The specific aim of this project is to develop a compact, convenient, and easy-to-use wireless brain-computer interface (BCI) system that allows limited but extremely valuable communications with the most severely head injured patients using their brain wave. The hardware components include a miniature 8-channel (19 or 38-channel) transmitter capable of acquiring, amplifying, digitizing, and transmitting EEG/ERP signal, a receiver that is attached to a commercial PC, and a laptop PC, which can be easily fitted in a typical laptop PC case. A software package is developed for delivering user defined auditory/visual stimuli, acquiring and extracting ERPs, and classifying ERPs corresponding to each stimulus continuously in real-time. The primary challenge of this project is to classify match/mismatch event related potential (ERP) accurately and in real-time. An innovative multi-domain dense array neuroinformatic methodology is developed to classify the ERPs. Results from the preliminary study clearly show the feasibility of real-time classifying match/mismatch ERPs at high accuracy. The proposed brain-computer interface system will be evaluated in normal subjects during this phase I study. Successful implementation of the BCI system could have clinical significance by setting up a communication channel between physician and the severely injured patients and practical use. ? ?