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

Stroke survivors showed widespread intermuscular coherence network impairments across all frequency bands compared to healthy adults before speed control.

• 39 stroke survivors and 34 healthy adults were analyzed during walking
• Surface electromyography was recorded from 14 muscles
• IMC networks were constructed across six frequency bands
• Graph theory metrics were computed from the IMC networks
• Impairments were observed across all frequencies prior to speed adjustment

After speed adjustment using ANCOVA, only the clustering coefficient in the Type IIb band (165-224.5 Hz) remained significantly reduced in stroke survivors.

• Walking speed was controlled via analysis of covariance (ANCOVA)
• The Type IIb frequency band spans 165-224.5 Hz
• The finding was confirmed in a speed-matched subsample validation
• All other frequency band differences became non-significant after speed control
• The clustering coefficient was the specific graph theory metric that survived speed adjustment

The Type IIb band clustering coefficient showed significant but moderate associations with clinical gait scores in stroke survivors.

• R² values ranged from 0.18 to 0.20 for associations with clinical gait scores
• The associations were described as 'significant but moderate'
• This suggests the metric has potential as a 'speed-robust quantitative biomarker'
• The metric was proposed for use in targeted assessment and rehabilitation monitoring

Speed-matched subsample validation confirmed the Type IIb band clustering coefficient as a speed-independent finding.

• A speed-matched subsample was used as a validation cohort
• The reduced clustering coefficient in the Type IIb band was confirmed in this validation sample
• This dual validation approach (ANCOVA and speed-matched subsample) strengthened the finding's robustness
• The metric was characterized as 'speed-independent' following this validation

The study aimed to identify neuromuscular biomarkers for post-stroke gait impairment by controlling for walking speed effects on IMC network metrics.

• Walking speed differences between stroke survivors and healthy adults can confound group comparisons
• Six frequency bands were analyzed for IMC network construction
• Graph theory metrics were computed to characterize network topology
• The study used both ANCOVA and speed-matched subsampling to address speed confounds
• 14 muscles were measured using surface electromyography during walking