Evaluation of EEG and MEG responses during Fine Motor Imagery from the same limb.

Motor imagery (MI) is an important brain-computer interface (BCI) paradigm. BCI systems based on fine MI can provide an intuitive control pathway of the outer device. Electroencephalography (EEG) is a widely used modality for MI due to its high temporal resolution and portability. Magnetoencephalography (MEG) has high spatial and temporal resolution, which has received more and more attention. This study designed four kinds of MI tasks of different joints from the same upper limb, including finger, wrist, elbow, and shoulder joints, and additionally added a resting task. The EEG and MEG signals of eight subjects were acquired synchronously. Analysis was conducted on the EEG and MEG data to find the time, time-frequency, and spatial difference between MI tasks of different joints from the same limb. The induced event-related desynchronization (ERD) in EEG signals at the electrode position of the left motor area are more broad and stronger in the alpha frequency band than that in MEG signals during fine MI tasks. From the topographical distribution, different MI tasks affects the area and intensity of the activated area, and topographical distribution of MEG signals in different MI tasks are more discriminative than that of EEG signals. Moreover, the analysis of movement-related cortical potentials (MRCP) showed that significant negative potentials were detected near the onset of the motor imagery events and there is a significant difference in temporal dimension between magnetoencephalogram and electroencephalogram signals. The work implies that there exist the separable differences between EEG and MEG during fine MI tasks, which can be utilized to build a multimodal classification method for fine MI-BCI systems.