Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Not just noise: Using aperiodic EEG activity to study listening effort in speech comprehension
Poster Session D, Saturday, October 26, 10:30 am - 12:00 pm, Great Hall 3 and 4
Sarah J. Woods1, Jack W Silcox1, Brennan R. Payne; 1University of Utah
Introduction Previously considered artifactual noise, aperiodic neural activity present in EEG has recently been shown to reflect an important and functionally relevant signal that is sensitive to individual differences and cognitive task manipulations. Given this sensitivity of aperiodic activity to cognitive factors, it could serve as neural index of listening effort, or the cognitive resources allocated to accomplish a listening task. In order to assess the potential of aperiodic EEG to index LE in speech comprehension, we conducted secondary analyses of two datasets in which participants listened to speech in quiet or in moderate background noise, shown to induce increases in listening effort. Methods We decomposed EEG power spectra into periodic (i.e., alpha oscillation) and aperiodic EEG (i.e., broadband 1/f spectral slope and offset) activity in young normal-hearing adults (Study 1, N = 31) and then replicated our results in an older adult sample with a variety of levels of hearing acuity ranging from normal hearing to moderate hearing loss (Study 2, N = 48). Participants in both groups listened to a total of 240 sentences with 120 presented in quiet and 120 presented with the addition of +3 dB SNR speech-shaped background noise while continuous EEG was recorded. Procedures for the two studies were nearly identical save three changes to better accommodate the older adult participants. First, older adults were pre-screened with the Montreal Cognitive Assessment. Next, young adult participants were presented with the stimuli binaurally and older adults were presented in their better-hearing ear. Finally, younger adults were presented with the stimuli at a fixed level of 50 dB HL for all participants whereas older adults were presented with the stimuli at a level 40 dB over their personal speech-reception threshold, up to a maximum level of 70 dB HL. Results In both samples, we found the spectral slope and offset to be sensitive to the acoustic challenge manipulation, such that the broadband slope flattened (i.e., decreased) and overall offset decreased with increasing acoustic challenge. In both groups, this aperiodic effect showed a strong pre-frontal distribution. In contrast, whereas aperiodic activity predicted acoustic challenge in both age groups EEG alpha power was only predictive in the younger group. Of note, periodic activity was only significant when accounting for the change in underlying aperiodic activity. Finally, we present how individual differences in age, sex, cognitive functioning and hearing acuity relate to both periodic and aperiodic activity in speech comprehension. Discussion Collectively, our findings suggest that aperiodic neural activity may be an important neural feature to study in effortful speech comprehension. Specifically, we argue that the change in aperiodic activity we observed across both studies may reflect compensatory effort-related shifts in the balance between excitatory and inhibitory synaptic inputs in prefrontal neural circuits, effects that are distinct from previously-reported increases in narrowband parietal alpha activity. These findings have implications for not only for understanding the cognitive processes used while comprehending speech under effortful conditions but also for the processing of EEG data more generally.
Topic Areas: Speech Perception, Methods