Presentation

Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams

Modulation of pupil dilation and alpha power during masked speech perception reveal distinct neural mechanisms contributing to listening effort

Poster B64 in Poster Session B and Reception, Thursday, October 6, 6:30 - 8:30 pm EDT, Millennium Hall

Sarah Villard¹, Tyler Perrachione¹, Sung-Joo Lim^1,2, Ayesha Alam¹, Gerald Kidd, Jr.¹; ¹Boston University, ²Binghamton University

Listening effort, or the cognitive effort exerted by an individual to process target speech under challenging conditions, is dissociable from task performance (Winn & Teece, 2021). Because many everyday communication situations involve background noise/speech (e.g., noisy restaurants, family gatherings), measuring listening effort under such conditions can provide important insight into the effect of masking on the listener. However, the study of listening effort is complicated by the finding that measurements of effort obtained through different physiological approaches (e.g., pupil size vs alpha power) are rarely correlated with one another (Alhanbali et al, 2019). This suggests that these psychophysiological indices may be sensitive to different mechanisms underlying the multidimensional construct of listening effort. The aim of this study was to compare the effort elicited by noise conditions producing primarily energetic masking (masking due to target-masker spectrotemporal overlap) to the effort elicited by an intelligible speech masking condition producing primarily informational masking (additional masking due to listener uncertainty/confusion), by examining changes in pupil size and alpha power. Fifteen young, normal hearing participants listened to target sentences while ignoring masker stimuli. Three different masking conditions were included: an intelligible speech masking condition; a speech-shaped, speech envelope-modulated noise masking condition; and a speech-shaped, unmodulated noise masking condition. We first used individual adaptive tracking to estimate the target-to-masker ratio (TMR) at which each participant could report 75% of the target words correctly, in each condition. Stimulus intensities were then held constant at these 75%-correct TMRs while both pupil size (via pupillometry) and neural alpha power (via EEG) were measured. This approach allowed for comparing these two physiological estimates of listening effort while controlling performance levels across participants and conditions. A 1 x 3 repeated measures ANOVA revealed a significant within-subject effect of condition on peak pupil size, F(2,28) = 7.79, p < 0.01. Post-hoc comparisons revealed that the speech masking condition elicited a significantly greater peak pupil size than either of the noise masking conditions (both p < 0.05). In terms of alpha power, a 1 x 3 repeated measures ANOVA revealed a significant within-subject effect of condition, F(2,26) = 4.53, p < 0.05. In contrast to the pupil size results, post-hoc comparisons revealed that the unmodulated noise masking condition elicited significantly greater alpha power than the speech masking condition (p < 0.01). Results suggest that pupil dilation may be more sensitive to the effort required to distinguish between highly confusable target and masker speech in a high-informational masking condition, whereas alpha power may be more sensitive to the effort required to perceptually “fill in” unavailable portions of the target signal in a high-energetic masking condition. Work supported by: NIH K99DC018829 (PI: Villard), NIH R01DC013286 (PI: Kidd), and a Plural Publishing Research Scholarship (Alam).

Topic Areas: Speech Perception, Perception: Auditory