POSSIBLE ERRORS IN MEASUREMENT OF SHOULDER AXIS ANGLE USING 3-D CINEMATOGRAPHY

Abstract

INTRODUCTlON - Two and three dimensional cinematography have been extensively employed to measure hip and shoulder horizontal axis angles in a bid to infer relative injury risks to the lower back in Cricket fast bowling. The Mixed technique has been defined using some of these measurement systems and recommendations have been made with respect to minimising the hip-to-shoulder separation an- gle and hence the torsional load placed on the lower lumbar vertebrae. However, it may be that these systems are not ideal in terms of quantifying the degree of spinal rotation due to problems with the method employed. The object of this study was to determine the magnitude of errors involved in horizontal shoulder axis determination owing to movements of the shoulder girdle. 10 athletic male volunteers performed an extensive warm-up and stretching routine and were seated on a high-backed chair and secured using straps around the lower abdominal region (to secure the pelvis) and the nipple line (to minimise lumbar and lower thoracic rotation). With the frontal plane parallel to the principal plane of per- formance, ranges of horizontal extension and flexion were measured for each subject to provide an index of shoulder joint and shoulder girdle flexibility. Subjects were also instructed to attain three different positions of each humerus/shoulder girdle: Position 1 -humerus abducted through 90°, Position 2 - maximal horizontal flexion, Po- sition 3 - maximal horizontal extension. These positions were repeated for both arms resulting in 9 combinations. Two Panasonic F15 video cameras were used to film each subject. The footage was digitised using an Acorn Archimedes microcomputer. Reconstruction of the two sets of 2- D image-space coordinates into 3-D real world coordinates was performed using a DLT algorithm. RESULTS - The results showed that al- though lumbar and lower thoracic rotation was prevented, the shoulder angles and hence the hip-to-shoulder separation angles suggested large degrees of vertebral rotation (Range 2-29" for all conditions). The degree of measured rotation varied across conditions and across subjects with the most flexible subjects showing large discrepancies from the expected score. The greatest deviation from the plane of performance (21.4 _+ 9") occurred with the left arm in position 2 and the right arm in position 3, and vice versa. These are similar positions to those adopted during fast bowling. CONCLUSIONS -The results question the use of this popular method of determining shoulder and hip-to-shoulder angles in order to accurately measure, or at least estimate the torsional stress experienced in the lower back. It is important to note that this error in measurement can not only lead to an overestimation of torsional stress, as shown in this study, it can also lead to an underestimation. Both can have serious implications for the athlete. Alternative methods of determination of spinal rotation may provide a more accurate measure especially in activities which involve large degrees of shoulder girdle movement.