A Multimodal fNIRS-EEG Dataset for Unilateral Limb Motor Imagery

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Unilateral limb motor imagery (MI) is crucial for upper-limb rehabilitation and brain-computer interfaces.
• Existing datasets lack multimodal recordings (EEG/fNIRS) for complex, multi-directional MI tasks.
• Higher spatial resolution is needed for precise control and rehabilitation applications.

Purpose of the Study:

• To introduce the MIND dataset, a public multimodal dataset for motor imagery.
• To provide simultaneous electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) recordings.
• To support research in unilateral, multi-directional motor imagery of the right upper limb.

Main Methods:

• Collected 64-channel EEG and 51-channel fNIRS data from 30 participants.
• Utilized a four-class directional motor imagery paradigm for the right upper limb.
• Analyzed spatiotemporal characteristics of EEG spectral power and fNIRS hemodynamic responses.

Main Results:

• The MIND dataset provides comprehensive EEG and fNIRS recordings for detailed analysis.
• Baseline classification summaries demonstrate task-related information in individual and combined modalities.
• The dataset facilitates evaluation of neuroimaging analysis and decoding methods for upper-limb MI.

Conclusions:

• The MIND dataset addresses the need for multimodal, multi-directional MI data.
• It enables advanced research into upper-limb motor control and rehabilitation.
• This resource will accelerate the development and comparison of brain-computer interface technologies.