Identification of Finger Movement from ECoG Signal Using Machine Learning Model

Epilepsy is a brain-related neurological disorder that increases a person's risk of experiencing repeated seizures. This study explores Electrocorticogram (ECoG) signals, which record brain electrical activity for epileptic patients in response to the need for innovative rehabilitation treatments. It uses Machine learning (ML) algorithms to detect finger flexion movements, which includes data preprocessing, feature extraction, and classification. The proposed approach includes the Discrete Wavelet Transform and Entropy, where these methods help extract features from ECoG data. Subsequently, several machine learning algorithms such as Logistic Regression (LR), Linear Discriminant Analysis (LDA), Random Forest (RF), and Support Vector Machine (SVM) are used for classification. A deep learning approach is also incorporated through the Convolutional Neural Network (CNN) model. The BCI Competition IV dataset 4, was utilized for training and evaluation, which involves finger motions recorded through ECoG. The collection includes data from three distinct subjects, resulting in a diversified set of ECoG recordings connected to finger flexion motions. In our work, the Random Forest (RF) classifier achieved the highest classification accuracy of 80%. This performance outperformed that of the other classifiers evaluated. This analysis contributes valuable insights into the effectiveness of the proposed machine learning approach in capturing finger flexion movements from ECoG signals.