Research Article
Automatic Recognition of Fear-Avoidance behavior in Chronic Pain Physical Rehabilitation
@INPROCEEDINGS{10.4108/icst.pervasivehealth.2014.254945, author={Min Aung and Nadia Bianchi-Berthouze and Paul Watson and Amanda Williams}, title={Automatic Recognition of Fear-Avoidance behavior in Chronic Pain Physical Rehabilitation}, proceedings={8th International Conference on Pervasive Computing Technologies for Healthcare}, publisher={ICST}, proceedings_a={PERVASIVEHEALTH}, year={2014}, month={7}, keywords={emotion protective behaviour machine learning motion capture electromyography body movement pain rehabilitation technology physical rehabilitation}, doi={10.4108/icst.pervasivehealth.2014.254945} }- Min Aung
Nadia Bianchi-Berthouze
Paul Watson
Amanda Williams
Year: 2014
Automatic Recognition of Fear-Avoidance behavior in Chronic Pain Physical Rehabilitation
PERVASIVEHEALTH
ACM
DOI: 10.4108/icst.pervasivehealth.2014.254945
Abstract
Physical activity is beneficial in chronic pain rehabilitation. However, due to psychological anxieties about pain and the percevied risk of injury, physical activity is often avoided by people with chronic pain. This avoidance is expressed through self protective body movement aimed at avoiding strain, particularly in painful areas. The detection of protective behaviour is crucial for effective rehabilitation advice and to enable a more normal lifestyle. Current technology to motivate physical activity in rehabilitation contexts does not address these psychological barriers. In this paper, we investigate the automatic recognition of a specific form of protective behaviour, guarding, common in people with chronic lower back pain. We trained ensembles of decision trees, Random Forests, on posture and velocity based features from motion capture and electromyographic data. Results show overall out of bag F1-classification scores of 0.81 and 0.73 for sitting to standing and one leg stand exercises respectively.