The discovery by David Kemp of otoacoustic emissions (OAEs) has led to two breakthroughs. That which influenced most people, clinicians as well as parents, stems from the use of OAEs as a non-invasive and easily serviceable tool for early auditory screening. The second one may be less readily acknowledged because researchers always tend to prefer when invasive experiments directly disclose measurable features of cochlear mechanics. Still, we have to remember that before OAEs, nobody thought of an active cochlea, whereas OAEs came out just one step ahead of mainstream researchers, when they eventually admitted that some cells were contributing to basilar membrane motion, just as Thomas Gold had predicted.
In the nineties, it was thought that because DPOAEs (for distortion-product OAEs) were comparatively "simple" looking emissions and yet provided insights into the secrets of cochlear micromechanics, their future as routine laboratory tools was ensured, even though many features of their behavior remained in want of full interpretations. Yet a decade later, the place devoted to DPOAEs in big scientific meetings has fallen, from several full–time sessions in a row, to just a few marginal posters in some dark corner. When implementing universal neonatal hearing screening, several countries just discard OAEs and go to recommending the use of automated ABRs.
Is it a good idea to discard the DPOAE tool just because few new concepts on their physiological meaning have come out recently? And why are clinicians reluctant, or afraid of using such a simple test whereby pairs of pure tones around 60 –70 dB SPL are swept in frequency, within a few seconds, almost irrespective of ambient noise? Perhaps the old (and actually all too naïve) dream that OAEs might allow automatic pure-tone audiograms to be plotted and that there was nothing more to it killed the goose that lays golden eggs… in about the same way that ABRs were largely neglected after a flurry of publications? If the analogy holds, the future of DPOAEs may be promising if we consider the second birth of ABR-derived complements: derived or stacked ABR, ASSR, FFR, speech-ABR, etc. I would therefore like to remind the readers that the interest of DPOAEs has been recently brought forward in a variety of experiments wherein building an audiogram was not the prime motivation. They will pardon me if they find a number of self-references, but after all this is meant to be a comment / editorial and not a scientific review.
So, it has been shown (Avan and Bonfils, Hear Res. 2005; 209: 68-75) that with DPOAEs, the frequency intervals corresponding to hearing loss and (presumably since the study was done in humans) OHCs damaged by noise exposure can be spotted with fine accuracy. The patients were workers who had been exposed to impulsive noise once, and thanks to yearly follow-up procedures at work, their pure-tone audiograms were available before and after exposure. The outcome of high-resolution Békésy audiometry was compared to that of so-called DPgrams in terms of corner frequencies between normal and pathological intervals. Two caveats were brought forward, the first one was that the degree of hearing loss could not be ascertained, and the second was that a number of patients had strictly normal DPOAEs despite NIHL: in their case, we speculated that damaged inner hair cells or neurons had to be at the origin of the hearing loss.
A little earlier, our team (in J Acoust Soc Am. 2003; 113: 430-4, already discussed in this site), elaborating on previous work by D.M. Mills, examined DPOAEs existing in both active and passive modalities (that is, DPOAEs generated by low vs. high intensity stimuli). We showed that in the last case (high intensities, for example 75-80 dB SPL were used in mice with negligible instrumental distortion), intervals with damaged OHCs could still be spotted with fine accuracy. Following this result, we proposed that DPOAEs probably have a single source in both active and passive modalities, and that this source has to do with OHCs, and probably with the question of whether their stereocilia bundles are still present with reasonably normal mechanical properties.
Along a similar line of thought, genetics has brought about a few surprises, e.g., Liberman, Zuo and Guinan have used their EPL model of prestin-null mice (in J Acoust Soc Am. 2004; 116: 1649-55) to show that the absence of somatic motility does not prevent DPOAEs to be still, present provided stimulus levels are set high enough. These DPOAEs remain physiologically vulnerable. The fact that higher stimulus levels are required might just be ascribed to the need for making up for the lack of cochlear amplifier associated with missing prestin, as it may be assumed that the nonlinear element at the origin of DPOAEs in OHCs needs a large enough input to produce sizeable distortion. It is nonetheless remarkable that this distortion is coupled to the motion of cochlear partition in such a way that it eventually comes out as DPOAEs. Lukashkin et al. (J Neurophysiol. 2004; 91: 163-71) raised a similar issue when they published that mice with a tectorial membrane detached from OHC (Tecta deltaENT / deltaENT mutants) still produce DPOAEs at high intensities. Genetics again: DPOAEs are of great help for spotting auditory neuropathies and synaptopathies, of which more varieties have been thoroughly explored recently (e.g., Delmaghani et al., Nat.Genet. 2006; 38: 770-8; Roux et al., Cell 2006; 127: 277-89).
Indeed, for purely audiometric purposes, DPOAEs turn out to be rather insensitive to hearing loss whenever stimulus levels exceed 60 dB SPL. However, from the previously mentioned reports, the use of louder stimuli enables the audiologist to gain interesting insights regarding the presence or absence of nonlinear mechanical elements inside the cochlea, irrespective of the presence or absence of cochlear amplifier. Admittedly, it is likely that these nonlinear mechanical elements are passive by themselves. Yet they play an essential part in a series of coupled elements acting in a loop. Their coupled function eventually results in an enhanced and filtered input to the sensory cells. Are not these sorts of data infinitely more interesting than the outcome of an automatic pure-tone audiogram (all the more since there are so many ways of obtaining accurate pure-tone audiograms, and so little practical use for them when not combined with other measurements)? I am sure that all the readers who concur will easily find other, even more clever ways to prevent their DPOAE equipment from collecting dust instead of useful data…
