Using Brain Signal Patterns and Neural Networks to Identify Movements from EEG

The proposed CNN architecture achieved an average accuracy of 97.65% for slow movements and 96.25% for fast movements in distinguishing eight different shoulder, wrist, and elbow movements using EEG signals.

• An EEG-based brain-computer interface system can enhance control of external prostheses by recognizing various movements through brain signals.
• The study utilized a combination of one-versus-rest common spatial pattern and convolutional neural network techniques to classify movements.
• Ten subjects participated, with EEG signals recorded during both fast and slow movement speeds.
• The CNN method significantly outperformed traditional classifiers such as KNN, SVM, and MLP in movement classification.
• These findings indicate potential improvements in the functionality of BCI systems for individuals with movement disabilities.

Simplified