Optimal parameters for measuring multiband auditory brainstem responses to continuous speech

Abstract Accurate clinical hearing assessment depends on efficient, engaging measures designed to evaluate ecologically relevant stimuli. Often brief tones or narrowband noise stimuli are used, providing a useful but limited snapshot of hearing function. Including dynamic speech offers a means to capture how the hearing system encodes complex sounds critical for everyday communication. Here we describe the optimal parameters for using audiobook continuous speech with the multiband peaky speech paradigm to measure frequency-specific auditory brainstem responses (ABRs) to standard audiological octave bands from 500–8000 Hz in each ear simultaneously. Using computational modeling and direct human ABR testing in adults with normal hearing, we demonstrate that continuous speech signals with a chirp phase profile and fundamental frequency (f0) lowered to the range of 90–110 Hz evoke the largest ABR wave V amplitudes. This amplitude boost occurs when any narrator’s f0 is lowered to this optimal range, but the largest responses occur for narrators with original f0s below 170 Hz. We also confirmed that different narrator speech stimuli with these optimized parameters can evoke similarly sized ABRs, but some minor differences remain for testing time. Ultimately, optimizing phase-f0 parameters substantially sped up the median testing time to obtain robust audiobook-based multiband ABRs to within 14 minutes, thereby making this paradigm more feasible for future research and clinical translation.