EEG Alpha Power and Pupil Diameter Reflect Endogenous Auditory Attention Switching and Listening Effort

Haro, Stephanie; Rao, Hrishikesh M.; Quatieri, Thomas F.; Smalt, Christopher J. · 2021 · Crossref

DOI: 10.1101/2021.07.29.453646

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the neural and physiological markers of endogenous auditory attention switching, addressing a gap in research regarding the cognitive effort required to shift focus between competing speech sources. While auditory attention decoding (AAD) has been extensively studied in sustained attention scenarios, real-world listening involves frequent, voluntary switches that impose additional cognitive load due to working memory limits. The authors aimed to characterize this effort using electroencephalography (EEG) and pupillometry, and to evaluate whether an AAD algorithm could accurately track these transitions without relying on explicit motor responses like button presses, which introduce artifacts. The experimental protocol involved ten participants listening to two competing audiobook talkers presented from left and right speakers. Three conditions were tested: "at-will" switching (listener-initiated), "directed" switching (pre-defined time), and "sustained" attention (no switch). To avoid motor artifacts, participants memorized the time of the switch or a control event using a visual clock. The researchers employed a regularized least-squares decoder to predict the attended talker’s speech envelope from EEG data. Additionally, they analyzed centrotemporal alpha power via event-related spectral perturbation (ERSP) and mean pupil diameter (MPD) as indices of listening effort. Statistical analyses included ANOVA to compare conditions and lateralization metrics to assess hemispheric differences based on talker location. The results demonstrated that the EEG-based decoder achieved a grand mean accuracy of 64.1% with a 5-second correlation window. The decoder output showed smooth transitions, with the correlation with the initial talker weakening below that of the secondary talker approximately 2.2 seconds after the switch, indicating a lag of roughly half the analysis window. Crucially, listening effort metrics significantly differentiated switch trials from sustained trials. Centrotemporal alpha power [F(2, 18) = 7.473, P = 0.00434] and mean pupil diameter [F(2, 18) = 9.159, P = 0.0018] were significantly different during switching conditions. Furthermore, alpha lateralization was modulated by the interaction between experimental condition and the timing relative to the switch [F(2,18) = 3.227, P = 0.0634], suggesting that spatial location influences neural responses during attention shifts. These findings indicate that expended listening effort, measurable via alpha power and pupil dilation, serves as a strong indicator of attentional switching. The study concludes that incorporating these effort-based features into decoding algorithms, alongside speech and location cues, could enhance the robustness of cognitively-controlled hearing aids. By capturing preparatory activity associated with endogenous switches, future systems may reduce the latency of auditory enhancement, improving listener experience in complex, multi-talker environments.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• neuro workload indices