Otoacoustic emissions (OAEs), low-level sounds produced by the healthy cochlea, require normal or near-normal outer hair cells (OHCs) to provide amplification of the backward traveling waves so the outgoing energy can be detected in the ear canal and, for some types of OAEs, to produce the nonlinearities that give rise to the emission itself. As most audiologists know, these low-level acoustic by-products provide an invaluable window into the otherwise inaccessible cochlea and a useful gauge of OHC health and hearing. 

FIGURE 1. The upper panel shows amplitude spectra for a distortion-type otoacoustic emission (cyan) and a reflection-type otoacoustic emission (red).
FIGURE 1. The upper panel shows amplitude spectra for a distortion-type otoacoustic emission (cyan) and a reflection-type otoacoustic emission (red). Distortion OAEs have smoother spectra and little fine structure once the reflection energy is removed, as in this example. In contrast, the stimulus-frequency OAE, a reflection emission generated by one low-level pure tone, has fine structure with many peaks and valleys due to its origin in back-scattered wavelets. The bottom panel shows a phase versus frequency function for each OAE type. OAE phase is used to classify emissions as distortion or reflection: Distortion OAE phase is relatively invariant across much of the frequency range (when a fixed f2/f1 is used), whereas reflection OAE phase rotates rapidly across frequency producing longer delays. These very different patterns of phase across frequency provide evidence that each OAE comes about by a distinct generation process within the cochlea. 

When OAEs were first discovered (Kemp, 1978), it was thought that each type of emission was basically a duplicate of the other; that is, although some were evoked with clicks and others with tones, we assumed that they all came about the same way and provided redundant information about the cochlea. However, in the last decade or so, we have come to understand that not all OAEs are alike (Shera and Guinan, 1999). There are two basic types of OAEs: nonlinear distortion and linear reflection. The familiar OAEs that audiologists use in the clinic generally fall into one or the other of these two categories, or include a mix of the two. 

Distortion-type emissions are created by nonlinearities in OHC transduction. Nonlinearity in the cochlea likely originates at the ion channels found on the tips of OHC stereocilia. A stimulus vibrates the basilar membrane within the cochlea and in doing so, displaces the OHC and stereocilia atop these specialized cells. The thin filaments attached to the tips of the stereocilia sway with the motion and pull open ion channels like trap doors, near the ciliary tip. Opening the ion channels [mechanoelectric transduction (MET) channels] allows current to flow into the cell and change the intracellular voltage, initiating OHC motility which powers the “cochlear amplifier.” Because there are only a limited number of MET channels, the intracellular voltage and resulting OHC force cannot increase in direct proportion to an increasing stimulus, but grows more slowly (i.e., compressively). As a result, the cochlear response to sound slows and saturates. We believe this cochlear compression is the principal nonlinearity gauged by distortion-type OAEs.

Reflection-type emissions are another animal. Reflection OAEs arise through a linear reflection process: that is, the back-scattering of incoming waves, much as waves might bounce back off of pier pilings in a large body of water. No biological membrane or array of cells can be perfectly smooth and uniform along its length—the cochlea is no exception. When a sound is presented to the ear, traveling waves are launched down the cochlear spiral on the basilar membrane. As they propagate, they encounter irregularities that disrupt the smooth forward flow of energy and give rise to back-scattered wavelets that turn energy around. This reflected energy travels back toward the base of the cochlea and some of it makes its way into the ear canal. The physics of this scattering process indicates that the strongest reflection occurs near the peak of the traveling wave (Zweig and Shera, 1995). When enough of these back-scattered wavelets sum in a coherent way, they produce an emission that is large enough to be recorded in the ear canal as a reflection OAE. 

How do familiar clinical OAEs fit into this taxonomy? As you might have suspected, the DPOAEs, which are evoked by two tones presented simultaneously near the overlap of the two traveling waves elicited by these two primary tones, are nonlinear distortion emissions, though they also include a small reflection component. (The DPOAE is really a mixed OAE but the distortion part is dominant under most clinical protocols.) Click-evoked or transient-evoked OAEs, stimulus-frequency OAEs (OAEs generated with one low level pure tone), and spontaneous OAEs are all reflection emissions. 

By now it is clear that these two classes of OAEs—distortion and reflection— come about by two different processes in the cochlea. Their distinct generation mechanisms produce distinct phase signatures, which is a convenient way of distinguishing them. The reflection-emission phase rotates rapidly across frequency, whereas distortion-emission phase is relatively invariant across frequency when a fixed  f2/f1 ratio is used as shown in the bottom panel of FIGURE 1. The top panel contrasts the amplitude spectra from a distortion (smooth, with little fine structure) versus reflection (much fine structure related to backscattering of wavelets) emission. For more details on generation mechanisms, see Shera and Guinan (1999) or Shera (2004).

Experimental Evidence

Are these two OAE types fundamentally different? Yes. Empirical evidence is abundant. First of all, their remarkably different phase responses provide evidence that the two OAEs come about in very different ways. Experimental manipulations also suggest different origins. For example, strong doses of aspirin, which impairs OHC motility, and sound-evoked activation of the medial olivocochlear reflex both impact reflection emissions more than distortion emissions (Abdala et al, 2009; Deeter et al, 2009;l Roa and Long, 2012).

Age also impacts the two OAE types differently: during maturation, distortion emissions appear to be nearly mature early in neonatal life (except for their phase at low-frequencies, but that is another story); reflection emissions, by contrast, show non-adult-like features in newborns. Ironically, they are bigger than adult OAEs, which may be partly due to middle ear immaturities (Abdala and Dhar, 2010, 2012; Abdala and Keefe, 2012). During aging, the distortion emission is more reduced with advancing age than is the reflection emission (Abdala and Dhar, 2012; Abdala et al 2017a). 

Finally, a strong piece of evidence can be found in genetic mutations that cause hearing loss. A genetically engineered mouse, the stereocilin mouse, lacks thread-like lateral links between stereocilia. This mouse does not have distortion-type OAEs but its sensitivity and tuning are near normal early in its life, suggesting that reflection emissions would be present if they had been measured (Verpy et al, 2008). Another transgenic mouse, the Ceacam16 mouse, has a ragged, porous tectorial membrane; this mouse produces abnormally high-level reflection emissions and normal distortion products (Cheatham et al, 2014). Thus, both human and mouse evidence tell us that the two OAE types can be independently affected and are indeed sensitive to distinct cochlear properties. 

Can the two OAE classes tell us different things about human cochlear function and dysfunction? We think so. Anecdotally, many audiologists observe that the click-evoked OAE becomes unmeasurable (and yields an “absent” test) with milder degrees of hearing loss than the DPOAE, which can be present with a moderate degree of loss. In this way, reflection emissions appear to be more sensitive to slight amounts of hearing loss than distortion emissions and perhaps, to broadened tuning. Why? Reflection emissions backscatter from the peak region of traveling waves in the cochlea, which is where the cochlear amplifier gain is strongest and tuning is sharpest. So, even a slight-mild hearing loss can reduce reflection emissions. 

The DPOAE, in contrast, is strongest at moderate stimulus levels where cochlear response growth is compressed and nonlinear distortion is created. A handful of studies offer bits of evidence consistent with click-evoked or stimulus-frequency OAEs (both reflection emissions) being more sensitive to slight amounts of hearing loss than the DPOAE (Gorga et al, 1993; Lapsley-Miller et al, 2004; Abdala and Kalluri, 2017). However, research is needed measuring both OAEs together in ears with varied degrees and etiologies of hearing loss to confirm and define these distinct sensitivities. 

If each of the two types of OAEs—distortion and reflection—offer unique and non-redundant information about the ear and provide a more complete picture of the hearing loss when considered together, why record only one in the audiology clinic? Considering both OAEs together may be maximally informative because we exploit the information offered by both. The common OAE clinical protocol at present seems to resemble the following: one OAE type (but not the other) is applied during hearing assessment at one stimulus level (~ 65–55 dB SPL for DPOAEs; ~ 80–86 dB pSPL for CEOAEs) across an abbreviated range of frequencies. Given what we now understand about OAEs, perhaps it is time to update this protocol.

Three Suggestions

FIGURE 2. The DPOAE shown here was recorded using swept-tones and analyzed with fine frequency resolution (approximately 500 points across frequency).
FIGURE 2. The DPOAE shown here was recorded using swept-tones and analyzed with fine frequency resolution (approximately 500 points across frequency). Because the DPOAE is comprised of both distortion and reflection components, the spectrum shows fine structure, i.e. peaks and valleys, when recorded with sufficient resolution (because the two components interfere with one another in constructive and destructive ways). This DPOAE spectrum generated with sweeping stimulus tones provides a much more detailed picture of cochlear distortion across frequency than a conventional clinical DP-gram. (Note: Once the DPOAE is separated from its reflection components, it produces a smoother spectrum, which is shown here as a thin gray line). 
  1. Record both OAE types together. 

    The idea of using both OAEs together moves beyond the rudimentary (but critical) goal of detecting a hearing loss. The two genetic mutations presented previously have been found in humans. An individual with a stereocilin deficiency or a Ceacam16 deficiency might walk into your audiology clinic tomorrow. If only one OAE class is impacted by these mutations, how will you identify the genetic hearing loss and refer for appropriate genetic testing and counseling by recording only a DPOAE (or CEOAE)? It is true that these mutations are relatively rare, but perhaps they are more common than we think—perhaps we have simply failed to identify them due to our limited testing protocols (much like disorders such as auditory neuropathy and “hidden hearing loss” went undetected for decades).  

    A thought experiment might help us consider the impact of this suggestion: a child comes into the clinic and has an absent click-evoked OAE. (The ABR exam has not yet been scheduled for this child.) The absent CEOAE tells you that some degree of hearing loss is present but little more. Thinking strategically, you record a DPOAE also. The DPOAE is present (low in amplitude perhaps, but measurable with adequate signal-to-noise ratio), which expands your understanding of the hearing loss. You have recorded an absent CEOAE, which is a reflection emission sensitive to even small amounts of hearing loss, but a present DPOAE, which is often measurable even with a moderate degree of hearing loss. What does this combination of results tell you? It tells you that the sensorineural hearing loss is more likely to be moderate in nature than profound. Profoundly hearing-impaired ears do not produce cochlear nonlinearities. The results of these two OAE tests combined have allowed you to estimate the degree of sensory loss.  

    Though not ready for prime time, there is also work afoot combining novel features of the SFOAE and DPOAE (e.g., emission strength and compression) into a merged profile to better characterize hearing loss and to distinguish among seemingly similar sensory hearing losses (Abdala and Kalluri, 2017). In this preliminary work, we have plotted a measure of OAE “strength” for both reflection (SFOAEs) and distortion OAEs, one against the other on a two-dimensional plot to define the relationship between the two emission types in any given ear. Interestingly, the normal relationship has a well-defined pattern or “cluster” and atypical patterns are emerging in hearing-impaired ears.  

    But questions remain: do ears with a similar dual reflection-distortion OAE profile have similar underlying pathologies? Conversely, can dissimilar dual-OAE profiles discriminate between two hearing losses with similar audiograms but different types of deficits? This work is currently in progress but by exploiting both OAE types in a combined profile, we hope to capture variance among hearing losses that is not currently detected by the audiogram.

  2. Record the DPOAE at higher stimulus levels. 
    In our thought experiment above, the DPOAE might have been initially absent when attempted at default stimulus levels of 65–55 dB SPL. You might then have presented slightly higher-level primary tones, 75–75 dB SPL perhaps, in an attempt to evoke a DPOAE. It is sometimes the case that DPOAEs in impaired ears are not measurable at default primary-tone levels but measurable at higher primary-tone levels. Of course, a threshold test such as an ABR or behavioral test will verify and confirm the estimate, but this combination of OAE results (non-measurable CEOAE and measurable DPOAE even at higher levels) lessens the likelihood of profound hearing loss. Note: the audiologist must know the system distortion levels of the equipment when presenting primary tones at higher levels (Siegel, 2002). 
  3. Look at Normative Data to Interpret OAEs. 
    If you deem an OAE absent or present and proceed no further, the diagnostic process is incomplete. An OAE recorded in the clinic is typically considered “present” if it is 3–6 dB above the noise floor recorded in the ear canal (the actual SNR criteria vary from clinic to clinic). This “present” diagnosis provides critical information but OAE levels should also be compared to those published in the literature or to the normative amplitudes generated in your own clinic, to determine whether the response falls within the range of levels a normal-hearing individual of the same age is expected to produce. As an example, newborns have click-evoked OAEs that are robust and broadband; it is not unusual to observe 10–20 dB SPL responses from neonatal ears. If you see a newborn with about 5 dB SPL CEOAE present from only 1–3 kHz, can you consider it normal? No. But it is likely to be normal for a 65-year-old adult. Awareness of OAE level trends across the human lifespan is important in determining normalcy. 

These three guidelines may help you utilize the rich and varied information OAEs offer about cochlear health and hearing.

To Infinity and Beyond

Recording both OAEs and presenting more than one primary-tone level takes time and in the clinic, time is money. How can an audiologist make up this added test time? OAEs can be recorded with rapidly swept tones (or chirps) rather than conventional single, discrete pure tones. Tones are swept continuously upward or downward at rapid rates (approximately one octave per second) many times and the average OAE is extracted from the sweeps after the test has been completed (Long et al, 2008; Kalluri and Shera, 2013; Abdala et al, 2015, 2017b). The entire test can take only minutes. And resolution of the swept-tone OAE is unparalleled because it is possible to estimate magnitude and phase offline at as many points along the frequency range as desired. The resulting waveform is not a gross clinical DP-gram but a complex, nuanced OAE spectrum as shown in FIGURE 2. It is a well-defined, intricate record of cochlear emissions across frequency, showing characteristic peaks and valleys in the DPOAE (unless it has been unmixed from its reflection parts) (Long et al, 2008; Abdala et al, 2015). 

An extension of the above innovation is a rapid swept-tone OAE program that interleaves the measurement of reflection and distortion OAEs for nearly simultaneous recordings of both, i.e., every other sweep includes either two tones for the DPOAE or a single probe for the SFOAE. The SFOAE is not a familiar OAE to most audiologists but it is the most simple form of reflection since it is evoked with one single low-level tone. Recording both SFOAEs and DPOAEs in a combined fashion will exploit the power offered by both. Our preliminary work suggests the relationship between the two emissions may hold diagnostic clues that neither OAE recorded alone can provide. 

Thirdly, we are hopeful that OAE phase and its derivative, group delay, will make its way into the clinic as a diagnostic tool in the near future. The variation of OAE phase with frequency provides a measure of the emission delay, which can be thought of as the latency of the response. In laboratory studies, the SFOAE delay has been reliably linked to measures of cochlear tuning (Shera et al, 2002). Other work has also shown DPOAE phase to be sensitive to changes in intra-cranial pressure—it may offer a non-invasive probe of this important neurological indicator (Voss et al, 2006). More translational research needs to be done to study and quantify how various auditory pathologies impact OAE phase, but it seems obvious that we should exploit the entire information package available in each OAE rather than only half of it. 

Finally, new calibration techniques have decreased the sometimes-excessive variability in OAE measures among subjects and within one subject over time, making serial monitoring for ototoxic drug treatments or noise exposure more precise. They are beyond the scope of this paper, but these calibration techniques mitigate the contaminating effects of ear canal acoustics on the stimulus level (Scheperle et al, 2008) and the OAE itself (Charaziak and Shera, 2017).

Over the years, OAEs have taught researchers much about the mechanical workings of the cochlea. Advanced signal processing has allowed us to do so rapidly and efficiently with enhanced resolution. However, the ultimate sign of progress is when these discoveries and innovations are applied in the audiology clinic to enhance the diagnoses of auditory pathology and positively impact the individual lives of those with hearing impairment. Inarguably, OAEs have yet to reach their full potential in this realm. Combining reflection and distortion OAEs to generate a comprehensive dual-OAE profile may be one step in that direction. Technical innovations on the horizon should facilitate this endeavor. In the end, our lofty goal is to fully extract and exploit the information these unique probes of cochlear health offer us to better understand and treat sensory hearing loss. 


Abdala C, Dhar S. (2010) Distortion product otoacoustic emission phase and component analysis in human newborns. J Acoust Soc Am 127:316–325.

Abdala C, Keefe DH. (2012) Morphological and Functional Development of the Ear. In L. Werner, A. Popper and R. Fay, (Eds.). Springer Handbook of Auditory Research: Human Auditory Development Springer: NY. pp. 19–60.

Abdala C, Dhar S. (2012) Maturation and aging of the human cochlea: a view through the DPOAE looking glass. J Assoc Res Otolaryngol 13:403–421. 

Abdala C, Luo P, Shera CA. (2015) Optimizing swept-tone protocols for recording distortion-product otoacoustic emissions in adults and newborns J Acoust Soc Am 138:3785–3799.

Abdala C, Kalluri R. (2017) Toward a joint reflection-distortion otoacoustic emission profile: Results in normal and impaired ears. J Acoust Soc Am 142:812–824.

Abdala C, Mishra SK, Williams TL. (2009) Considering distortion product otoacoustic emission fine structure in measurements of the medial olivocochlear reflex. J Acoust Soc Am 125:1584–1594.

Abdala C, Ortmann AO, Luo P, Shera CA. (2017a) Shifts in the relationship between reflection- and distortion-source OAEs during aging. To be published.

Abdala C, Guardia YC, Shera CA. (2017b) Swept-Tone Stimulus-Frequency OAEs:  Normative Findings and Methodological Considerations. Under Review J Acoust Soc Am. 

Charaziak KK Shera CA. (2017) Compensating for ear-canal acoustics when measuring otoacoustic emissions. J Acoust Soc Am 141. 

Cheatham MA, Goodyear RJ, Homma K, et al. (2014) Loss of the tectorial membrane protein CEACAM16 enhances spontaneous, stimulus-frequency, and transiently evoked otoacoustic emissions. J Neurosci 34:10325–10338.

Deeter R, Abel R, Calandruccio L, Dhar S. (2009) Contralateral acoustic stimulation alters the magnitude and phase of distortion product otoacoustic emissions. J Acoust Soc Am 126(5):2413–24. 

Gorga MP, Neely ST, Bergman BM, et al. (1993) A comparison of transient-evoked and distortion product otoacoustic emissions in normal-hearing and hearing-impaired subjects. J Acoust Soc Am 94:2639–2648.

Kalluri R, Shera CA. (2013) Measuring stimulus-frequency otoacoustic emissions using swept-tones. J Acoust Soc Am 134:356–368. 

Kemp DT. (1978) Stimulated acoustic emissions from within the human auditory system. J Acoust Soc Am 64:1386–1391.

Lapsley-Miller JA, Marshall L, Heller LM. (2004) A longitudinal study of changes in evoked otoacoustic emissions and pure-tone thresholds as measured in a hearing conservation program. Intl J Audiol 43:307–322. 

Long GR, Talmadge CL, Lee J. (2008) Measuring distortion product otoacoustic emissions using continuously sweeping primaries. J Acoust Soc Am 124:1613–1626. 

Rao A, Long GR. (2011) Effects of aspirin on distortion product fine structure: Interpreted by the two-source model for distortion product otoacoustic emission generation.

Scheperle RA, Neely ST, Kopun JG, Gorga MP. (2008) Influence of in situ, sound-level calibration on distortion-product otoacoustic emission variability. J Acoust Soc Am 129:792–800. 

Shera CA, Guinan JJ. (1999) Evoked otoacoustic emissions arise by two fundamentally different mechanisms: A taxonomy for mammalian OAEs. J Acoust Soc Am 105:782–798.

Shera CA, Guinan JJ, Oxenham AJ. (2002) Revised estimates of human cochlear tuning from otoacoustic and behavioral measurements. Proc Natl Acad Sci USA 99:3318–3323.

Shera CA. (2004) Mechanisms of mammalian otoacoustic emission and their implications for the clinical utility of otoacoustic emissions. Ear Hear 25:86–97. 

Siegel J. (2002) Calibrating Otoacoustic Emission Probes. In Robinette, M. and Glattke T. (Eds), Otoacoustic Emissions: Clinical Applications Thieme: New York, pp.416–441.

Verpy E, Weil D, Leibovici M, et al. (2008) Stereocilin-deficient mice reveal the origin of cochlear waveform distortions. Nature 456:255–258. 

Voss S, Horton N, Tabucchi T, Folowosele F, Shera C. (2006) Posture-induced changes in distortion-product otoacoustic emissions and the potential for noninvasive monitoring of changes in intracranial pressure. Neurocrit Care 4:251–257. 

Zweig G, Shera CA. (1995) The origin of periodicity in the spectrum of evoked otoacoustic emissions. J Acoust Soc Am 98:2018–2047.   

Share this