Software for stochastic musculoskeletal modeling

Period of Performance: 02/10/2008 - 02/09/2009

$133K

Phase 1 STTR

Recipient Firm

Grizzly Moose, LLC
Ann Arbor, MI 48103
Principal Investigator

Abstract

DESCRIPTION (provided by applicant): Introduction: Computational models of human joint mechanics have become common in musculoskeletal science. Although existing deterministic models have provided significant insight into musculoskeletal mechanics, they cannot explain some important phenomena. The discrepancy between model predictions and experimental observations can be resolved by treating model parameters as random variables (a stochastic model). Thus, there is a need for software to implement stochastic biomechanical models. Technological Innovation: A software toolbox will be developed that will aid biomechanists in developing stochastic models describing the function of bones, joints, muscles, ligaments, and tendons. This software will be novel because it moves beyond the traditional deterministic approach to muscle force prediction modeling available today. Long-Term Goal: The long-term goal of this project is to develop software for stochastic musculoskeletal modeling. Phase I: The software will be based on calling dynamic modeling software from MATLAB. Phase I activities will prove feasibility of this approach and develop software architecture for the project. Therefore, there will be three specific aims in Phase I: SA1: Develop structure for repeatedly calling AnyBody from MATLAB, and use it to analyze three-dimensional mechanics of arm elevation. Criteria for Acceptance: Successfully perform stochastic simulation of the human shoulder under a range of parameter conditions. SA2: Estimate the precision of parameter covariance estimates required to perform three- dimensional biomechanical simulation. Criteria for Acceptance: Generate the deviation of muscle force predictions as a function of bootstrap sample size. SA3: Develop software architecture. Criteria for Acceptance: Describe the software architecture using the Unified Modeling Language. Phase II will consist of coding and validating the software and Phase III will consist of commercialization. Commercial Opportunity: The proposed software will provide the musculoskeletal modeling community with a powerful tool for analyzing the effect of variability, which is important for robust design of orthopaedic prosthetic components and injury prevention research.