Dr. Ross on Hearing Loss
Some Reflections on Digital Signal Processing (DSP) Hearing Aids
by Mark Ross, Ph.D.
So-called digital hearing aids are an outgrowth of the computer revolution that is transforming our entire society. Today, digital technology has been applied to hearing aids in two distinct formats: Digitally programmable analog signal-processing (ASP) instruments and digitally programmable digital signal processing (DSP) Digital encoding in the latter technology is proving to be an economical and versatile method of manipulating and performing all sorts of electronic functions. It has taken a while for engineers to incorporate digital signal processing (DSP) in an ear-level hearing aid, but they finally succeeded about two years ago. (A DSP body-worn hearing aid was available ten years ago, but never met with commercial success.) This truly represents an impressive technological breakthrough, the implications of which should be becoming increasingly apparent in the next few years. None of the cautionary comments I will be making below is intended to denigrate in any way the commendable technological achievement that these developments signify.
A digital signal processing hearing aid operates in a fundamentally different manner from an analog hearing aid or a digitally programmable analog signal processing hearing aid, but - and this is crucial - the output from a DSP aid is still amplified acoustic energy. In other words, people still listen to amplified sounds and not binary impulses. The sounds that reach the microphone of a digital hearing aid are converted by an electronic circuit, termed an analog to digital (A/D) converter, into discrete positive and negative electrical charges. This can perhaps best be visualized as a series of binary digits. That is, at any instant in time the speech wave is coded into a series of numbers (representing these electrical charges) all of which are labeled either zero or one. For example, an eight bit analysis may be processed as the binary digits: 11011100. Each /0/ and each /1/ represents whether some preset criterion has or has not been met (such as the input sounds exceeding a certain intensity at a certain frequency at a particular instant in time). The quality of the resulting amplified signal depends upon the rapidity and the resolution of the sampling rate; generally, within limits, more is better for higher quality reproductions.
By converting ongoing speech signals to binary electrical impulses, it is possible to manipulate them through a digital signal processor (just as in a computer) in ways that cannot be accomplished with analog hearing aids. These manipulations (termed "algorithms") determine exactly how the speech signal is to be processed as it proceeds from the A/D converter to the converse activity, the digital to analog (D/A) converter that changes the digital signals back into audible sounds. For the hearing aid user, it is the content of these algorithms that matters most. It is these algorithms (i.e. speech processing strategies) that determine the modifications that will be made to sound signals as they are delivered to the ears of a listener. While it is not really necessary for consumers to comprehend all the technical complexities of digital hearing aids, they do have to be assured of the listening benefits provided by a specific algorithm incorporated in a specific digital hearing aid.
Currently, according to a recent special issue of The Hearing Journal, sixteen manufacturers have introduced, or will soon be introducing, digital signal processing hearing aids. The new aids come in all sizes and styles, ranging from behind-the-ear models (BTE) to completely-in-the-canal (CIC). All of them are capable of being modified electroacoustically in any number of creative ways. Among the highlighted special features are multiple memories, multi-band noise reduction, multi-microphones, low battery warnings, feedback reduction, and digital cell phone compatibility. Given the high degree of competitiveness in this area, we can expect to see additional features periodically introduced by manufacturers. This represents quite a technological feat and one in which hearing aid companies may justly take pride. Still, the question remains, has it been demonstrated that this technological capacity has produced improved speech perception scores above and beyond that possible with advanced analog hearing aids?
Right now, in terms of the currently available research the jury is still out. As I read through the articles in this special issue of The Hearing Journal, I could sense a bit of unease from the authors about the paucity of clinical studies attesting to and supporting the objective superiority of digital hearing aids. Nobody doubts their technical sophistication, or the fact that they are first-rate instruments. Nobody currently using a digital hearing aid need fear that they have been fit with an inferior instrument. Moreover, all can be reprogrammed and "tuned-up" as required or requested to try to ensure optimal satisfaction. What makes some people uneasy (and I'm one of them) is the "hype" associated with their introduction and marketing. There is an implication of universal applicability and assured superiority that has not yet been convincingly demonstrated in careful, comparative research.
This is addressed in one of the articles, by David Fabry an audiological researcher from the Mayo Clinic, who discusses the "facts, myths, and leaps of faith" associated with digital hearing aids. After commenting that the "facts" regarding improved speech perception are equivocal, and exposing some of the "myths" (e.g. no need for a volume control, aids will last longer, cost not an important factor), he concludes with his "leap of faith" (with which I concur) that digital aids will eventually and unambiguously improve the speech perception capabilities of more people in more listening situations than is now possible with even the best analog hearing aid. Considering the pace of recent developments, it is quite likely that digital hearing aids may soon include such features as the elimination of acoustic feedback, superior binaural listening, better hearing in different types of noise, and enhanced speech feature processing. In the meantime, however, we have to deal with what is available now.
In another article in this same issue, Ruth Bentler from the University of Iowa describes ongoing studies that are designed to investigate the influence of marketing "hype" upon a person's hearing aid preferences. These studies seemed so provocative that I contacted her directly to make sure that I had my facts right. What she and her colleagues did was to alternately try two different digital aids, adjusted according to the manufacturer's specifications, on their subjects for a month each. For the purposes of the study, however, one of the aids was labeled "digital" and the other "analog". At the conclusion of the trial, they were asked which one they preferred. When fitting the "digital" aid on the people, she recited the manufacturer's marketing appeals (CD sound quality, DSP, microprocessor, etc.). With just a few exceptions, the subjects preferred the sound quality of the aid that was labeled "digital" (as indeed it was) over the aid labeled "analog" (which was actually another digital aid). Since the aids were rotated among the subjects, both aids took turns being labeled analog and digital. Clearly, expectations and belief influenced the judgment of the subjects.
This is seen even more clearly in another section of the study. Twenty of the subjects took turns wearing a presumed "digital" (DSP) and presumed "analog" hearing aid. In reality, the subjects wore the exact same digital (DSP) instrument for both conditions. That is, the investigators would ostensibly switch aids after a month of wear, but in reality the subjects would be "switched" right back to the original aid. Again, except for a few exceptions, the subject's overwhelmingly preferred the "digital" (DSP) hearing aid over the "analog" aid. The only difference in these two conditions was their belief that one of the aids represented the cutting edge of modern hearing aid technology. It is interesting to note that speech perception scores under the two conditions were similar. Evidently, the subjective impressions and expectations did not translate into improved comprehension of speech.
What are we to make of this? This research was not conducted to discredit a person's personal judgment about performance with a hearing aid. When somebody relates specific instances in which one type of aid performs noticeably better than another, he or she must be believed. All of us who wear hearing aids make these kinds of judgments all the time. They cannot simply be dismissed. If we could not trust the evidence of our own ears, we could hardly make any kind of comparative judgment regarding hearing aids. In other words, we have no choice but to trust ourselves to make choices that are correct for us.
On the other hand, as these studies show, we can be swayed by our expectations and by clever marketing appeals. This is hardly surprising; every time we read a paper or magazine, watch TV and listen to the radio, we are inundated with messages that affect what we purchase, who we vote for, and indeed how we think. When these messages concern something so central to our lives as our functional hearing ability, no wonder we find ourselves susceptible to their influence.
I don't meant to rain on anyone's parade, but we do need evidence regarding performance of digital hearing aids that is not potentially contaminated by hopes, expectations, or self-interest, on the part either of the consumer, the hearing aid dispenser, or the manufacturer. Our society has an almost supernatural belief in the possibilities of technology. If it's newer, it's got to be better. Telling someone that the hearing aid he/she is going to be trying is a newly developed all-digital "computer" hearing aid can't help but influence performance. I don't know of any hard of hearing person who does not want to hear better, and most of us (including me) believe that some improvement in our ability to hear has got to be possible. It is true that these new hearing aids are technically more sophisticated than previous generations of hearing aids. However, and still, this does not mean that greater technical complexity automatically translates in superior listening performance. This has to be directly demonstrated, and it has to be demonstrated in as objective a fashion as we can contrive. It may take some doing, but it can be done and is being done in a number of centers right now.
What makes the situation so problematical for consumers is the fact that some of the biggest "believers" are those professionals who actually dispense these hearing aids. They are now spending a great deal of their time and energy trying to master this impressive new technology. After learning all (or something) about "digital signal processing" and how to program these aids on people, it is understandable that some audiologists would be, at the least, unconsciously biased in their favor. They, too, are going to see what they want and hope to see. Confounding this positive predisposition regarding the potential performance of the hearing aids, is the fact that they are much more expensive than even advanced analog aids, with a consequently much higher profit margin. We hate to think that this factor influences hearing aid recommendations, but it would be naive to dismiss that possibility entirely. It is difficult for a consumer or an audiologist to be completely objective about such a high-tech, expensive instrument, one that draws so heavily on the language and expectations of modern science. We both need to believe: the consumer that he or she is hearing better, the audiologist that he or she is providing the best hearing aid services available. This is the reason why objective evidence is so important, not to challenge our hopes and expectations, but to support them.
Right now, as I interpret the available research, I'd exercise a little extra caution before purchasing a digital signal processing hearing aid. While there is no question that these are excellent hearing aids, at the present time I would suggest that anyone content with his or her current hearing aids stick with them for a while. Anybody contemplating purchasing DSP hearing aids, because their present units are unsatisfactory or because a trusted audiologist is making this recommendation, should be sure to give the aids a fair trial (60 days would be my recommendation). Experienced hearing aid users can compare the DSP aids to their previous models, perhaps "stopping over" and trying advanced analog aids before making a final decision. Anybody already wearing digital signal processing hearing aids can expect to hear at least as well with them as with the high-end analog or digitally programmable analog hearing aids (in itself a major improvement over previous generation of analog hearing aids). Because even the most knowledgeable hearing aid dispenser may still be in a "learning curve" in fitting DSP hearing aids, people trying these aids should be prepared, and not hesitate, to return often for "tune-ups". These instruments are capable of an awesome variety of electroacoustic adjustments, but these cannot be made unless hearing aid users work closely with their audiologists. Whatever kind of hearing aid is used, the "partnership" relationship between the user and the dispensing audiologist is still the pre-eminent consideration in ensuring optimal hearing aid usage.
This paper was supported in part by grant #H133E980010 from the US Department of Education, NIDRR, to the Lexington Center. Some of this material was adapted from articles appearing in Hearing Loss: the Journal of Self-Help for Hard of Hearing People.