Face coverings and speech recognition (Vos et al., 2022)
figureposted on 2022-03-10, 17:36 authored by Teresa G. Vos, Matthew M. Dedmon, Emily Buss, Meredith A. Rooth, Margaret E. Richter, Margaret T. Dillon
Purpose: The purpose of this study was to assess the influence of talker protective face coverings on sentence recognition in noise for cochlear implant users.
Method: The AzBio sentences were recorded in three conditions: (a) without any face covering (uncovered), (b) with an N95 mask, or (c) with an N95 mask plus face shield. Target sentences were presented at 60 dB SPL, and the 10-talker masker was presented at 50 dB SPL (10 dB signal-to-noise ratio. Speech recognition for these auditory stimuli was compared across conditions for 21 adult subjects with at least 6 months of cochlear implant (CI) use.
Results: Significant deterioration in sentence recognition was observed for the N95 plus face shield (Mdn = 27% [IQR: 14%–35%]) compared with the N95 (Mdn = 72% [IQR: 55%–78%]) condition and for the N95 compared to uncovered (Mdn = 86% [IQR: 68%–91%]) condition.
Conclusions: Talker protective face coverings have a significant influence on speech recognition in noise for CI users. More research is needed to understand the influence of visual cues with protective face coverings that include a plastic component allowing visualization of the talker’s face.
Supplemental Material S1. The long-term average speech spectrum for the AzBio sentences produced with and without face coverings and the 10-talker masker (7 lists).
Vos, T. G., Dedmon, M. M., Buss, E., Rooth, M. A., Richter, M. E., & Dillon, M. T. (2022). Effect of protective face coverings on sentence recognition in noise for cochlear implant patients. American Journal of Audiology. Advance online publication. https://doi.org/10.1044/2021_AJA-21-00173