Validity and reliability of shoulder kinematics in typically developing children and children with hemiplegic cerebral palsy

Shoulder motion has been mainly analysed in children based on thoraco-humeral (TH) joint kinematics, excluding the scapula-thoracic (ST) and gleno-humeral (GH) joints. In order to measure 3D scapulo-humeral motion using an optoelectronic system, we propose a protocol based on an acromion marker cluster (AMC), a functional method to determine the gleno-humeral rotation centre and different Euler sequences. This study investigated the validity of the AMC compared to the palpation of anatomical landmarks with a scapula locator, assessed the intra-session repeatability and the ability to discriminate differences of such a protocol in 10 typically developing children (TD) and 10 children with hemiplegic cerebral palsy (HCP) during 6 different tasks (flexion, abduction, horizontal abduction, hand to head, hand to controlateral shoulder and hand to back pocket). For both populations, the AMC method showed an overall Root Mean Square Error (RMSE) of 5.5°. The AMC method under-estimated the protraction/retraction of the scapula during abduction. The within-session reliability was good to excellent for all tasks except the hand to back pocket task. The YXY recommended Euler sequence for TH and GH joints resulted in gimbal lock for most of the tasks whereas the XZY sequence could be used for most of the tasks and most of the children.