The Brain–Muscle Intelligence Systems group works at the intersection of neural engineering, neuromuscular signal processing, and rehabilitation robotics. We treat the brain and the musculoskeletal system as one coupled control loop, and we build the sensing, decoding, actuation, and evaluation technologies needed to close that loop in real time for clinical and everyday use.
Our work runs from signal to system: from a single electrode contact, to a decoding algorithm, to a worn robot, to a standard that lets the whole field compare results.
Research areas
- Neural signal acquisition and intelligent decoding.
Robust acquisition and decoding of EEG, fNIRS, and evoked potentials. We develop channel-selection strategies based on normalized mutual information centrality across refined frequency bands, entropy-based feature families, and nonlinear network models of working memory and cognitive load. A current focus is reducing the number of channels a usable BCI requires. - Brain–muscle fusion and neuromuscular interfaces.
EEG alone is noisy; EMG alone is incomplete. We fuse cortical and muscular signals to decode movement intent, contraction force, and fatigue. Recent work uses fusion-entropy features to distinguish isometric contraction levels under the same action, a capability that matters for prosthetic and assistive control. - Neurofeedback and closed-loop neuromodulation.
Neurofeedback turns brain activity into a training signal. We design protocols for motor learning, surgical skill acquisition, attention, and cognitive compensation, and we study why some learners respond and others do not — including the role of low-frequency oscillation and cerebral autoregulation as revealed by fNIRS. - Human–robot shared control.
Assistive robots should neither ignore the user nor override them. We build shared-control frameworks in which decoded intent, residual muscular function, and machine autonomy are combined with explicit arbitration, applied to upper-limb exoskeletons and assistive wheelchair platforms. - Wearable monitoring and neurofunctional assessment.
Lightweight, long-duration wearable sensing for real-world measurement: cerebral oxygenation, muscle activation, balance, and vital signs. - Standardization, benchmarking, and reproducibility
We contribute to international standardization in brain–computer interfaces and wearable devices, and we publish the datasets, metrics, and reference implementations that make results comparable across laboratories.