Calf muscle model parameters in children with cerebral palsy compared to typically developing children

Abstract

Background: In children with cerebral palsy(CP), muscle-tendon structures are altered. While interventions exist to treataltered structures, the selection of the most suitable treatment is verycomplex with highly variable outcomes. Musculoskeletal models have thepotential to support clinical decision making. However, a known limitation isthe translation of altered muscle-tendon parameters into musculoskeletalmodels. The aim of this study was to estimate the intrinsic calf musclesproperties using neuromuscular simulations of a passive ankle rotation intypically developing children and children with CP. With these simulations wedetermine to what extent optimal fiber length, tendon slack length, stiffness,and strain differ between typically developing (TD) children and CP children.  Methods: Experimental data was collected onthirteen children with spastic CP (6 diplegia, 7 hemiplegia, age 11.6 ± 3.1years) and 17 TD children (age 10.4 ± 3.3 years) during a slow passive anklerotation. Ankle angle, external forces applied on the ankle, and medialgastrocnemius fascicle length were measured. An OpenSim model with four musclesaround the ankle, GASM, GASL, SOL, and TA, was used to simulate passive anklerotation experiments. Optimal fiber length, tendon slack length, stiffness atlow force, strain at zero force, and strain at maximum force were optimized tomatch the measured ankle moment-angle curves and the GASM fascicle length-anglecurves.  Results: The ankle moment-anglecurves could be successfully matched in both CP and TD (RMSE < 0.42 Nm) byoptimizing individual calf muscle-tendon parameters. The fascicle length-anglecurves could be predicted much better in CP children by optimization, howeverrelatively large residuals remained (RMSE 0.40 cm). These simulations revealthat children with CP have a shorter normalized optimal fiber length and alonger triceps surae normalized tendon slack length compared to TD children.Also, CP triceps surae was found to be stiffer and undergoes less fasciclestrain compared to TD children. Further, the triceps surae passive fiberforce-length curve in CP children is engaged at shorter fiber lengths whencompared to that of TD children.  Conclusion:Simulations show that intrinsic calf muscle-tendon properties aresystematically different between CP and TD children. However, large variancesin fascicle lengthening muscle-tendon parameters in CP children exist. Futureresearch should attempt to better match groups in terms of age, height, andweight. A next step would be to apply these optimized parameters to simulationsof CP gait. This would help identify to what extent altered muscle propertiesaffect gait in children with CP and subsequent treatment decisions.