Intrasubject variability of early markers of sensorineural hearing damage

Abstract

The pursuit of noninvasive early markers for sensorineural hearing loss (SNHL) has yielded diverse measures of interest. However, comprehensive studies evaluating the test-retest reliability of multiple measures and stimuli within a single study are scarce, and a standardized clinical protocol for robust SNHL-markers remains elusive. To address these gaps, this study covers the intra-subject variability of potential EEG-biomarkers for cochlear synaptopathy (CS) and other SNHL-markers to determine their clinical suitability. Fifteen normal-hearing young adults underwent repeated measures of (extended high-frequency) pure-tone audiometry, speech-in-noise intelligibility, distortion-product otoacoustic emissions (DPOAEs), and auditory evoked potentials; comprising envelope following responses (EFR) and auditory brainstem responses (ABR). Results show high reliability in pure-tone audiometry, whereas the matrix sentence-test showed a significant learning effect. DP-grams and input-output functions’ reliability varied across three evaluation methods with distinct SNR-based criteria for DPOAE-datapoints. EFRs demonstrated superior test-retest reliability compared to ABR-amplitudes. Our findings underscore careful interpretation of presumed noninvasive SNHL measures. While we confirm the robustness of tonal-audiometry, we found a confounding learning effect in longitudinal speech audiometry. DPOAE variability underscores the need for consistent ear probe replacement and meticulous measurement techniques and renders I/O-functions unsuitable for clinical application. As potential EEG-biomarkers of CS, EFRs are favored over ABR-amplitudes.