Lowering narrator fundamental frequency optimizes brainstem response testing using audiobooks
Auditory Brainstem Responses (ABR) are used to assess hearing loss across different frequency ranges in infants and young children by recording the brain’s activity when sound is presented. Currently, ABRs are evoked using very brief, boring non-speech sounds. While effective and fast, these brief sounds do not reflect the brain’s response to everyday speech and are not engaging for young children. To address this issue, audiobooks can be modified to obtain similar ABRs as the non-speech audio across frequencies, called multiband peaky speech. But there are different parameters that affect the size of these speech-evoked ABRs and consequently testing time. This study tested variations of the narrator’s fundamental frequency (f0) to determine if shifting any narrator’s voice, whether initially high or low, to a low f0 produces the largest ABRs in the shortest recording time. We used two audiobooks - The Wizard of Oz with a low f0 narrator (125 Hz) and Toto’s Merry Winter with a high f0 narrator (183 Hz). We presented these two stories with the unaltered voices and the voices shifted down to an f0 of 100 Hz. Twenty-eight adults with normal hearing were recruited. They listened to a total of 2 hours 20 minutes of the four intermixed audiobook stimuli while their ABRs were recorded. Measures included the ABR amplitudes and the time needed for each ABR to reach a set signal-to-noise ratio (SNR) criterion that corresponds with visually identifiable responses. Both stories with a shifted f0 of 100 Hz had the largest amplitudes, but the ABRs from the Toto story – that required a larger f0 shift down to 100 Hz – were noisier and therefore took longer to reach the criterion SNR. The Wizard story – that had a smaller shift in f0 down to 100 Hz – evoked ABRs that reached the SNR criterion for all tested frequencies in the shortest time, with 80% of participants obtaining good ABRs within 35 minutes compared to 70% for Toto. In summary, the quickest and most effective stimuli for ABR testing with audiobooks uses a narrator with an originally low f0 that is slightly shifted down to a 100 Hz f0.