Abstract:Coping strategies of patients with obstetric brachial plexus palsy (OBPP) are highly individual. Up to now, individual movement performance is assessed by visual observations of physicians or therapists - a procedure, which is highly subjective and lacks objective data. However, objective data about the individual movement performance are the key to evidence-based and individualized treatment. In this paper, a new approach is presented, which provides objective information about the upper extremity movement performance of patients with OBPP. The approach is based on the use of accelerometers in combination with a classification procedure. The movement performance of 10 healthy volunteers and 41 patients with OBPP has been evaluated by experienced physiotherapists and has been assigned to one of 4 categories representing the Mallet Scale (MS) IV to I. Three triaxial-accelerometers were placed at chest, upper arm and wrist of the affected side of the patient. Acceleration signals have been recorded during repetitive movements with relevance regarding daily life. Here, especially the results from the “hand to mouth” task are presented. From the 9 recorded acceleration signals 13 relevant features were extracted. For each of the 13 features 4 thresholds have been determined distinguishing best between the 4 patient categories of the MS and the healthy subjects. With respect to the thresholds each feature value has been assigned to the discrete numbers 0, 1, 2, 3 or 4. Afterwards, each discrete number has been weighted by a factor regarding the correlation between the feature’s value and the MS score. The resulting weighted discrete numbers of all 13 features have been added resulting in a score, which quantifies the individual upper extremity movement performance. Based on this score the movement performance of each patient has been assigned to the classes “very good”, “good”, “regular” and “bad”. All movements of the 10 healthy volunteers were classified as “very good”. The movement performance of two patients MS IV were classified as “very good” as well and the movements of the other 16 patients as “good”. The movements of the entire group of MS III patients fell into the class “regular”. Just one MS II patient was assigned to the class “regular” while the others were classified as “bad”. It was not possible to classify the movements of MS I patients. This was mainly due to the fact that none of these patients MS I was able to complete the task successfully. The developed approach demonstrated its ability to quantify the movement performance of upper extremity movements based on accelerometers. This provides an easy to use tool to assess patient’s movement strategies during daily tasks for diagnosis and rehabilitation.
Capture, learning, and classification of upper extremity movement primitives in healthy controls and stroke patients
