Classification of fNIRS Signals for Decoding Right- and Left-Arm Movement Execution Using SVM for BCI Applications
Brain–computer interface-based systems help people who are incapable of interacting with the external environment using their peripheral nervous system. BCIs allow users to communicate purely based on their mental processes alone. Signals such as fNIRS corresponding to the imagination of various limb movements can be acquired noninvasively from the brain and translated into commands that can control an effector without using the muscles. The present study aims at classifying Right-Arm and Left-Arm movement combination using SVM. The study also aims at analyzing the efficacy of two different features, namely average signal amplitude and the difference between the average signal amplitudes of ΔHbO and ΔHbR on the accuracies obtained. The combination of these two features is also explored. The results of the study indicate that chosen features yield average accuracies between 70 and 76.67% calculated for all the subjects. The difference of mean amplitudes of ΔHbO and ΔHbR is investigated as one of the features for fNIRS-BCI application, and it yields an average accuracy of 70%. It indicates the possibility of using this feature for evaluating the binary BCI system for practical communication use. Two-feature combination improved the average accuracy value from 70 to 76.67%. The results obtained from the study suggest that distinct patterns of hemodynamic response arising out of Right-Arm and Left-Arm movements can be exploited for the development of BCI which are best described by the features used in the present study.
KeywordsFunctional near-infrared spectroscopy Brain–computer interface Support vector machine
- 4.Sitaram R, Zhang HH, Guan CT, Thulasidas M, Hoshi Y, Ishiawa A et al (2007) Temporal classification of multichannel near-infrared spectroscopy signals of motor imagery for developing a brain-computer interface. Neuroimage 34:1416–1427. https://doi.org/10.1016/j.neuroimage.2006.11.005NCrossRefGoogle Scholar
- 6.Buccino AP, Keles HO, Omurtag A (2016) Hybrid EEG-fNIRS asynchronous brain-computer interface for multiple motor tasks. PLoS ONE 11(1)Google Scholar
- 9.Strait M, Scheutz M (2014) What we can and cannot (yet) do with functional near infrared spectroscopy. Front Neurosci 8Google Scholar
- 10.Naseer N, Qureshi NK, Noori FM, Hong KS (2016) Analysis of different classification techniques for two-class functional near-infrared spectroscopy-based brain-computer interface. Comput Intell Neurosci, 5480760, https://doi.org/10.1155/2016/5480760