Evoked potential threshold estimation can be made truly objective by using statistically based methods. In general, time domain analysis is preferable for responses which are impulsive (temporally narrow, spectrally broad), whereas frequency-domain analysis is more appropriate for tonal responses (spectrally narrow, temporally broad). In both time and frequency domains, methods comparing evoked potential power to noise models are robust and powerful. For spectrally narrow responses such as steady-state evoked potentials, the performance of magnitude-squared coherence, the ratio of grand average power to mean subaverage power, is superior to that of other methods for objective response detection.
Address reprint requests to Robert A. Dobie, M.D., Department of Otolaryngology-Head and Neck Surgery, The University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, TX 78284-7777.
Presented at the Committee on Hearing, Bioacoustics and Biome-chanics Conference on Objective Audiometry, Washington, DC, June 1-2,1992.
© Williams & Wilkins 1993. All Rights Reserved.