Characterizing P300 Speller Performance: The Effects of Stimulus Timing Features and Search for a Physiological Biomarker

Copyright Information: All rights reserved unless otherwise indicated. Contact the author or original publisher for any necessary permissions. eScholarship is not the copyright owner for deposited works. Learn more at This thesis presents two studies that focus on improving P300 speller system efficiency by examining fundamental stimulus timing parameters and the possibility of a physiological biomarker to identify user performance. By analyzing four distinct stimulus presentation features, the first study reports that both stimulus-off time and interstimulus interval can significantly affect both accuracy and characters per minute. Optimization of stimulus timing parameters should not only consider accuracy rates, but also bit rates (i.e. characters per minute), considering that reductions in accuracy may be more than offset than the time saved resulting in overall improvements in system efficiency. Subsets of the entire group were compared to show iii consistency of performance trends despite great variance among subjects. The next study explored the possibility of an EEG biomarker that distinguished performer differences. Multiple extensive analyses showed no strong correlation between power in the EEG signal and user performance in spelling the temporally associated letter or word. iv The thesis of Jessica Lu is approved. Figure 1. Average classification accuracy for all subjects grouped by (A) stimulus timing pattern, (B) ISI*, (C) ISI, (D) flash duration, and (E) ratio of flash duration to ISI. SEM error bars are included in (A).. Figure 3. Average characters per minute for all subjects grouped by (A) stimulus timing pattern, (B) flash duration, (C) ISI*, (D) ISI, and (E) ratio of flash duration to ISI. SEM error bars are included for the largest CPM point of each average trend line.. Figure 5. Average group waveforms at Pz for each stimulus timing pattern. Waveforms for targets are represented by the solid black line and waveforms for non-targets are represented by Figure 6. Average classification accuracy (AD) and CPM (E-H) under the 32-channel set (represented by the solid black line) compared to the 8-channel set (represented by the dashed black line) for each of the four patterns.. Figure 8a and b. Sections highlighted in gray indicate the time in the EEG signal used to vii Figure 9. Power spectral density estimate in channel Cz of the top three performers (solid black) compared to the bottom three performers (dotted black). The low gamma (30-70Hz) frequency Figure 10. Gamma power in electrode Cz along the x-axis is plotted against a measurement …