Otoacoustic emissions, first discovered by David Kemp in 1977, are sounds generated from the cochlea, the fluid-filled portion of the inner ear that contains the sensory organ of hearing. Emissions may be spontaneous, that is, they may occur without any prior stimulation, or they may be evoked, meaning that they are triggered by some other sound stimulus. Evoked otoacoustic emissions were originally known as "Kemp echoes" to signify the way in which the stimulus precedes the emission; however, what makes them so unusual is that they are no ordinary echoes. Evoked otoacoustic emissions do not merely reflect back almost instantaneously the sound that is used as the stimulus, with a duration close to that of the original signal. Rather, evoked otoacoustic emissions take far longer than conventional echoes to be produced; their duration is also far longer than what the duration of the original signal would suggest; and, finally, the "echo" itself is acoustically quite different from the stimulus. It is akin to standing before a canyon and yelling your name, except that it takes longer for the echo to be produced (as if the canyon were further away than it appears); the echo that does eventually come actually sounds like someone else's name; and the echo lasts longer than an ordinary reverberation (reference).

There are several different classes of otoacoustic emissions (OAEs). Spontaneous otoacoustic emissions (SOAEs) are present continuously without being evoked by some external stimulus, and are found in approximately 80 percent of healthy ears (reference). In contrast, evoked otoacoustic emissions (EOAEs) may be detected in virtually all healthy ears, and are only present when elicited by various stimuli. Transient evoked otoacoustic emissions (TEOAEs) respond to short acoustic stimuli such as clicks, tone bursts, and sinusoids. Stimulus frequency otoacoustic emissions (SFOAEs) are evoked by a low-intensity constant tone, which produces a response at the same frequency as the stimulus. Distortion product otoacoustic emissions (DPOAEs) result from the simultaneous presentation of two tones of differing frequency, leading to an emission of yet another frequency. The emission is called distorted because its frequency is not present in the stimulus, instead an intermodulation-distortion response produced by the cochlea. The input stimuli are known as the primary tones, f1 and f2, with f2 typically about 1.2 times greater than f1 in frequency. There are several distortion-product frequencies at which emissions are produced; the largest evoked from mammalian ears is at the frequency that is the difference between twice f1 and f2, or 2f1 - f2. This product is known as the cubic difference tone, and is usually detected at a level of 5 to 30 dB SPL, no matter how high the level of the primary tones (reference).

The available evidence suggests that the various classes of OAEs are all closely related, rather than distinct, phenomena, with the same mechanical processes responsible for each. These processes originate in the outer hair cells of the cochlea, as evidenced by reduced or absent emissions in individuals with damaged outer hair cells: those with hearing losses exceeding 30-45 dB HL in the frequency range 1-4 kHz do not display EOAEs (Kraus & McGee 1992). Thus the cochlea contains not just passive but active mechanisms as well. Not only do the hair cells in the cochlea act as mechanoreceptors and respond to a mechanical stimulus by transducing it into neural impulses that are processed by the auditory nerve and then perceived by the brain as sound. In addition, to produce OAEs, the cochlea contains a mechanically active component that causes a vibration of the basilar membrane that is transmitted toward the base of the cochlea, through the middle ear, and into the external auditory meatus. It is believed that the physiological basis for OAES is the motility, or contractile properties, exhibited by the outer hair cells. That the cochlea displays active mechanisms was theorized by Thomas Gold in 1948 but remained unproven--indeed, was scarcely taken seriously--until Kemp's discovery of OAEs almost thirty years later.

1. History

Long before Kemp discovered OAEs, the work of others had pointed in the direction of the cochlea's active properties, beginning with research into the how the ear encodes frequency. In the nineteenth century, Hermann Helmholtz drew upon Ohm's acoustic law, which states that the ear can resolve a complex tone into a Fourier series, to formulate a theory of hearing based on resonance. He postulated that within the cochlea, the fibers of the organ of Corti and the basilar membrane are responsible for assigning frequency-specific information to the brain according to the place along the basilar membrane that is maximally displaced or resonated. This resonance theory assumes that a particular area of the basilar membrane resonates in sympathy with the frequency of the input stimulus. While the theory offered some attractive explanations for frequency sensitivity, problems were also apparent. If any one place on the basilar membrane is excited, for example, other places must also be stimulated, and so the sharp frequency selectivity exhibited by the human ear would be lost. In addition, a resonator that is so frequency selective must also be heavily damped, meaning that the resonators would have to be relatively slow to reach maximum amplitude in vibration and to return to their resting state. If this were so, rapid changes in frequency as well as amplitude would be more difficult to hear than they actually are (Buser 1992). Furthermore, such a theory relies on the model of a simple resonant system, in which the phase differences between input and output may not exceed 180 degrees; however, along the basilar membrane the observed phase changes greatly surpass that (Pickles 1982).

Beginning in the late 1920s, the work of Georg von Bekesy lent support to the notion of a traveling wave along the basilar membrane, and much of his research was consistent with or improved upon Helmholtz's idea that the area resonated along the basilar membrane is critical in frequency resolution. von Bekesy used microscopic and stroboscopic techniques to discover that the stapes generates a pressure variation through the oval window that results in displacement of endolymph in the scala media of the cochlea, leading to a displacement of the round window. The effect of the displacements creates a transverse wave that moves from the base to the apex of the basilar membrane with a change in amplitude and a reduction in propagation speed as the wave travels. The traveling waves produced by sounds of different frequencies reach their maximum amplitude at different places along the basilar membrane: low frequency sounds reach a peak near the apex, and high frequency sounds near the base. The problem of frequency resolution remained, however, because this explanation predicts that human frequency resolution is worse than actually shown in practice. Specifically, the loss of energy in the traveling wave, as postulated by von Bekesy, should adversely affect the cochlea's sensitivity and precision, contrary to what is observed in auditory behavior. von Bekesy's solution was the conjecture that some neural mechanism must exist to heighten the subjective response of the hearer, and for decades his version of the traveling wave was generally accepted by auditory physiologists as a satisfactory explanation of cochlear sensitivity (Gold 1988; Kemp 1997). In recognition of his groundbreaking work, von Bekesy was posthumously awarded the Nobel Prize in medicine and physiology in 1961.

The acceptance of von Bekesy's theory did not come in the absence of ideas to the contrary. In 1948 Gold, a biophysicist, together with biologist R. J. Pumphrey, reasoned that the loss of energy consistent with von Bekesy's findings was too great given the precision of auditory function. The human auditory system, Gold argued, must contain some active feedback mechanism to reduce the amount of energy lost. He theorized that the cochlear response is enhanced by electrochemical energy taken from the endocochlear potential and converted into mechanical vibration in sync with the stimulus vibration, making up for the loss of energy (Kemp 1997). The active mechanism that he envisioned was missed by von Bekesy because he experimented with cochleas taken from cadavers, Gold said, and the cochlea decays too rapidly after death to reveal these processes. Gold went even further and speculated that the feedback mechanism that he conceived could become unstable and generate a spontaneous ringing--thus presaging spontaneous otoacoustic emissions (Bright 1997). But at the time there was little interest in Gold's work, and he eventually left the field for work in cosmology.

Gold had attempted to find the spontaneous emissions his research had predicted, but failed because he believed that they would be associated with strong tonal tinnitus (Kemp 1997). In fact, tinnitus has been shown to be infrequently related to spontaneous emissions. He also did not foresee the discovery of evoked emissions or their coexistence with the normal functioning of the cochlea. Nevertheless, he essentially predicted that there should be motility in the organ of corti, which was eventually discovered in the outer hair cells in the late 1970s. This discovery was prompted by the findings of Kemp and others that otherwise could not explain aspects of behavioral sensitivity and loudness enhancement, until it was demonstrated that the cochlea indeed emitted energy that could be recorded in the external auditory meatus as acoustic vibrations, using specialized techniques and equipment (Probst 1991).