TY - JOUR T1 - EEG correlates of P300-based brain-computer interface (BCI) performance in people with amyotrophic lateral sclerosis. JF - Journal of neural engineering Y1 - 2012 A1 - Mak, Joseph N. A1 - Dennis J. McFarland A1 - Theresa M Vaughan A1 - McCane, Lynn M. A1 - Tsui, Phillippa Z. A1 - Zeitlin, Debra J. A1 - Sellers, Eric W. A1 - Jonathan Wolpaw KW - User-Computer Interface AB - The purpose of this study was to identify electroencephalography (EEG) features that correlate with P300-based brain-computer interface (P300 BCI) performance in people with amyotrophic lateral sclerosis (ALS). Twenty people with ALS used a P300 BCI spelling application in copy-spelling mode. Three types of EEG features were found to be good predictors of P300 BCI performance: (1) the root-mean-square amplitude and (2) the negative peak amplitude of the event-related potential to target stimuli (target ERP) at Fz, Cz, P3, Pz, and P4; and (3) EEG theta frequency (4.5-8 Hz) power at Fz, Cz, P3, Pz, P4, PO7, PO8 and Oz. A statistical prediction model that used a subset of these features accounted for >60% of the variance in copy-spelling performance (p < 0.001, mean R(2)?= 0.6175). The correlations reflected between-subject, rather than within-subject, effects. The results enhance understanding of performance differences among P300 BCI users. The predictors found in this study might help in: (1) identifying suitable candidates for long-term P300 BCI operation; (2) assessing performance online. Further work on within-subject effects needs to be done to establish whether P300 BCI user performance could be improved by optimizing one or more of these EEG features. VL - 9 UR - http://www.ncbi.nlm.nih.gov/pubmed/22350501 ER - TY - JOUR T1 - A brain-computer interface for long-term independent home use. JF - Amyotrophic lateral sclerosis : official publication of the World Federation of Neurology Research Group on Motor Neuron Diseases Y1 - 2010 A1 - Sellers, Eric W. A1 - Theresa M Vaughan A1 - Jonathan Wolpaw KW - User-Computer Interface AB - Our objective was to develop and validate a new brain-computer interface (BCI) system suitable for long-term independent home use by people with severe motor disabilities. The BCI was used by a 51-year-old male with ALS who could no longer use conventional assistive devices. Caregivers learned to place the electrode cap, add electrode gel, and turn on the BCI. After calibration, the system allowed the user to communicate via EEG. Re-calibration was performed remotely (via the internet), and BCI accuracy assessed in periodic tests. Reports of BCI usefulness by the user and the family were also recorded. Results showed that BCI accuracy remained at 83% (r = -.07, n.s.) for over 2.5 years (1.4% expected by chance). The BCI user and his family state that the BCI had restored his independence in social interactions and at work. He uses the BCI to run his NIH-funded research laboratory and to communicate via e-mail with family, friends, and colleagues. In addition to this first user, several other similarly disabled people are now using the BCI in their daily lives. In conclusion, long-term independent home use of this BCI system is practical for severely disabled people, and can contribute significantly to quality of life and productivity. VL - 11 UR - http://www.ncbi.nlm.nih.gov/pubmed/20583947 ER - TY - JOUR T1 - A comparison of classification techniques for the P300 Speller. JF - Journal of neural engineering Y1 - 2006 A1 - Krusienski, Dean J. A1 - Sellers, Eric W. A1 - Cabestaing, François A1 - Bayoudh, Sabri A1 - Dennis J. McFarland A1 - Theresa M Vaughan A1 - Jonathan Wolpaw KW - Normal Distribution AB - This study assesses the relative performance characteristics of five established classification techniques on data collected using the P300 Speller paradigm, originally described by Farwell and Donchin (1988 Electroenceph. Clin. Neurophysiol. 70 510). Four linear methods: Pearson's correlation method (PCM), Fisher's linear discriminant (FLD), stepwise linear discriminant analysis (SWLDA) and a linear support vector machine (LSVM); and one nonlinear method: Gaussian kernel support vector machine (GSVM), are compared for classifying offline data from eight users. The relative performance of the classifiers is evaluated, along with the practical concerns regarding the implementation of the respective methods. The results indicate that while all methods attained acceptable performance levels, SWLDA and FLD provide the best overall performance and implementation characteristics for practical classification of P300 Speller data. VL - 3 UR - http://www.ncbi.nlm.nih.gov/pubmed/17124334 ER - TY - JOUR T1 - A P300 event-related potential brain-computer interface (BCI): the effects of matrix size and inter stimulus interval on performance. JF - Biological psychology Y1 - 2006 A1 - Sellers, Eric W. A1 - Krusienski, Dean J. A1 - Dennis J. McFarland A1 - Theresa M Vaughan A1 - Jonathan Wolpaw KW - Amyotrophic Lateral Sclerosis KW - brain-computer interface KW - electroencephalogram KW - event-related potentials KW - P300 KW - Rehabilitation AB - We describe a study designed to assess properties of a P300 brain-computer interface (BCI). The BCI presents the user with a matrix containing letters and numbers. The user attends to a character to be communicated and the rows and columns of the matrix briefly intensify. Each time the attended character is intensified it serves as a rare event in an oddball sequence and it elicits a P300 response. The BCI works by detecting which character elicited a P300 response. We manipulated the size of the character matrix (either 3 x 3 or 6 x 6) and the duration of the inter stimulus interval (ISI) between intensifications (either 175 or 350 ms). Online accuracy was highest for the 3 x 3 matrix 175-ms ISI condition, while bit rate was highest for the 6 x 6 matrix 175-ms ISI condition. Average accuracy in the best condition for each subject was 88%. P300 amplitude was significantly greater for the attended stimulus and for the 6 x 6 matrix. This work demonstrates that matrix size and ISI are important variables to consider when optimizing a BCI system for individual users and that a P300-BCI can be used for effective communication. VL - 73 UR - http://www.ncbi.nlm.nih.gov/pubmed/16860920 ER -