Channels Selection for Pattern Recognition of Five Fingers Motor Imagery Electroencephalography Signals

Abstract

Research about pattern recognition on electroencephalography (EEG) signal of finger motor imagery (MI) plays a critical role in Brain-Computer Interfaces (BCI) based hand prosthetics development. However, the previous research still used irrelevant channels to finger MI. This work proposed optimal EEG channels combination for five-finger MI. It is achieved by subject-dependence channel selection using One versus Rest Common Spatial Pattern (CSP-OVR) combined with sequential searching algorithms due to specific neural activation areas of MI. Optimal channels combinations are of great importance to reduce channels number. It supports the development of practical BCI-based hand prosthetics that can help hand handicapped to do daily activities easier. Experimental results show 4 out of 19 channels are relevant to five-finger MI with 0,6% accuracy degradation compared with EEG-MI pattern recognition using 19 channels. This result is better than the Principal Component Analysis (PCA) channel selection method that only selects 11 out of 19 channels with 1 % accuracy degradation.

EEG-based finger movement classification with intrinsic time-scale decomposition

“…Table 10 presents a comparison of the suggested study to relevant prior studies. Clearly, both subject-dependent (Kaya et al, 2018 ; Anam et al, 2019 , 2020 ; Kato et al, 2020 ; Mwata-Velu et al, 2021 , 2022 ; Azizah et al, 2022 ) and subject-independent (Kaya et al, 2018 ; Zahra et al, 2022 ) studies were conducted for FM classification in literature. In general, the highest performance values were achieved in subject-dependent classification as in our study.…”

“…Thus, a comparison between the results of this study (Anam et al, 2020 ) and ours would not be meaningful. On the other hand, some of these prior studies (Mwata-Velu et al, 2021 , 2022 ; Azizah et al, 2022 ) performed channel reduction. In these studies, four out of all 19 channels were defined as effective channels and used for the feature extraction stage.…”

“…The classification performance was found to be 100% for training accuracy for each subject and the test accuracy performances ranged between 51.00 and 56.00%. In 2022, Azizah et al (2022) carried out a subject-dependent FM classification study. They performed channel selection based on One vs. Rest Common Spatial Pattern (CSP-OVR) and four out of 19 EEG channels were defined as relevant channels in their study.…”

“…The raw EEG time series (Kaya et al, 2018 ; Mwata-Velu et al, 2022 ; Zahra et al, 2022 ), different amplitude-based, and statistical-based EEG signal features (Degirmenci et al, 2024 ) were utilized to examine the effectiveness of the time domain. As for the spectral-domain features, different frequency-domain [Fourier transform (Kaya et al, 2018 )] and time-frequency domain [Wavelet transform (Yahya et al, 2019 ), Short-time Fourier transform (Azizah et al, 2022 ), Empirical mode decomposition (Mwata-Velu et al, 2021 )] representation algorithms and their various versions were investigated to classify FM. Common spatial pattern (Anam et al, 2019 , 2020 ) and its different versions (Kato et al, 2020 ) are one of the most experimented methods for the analysis of spatial domain in FM classification.…”

“…In 2022, Azizah et al ( 2022 ) carried out a subject-dependent FM classification study. They performed channel selection based on One vs. Rest Common Spatial Pattern (CSP-OVR) and four out of 19 EEG channels were defined as relevant channels in their study.…”