A 2-Block Architecture for Real-Time EEG Gait Decoding: A Pilot Study
2026-08-03 • Machine Learning
Machine LearningHuman-Computer Interaction
AI summaryⓘ
The authors developed a new brain-computer interface system to control lower-limb exoskeletons using brain signals (EEG). Their design includes a special feature extractor that cleans the data and a decoder that identifies four different walking states. They showed their method works better than previous ones and can predict walking intentions in real time. In tests, the system successfully detected when a person wanted to start walking about half the time, demonstrating it could be useful for controlling exoskeletons.
Electroencephalography (EEG)Brain-Computer Interface (BCI)Lower-limb exoskeletonMotion artifactsFeature extractionPolynomial Time-Varying Layer (PolyTVL)LSTMGait classificationReal-time prediction
Authors
Shantanu Sarkar, Saurabh Prasad, Jose L. Contreras-Vidal
Abstract
Closed-loop lower-limb exoskeleton control via Electroencephalography (EEG) remains limited by motion artifacts, low signal-to-noise ratio, and binary gait formulations that fail to capture full cortical gait complexity. We propose a 2-block Brain-Computer Interface (BCI) architecture: a trainable session-specific Feature Extraction Block with real-time artifact suppression and multi-domain feature extraction, coupled with a Decoder Block built on a novel Polynomial Time-Varying Layer (PolyTVL)+LSTM for four-state gait classification (Stand, Initiate, Execute, Terminate). Ablation confirmed v01 (PolyTVL+LSTM) outperformed all variants (validation MCC: 0.435, gap: 0.187), with consistent EEG feature discriminability across ROIs and sub-bands (p<0.05). Closed-loop deployment with v01 achieved 55.3% (Rex-assisted) and 52.7% (volitional) gait initiation success, with a mean prediction time of 70.5~ms (+/-41.5), validating real-time feasibility in this pilot study.