Analysis of Auditory Evoked Potentials

Dobie, Robert

Abstract

The primary long-term objective of this research program continues to be the development of more efficient and accurate auditory evoked potential (AEP) tests, especially for threshold estimation in children and others who can't cooperate for conventional audiometry. Most of the analytic techniques under study are not unique to AEPs, however. Thus, this work is broadly relevant to evoked potentials, regardless of sensory modality,k and to signal processing for other types of evoked response (e.g. acoustic reflex or evoked oto-acoustic emissions). In a series of experiments using normal-hearing subjects and patients with hearing loss, scalp-recorded AEPs will be obtained and analyzed to answer the following questions: 1.Can coherence analysis be optimized by smoothing across data segments or frequency? 2.Will weighted averaging and/or Wiener filtering improve AEP and coherence estimates? 3.In normal subjects, is objective response detection using coherence analysis more accurate than human observers' judgments in detecting low- level AEPs? 4.In patients with hearing loss, are threshold estimates from coherence analysis more accurate than estimates from observer inspection of AEPs? Will high-pass masking improve low-frequency threshold estimates? 5.Can a single AEP (to a pseudo-random noise stimulus) be used to estimate responses to different spectral regions of the stimulus? 6.Can AEPs to complex AM tones be used to estimate system response to different modulation frequencies? 7.Will chloral hydrate sedation differentially affect AEP response to low modulation frequencies?