Heterogeneous structural deficits across lower-leg agonist– antagonist muscle groups in children with cerebral palsy: a cross-sectional imaging study

Abstract Structural alterations in skeletal muscle in cerebral palsy (CP) directly contribute to joint stiffness and impaired motor control, and govern the outcome of rehabilitative interventions. Diffusion tensor imaging (DTI) has established architectural alterations in individual lower-limb muscles, but previous work has focused on a small subset of plantarflexor muscles in isolation, leaving their antagonists uncharacterized despite the reciprocal relationships between muscle pairs crossing the same joint. Because imbalance between such pairs is linked to gait limitations, characterizing whether architectural deficits in CP are proportional across agonist-antagonist pairs, or instead reflect atypical antagonist co-activation, is important. In this study, MRI and DTI data from nine children with CP (10.6 ± 2.2 years, GMFCS I–III) and nine age- and sex-matched typically developing controls from a previous study were used to assess muscle size and architecture in seven muscles and muscle groups in the leg. Volume and muscle length were determined from segmented surface meshes; physiological cross-sectional surfaces (PCSS), fascicle length, and angle were computed from DTI tractography; fractional anisotropy and mean diffusivity were determined from diffusion tensor data. Normalized muscle volume of examined muscles was 38.8% lower in CP (g = − 2.71) and PCSS 27.2% lower (g = − 2.18, both p < 0.001). Volume (− 29.2% to − 41.3%) and muscle length (− 9.5% to − 12.2%, all p_fdr ≤ 0.004) were reduced in all four functional muscle groups of the leg, but the composition of that loss differed: dorsiflexor volume fell almost entirely through cross-sectional thinning (PCSS − 26.6%, fascicle length − 0.9%), the invertor almost entirely through fascicle shortening (− 3.4%, − 26.1%), and the plantarflexors through both (− 28.6%, − 31.5%). Plantarflexor:dorsiflexor and invertor:evertor size ratios did not differ, though the sample supports only effects of g ≥ 1.41. Individually, 33% of CP participants exceeded |z| > 2 on invertor:evertor ratio, driven by an undersized evertor, and volume z-scores exceeded |z| > 2 in 41% of muscle-participant combinations. Fascicle orientation was more uniformly aligned in tibialis anterior, and more dispersed in posterior soleus. Muscle deficits in CP were pervasive but heterogeneous, with clinically relevant imbalance emerging at the individual level despite proportional group averages.