Welcome. I am Jiangbo Pu, Professor at the Institute of Biomedical Engineering, Chinese Academy of Medical Sciences & Peking Union Medical College, where I lead the Brain–Muscle Intelligence Systems group.
Our group studies:
how the brain and the musculoskeletal system speak to each other and how machines can join that conversation.
We build closed-loop systems that read neural and muscular signals in real time, decode intention and effort, and return meaningful feedback to the user through neurofeedback, assistive robots, and wearable devices. The goal is simple to state and hard to achieve: restore, augment, and quantify human motor and cognitive function.
Research vision
Human movement is not produced by the brain alone, nor by muscles alone. It emerges from a continuous loop, cortical command, spinal relay, muscle activation, sensory return. Most brain–computer interfaces listen to only half of that loop. Our vision is to treat the brain–muscle loop as a single intelligent system:
- Read the loop. Fuse electroencephalography (EEG), functional near-infrared spectroscopy (fNIRS), and high-density surface electromyography (HD-sEMG) into one decoding framework, rather than treating them as separate modalities.
- Close the loop. Use neurofeedback and neuromuscular stimulation to turn decoded states into training signals that drive genuine neuroplastic change.
- Act on the loop. Couple decoded intent to wearable rehabilitation robots — upper-limb exoskeletons and assistive wheelchairs — through shared control that respects both human agency and machine capability.
- Make the loop measurable and comparable. Contribute open datasets, reference implementations, and international standards so that results in this field can be reproduced and trusted.
Highlights
- Brain–muscle fusion decoding. By combining entropy-based EEG features with muscle activation patterns, our group has shown that force level during isometric contraction can be identified at the single-action level — a step toward fine-grained, intuitive human–machine interfaces.
- Wearable robots for rehabilitation. Our hardware line includes an upper-limb exoskeleton with a patented joint-protection mechanism (also granted in Italy) and assistive wheelchair platforms designed for shared human–robot control.
- Neurofeedback that trains the motor system. We develop EEG neurofeedback protocols for motor learning and skill acquisition, including work showing how alpha-band oscillation relates to performance in laparoscopic surgical skill training.
- Standards that shape the field. Prof. Pu serves as Convenor of ISO/IEC JTC 1/SC 43/WG 9 and Project Leader of ISO/IEC 8663:2025, Information technology — Brain–computer interfaces — Vocabulary, the first ISO/IEC international standard in brain–computer interfaces.
- Open resources. We release datasets, decoding toolboxes, and reference hardware designs so that other labs can build on our work. See [Software] and [Hardware].