Muscle synergy structure and gait patterns in children with spastic cerebral palsy

Abstract

Aim: To determine if muscle synergy structure (activations and weights) differs between gait patterns in children with spastic cerebral palsy (CP).

Method: In this cross-sectional study, we classified 188 children with unilateral (n=82) or bilateral (n=106) spastic CP (mean age: 9y 5mo, SD: 4y 3mo, range: 3y 9mo-17y 7mo; 75 females; Gross Motor Function Classification System [GMFCS] level I: 106, GMFCS level II: 55, GMFCS level III: 27) into a minor deviations (n=34), drop foot (n=16), genu recurvatum (n=26), apparent equinus (n=53), crouch (n=39), and jump gait pattern (n=20). Surface electromyography recordings from eight lower limb muscles of the most affected side were used to calculate synergies with weighted non-negative matrix factorization. We compared synergy activations and weights between the patterns.

Results: Synergy structure was similar between gait patterns, although weights differed in the more impaired children (crouch and jump gait) when compared to the other patterns. Variability in synergy structure between participants was high.

Interpretation: The similarity in synergy structure between gait patterns suggests a generic motor control strategy to compensate for the brain lesion. However, the differences in weights and high variability between participants indicate that this generic motor control strategy might be individualized and dependent on impairment level.

Figure 1

The different gait patterns identified in a previous systematic review, their sagittal plane joint angles, and average stance and swing duration (% gait cycle). The mean curve (1 SD) of reference sample of typically developing (TD) children is plotted in grey. On the bottom of the figure, stance and swing duration for the typically developing children are represented in light and dark grey respectively. The mean curve and mean stance phase duration of the children with cerebral palsy (CP) are plotted in black. The true equinus gait pattern was excluded because it did not meet the sample size requirements.

Figure 2

Synergy activations and weights per gait pattern. Synergy weights for each child and synergy activations for each of the four gait cycles per child are represented in grey and as the mean activation and weight for each pattern (black) as a function of the normalized gait cycle. Significant differences between the individual patterns (post hoc analyses) are indicated with the numbers of the significantly different patterns. REF, rectus femoris; VAL, vastus lateralis; BIF, biceps femoris; MEH, medial hamstrings; TIA, tibialis anterior; GAS, gastrocnemius; SOL, soleus; GLU, gluteus medius.