Estimating muscle forces in patients with cerebral palsy during walking using optimization approaches

Date

2023-12-5

Author

Kidwai, Alina Nawab

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

146
views

53
downloads

Cite This

In vivo muscle force measurement is ethically questionable and practically infeasible since it requires invasive techniques which can impact nerve and muscle function. Therefore, different theoretical approaches have been proposed to calculate muscle forces, especially based on optimization techniques. Muscle force predictions are sensitive to subject population owing to diversity in musculoskeletal properties and neural control. Thus, they need to be compared in context, e.g., of specific movement disorders. Several people with cerebral palsy (CP) exhibit the crouch gait pattern characterized by excessive knee and hip flexion and ankle dorsiflexion, which leads many to lose walking ability. Muscle force estimates for this population can aid clinical decision-making. Several studies have compared optimization methods to EMG-informed methods (which employ more experimental information than the former) for children with CP and healthy children and deemed the latter superior and more reflective of the altered neural control in CP. However, these studies have not addressed crouch gait, including its severity levels. Nor have optimization methods, which are easier to implement, been compared in this context. Therefore, this thesis compared two commonly used optimization methods, static optimization (SO) and computed muscle control (CMC), in terms of the validation performance of their force predictions for this patient group. Unlike SO, CMC allows EMG-based constraints on muscle activity predictions and models passive muscle forces and the history dependency of activation-contraction dynamics. Both techniques use a phenomenological cost function known to work well for healthy level walking. Given the departure of CP patients' muscle activity from normal that may not be easily predicted without incorporating information on this departure into the optimization, the thesis hypothesized that CMC-predicted muscle activity would be closer to EMG data. Additionally, since muscle activity is increasingly pathological in higher crouch severity levels, CMC was expected to be increasingly superior to SO for moderate and severe crouch. Using an open-access motion analysis dataset of nine subjects, uniformly divided into mild, moderate, and severe crouch, OpenSim's SO and CMC tools were used to estimate muscle forces and activity. Validation through a quantitative comparison of predicted muscle activity with EMG revealed no superiority of CMC over SO nor any major improvements in validation performance with crouch severity. Given that neither provided excellent validation performance, EMG-informed methods may be better suited for clinical implementation.

Subject Keywords

Cerebral palsy, Crouch gait, Muscle forces, Optimization, OpenSim

URI

https://hdl.handle.net/11511/107771

Collections

Graduate School of Natural and Applied Sciences, Thesis