Lower gamma band in the classification of left and right elbow movement in real and imaginary tasks

• Published in: Journal of the Brazilian Society of Mechanical Sciences and Engineering Vol. 41; no. 2; pp. 1 - 15
• Main Authors: Veslin, E. Y., Dutra, M. S., Bevilacqua, L., Raptopoulos, L. S. C., Andrade, W. S., Pereira, A. S., Fiorani, M., Soares, J. G. M.
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01-02-2019; Springer Nature B.V
• Subjects: Algorithms; Classification; Cognitive tasks; Discrimination; Engineering; Feature extraction; Intervals; Mechanical Engineering; Technical Paper; Rehabilitation robotics; Electroencephalography; Gamma band; Brain–computer interfaces; Machine learning
• ISSN: 1678-5878; 1806-3691
• DOI: 10.1007/s40430-019-1585-2

In this article, the activity of lower gamma band was used to classify right and left elbow movements performed from real and imaginary tasks in two different cognitive states: the preparation and the movement execution. Discriminability maps were used both to generalize the signal behavior of all the volunteers and to select time intervals of high discrimination among classes. The features extracted from chosen intervals were tested in eight different classification algorithms. To improve classes discrimination, LDA was used for dimensional reduction. Algorithms were tested using a fivefold cross-validation. The results showed similar signal activity for both real and imaginary actions, obtaining closer classification responses among all volunteers. The tested algorithm gave a mean classification success up to 70%, with minor differences between the type of task and the cognitive state.