2-D Cursor Movement using EEG

The demonstration will show a non-invasive brain-computer interface device used to control continuous cursor movements in two dimensions. The decision of which direction to move the cursor is made by analyzing ERP components in the time-series EEG signal, using the continuous wavelet transform for time-scale analysis. These ERP components are observed as the result of a stimulus (the cursor moving), and are recorded for 300-1000 milliseconds after the stimulus. Details 1) The moving cursor on the screen is continually observed by the subject, and the results are recorded in 300-1000 ms windows, with a new window being generated every ~100 ms. 2) The EEG trace has filters applied. 3) Important features of the EEG signal are extracted using the continuous wavelet transform for time-scale analysis. 4) The important features are used to determine the new direction for the cursor to proceed. This approach is novel in the area of 2-D EEG movement control because: - The training time for a user is minimal, no more than a few minutes. Other techniques for 2-D movement involve the subject learning to evoke potentials in the EEG trace over a period of days or weeks. - Physical movement analogues, such as imagined movements of the feet or hands, are not employed. - The system will work with low-cost, low-channel EEG hardware. The system will be shown with the Emotiv EPOC hardware, a 14-channel EEG headset, costing approximately $300 USD.