Children With Cerebral Palsy Have Greater Stride-to-Stride Variability of Muscle Synergies During Gait Than Typically Developing Children: Implications for Motor Control Complexity

Abstract

Background: There is mounting evidence that the central nervous system utilizes a modular approach for neuromuscular control of walking by activating groups of muscles in units termed muscle synergies. Examination of muscle synergies in clinical populations may provide insights into alteration of neuromuscular control underlying pathological gait patterns. Previous studies utilizing synergy analysis have reported reduced motor control complexity during walking in those with neurological deficits, revealing the potential clinical utility of this approach.

Methods: We extracted muscle synergies on a stride-to-stride basis from 20 children with cerebral palsy (CP; Gross Motor Function Classification System levels I-II) and 8 children without CP, allowing the number of synergies to vary for each stride. Similar muscle synergies across all participants and strides were grouped using a k-means clustering and discriminant analysis.

Results: In total, 10 clusters representing 10 distinct synergies were found across the 28 individuals. Relative to their total number of synergies deployed during walking, synergies from children with CP were present in a higher number of clusters than in children with typical development (TD), indicating significantly greater stride-to-stride variability. This increased variability was present despite reduced complexity, as measured by the mean number of synergies in each stride. Whereas children with CP demonstrate some novel synergies, they also deploy some of the same muscle synergies as those with TD, although less frequently and with more variability.

Conclusion: A stride-by-stride approach to muscle synergy analysis expands its clinical utility and may provide a method to tailor rehabilitation strategies by revealing inconsistent but functional synergies in each child with CP.

Keywords: cerebral palsy; electromyography; gait; motor function; muscle synergy.

Figures

Fig. 1.

Mean number of synergies per stride (A), total number of synergies (B), number of clusters (C), and normalized cluster number of muscle synergies (D) in five strides. Data are presented as means and standard errors. An asterisk (*) indicates a significant difference (p

Fig. 2.

Distribution of similar synergies across individual participants. Muscle synergies in children with typical development (TD) and cerebral palsy (CP) were extracted from individual five strides, and similar synergies were assigned into the same clusters. Row and column are identification (ID) of subjects and clusters, respectively. Note that the same cluster ID between subjects indicates similar synergies. Cluster ID is ordered by most prevalent assignment of synergies in children with TD such that no synergies for children with TD were assigned to cluster 10. Brightness scale presents the number of synergies in a cluster.67×97mm (600 × 600 DPI)

Fig. 3.

Muscle synergies assigned to individual clusters. The gait cycle is defined based on heel strike of dominant leg. Line (left) and bar (right) plots present synergy activations and structures, respectively. Each cluster number is corresponding to the cluster ID in Fig. 2. Abbreviations: Nsyn, the number of synergies assigned to each cluster; ICC, intraclass correlation coefficient (3,1) of synergy structures within each cluster; TA, tibialis anterior; MG, medial gastrocnemius; RF, rectus femoris; MH, medial hamstring.85×107mm (600 × 600 DPI)

Fig. 4.

Samples of muscle synergies extracted from 5 individual strides in CP 15 and TD 4 (refer to Figure 3). Line and bar plots present synergy activations and structures, respectively. Synergies in TD 4 were assigned into four clusters indicating low stride-to-stride variability while those of CP 15 were assigned into ten clusters indicating high variability. Data are presented with means and standard errors. Abbreviations: Nsyn, number of synergies; TD, typical development; CP, cerebral palsy; TA, tibialis anterior; MG, medial gastrocnemius; RF, rectus femoris; MH, medial hamstring.150×106mm (600 × 600 DPI)

Fig. 5.

Variance (A, B) and ICC (C, D) values of EMG (A, C) and kinematic variables (B, D) across five strides in children with TD (dark gray) and CP (light gray). Data are presented with means and standard errors. Abbreviations: TD, typical development; CP, cerebral palsy; EMG, electromyography; ICC, intraclass correlation coefficient; N, non-dominant side; D, dominant side; TA, tibialis anterior; MG, medial gastrocnemius; RF, rectus femoris; MH, medial hamstring. 311×168mm (300 × 300 DPI)