IPSIHAND AN EEG BASED BRAIN COMPUTER INTERFACE FOR MOTOR REHABILITATION MOTIVATION 900,000 individuals in US with severe difficulty grasping Causes: Stroke Traumatic Brain Injury Spinal Cord Injury Neurodegenerative diseases MOTIVATION 1.Restore hand control 2.Provide novel rehabilitation therapy RECORDING TECHNIQUES Vs. SCREENING PROCEDURE 2 Conditions: Left Hand Movement Rest Look for change in EEG signal between conditions SCREENING DATA CURRENT SIGNAL PROCESSING 1.Band-Pass Filtering 2.Spatial Filtering 1.Raw 2.Common Average 3.Bipolar pick an electrode 3.Autoregressive Spectral Estimation 4.Feature Selection 5.Control Signal Normalization, Adaptation 0 mean, unit variance Adapted to buffer of previous data ALTERNATIVE ADAPTATION TECHNIQUES Least Mean Squares Linear Regression PERFORMANCE RESULTS Actual Hand MovementImagined Hand MovementLeft vs. Right CHALLENGES, POSSIBLE PROJECTS Signal Processing (subject specific, non-stationary) Spatial Filter optimization Feature identification Adaptation Automation Hardware & Software Miniaturization embedded platforms Power Consumption Comparison with alternative BCI software platforms (OpenViBE) Performance 2D, 3D control Latency reduction Accuracy gain Alternative actuators