[home] [Personal Program] [Help]

---

Session: Neuro-muscular – lower extremities 2
Session starts: Friday 25 January, 13:00
Presentation starts: 14:15
Room: Lecture room 536

---

Kirsten Veerkamp ()
Wouter Schallig ()
Jaap Harlaar ()
Claudio Pizzolato ()
Christopher Carty ()
David Lloyd ()
Marjolein van der Krogt ()

Abstract:
Cerebral palsy (CP) is one of the most common motor disorders among children, often resulting in problems with gait [1]. Neuro-musculoskeletal modelling can provide insight into the aberrant muscle function during gait in these patients. When estimating muscle activations and forces, static optimization approaches are commonly used to solve for the redundancy of the system, combined with a generic model (e.g. [2]). However, these generic approaches do not account for disturbed motor control and muscle weakness in CP. The aim of this study was to evaluate different forms of neuro-musculoskeletal model personalization and optimization to estimate musculotendon forces during gait in children with CP. Nine children with CP (GMFCS I-II) and nine typically developing (TD) children participated. Data collection included 3D-kinematics, ground reaction forces, and electromyography (EMG) of eight lower limb muscles. Four different methods were used to estimate muscle activations and musculotendon forces of a scaled-generic musculoskeletal model for each child, i.e. I) static optimization that minimized activation squared (SO), II) SO with maximum isometric muscle forces scaled to body mass (SO_MIF), III) an EMG-assisted approach using optimization to minimize summed-activation squared while reducing tracking errors of experimental EMG-linear envelopes and inverse dynamics joint moments ([3, 4], EA), and IV) EA with musculotendon model parameters first personalised by calibration ([4, 5], EA_CAL). Model performance was assessed by comparing each model’s estimated activations and joint moments to the corresponding experimental measures. All models’ agreements with EMG were significantly worse for CP than for TD, possibly due to the aberrant motor control in CP. SO and SO_MIF showed the poorest agreement with EMG in both pattern and amplitude. Compared to SO and SO_MIF, EA had substantially higher agreement with EMG, especially with CP, with only minor decrement in joint moments predictions. Overall, EA_CAL performed best. It is expected that a model more consistent with experimental measures is more likely to yield more physiologically results. Hence, this study highlights the added value of calibrated EMG-assisted modelling, in TD children and even more so in children with CP. References [1] Graham et al., Nature Reviews Disease Primers (2016) [2] Steele et al., Gait & Posture (2012) [3] Pizzolato et al., Journal of Biomechanics (2015) [4] Sartori et al, Journal of Biomechanics (2014) [5] Lloyd & Besier, Journal of Biomechanics (2003)