Leveraging Imaging and Computational Modeling to Characterize Sex-Specific Musculoskeletal Form and Function

Abstract

The organization of skeletal muscle influences how force is generated and transmitted to produce joint torque. However, isolating muscle morphology contributions to functional capacity remains difficult, as torque reflects interacting effects between muscle size, internal architecture, mechanical leverage, activation, and joint position. Understanding these relationships therefore requires characterization of muscle structure across multiple scales, from internal architecture and three-dimensional morphology to muscle groups, while accounting for characteristics such as sex and body size. This is particularly important for musculoskeletal model development, which has utilized scaled measures from generalized reference datasets based on small sample sizes, cadaveric measurements, or predominantly male cohorts. These generalized measures may not capture participant-specific anatomy or account for sex. Advances in imaging scan times, automated skeletal muscle segmentation, and shape modeling capable of accounting for allometric variation between sexes provide new opportunities to advance our understanding of skeletal muscle. This dissertation leverages a large, sex-balanced cohort pairing structural and functional measurements across a broad range of body sizes to investigate sex-specific differences in lower-limb muscle form and function. First, standardized functional assessments of the ankle, knee, and hip across multiple joint positions were paired with MRI-derived muscle volumes to characterize relationships between muscle-group volume and joint torque and determine the influence of sex. Second, joint torque was normalized by muscle-group volume to characterize morphological efficiency. Torque density identified sex and position-dependent differences, while providing a standardized measure for comparisons across muscle groups and exertions. Third, participant and muscle characteristics were integrated into the Individualized Maximal Torque (IMT) framework, which selects the best validated torque-prediction model according to available measurements for an individual. Muscle volume remained one of the most informative characteristics, with selected subsets of participant and muscle characteristics improving predictions compared with models incorporating all available measures. Fourth, participant-specific diffusion tensor imaging-derived fiber orientations and muscle geometry were used to investigate geometric fiber-orientation derived cross-sectional area (GOCSA). GOCSA was related to maximal torque but remained strongly associated with muscle volume. Finally, statistical shape modeling characterized three-dimensional muscle morphology and identified sex-associated regional differences independent of allometry. Collectively, these studies demonstrate the importance of accounting for sex across multiple scales of muscle structure and contribute uses of muscle volume in relation to torque. Participant-specific architecture and three-dimensional morphology provide further complementary information, through the identification of sex differences beyond size alone. Together, this work progresses muscle characterization from generalized scalar measures towards increasingly individualized representations of muscle architecture, shape, and functional capacity.