Comparison of Multi-Band Intermuscular Coherence Complex Network Metrics Between Stroke Survivors and Healthy Adults Before and After Speed Control

OBJECTIVE: This study aimed to identify robust neuromuscular biomarkers for post-stroke gait impairment by comparing intermuscular coherence (IMC) networks between stroke survivors and healthy adults while controlling for walking speed effects. METHODS: We analyzed surface electromyography (sEMG) from 14 muscles in 39 stroke survivors and 34 healthy adults during walking. IMC networks were constructed across six frequency bands, and graph theory metrics were computed. Walking speed was controlled via analysis of covariance (ANCOVA) and validated in a speed-matched subsample. RESULTS: Stroke survivors showed widespread network impairments across all frequencies. After speed adjustment, only the clustering coefficient in the Type IIb band (165-224.5 Hz) remained significantly reduced in patients, which was confirmed in speed-matched validation. It showed significant but moderate associations with clinical gait scores (R ${}^{2} = 0.18$ -0.20), suggesting a potential speed-robust quantitative biomarker. CONCLUSION: The Type IIb band clustering coefficient represents a speed-independent biomarker of impaired neuromuscular coordination in stroke, with potential applications in targeted assessment and rehabilitation monitoring.