作者: Adham Atyabi , Martin Luerssen , Sean P. Fitzgibbon , David M. W. Powers
关键词:
摘要: Dimension reduction is an important step toward asynchronous EEG based BCI systems, with EA Feature/ Electrode Reduction (FR/ER) methods showing significant potential for this purpose. A PSO approach can reduce 99% of the data in manner while demonstrating generalizability through use 3 new subsets features/electrodes that are selected on best performing subset validation set, testing and most commonly used swarm. This study focused applying generated from 4 subjects a 5th one. Two schemes implemented i) extracting separate feature/electrodes each subject (out subjects) combining final products together subject, ii) concatenating preprocessed desired subject. The results indicate feasibility generating feature/electrode indexes task specific be subjects.