Abstract
Recently motor imagery (MI) based Brain-Computer Interface (BCI) for lower limb rehabilitation is gaining attention. Feature extraction and dimensionality reduction are crucial signal processing blocks that determine the performance of a BCI system. In this work, various features, that are, band power (BP) features, autoregressive (AAR) parameters and Hjorth (HJ) parameters, widely used in BCI research are studied for their efficacy in discriminating MI brisk walking activity from the idle state. Feature transformation (FT) techniques, a type of dimensionality reduction techniques, namely Principal Component Analysis (PCA), Locality Preserving Projections (LPP) and Local Fisher Discriminant analysis (LFDA) are then applied on the extracted features to map them into a lower dimensional subspace. Ten-fold cross-validation is used to choose the dimension of the projection subspace. In a group of five novice users, it is observed that none of these features separately or all taken together represented the activity well. On using FT techniques, the discriminability of the fused features improved. Among the three techniques, LFDA performed the best showing an average increase in classification accuracy (26.9%), sensitivity (37.6%) and specificity (26.2%) over the average values obtained when no FT technique are used for the group of five subjects.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.