Multimodal Data Processing and Brain Neural Information Mining Systems

In theory research domain, including of specific information extraction, information fusion-based decision,  evaluation and prediction modeling, etc.

In application research domain, the structure MRI (sMRI), functional MRI (fMRI), diffusion tensor MRI (DTI) and electroencephalogram (EEG) signals, which acquired both from patients and health persons, are combined for finding the biomarkers of treatment effect, and for establishing the models of neurorehabilitation evaluation and prediction, exceptionally used for individual patient.

The network science, including graphic and topological theories, are used for building both static and dynamic brain network model, and  for guiding topological data analysis and spatio-temporal measurement combined  structural and functional features of brain. The inverse EEG methods for source reconstruction or source localization are also studied for making the joint brain network analysis of sMRI, fMRI, and EEG (both electrode and source space). Evidence theory, statistical learning theory, deep learning models, etc. are applied for designing the multimodal information fusion systems, and for making the decision of neurorehabilitation predictors. Some research fruits have been used to study the altered brain volume, microstructure metrics and functional connectivity features in multiple system atrophy; zolpidem-related alterations and prognostic value in disorders of consciousness; Brain-Computer Interface Training in Stroke Patients; Microstate-specific functional connectivity and stroke-related alterations in Brain dynamics; Non-concomitant cortical structural and functional alterations in sensorimotor areas following incomplete spinal cord injury, etc.