• Citer

Intelligent diagnostic system for low back pain using dynamic motion characteristics

par Bishop, Jeffrey;Ananthraman, Santosh S.K.;McIntyre, Donald D.R.;Szpalski, Marek ;Pope, Malcolm M.H.
Référence Journal of intelligent systems, 8, 1-2, page (185-201)
Publication Publié, 1998

Article révisé par les pairs

• Détails
• Contenu

Résumé :

The purpose of this study is to show that people with low back pain move differently from people without low back pain and that an intelligent diagnostic system can be developed to classify low back pain status based on features of dynamic motion. Low back pain (LBP) is a problem of enormous medical and economic importance. Lifetime prevalence for low back pain is 75 percent, yet precise diagnosis unknown in 80-90 percent of patients. Accurate evaluation of LBP is very difficult. An artificial neural network has been developed for analysis of trunk dynamic motion as a diagnostic aid for LBP. A neural network approach was selected because the effect of LBP on back movement is very hard to model based on first principles and can be considered a non-linear multiple input multiple output (MIMO) system. A radial basis function network is trained using motion features extracted from data collected from patients who have back pain that has been classified by a clinician. Pain classification of up to 86% accuracy is achieved by the system.