Title: Interpreting the Tilt-Torsion Method to Express Shoulder Joint Kinematics

Abstract: Background: Kinematics is studied by practitioners and researchers in different fields of practice. It is therefore critically important to adhere to a taxonomy that explicitly describes positions and movements. However, current representation methods such as cardan and Euler angles fail to report shoulder angles in a way that is easily and correctly interpreted by practitioners, and that is free from numerical instability such as gimbal lock (GL). Methods: In this paper, we comprehensively describe the recent Tilt-Torsion (TT) method and compare it to the Euler YXY method currently recommended by the International Society of Biomechanics. While using the same three rotations as Euler YXY (plane of elevation, elevation, humeral rotation), TT reports humeral rotation independently from the plane of elevation. We assess how TT can be used to describe shoulder angles (1) in a simulated assessment of humeral rotation with the arm at the side, which constitutes a GL position, and (2) during an experimental functional task, with 10 wheelchair basketball athletes who sprint in straight line using a sports wheelchair. Findings: In the simulated GL experiment, TT provided both humeral elevation and rotation measurements, contrary to the Euler YXY method, despite both methods sharing the same GL positions. During the wheelchair sprints, humeral rotation ranged from 14{\deg} (externally) to 13{\deg} (internally), which is consistent with typical maximal ranges of humeral rotation, compared to 65{\deg} to 50{\deg} with the Euler YXY method. Interpretation: Based on our results, we recommend that shoulder angles be expressed using TT instead of Euler YXY.