P300-based brain computer interface: Reliability and performance in healthy and paralysed participants

• Published in: Clinical neurophysiology 2006-03, Vol.117 (3), p.531-537
• Main Authors: Piccione, F., Giorgi, F., Tonin, P., Priftis, K., Giove, S., Silvoni, S., Palmas, G., Beverina, F.
• Format: Article
• Language: English
• Subjects: Adult; Aged; BCI; Behavioral psychophysiology; Biofeedback, Psychology; Biological and medical sciences; Brain - physiopathology; Electrodiagnosis. Electric activity recording; Electroencephalography - methods; Electrophysiology; ERP; Event-Related Potentials, P300 - physiology; Female; Fundamental and applied biological sciences. Psychology; Humans; ICA; Investigative techniques, diagnostic techniques (general aspects); Locked-in syndrome; Male; Medical sciences; Middle Aged; Nervous system; Neural network; P300; Paralysis - physiopathology; Pattern Recognition, Visual - physiology; Photic Stimulation - methods; Psychology. Psychoanalysis. Psychiatry; Psychology. Psychophysiology; Reproducibility of Results; Signal Processing, Computer-Assisted; User-Computer Interface
• ISSN: 1388-2457; 1872-8952
• DOI: 10.1016/j.clinph.2005.07.024

This study aimed to describe the use of the P300 event-related potential as a control signal in a brain computer interface (BCI) for healthy and paralysed participants. The experimental device used the P300 wave to control the movement of an object on a graphical interface. Visual stimuli, consisting of four arrows (up, right, down, left) were randomly presented in peripheral positions on the screen. Participants were instructed to recognize only the arrow indicating a specific direction for an object to move. P300 epochs, synchronized with the stimulus, were analyzed on-line via Independent Component Analysis (ICA) with subsequent feature extraction and classification by using a neural network. We tested the reliability and the performance of the system in real-time. The system needed a short training period to allow task completion and reached good performance. Nonetheless, severely impaired patients had lower performance than healthy participants. The proposed system is effective for use with healthy participants, whereas further research is needed before it can be used with locked-in syndrome patients. The P300-based BCI described can reliably control, in ‘real time’, the motion of a cursor on a graphical interface, and no time-consuming training is needed in order to test possible applications for motor-impaired patients.