An EEG-EMG-kinematics dataset from wrist pointing tasks for biomarker research in neurorehabilitation

This work presents a multimodal dataset containing synchronized electroencephalography (EEG), electromyography (EMG), and kinematic recordings acquired during wrist motor tasks performed with a three degree of freedom robotic exoskeleton (BiomechWrist) coupled to a virtual interface. Designed as a normative baseline and benchmark resource for studying electrophysiological biomarkers and motor performance in healthy individuals, the dataset includes recordings from 45 healthy participants, each completing 320 trials of standardized wrist movements. The exoskeleton operated in transparent mode (actuators de-energized) to capture voluntary movements through high resolution encoders. Data are formatted according to the Brain Imaging Data Structure (BIDS) standard and follow FAIR principles, comprising raw biosignals, encoder trajectories, event markers, and derived performance metrics. To assess data quality, we provide subject level validation analyses, including power spectral density (PSD) and event related desynchronization/synchronization (ERDS) for EEG, as well as an EMG-Kinematic coupling analysis through Electromechanical Delay (EMD), and kinematic trajectory evaluation with performance metrics (accuracy, execution time, trajectory efficiency). This dataset supports research on wrist rehabilitation technologies and biomarker driven neuromodulation therapies, while also enabling studies in biosignal processing, artifact removal, machine learning for motor intention decoding, and the development of brain computer interfaces (BCI) and assistive devices targeting wrist mobility.