Quantitative comparison of current models for trunk motion in human movement analysis

Background: A number of different models for human trunk kinematics during locomotion have been proposed, though mainly addressing specific clinical questions rather than general populations. These differ considerably for the skeletal segments considered, marker-set, anatomical axis and frame definitions, and joint conventions. The scope of the present study is to compare quantitatively these models on the basis of the same motion. Methods: Ten subjects were analysed, instrumented with a single comprehensive marker-set of 14 markers identified from the union of the corresponding from eight current models for trunk kinematics. Activities of daily living (walking, chair rising/sitting, step-up/down), elementary trunk movements (flexion, bending and axial rotation), and isolated motion of the shoulders, both synchronous and asynchronous were collected. Resulting rotations in the three anatomical planes, both in the laboratory and in the pelvis reference frames, were calculated. Findings: In addition to the expected bias between the rotation angle time-histories, very different patterns and range of motion were found between the models. In chair rising/sitting, and in the laboratory global frame, the range of flexion averaged over the subjects was measured by the different models in the full scale from about 28° to 44°. In elementary trunk rotation and in the pelvis anatomical reference frame, three models measured about 10° excursion of the coupled bending motion, other two about 38° and 49° on average. Interpretation: In trunk kinematics analysis, it is recommended that all models, both in terms of markers involved and of reference frame definitions, are understood carefully before interpreting the results in clinical decision making.