Covariation Within and Among Joint Articular Surfaces of the Upper and Lower Limb

Abstract

Joints are complex skeletal structures that must accommodate the range of motion required by limbs while maintaining stability in those actions. Research has shown how joint articulations relate to body size and general movement and has endeavored to link these to growth and development. Yet lasting questions remain about how morphological covariance manifests in joint surfaces and how this relates to articular conformity and complexity of function. This study employs 3D geometric morphometrics to examine these topics in a sample of human skeletons, focusing on the shoulder, elbow, hip, and knee. The goals of this study are to 1) establish the magnitude and pattern of morphological covariance between articulating joint surfaces (e.g., glenoid fossa and humeral head), and 2) determine how this compares to covariance between homologous surfaces (e.g., humeral and femoral heads). We hypothesize that, with size-corrected data, there will be comparable covariance between the articulating components of all joints, given their shared developmental processes and collective formation through chondral modeling. To a lesser degree, we also expect homologous joints of the upper and lower limbs to exhibit covariance as they are accommodating similar actions, albeit with varying ranges of motion. Computed-tomography scans for a sample of twenty skeletons were segmented to produce 3D virtual surfaces. Eighty-three landmarks were digitally placed on the articular surfaces of each skeleton, and raw landmark data were subjected to a Generalized Procrustes Analysis; t-tests confirmed that repeated trials of landmark placement exhibit high intra-observer consistency (p > 0.05). To test for covariance between landmarks of articulating surfaces, a two-block partial least squares (PLS) analysis was conducted for each joint. Preliminary results indicate moderate-to-high correlation coefficients for the knee (r-PLS = 0.78), elbow (r-PLS = 0.74), and hip (r-PLS = 0.67); the shoulder exhibits comparatively low covariance (r-PLS = 0.37; p > 0.05 for all joints). However, a PLS comparison between joints indicates no joint is significantly more or less covariant than the others. Between pairs of homologous surfaces, PLS analysis indicates a pattern of collectively stronger covariance than is observed between articulating surfaces, with the humeral trochlea-distal femur exhibiting the strongest covariance (r-PLS = 0.84), followed by ulnar trochlear notch-tibial plateau (r-PLS = 0.74), glenoid fossa-acetabulum (r-PLS = 0.71), and the least correlated homologous pair, the humeral head-femoral head (r-PLS = 0.51). Together, these results support our hypothesis that the functional need for articular fit should produce comparable morphological covariance between the opposing joint surfaces of the shoulder, elbow, hip, and knee; however, it is notable that covariance between homologous surfaces appears to be stronger than between articulating surfaces, indicating that serially homologous limb development may predominate over conformity in the production of coordinated variation.