Novel high-density electromyography analysis techniques and neuromuscular models for robotic assessments: Towards a better understanding of Cerebral Palsy and Botulinum Neurotoxin-A treatment

Abstract

Cerebral Palsy (CP) is a common paediatric condition affecting cognitive and motor functions. It develops due to a brain lesion, which is non-progressive, whereas muscle pathologies progress over the lifespan of individuals with CP. In this context, muscles frequently have an increased tone, are shorter and stiffer compared to muscles of typically developed children. Depending on the severity of CP, costly treatments such as Botulinum Neurotoxin-A (BoNT-A) injections, casting and surgery can be necessary and families of those affected face various social burdens. Currently, there is no cure for the condition highlighting the need for more research to improve our understanding and generate new ideas for treatment strategies in CP. Children aged 5-15 years with and without CP have been recruited for the studies to better understand between-group differences in strength and coordination. Furthermore, a muscle model was developed to identify subject-specific muscle parameters (e.g. stiffness) during isometric contractions non-invasively. High-density electromyography (hdEMG) data was recorded from the upper (Biceps and Triceps brachii) and lower limbs (Gastrocnemius medialis and Tibialis anterior) to develop a novel method analysing muscle activation on a motor unit level pre and post BoNT-A treatment. Elbow flexion rapid force development was significantly lower in CP by approximately 60%. Force fluctuations during sustained isometric contractions were more regular and smoother in children with CP, while their accuracy of tracking a predefined force ramp on a computer screen was more than three times lower compared to the control group. The novel method developed to estimate the number of merged MUs in spike trains resulted in an overall good accuracy with a median difference of 0.29 MUs in comparison to the true number of merged MUs. The findings provide novel insights about differences in muscle force generation and coordination during isometric contractions in children with CP. In addition, the muscle model developed successfully identified subject-specific muscle parameters and might be useful to evaluate treatment effectiveness over time. Another new merged MU estimator developed provides an extremely promising tool to evaluate MU firing properties in decomposed hdEMG signals in children with CP for future studies.