BCIwiki (bciwiki.com) — Researchers at the Indian Institute of Technology Delhi, the All India Institute of Medical Sciences in New Delhi and the University of Queensland recorded EEG from ten healthy participants performing bicep curls. They classified five resistance conditions with 97.2% accuracy in within-subject testing. The paper appeared in IEEE Journal of Biomedical and Health Informatics on October 20, 2026.
Participants performed unilateral bicep curls under different loads. Analyses of movement-related cortical potentials and frequency-specific EEG power found differences across resistance conditions, with significant changes before movement, at peak contraction and during execution.
A TTCN neural network extracted EEG features for a k-nearest neighbors classifier. Within-subject leave-one-trial-out testing, which reserves one trial from the same person for each test, achieved 97.2% accuracy. Leave-one-subject-out testing reserved a person excluded from training and reached a maximum accuracy of 93.0%. Errors mainly occurred between adjacent resistance conditions.
The findings suggest that EEG contains information about movement resistance that could inform future monitoring of motor effort. Evidence currently comes from healthy participants; use in patient rehabilitation remains to be tested.