Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Neural oscillations as markers of prosodic and planning boundaries
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
This poster is part of the Sandbox Series.
Jungyun Seo1, Jelena Krivokapić1, Jonathan Brennan1; 1University of Michigan
Prosodic phrase boundaries are indicated by acoustic features including pauses along with lengthening of surrounding segments and boundary tones. Electrophysiological studies show that phase angles in the delta band (1-4 Hz) are more synchronized when there are prosodic boundaries, reflecting phase reset (Kayser et al., 2015; Chalas et al., 2023). However, pauses in speech is also related to planning, such that pauses are longer when preceding structurally complex or longer phrases to allow speakers more time to plan that material (e.g., Fuchs et al., 2013; Ferreira, 1991; Krivokapić, 2012). Are pauses that reflect prosodic structure (prosodic boundaries) processed differently from pauses that reflect planning (i.e., planning boundaries, see also Ferreira, 2007)? The present EEG study tests whether phase-reset occurs similarly at these different kinds of boundaries. Stimuli sentences consist of two phrases separated by a pause that was manipulated to reflect either a prosodic phrase boundary or planning boundary; these were recorded by a male L1 American English speaker. Prosodic phrases were induced via syntactic structure; planning boundaries were induced by manipulating task demands (the speaker determined which word to use at the beginning of the second phrase by counting dots presented on the screen.) All boundaries were produced with lengthening at the last word of the first phrase and pauses; pauses were further manipulated to be short (200 ±10 ms) or long (500 ±10 ms). 164 utterances were recorded for 41 target word sequences in each of the 4 conditions (2 boundary x 2 pause lengths). Participants (N=13; to date, target = 30, L1 American English speakers) are instructed to listen to stimuli and answer comprehension questions for 25% of the total stimuli during EEG recording. To measure phase reset around boundaries, we calculate inter-trial phase clustering (ITPC) in frequency bins from 1 to 10 Hz. This yields a spectro-temporal time-series where higher values indicate more phase synchrony. Given that phase resetting is typically characterized by increased phase consistency without a corresponding increase in power, we will conduct a power analysis to examine whether the higher ITPC in a given condition is accompanied by changes in power. The phase of neural oscillations would be aligned at prosodic (structure) boundaries and phase reset would occur at prosodic boundaries (Kayser et al., 2015; Chalas et al., 2023). For planning boundaries, if phase resets occur whenever listeners encounter pauses regardless of differences in two boundaries, the timing of phase resets will be similar across different boundaries, indicating that they are processed similarly. Alternatively, planning boundaries might show earlier or later starting of the phase reset as listeners process planning boundaries differently. This study contributes to understanding how the brain responds to prosodic information in the service of speech understanding. It further contributes to the integration of structural linguistic representation of boundaries with non-linguistic factors of speech planning in models of prosody.
Topic Areas: Prosody, Speech Perception