The two major classes of OAE technology, TEOAE and DPOAE, differ fundamentally in the condition of the cochlea which they observe. In TEOAE testing the OAE sound is recorded during the silence between brief stimuli - so that the relaxed status of the outer haircells is observed. As most of the cochlea is excited by a click, reports are simultaneously received from multiple sections of the organ of Corti. This doesn’t blur the picture because each section responds at its own characteristic frequency. Signal processing can easily separate the response from each part. A 20ms sweep allows a frequency resolution of 50Hz, or 20 points per octave from 1 to 2kHz. TEOAEs therefore test many parts of the cochlea individually and simultaneously in a functional state close to threshold stimulation. With DPOAEs a more restricted part of the cochlea is more intensely stimulated and continuously driven so that the outer haircells are observed in their ‘working state’. The width of the region tested is not defined by the precision of the pure tone frequency but by the natural bandwidth of the cochlea. The stimulated region is quite extensive (see ‘OAEs and the cochlea’). Only one or two regions of the cochlea can be simultaneously observed by DPOAEs. This is primarily because modern day transducers produce more distortion than the cochlea if fed with multiple tones. DPOAE measurements must therefore be repeated at several frequencies to get a balanced overall picture. To match the cochlea’s natural bandwidth for processing, a 3 points per octave DPOAE resolution is required as a minimum. Higher resolution is desirable so as to overcome the misleading effects of standing wave interference within the cochlea which occur with pure tones.

Both DP and TEOAE views of cochlear function are valuable and complementary. Each technology has different advantages and disadvantages. TEOAE technology has the advantages of sensitivity, frequency resolution and speed, but it fails to recover OAEs in adults much above 4kHz. This is due to the shorter latency of high frequency OAEs. DPOAE technology has the advantage of superior detection of high frequency OAEs but it suffers from lower frequency resolution and lower noise immunity at low frequencies. The technique is unable to capture primary OAE energy but a more serious practical drawback is the dependance of DPOAE on the precise stimulus configuration (frequency and level ratios).

Frequency specificity is very important to cochlear function but is often misrepresented in OAE literature. It is the frequency specificity of the cochlea that is important and not that of the stimuli. Clicks or tones are therefore equally suitable stimuli with which to observe the cochlea. All OAEs are highly frequency specific in that each frequency component of an OAE can be directly traced to a frequency component in the stimulus. What is desirable and is often assumed to be true of OAEs is that the response obtained to a specific stimulus frequency tells about the status of a particular PLACE in the cochlea. This is probably true only in a very broad sense.

The relation between OAE levels and auditory threshold - or rather the lack of it - has already been discussed. In the early days of DPOAE research it was common to define a ‘DPOAE threshold’ as the stimulus level at which the OAE equalled the noise present in the instrument. OAEs do not have a threshold and this measure is unsafe. Threshold is a property of the inner haircells and nerve synapase which play no part in the creation of OAEs. A related and more meaningful measure is the growth rate of DPOAE with stimulus level which appears to steepen as auditory threshold is elevated. Observations must however be averaged over a range of stimulus frequencies and ratios.

The concept of ‘passive’ and ‘active’ DPOAE responses arose from animal observations and should be applied with caution to clinical work. Human ears stand much higher levels of stimulation than rodents and ‘passive’ cochlear responses are very unlikely in response to stimuli of 70dBspl and below. What is more likely is passive DP generation in the probe and instrumentation. DPOAE systems should be checked in a test cavity, but a more powerful test is to measure the latency of DPOAE found in the ear. Latencies of 3ms or greater are highly indicative of a cochlear origin, and lower latencies of instrumentation distortion.

Finally, calibration. In the clinic OAE systems are used as function detectors, not as measuring systems - but calibration is still important to ensure proper operation and data comparability. OAE systems display sound levels on screen, so the microphone calibration can be quickly checked against a calibrated sound level meter. Stimulus calibration presents special problems due to standing waves in the ear canal. The sound level at the drum cannot be accurately set from measurements at the probe. The problem becomes serious above 5kHz in adults. It is less important in infants. Additionally, the decibel level of the OAE also depends strongly on ear canal acoustics.

This article has been extracted from the publication ‘Understanding & Using Otoacoustic Emissions’, written by Professor David Kemp and published by Otodynamics, and is reproduced with the author’s permission. Copyright remains with the author.