). However, and this is a malor consideration, if the noise and the speech both arrive from a frontal direction (like talking to somebody with a band playing behind the person's back), there would be no advantage to a directional microphone. Which is why it is recommended that people wearing hearing aids with directional microphones always try to position themselves with the major noise sources to their rear (turn the person around so your back is to the band). In other words, the speech and noise sources have be spatially separated in order for the directional microphones to work, Sometimes, however, such a spatial separation is not possible. This is where another development in noise reduction circuits comes in.

The newer generation of noise reduction hearing aids require that they incorporate multiple bands across frequency and a circuit within each band that is supposed to distinguish between sounds that are predominately speech and those that are predominately noise. This is quite a technological challenge, since a desired speech signal and an unwanted noise sig­nal often share acoustical charac­teristics, for example when the noise is a bunch of other people talking at a party.

Hearing aid users will agree that noise reduction systems are an extremely important topic to research. For many such people, understanding speech in the pres­ence of noise essentially defines their hearing problems. If this can be improved, then a malor step for­ward will have been taken to reduce the impact of a hearing loss upon people's lives.

The question we have to ask at his point is how well these noise reduc­tion systems actually work in improv­ing speech perception in the pres­ence of noise. So far; in my mind, the jury is still out (but still deliberat­ing!I. There is some good evidence that people like having this type of circuit; subjective impressions are often positive. But there is less evi­dence that a noise reduction circuit can actually improve speech percep­tion, although several recent studies offer some tantalizing hints that, at least under certain circumstances, they can.

Sonic Innovations (SI reported a study last year in which such improvements were found. The SI hearing aids employ a nine-band system, with each band centered on one of the audiometric frequen­cies. In this study, the speech spec­trum noise and speech were both delivered from the same frontal source, with and without their noise reduction system operating.

Their results indicate about 15 percent improvement in speech perception in noise when the noise reduction system was activated. In a study conducted with SI hearing aids at Washington University, Valente reported results in a simi­lar direction, though somewhat less optimistic ~this is why studies have to be replicated).

The Phonak company also report­ed a study recently in which the efficacy of their 20-band noise reduction system was evaluated. As was the case with the Sonic Innovation study, the noise and speech stimuli were presented to the subjects from the same frontal source. Instead of using broadband noise, as used by the Sonic Innovation study, this one employed narrower bands of noise, meant to simulate such noise sources as air conditioning fans or car noise. Two such octave-band noises were used, one centered at 250 Hz and the other at 500 Hz. Under these rather restricted con­ditions, using the noise reduction system improved speech scores by about 20 percent.

This is a "hot" topic in heaiing aid research and we can, hopelully, anticipate additional technical development in the area. Oticon, for example, has just reported the introduction of a new hcaring aid they term "ADAPTO." They report that this aid contains a "voice find­er" circuit that "prioritiles the human voice over other sounds in the environment." The company has clearly made a major research and development investment in this new aid (it also includes additional features, such a new feed-back reduction system). As they describe the aid, it does appear quite intriguing and potentially very useful. As with any other reported hearing aid development, I look forward to reading clinical research studies that validate the technical descriptions.

This column is supported, in part, by GRANT #H133E980010 from the U.S. Department of Education, NIDRR, to the Lexington Center. Mention of products or companies by the author does not indicate SHHH endorsement, nor should exclusion suggest disapproval. Since everyone's communication problems and needs vary, SHHH suggests consulting with your hearing health professional.