Single Trial Eeg Decoding of Arm Extension and Flexion: A Case Study

Abstract

Brain-Machine Interfaces (BMIs) transcribe brain signals into commands that can be used for assistive devices and rehabilitation devices, such as prostheses, and exoskeletons. In this sense, decoding the cortical activity associated the execution or intention to execute motor actions may provide important elements to design future BMIs for controlling assistive devices. In this paper, we explored the hypothesis of using EEG signals in single trial to classify arm flexion and extension based on endogenous movement intention. One female subject, age of 20 years, right-handed, participated in the experiment. The subject executed the maximum flexion, subsequent to the maximum extension, according to her will. Each movement was repeated 50 times. EEG signals were epoched from -600 ms to 100 ms relative to the movement onset (zero mark). Single trial epochs of arm flexion and extension were classified using support vector machine (SVM). The best classifier performance was 73.78% (average of five holdout 10-fold cross-validation was 68.71%). The results are promising and show that EEG signals could be used to classify arm flexion and extension elicited by endogenous movement intention.