Improving longitudinal P300-BCI performance for people with ALS using a data augmentation and jitter correction approach

Abstract

ABSTRACT The P300 Brain-Computer Interface (BCI) is a well-established communication channel for severely disabled people. However, variation in P300 latency, or latency jitter, is both increased in people with amyotrophic lateral sclerosis (ALS) and negatively associated with BCI performance. In this study, we proposed an augmentation and correction (A/C) characterization scheme with data augmentation and correction for jitter, both relying on time-shifted responses with individualized parameters determined based on latency jitter. We tested this approach offline on longitudinal data collected from six participants with ALS. While our longitudinal analysis showed decreased BCI performance and increased latency jitter over time with both our proposed characterization scheme and conventional methods, our proposed A/C characterization scheme significantly improved character selection accuracy, required for usability, along with recall and F-scores, showing the effectiveness of our proposed approach. These results should inform further work on improving longitudinal BCI performance and reliability for people with ALS.