EEG Responses to Violations in Hierarchical Auditory Sequences
Author(s)
Pinson, Dessi
Advisor(s)
Editor(s)
Collections
Supplementary to:
Permanent Link
Abstract
The brain is thought to constantly generate predictions about incoming sensory input. One way to study this process is through the auditory system. Violations of expected patterns produce measurable neural responses, such as mismatch negativity (MMN). However, most prior work has focused on simple deviations rather than structured, multi-level sequences. The goal of this study was to examine how predictive processing operates in hierarchical levels.
Participants (N=12) listened to structured auditory sequences made up of tones, motifs, and sequences. Violations were added at three levels (tone, motif, and sequence). EEG data were recorded, and event-related potentials (ERPs) were measured around each violation. Mean amplitudes were calculated in early (100–250 ms) and later (300–500 ms) time windows. A repeated measured ANOVA was used to compare the conditions.
Tone violations showed the clearest ERP responses. Motif and sequence violations were more variable and not as strong. There were some differences between conditions, but they were not statistically significant. Early responses were in the expected MMN time window. Later responses did not show clear differences across conditions.
These results suggest that predictive processing may occur across multiple hierarchical levels. However, lower-level predictions appear to produce more robust neural responses than higher-level predictions. This study adds to previous research by examining prediction in structured auditory sequences and provides a basis for studying hierarchical processing in the brain.
Sponsor
Date
2026-05
Extent
Resource Type
Text
Resource Subtype
Undergraduate Thesis