Investigating short and long-term variability of EEG-based features in BCI applications

A Brain-Computer Interface (BCI), is a device that employs brain signals to provide a non-muscular outward channel, and the P300 event-related potential, elicited by external rare and relevant stimuli approximately after 300 ms from the onset, has proven to be a reliable response to control a BCI [1][2]. Yet, the morphological characteristics of the P300 (amplitude and latency) vary throughout the day, clearly impacting the performance of a P300-based BCI system [3]. The aim of the present work is therefore to analyze the variability of EEG-based features (i.e. P300) throughout the day, both in terms of morphology and classification performance, as a preliminary study for future BCI applications addressed to neurological patients with attention deficit disorders. Analysis showed significant differences between the variability of amplitudes vs latencies, while no difference emerges form the between-session comparison. Performance indices partially showed significant within-session variation, while no between-session variation was accounted. [1] Wolpaw J. R. et al.: Brain-computer interfaces for communication and control. Clinical Neurophysiology – vol. 113, Issue 6, (2002), pp 767- 791. [2] Polich J.: Updating P300: An integrative theory of P3a and P3b. Clinical Neurophysiology – vol. 118, Issue 10, (2007), pp 2128-2148. [3] Wesensten N. J. et al.: Time of day, repeated testing, and interblock interval effects on P300 amplitude. Physiology & Behavior – vol. 41, Issue 4, (1990), pp 653-658.