TPCNet: A Temporal Periodicity Convolutional Network for motor imagery EEG decoding in stroke patients

Key Points

• Motor imagery EEG decoding achieved an accuracy of 87% in stroke patients, showing promise for rehabilitation integration.
• The analysis focused on motor imagery, using EEG to assess temporal periodicity as a key feature for network training.
• Observational analysis involved applying a novel convolutional network to EEG data from stroke patients to decode motor imagery.
• The findings may enable more effective rehabilitation strategies, though further validation in clinical settings is necessary.