Session: 17-01-01: Research Posters

Paper Number: 150597

150597 - Streamlining Muscle Force Estimation: Simplified Cost Functions Meet Advanced Muscle Models 

This study develops a novel strategy for muscle force estimation by integrating simplified cost functions with advanced muscle models, thereby demonstrating significant advantages over conventional complex methods. This methodological synergy provides substantial improvements over traditional, complex computational methods, which often rely on dense data inputs and extensive processing. Utilizing elbow flexion as a representative case study, we demonstrate how this integrated approach can accurately predict muscle forces without the dependency on electromyography (EMG) data, thereby streamlining the computational process significantly.
The methodology employed in this study involves a combination of computational modeling and optimization strategies. We developed a biomechanical model of the arm using MATLAB, representing the upper arm and forearm as linked segments with a single degree of freedom (DoF) at the elbow joint. The model includes the three primary flexor muscles of the elbow: the biceps brachii, brachialis, and brachioradialis. We evaluated three muscle models of increasing complexity: (a) a basic force model with two contact points, (b) a Hill-type muscle model incorporating active force-length and force-velocity relationships, and (c) a Hill-type model that also includes passive force-length relationships. These muscle models were paired with the minimization of three different cost functions that do not require EMG data: (a) total muscle force, (b) total work performed by the muscles, and (c) total stress developed in the muscles, where stress is defined as force per physiological cross-sectional area (PCSA) of the muscle. The choice of cost function was based on its suitability for the specific muscles and tasks involved. The efficacy of each model and cost function combination was evaluated by comparing the computed muscle and arm moments with established elbow flexion data.
The empirical results from our study indicate that simplified cost functions, when skillfully incorporated with well defined muscle models, can produce highly precise estimations of muscle forces. This challenges the entrenched belief in the field that achieving high fidelity in muscle force estimations invariably requires complex and computationally intensive cost functions. Instead, our findings advocate for a balanced approach where strategic anatomical and biomechanical modeling is combined with streamlined computational strategies, achieving robust and reliable estimations without the traditional overhead.
Moreover, this novel approach significantly boosts the accessibility and applicability of muscle force estimation techniques, making them more viable for widespread use in clinical diagnostics, therapeutic settings, and sports science. By providing a tool that is both effective, fast and more user-friendly, our methodology promises to revolutionize rehabilitation protocols and optimize athletic training programs, which rely heavily on accurate and timely muscle force data.
In conclusion, this study not only broadens the practical applications of muscle force estimation but also catalyzes innovative directions in biomechanical research. Demonstrating that less complex computational methods can yield results comparable to, if not surpassing, more intricate systems, it encourages the scientific community to reassess the efficiency and practicality of current muscle modeling techniques. This approach not only makes muscle force estimation more accessible and effective but also encourages its adoption across a broader spectrum of applications, potentially transforming standard practices in both clinical and sports performance analysis.

Presenting Author: Muhammad Hassaan Ahmed University of California Merced

Presenting Author Biography: In 2018, I earned my B.E. degree in Mechanical Engineering from the National University of Science and Technology (NUST), Pakistan, securing GPA-based scholarships throughout all eight semesters. During my academic journey, I also served as a Research Assistant (RA) at the National Centre of Robotics and Automation (NCRA). My diverse research interests encompass biomedical devices, robotics, and computational dynamics, with a particular focus on microscope biological systems.
The crux of my research revolves around investigating the impact of constitutive laws on slender structures, modeled as continuum beams or rods, and their dynamics of deformation in bending and torsion. Specifically, I will delve into scenarios where the constitutive law exhibits non-linear and/or non-homogeneous characteristics. This exploration is propelled by a fundamental question: how do the features of non-linearity and non-homogeneity in constitutive laws influence the dynamics of deformation in biological filaments, subsequently shaping their biological activity or functions?
Although the exact application of this study may be elusive, I plan to define mechanics-based problems that intricately address engineering and mathematical challenges. This strategic approach serves as a foundational stepping stone toward achieving the overarching goal of understanding the intricate relationship between the structure and function of biological filaments.
Moreover, my research aims to develop inverse methods for identifying the constitutive laws governing slender filaments. To achieve this, I will leverage experimental data from real systems, employing it to discern the constitutive laws. The accuracy of these identified laws will be rigorously validated through molecular dynamics simulations on software platforms, ensuring the reliability and applicability of the research outcomes.

Authors:

Muhammad Hassaan Ahmed University of California Merced
Jacques-Ezechiel N’guessan University of California Merced
Ranjan Das University of California Merced
Matthew Leineweber BIOMOTUM, Inc.
Sachin Goyal University of California Merced