Using Machine Learning to Understand the Relationships Between Audiometric Data, Speech Perception, Temporal Processing, And Cognition
Rana M. Khalil, Alexandra Papanicolaou, Renee Ti Chou, Bobby E. Gibbs, Samira Anderson, Sandra Gordon-Salant, Michael P. Cummings, Matthew J. Goupell
Abstract
Aging and hearing loss cause communication difficulties, particularly for speech perception in demanding situations, which have been associated with factors including cognitive processing and extended high-frequency (>8 kHz) hearing. Quantifying such associations and finding other (possibly unintuitive) associations is well suited to machine learning. We constructed ensemble models for 443 participants who varied in age and hearing loss. Audiometric, perceptual, electrophysiological, and cognitive data were used to predict speech perception in noise, reverberation, and with time compression. Speech perception was best predicted by variables associated with audiometric thresholds (including new across-frequency composite variables) between 1–4 kHz, followed by basic temporal processing ability. Cognitive factors and extended high-frequency thresholds had little to no predictive ability of speech perception. Future associations or lack thereof will inform the field as we attempt to better understand the intertwined effects of speech perception, aging, hearing loss, and cognition.
BibTeX
@inproceedings{icassp2023_usingmachinelear,
title = {Using Machine Learning to Understand the Relationships Between Audiometric Data, Speech Perception, Temporal Processing, And Cognition},
author = {Rana M. Khalil and Alexandra Papanicolaou and Renee Ti Chou and Bobby E. Gibbs and Samira Anderson and Sandra Gordon-Salant and Michael P. Cummings and Matthew J. Goupell},
booktitle = {ICASSP 2023},
year = {2023}
}