Cluster-based muscle synergy analysis scheme for assessing crawling motor function in children with cerebral palsy

This study aims to evaluate the crawling motor function in children with cerebral palsy (CP) using a cluster-based muscle synergy analysis scheme.

Surface electromyography (sEMG) signals were recorded from 26 muscles across the body in 14 typically developing (TD) subjects and 10 children with CP while they performed eight prescribed crawling modes. The sEMG signals were preprocessed, and muscle synergies were extracted using a non-negative matrix factorization (NNMF) algorithm. A hierarchical clustering algorithm, incorporating synergy similarity constraints, was employed to cluster synergies from TD subjects performing the same crawling mode, identifying common synergies within each mode. A subsequent clustering process revealed common synergies across different modes. Using the common synergies of TD subjects as a benchmark, four evaluation metrics based on synergy similarity were developed to assess the crawling motor function in children with CP.

The analysis successfully extracted common synergies within and across crawling modes in TD subjects. Under the condition of auditory cueing, children with CP showed a significantly lower number and similarity of common synergy structures while maintaining a comparable number of common recruitment curves relative to the TD subjects.

The cluster-based muscle synergy analysis scheme effectively assesses the crawling motor function in children with CP.