Analysis of cerebral palsy gait based on movement primitives

Shahab Tavasoli1, Marzieh Tavasoli1, Mahya Shojaeefard1

  • 1Mechanical Engineering Department, Sharif University of Technology, Tehran, Iran.

Insights

Children with cerebral palsy exhibit altered gait patterns with reduced complexity and higher variability due to inconsistent kinematic modularity. Three basic movement primitives underpin their gait, with others ensuring smooth transitions.

Area of Science:

  • Biomechanics
  • Pediatric Neurology
  • Movement Science

Background:

  • Cerebral palsy (CP) is the most common childhood motor disorder.
  • Previous research focused on motor modularity in CP gait, but kinematic modularity remains understudied.

Purpose of the Study:

  • To investigate the kinematic modularity of gait in children with cerebral palsy (CP).
  • To compare kinematic gait modules between typical development (TD) children and various CP subtypes.

Main Methods:

  • Gait kinematics were captured and analyzed in 188 children with CP (grouped as True, Jump, Apparent, Crouch) and 13 TD children.
  • Non-negative matrix factorization (NMF) was employed to extract kinematic modules.
  • Movement primitives were clustered and activation profiles compared for similarity.

Main Results:

  • The number of kinematic modules varied: 3 for Crouch CP, 4 for other CP groups, and 5 for TD children.
  • CP groups showed significantly higher variability in kinematic modules and co-activation compared to TD children (P < 0.05).
  • All groups shared three temporally matched movement primitives, though their structures differed.

Conclusions:

  • Gait in children with CP is characterized by lower complexity and higher variability, stemming from reduced and inconsistent kinematic modularity.
  • A set of three basic movement primitives can generate overall gait kinematics, as seen in the Crouch CP group.
  • Additional movement primitives contribute to smooth transitions between these basic primitives, particularly in more complex gait patterns.
Abstract