Quantitative assessment of methods for the detection of otoacoustic emissions.

Objective: To compare the performance of various types of signal detector commonly used to detect otoacoustic emissions (OAEs).

Method: The signal detectors were tested with signals in various types of noise, including noise with various amplitude distributions. Commonly utilized transient evoked OAE and distortion product OAE detectors were analyzed.

Results: In both cases it was found that detector performance increased as the measurement bandwidth was increased. Noises of different amplitude distributions were found to affect the performance of both types of detectors. The effect of different amplitude distributions decreased as the measurement bandwidth was increased. For a given power per spectral line, the transient evoked OAE detector could detect signals 6 dB farther into gaussian noise than the distortion product OAE detector.

Conclusions: An analysis of the results of practical methods of OAE measurement show that distortion product OAEs can be measured with higher power per spectral line than transient evoked OAEs. This means that in adult humans distortion product OAEs can be reliably detected in about 1/14th of the time of transient evoked OAEs. In neonatal humans distortion product OAEs can be reliably detected in about two-thirds of the time of transient evoked OAEs.