Effect of time window on MVC reference for quantifying spastic cocontraction in spastic paresis

In hemiparesis, spastic cocontraction is typically quantified by normalizing electromyographic activity to the root mean square (RMS) values obtained during maximal voluntary contraction (MVC) of the cocontracting muscle when it acts as an agonist. However, the selection of the RMS time window and the use of filtering procedures vary widely across studies, limiting comparability. In this study, we evaluated the effects of window size and low-pass filter cutoff frequency (ƒc) on the RMS values obtained during MVC in chronic hemiparesis.

Participants with stroke-induced hemiparesis and gastrocnemius spasticity (X_V1-GAS-X_V3-GAS >5°) were tested in a seated position on an isokinetic ergometer, with the knee extended and the ankle positioned at 90°. Surface electromyography (EMG) was recorded from the medial and lateral gastrocnemius (MG and LG), soleus (SL), and tibialis anterior (TA) during standardized isometric plantar flexor and dorsiflexor MVCs. RMS values were computed using sliding windows ranging from 5 to 1,250 ms (in increments of 5 ms) and analyzed across low-pass filter cutoff frequencies between 6 and 100 Hz.

A total of 20 participants with hemiparesis (age: 56.4 ± 7.0 years and time since lesion: 7.8 ± 5.7 years) were included. Regardless of muscle type, experimental RMS curves as a function of window size adhered to a first-order model, with high consistency across trials (R ² ≈ 0.90, RMSE ≈ 8%). RMS values stabilized beyond 168.3 ms (time constant τ = 33.6 ms; 5τ threshold). In contrast, low-pass filtering caused a marked loss of amplitude, with >40% reduction in RMS magnitude at cutoff frequencies below 10 Hz, despite preservation of signal shape.

These findings demonstrate that reliable EMG normalization in spastic hemiparesis requires a minimal RMS window of approximately 200 ms during MVC and that unsmoothed EMG should be used to preserve signal amplitude. Methodological standardization using these parameters can improve the validity, reproducibility, and comparability of cocontraction indices across studies and may facilitate their application in clinical assessments and rehabilitation research in spastic paresis.