Undergraduate Honors Theses

Thesis Defended

Spring 2016

Document Type

Thesis

Type of Thesis

Departmental Honors

Department

Psychology & Neuroscience

First Advisor

Tim Curran

Second Advisor

Randy O'Reilly

Third Advisor

Willem Schreüder

Abstract

DTW Barycenter Averaging (DBA) has proven to be a useful tool for calculating consensus sequences, but it has not yet been applied to real electroencephalography (EEG) data. This study tests DBA on real EEG sequences using the modification proposed by Kotas et al. (2015), and proposes several further modifications to the initial sequence selection process to improve the method’s efficacy for EEG analysis. Errors in peak latency and peak amplitude measures for a single EEG component, namely N250, were measured to test each method. Three of the proposed DBA variations produced consensus sequences that were significantly more accurate for replicating features of single-trial N250 components than the widely-used Event-Related Potential (ERP) technique. Potential implications include the uncovering of previously obscured effects in EEG data and providing more accurate descriptions of the prototypical electrophysiological responses to external events.

dba_for_eeg.m (11 kB)

Share

COinS