Mental States
In this project, we recorded EEG signals when human participants performed mental resource-required tasks and investigated brain activities associated with mental states. We also detected the levels of mental states using machine learning methods.
The illustration of intraclass correlation coefficient computation of the measures for mental fatigue.
Brain Diseases and Diagnosis
In the project, we attempt to explore diverse data, including EEG, DTI and fMRI, for understanding brain diseases and revealing disease-related characteristics. Based on the findings derived from the data analysis, we also develop machine learning methods to diagnose brain diseases automatically.
Consensus connections and important regions associated with schizophrenia
The coupling between functional connectivity and structural connectivity in schizophrenia
(A) The significant nodal metrics between groups at Execution. The red spheres indicate the regions with significantly (p< 0:05, after FDR correction) higher value for TD while the blue spheres represent the regions with the opposite case. (B) The significant nodal metrics between phases for TD. The red spheres indicate the regions with significantly (p< 0:05, after FDR correction) increasing value at Execution relative to Pre-Stimulus while the blue spheres represent the regions with the opposite case. No significantly changing nodal metrics (after correction) were found for HFA.
Ageing & Dementia
This project investigates the factors that contribute to brain ageing and dementia, and finds manifestations from the perspective of neuroimaging. In addition, a number of interventions are adopted to explore their effects on the mitigation of ageing.
Flowchart of participants through trial
Exoskeleton-Aided Walking
In this project, we aim to understand the effect of an exoskeleton on the brain activities when the walking is assisted by the exoskeleton. We recorded both EEG that reflects brain activities and EMG that reflects muscular activities and performed a comprehensive investigation. We revealed the relationships between brain activities and muscular activities. We also localised the relevant regions that were involved in the exoskeleton-aided walking, as well as the effect of the exoskeleton on the brain activities.
Brain-Computer Interfaces
In this project, new effective training paradigms were proposed to facilitate the calibration of brain-computer interface systems. A few practice-oriented prototypes of BCI systems were designed and developed.
Brain-Driven Wheelchair
Bilateral Training Paradigm
Active Training Paradigm
Multi-Person Car Racing
Deep Learning Models & Other Methods
In the project, we propose advanced deep learning models and other methods which suit to recognise/classify neurophysiological signals.
Multi-Kernel CapsNet
Joint Label-Common and Label-Specific Feature Exploration Model
Spatio-Spectral Representation Learning