Musculoskeletal Model Personalization Affects Metabolic Cost Estimates for Walking

Area of Science:

• Biomechanics and Human Movement Science
• Rehabilitation Engineering
• Computational Physiology

Background:

• Metabolic cost is a key metric for human performance, increasingly assessed using musculoskeletal models.
• Existing musculoskeletal models require improved accuracy in metabolic cost predictions for practical applications.
• Personalized musculoskeletal models show promise, but their impact on metabolic cost estimation is unevaluated.

Purpose of the Study:

• To investigate the effect of musculoskeletal model personalization on metabolic cost of transport (CoT) estimates in post-stroke walking.
• To compare three common metabolic cost models across different personalization levels of musculoskeletal models.
• To assess the accuracy of CoT predictions against experimental data and clinical measures of walking impairment.

Main Methods:

• Analyzed walking data from two male stroke survivors with right-sided hemiparesis.
• Implemented three musculoskeletal modeling approaches: scaled generic (SOGen), personalized EMG-driven (SOCal, EMGCal).
• Calculated CoT using three metabolic cost models (Umberger et al., 2003; Bhargava et al., 2004; Umberger, 2010) and compared with clinical data.

Main Results:

• The SOCal and EMGCal approaches, using the Bhargava metabolic cost model, accurately reproduced experimental trends between CoT and post-stroke walking asymmetry.
• Generic models and other metabolic cost models showed less accurate predictions of CoT in relation to clinical measures.
• Magnitude differences between predicted and measured CoT suggest potential for parameter tuning in metabolic cost models.

Conclusions:

• Personalized musculoskeletal models, particularly EMG-driven approaches, are crucial for accurate metabolic cost of transport estimation in post-stroke gait.
• The Bhargava metabolic cost model, when used with personalized models, shows significant potential for clinical gait analysis.
• Accurate CoT predictions via refined computational simulations can aid in reliably assessing human performance and predicting surgical/rehabilitation outcomes.