Quantifying structural properties of forearm flexor muscles in individuals with hemiparetic cerebral palsy using diffusion tensor imaging

Abstract

This study investigated diffusion tensor imaging (DTI) derived macro- and micro-structural musculoskeletal adaptations in forearm flexor muscles in individuals with hemiparetic cerebral palsy (HCP) and typically developing (TD) individuals, and their relationship to reduced grip strength. In 14 individuals with HCP and 16 TD individuals, T1-weighted and diffusion-weighted magnetic resonance images of both forearms were acquired, and maximum grip strength was measured. In two forearm flexors, muscle volume, DTI-based diffusivity metrics, and probabilistic tractography derived fascicle architecture was estimated. Linear mixed-effects models evaluated interlimb differences in structural parameters and their impact on grip strength. In the HCP group, paretic muscles showed significant reductions in volume, diffusivity values, fascicle lengths, and physiological cross-sectional area as compared to nonparetic forearm and TD participants. Furthermore, reduced muscle volume and diffusivity together explained 62% of the grip strength deficit. These findings demonstrate that decreased muscle volume and altered microstructure, as indicated by reduced diffusivity, contribute significantly to functional impairments in HCP. DTI-based diffusivity metrics non-invasively reveal crucial insights into pathophysiological changes in muscle tissue, such as muscle atrophy and fibrosis. Future therapies should focus on both muscle macro- and micro-structural adaptations as targets to improve motor function in HCP.

Keywords: cerebral palsy; diffusion tensor imaging; diffusivity metrics; muscle architecture; skeletal muscle.

FIGURE 1

Pipeline of MR image data acquisition and analysis. (a) Participant positioning in 1.5 T MR scanner with 2 × 18 channel body coil around the forearm, with the elbow at 160° flexion and the hand and fingers placed in an MRI‐safe orthosis to hold the wrist at 0°, fingers at 0°, and forearm pronation‐supination at 0°. (b) MR data preprocessing, including segmentation of regions of interest from T1 images and diffusion tensor estimation from dMR volumes to calculate bone length, muscle volume, and whole‐volume diffusivity metrics. (c) Example of muscle fascicles reconstructed from valid tracts, shown with 3D bone and muscle surfaces, used to calculate fascicle architecture metrics.

FIGURE 2

Anthropometric measures derived from T1‐weighted image segmentations. (a) Bone length (mm), averaged between the ulna and radius bone lengths. (b) Muscle volume (mm³), averaged between the FCR and FDP muscle volumes. (a, b) Top: HCP group, bottom: TD group, green: TD dominant (D) or HCP non‐paretic (NP) arm, pink: TD nondominant (ND) or HCP paretic (P) arm. Black boxplots represent distribution of group data, while gray lines represent individual participant data paired between arms. P‐values shown represent significance of interlimb difference determined from linear mixed effects models. ns, no significant difference; *p < 0.05, **p < 0.01, and ***p < 0.001.

FIGURE 3

Probabilistic tractography derived measures. (a) Fascicle lengths normalized to bone length, (b) fascicle pennation angles (°), (c) physiological cross‐sectional area (PCSA) (mm²), and (d) fascicle count, averaged between the FCR and FDP muscles. (a–d) Top: HCP group, bottom: TD group, green: TD dominant (D) or HCP non‐paretic (NP) arm, pink: TD nondominant (ND) or HCP paretic (P) arm. Black boxplots represent distribution of group data, while gray lines represent individual participant data paired between arms. p‐values shown represent significance of interlimb difference determined from linear mixed effects models. ns, no significant difference, *p < 0.05; **p < 0.01; ***p < 0.001.

FIGURE 4

DTI derived diffusivity metrics. (a) Mean diffusivity (MD) (mm²/s), (b) axial diffusivity (AD) (mm²/s), (c) radial diffusivity (RD) (mm²/s), and (d) fractional anisotropy (FA), averaged between the FCR and FDP muscles. (a–d) Top: HCP group, bottom: TD group, green: TD dominant (D) or HCP non‐paretic (NP) arm, pink: TD nondominant (ND) or HCP paretic (P) arm. Black boxplots represent distribution of group data, while gray lines represent individual participant data paired between arms. p‐values shown represent significance of interlimb difference determined from linear mixed effects models. ns, no significant difference; *p < 0.05, **p < 0.01; ***p < 0.001.

FIGURE 5

Relationships between percent deficits (PD) (%) in grip strength and muscle macro‐structural and micro‐structural measures. (a) Grip strength PD versus muscle volume PD, (b) grip strength PD versus mean diffusivity (MD) PD, (c) grip strength PD versus normalized fascicle length PD, and (d) grip strength PD versus physiological cross‐sectional area (PCSA) PD. (a–d): Red dots represent individuals with HCP and green dots represent TD individuals. Distributions of x‐ and y‐data, by group, are shown at the top and right of each graph, respectively. Linear regression lines fitted to the data are shown in black, with 95% confidence intervals shown in light gray. Pearson correlation coefficients (r) and associated p‐values are shown for each comparison. ns, no significant difference; *p < 0.05; **p < 0.01; ***p < 0.001.

FIGURE 6

Schematic of adaptations in forearm musculoskeletal macrostructure and microstructure following hemiparetic cerebral palsy, and relationship to grip strength, a functional outcome. A *indicates parameters that showed significant correlations to grip strength. Possible clinical implications of this work are shown. Created using Biorender.com.