Denoising and Decoding the Time Course of Single-Joint Movements from Electroencephalography Cortical Currents

Although electroencephalography (EEG) sensor data have shown that slow cortical potentials correlate with motor neural activities, it remains unclear whether human arm trajectories can be reliably reconstructed from the cortical currents estimated from EEG sensor data. We addressed this gap by measuring scalp EEG data while participants performed single-joint movements. We simultaneously estimated the cortical and extra-brain source currents from the EEG data and reconstructed trajectories from the cortical currents using a sparse linear regression method. Test datasets demonstrated good performance, and the weighted values were distributed on multiple cortical areas related to reaching movements. These results indicate that our method reconstructed trajectory information from EEG cortical currents.