Loading...

Merging and Fractionation Analysis of Muscle Synergy Structures in Groups of Cerebral Palsy Gait

Taghi Mohammadi, Mohammad Reza | 2024

237 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57076 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Farahmand, Farzam
  7. Abstract:
  8. Cerebral palsy is the most common motor disorder among children, resulting from damage to the motor areas of the brain during fetal development or early infancy. Studying motor control changes in individuals with cerebral palsy helps better understand their motor abnormalities. In recent years, various methods for analyzing motor control have evolved, including synergy analysis. The current study aims to thoroughly examine the differences in muscle synergies between children with cerebral palsy and healthy children during walking. In the initial phase of the study, muscle synergies were compared among different groups of children with cerebral palsy (based on Rodda classification) and healthy children. The goal was to elucidate differences using a fusion algorithm (linear combination). For this purpose, EMG data from 9 lower limb muscles of 80 cerebral palsy patients in four groups (jump, true, apparent, and crouch) and 5 healthy individuals during gait were utilized. Synergies were extracted using non-negative matrix factorization, considering the most frequent number in each group based on the VAF criterion. Subsequently, hierarchical clustering and k-means clustering algorithms were applied to identify characteristic muscle synergies for each group. Through the application of the merging algorithm on the weight structures, integrated synergies were identified. The results indicate that children with cerebral palsy have fewer muscle synergies compared to healthy individuals, suggesting a lower complexity in their motor control system. It was also observed that in the jump, apparent, and crouch groups, a linear combination of multiple sequential muscle synergies can reconstruct weight structures over time. However, for the true group, a specific temporal sequence of integrating synergies was not identified. Given the absence of distinct synergy integration patterns for the Rodda groups, the study aimed to use synergy as criteria for classifying cerebral palsy patients further. After extracting muscle synergies from 345 cerebral palsy patients, a three-stage classification algorithm based on equal synergy number (motor complexity similarity), similar synergy activation profiles (motor command synchrony), and similar synergy weight structures (motor structure similarity) was applied to categorize the lower limbs of patients. The results demonstrate that, based on this algorithm, 329 limbs (95.3%) of cerebral palsy children could be classified into 5 different groups with similar synergy features. Kinematic data comparison among the groups indicates that in groups with fewer synergies, joint angle oscillations in the hip, knee, and ankle during walking are weaker, and the kinematics are closer to the healthy group in groups with three muscle synergies. This study suggests a novel classification method for children with cerebral palsy based on motor control performance, which could be used for designing more effective rehabilitation strategies for these individuals
  9. Keywords:
  10. Cerebral Palsy ; Gait Analysis ; Clustering ; Merging ; Muscle Synergy ; Non-Negative Matrix Factorization (NMF)

 Digital Object List

 Bookmark

No TOC