An optimization-based method for prediction of lumbar spine segmental kinematics from the measurements of thorax and pelvic kinematics.

Given measurement difficulties, earlier modeling studies have often used some constant ratios to predict lumbar segmental kinematics from measurements of total lumbar kinematics. Recent imaging studies suggested distribution of lumbar kinematics across its vertebrae changes with trunk rotation, lumbar posture, and presence of load. An optimization-based method is presented and validated in this study to predict segmental kinematics from measured total lumbar kinematics. Specifically, a kinematics-driven biomechanical model of the spine is used in a heuristic optimization procedure to obtain a set of segmental kinematics that, when prescribed to the model, were associated with the minimum value for the sum of squared predicted muscle stresses across all the lower back muscles. Furthermore, spinal loads estimated using the predicted kinematics by the present method were compared with those estimated using constant ratios. Predicted segmental kinematics were in good agreement with those obtained by imaging with an average error of ~10%. Compared with those obtained using constant ratios, predicted spinal loads using segmental kinematics obtained here were in general smaller. In conclusion, the proposed method offers an alternative tool for improving model-based estimates of spinal loads where image-based measurement of lumbar kinematics is not feasible.