Abstract
In this paper, we present an optimal channel selection method to improve common spatial pattern (CSP) related features for motor imagery (MI) classification. In contrast to existing channel selection methods, in which channels significantly contributing to the classification in terms of the signal power are selected, distinctive channels in terms of correlation coefficient values are selected in the proposed method. The distinctiveness of a channel is quantified by the number of channels with which it yields large difference in correlation coefficient values for binary motor imagery (MI) tasks, rather than by the largeness of the difference itself. For each distinctive channel, a group of channels is formed by gathering strongly correlated channels and the Fisher score is computed using the feature output, based on the filter-bank CSP (FBCSP) exclusively applied to the channel group. Finally, the channel group with the highest Fisher score is chosen as the selected channels. The proposed method selects the fewest channels on average and outperforms existing channel selection approaches. The simulation results confirm performance improvement for two publicly available BCI datasets, BCI competition III dataset IVa and BCI competition IV dataset I, in comparison with existing methods.
Highlights
Brain-computer interfaces (BCIs) provide non-muscular communication between humans and computer using brain signals
The various extensions of the CSP which overcome a frequency band dependency problem have been proposed such as filter-bank CSP (FBCSP) [7], sub-band regularized CSP (SBRCSP) [8], filter-bank regularized CSP (FBRCSP) [9], sparse filter band common spatial pattern (SFBCSP) [10], and filter band combined with Tikhonov regularization CSP (FB-TRCSP) [11]
AND DISCUSSION we present the experimental results for BCI competition III Dataset IVa and BCI competition IV Dataset I
Summary
Brain-computer interfaces (BCIs) provide non-muscular communication between humans and computer using brain signals. Electroencephalogram (EEG)-based BCIs, which directly translate the intent reflected by EEG signals into a control command, have been used due to its high temporal resolution and non-invasiveness [1], [2]. Motor imagery (MI) is an area of active research in EEG-based BCIs, as the power of motor-relevant cortex-related EEG signals is decreased or increased during imaging of body movements; these changes are known as event-related desynchronization (ERD) or event-related synchronization (ERS) [3]. The temporally constrained sparse group spatial pattern (TSGCSP) [12] and sparse group representation model (SGRM) [13] is proposed to overcome a time period dependency and subject-dependency problem, respectively and shows improved performance for MI-classification
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: 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.
