NOURISH Programme
Artificial intelligence and big data have changed the world in an unprecedented way. Brain-computer interface (BCI) technology will be encompassed in the next wave of artificial intelligence and data science to create new digital health solutions for mental health and brain disorders.
Stroke is the top cause of long-term disability in Singapore (63% of stroke patients have some disability at three months) and the burden of stroke is expected to increase dramatically owing to Singapore’s rapidly ageing population and stroke risk factors. Current gold-standard interventions for stroke are either drug-based or therapies limited by the target functionalities. State-of-the-art research on BCI-based stroke rehabilitation and cognitive training primarily focus on upper limb rehabilitation and motor function only. Little research has been done in incorporating motor, cognition, and emotion training in an integrated solution for treating brain and boost the patient outcome.
In this programme, we propose to develop a digital healthtech solution to address the healthcare challenges: Next-Generation Brain-Computer-Brain Platform – A Holistic Solution for the Restoration & Enhancement of Brain Functions (NOURISH). The NOURISH programme proposes to develop the best-in-class, holistic solution for the restoration and enhancement of motor, cognitive and emotional functions in one system. The NOURISH platform would offer a non-drug based, cost-effective and clinically feasible solution for effective stroke rehabilitation and mood therapy.
The programme will capitalize on powerful and advanced machine learning algorithms including deep learning, deep transfer learning and deep reinforcement learning for source-space, multi-view and multimodal brain state decoding as well as neuroimaging-based intervention efficacy prediction and monitoring.
The integrated BCB features AI-powered digital health technology for high-resolution brain signal decoding, intelligent neurofeedback and personalized treatment. It would be clinically validated for stroke therapy and can be extended to other use-cases including mental health and brain training. We anticipate that the research outcomes from this project will generate significant scientific, societal and potentially economic impact.
BCI for Upper Limb Rehabilitation
Distal Upper Extremity Motor training using robotic exoskeleton and EEG
Distal Upper Extremity Motor training using robotic exoskeleton and EEG
Distal Upper Extremity Motor training using robotic exoskeleton and EEG
Developed a BCI system to initiate and control motor attempts of the distal upper extremities (UE). A RCT is currently underway to compare the usability, safety, and efficacy of this system for stroke rehabilitation of the distal upper limb with conventional occupational therapy.
Decoding EEG during Continuous Movement of distal upper extremity
Distal Upper Extremity Motor training using robotic exoskeleton and EEG
Distal Upper Extremity Motor training using robotic exoskeleton and EEG
Developed deep learning models for the prediction and continuous reconstruction of hand movement trajectories from EEG data. The system will provide real-time control to robotic exoskeleton devices, facilitating fine motor neurorehabilitation.
Decoding Emotion and Cognition from EEG
Developed novel neuropsychologically plausible architectures for emotion recognition and attention decoding. These include a multi-scale CNN architecture, TSception, which learns temporal dynamics and spatial asymmetry from EEG data and a graph neural network, LGGNet, which integrates activities from various brain functional areas to aid in decision-making.