Using EEG to Decode Upper Limb Movement: Elbow Flexion or Extension

The movement of prosthetic limbs is controlled through invasive or non-invasive interface methods. In this study, we analyze the difference between low-frequency electroencephalography (EEG) signals of two contrary movements, elbow flexion or extension. We use a public dataset consisting of 15 healthy subjects executing six different upper-limb movements. We extract the part of elbow movement part and perform a two-class classification. We use three models, logistic regression, KNN, and random forest model to obtain the accuracy of 95.1%, 94.7% and 92.6% on average respectively. This accuracy is an improvement based on the previous research at the same time.