Learning how Sheep Move: Developing a General Inverse Kinetic Model for Sheep

Introduction: Whole body inverse kinetic models of humans are well developed. However, for animal models of human disease, the prevalent tools in human biomechanics do not exist. This work explores the development of whole body biomechanical models of quadrupeds, and presents preliminary results in the development of a whole body biomechanical model for sheep in OpenSIM, for use with motion capture data. We present statistical and machine learning models for predicting the hip joint centers from anatomical measurements and landmarks, as well as models for predicting the mass and inertial properties of sheep body segments.

Methods: CT scans from 16 sheep of varying ages, weight, sex, and phenotypes were acquired and the data used to calculate the known hip joint center by sphere fitting the femoral head. Anatomical measurements and additional subject information were used to create models to estimate the hip joint centers in absence of CT data. Then, the mass and moments of inertia for each body segment were estimated from the CT scans, and predictive models applied in the absence of CT scans.

Results: Hip joint centers were predicted with much greater accuracy than previous methods, with errors on the order of a few millimeters, depending on the animal. Mass and inertial properties were predicted with less accuracy, with errors typically within 10%, but in some cases exceeding 20%.

Conclusions: This work represents a significant new set of data for biomechanical models of sheep, being the first comprehensive study to include data from multiple animals. However, our data set is still limited, and would significantly benefit from a larger set of animals to be included. Additionally, sensitivity analysis on the models produced using this data will need to be performed to determine the extent to which errors in the various parameters affect final kinetic analyses.