Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Aperiodic neural activity as a putative index of listening effort for unfamiliarly accented speech
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
This poster is part of the Sandbox Series.
Sarah J. Woods1, Mel Mallard2, Kristin Van Engen2, Brennan R. Payne1; 1University of Utah, 2Washington University in St. Louis
Once considered a source of noise in the electroencephalography (EEG) signal, aperiodic (i.e., non-oscillatory) neural activity has been recently reconsidered to be a meaningful and dynamic signal itself. Changes in aperiodic activity – such as a flattening of the spectral slope – have been associated with increased age, decreased cognitive capacity, and increased task demands. For example, in our prior work we showed that listening to speech-in-noise was associated with reductions in the aperiodic slope and offset, compared to speech in quiet. For this reason, aperiodic EEG activity has emerged as a putative index of listening effort. Listening effort refers to the cognitive effort of speech processing triggered by challenging listening conditions. These conditions can vary greatly in the qualitative nature of their challenges. For example, background noise is a speech-extrinsic factor that reduces listeners’ access to the signal, whereas an unfamiliar accent is a speech-intrinsic factor that causes systematic deviations from listeners’ acoustic-phonetic expectations. These two types of listening challenge similarly impede access to accurate linguistic representations and require the listener to apply additional effort in order to achieve understanding; however, these two challenges’ exact neurocognitive impacts may differ. To address this question, this novel project extends the line of research on aperiodic neural activity as an index of listening effort into the realm of unfamiliarly accented speech. Our goal is to determine whether unfamiliarly accented speech yields similar or distinct aperiodic signatures to those previously seen for speech in noise. In our study, participants listen to and repeat aloud Hearing In Noise Test (HINT) sentences, after a brief delay, recorded by two male English speakers: one with an American accent (familiar) and one with a Mandarin Chinese accent (unfamiliar). We will record simultaneous pupillometry and EEG during listening, collect subjective measures at the end of the experiment, and score each spoken response for intelligibility via repetition accuracy. These methods will allow us to compare aperiodic measurements with previous work with this sentence set which showed disassociation between intelligibility and listening effort, as measured by pupil dilation, such that even sentences that were highly intelligible but unfamiliar were still more effortful to process. First, we expect that unfamiliarly accented speech will be associated with a flattening in the spectral slope of listeners’ aperiodic neural activity, indicating greater effort in processing the speech. Second, we expect to replicate prior work showing that listening to an unfamiliar accent results in increased task-evoked pupillary dilation. Finally, we will explore inter-correlations among measures of listening effort (pupil dilation, aperiodic activity, subjective effort, intelligibility), expecting to find that our physiologic indices will trend together, and that subjective effort will be a better predictor of these physiologic measures than intelligibility alone. In sum, the results of this work will provide us a better understanding of the psychophysiological processes that support communication across challenging listening conditions. Preliminary results will be presented.
Topic Areas: Speech Perception,