Dynamic assessment of listening effort by EEG alpha oscillations during an adaptive speech-in-noise test

Abstract

Abstract The task of speech recognition in noisy environments can be cognitively demanding for both hearing-impaired and normal-hearing individuals. Recent research emphasized brain oscillations, particularly alpha rhythms (8-13 Hz) from the electroencephalogram (EEG), as potential markers of processing auditory information in challenging listening contexts. However, most studies examined fixed listening demand with meaningful speech. Our study explores how alpha rhythms may quantify listening effort during an adaptive speech-in-noise task with reduced semantic complexity. For analysis, 11 participants with normal hearing were selected, and the individual speech recognition threshold was used to define two conditions with different task demands. Within each condition, multiple auditory stimuli were presented at different signal-to-noise ratios. The difference in listening demand between these conditions was reflected by reaction time and speech recognition performance. A cluster-based permutation test on EEG time-frequency data identified a fronto-central cluster in the alpha band. In-depth analysis of EEG dynamics revealed a significant difference in post-stimulus alpha event-related synchronization in a fronto-central cluster and a near-significant difference at central electrodes, providing preliminary support to the inhibition hypothesis. These findings further validate EEG-derived measures towards an accurate quantitative assessment of short-term neural responses in the context of a simple adaptive speech-in-noise test.