Evaluation of MS Predictor Factors with Emphasis on Neurophysiological Indices of Physical Activity
Oral Presentation
Paper ID : 1528-12THCONG
Oral / Poster Presentation File: 1528-12THCONG .mp4
Authors
1PhD Student in Exercise Physiology, Department of Exercise Physiology, Faculty of Sport Sciences, University of Mazandaran, Babolsar, Iran
22. Associate Professor, Department of Exercise Physiology, Faculty of Sport Sciences, University of Mazandaran, Babolsar, Iran
Abstract
Multiple sclerosis (MS) is a debilitating disease of the nervous system that costs of treatment of the problems involved with the disease. Early awareness can make the treatment more effective and more informed about the status of the disease.
The purpose of this study was to predict and diagnose early MS based on movement-dependent neurophysiological variables.
This is a cross-sectional analytical study. 110 men and women with and without MS disease Mazandaran province participated in this study. Three groups were classified. Group I: 72 patients with MS and 38 healthy control group, Group II: the active control group (n = 18), non-active control (n = 20), MS active (n = 27) and non-active MS (n = 45), Group III: 46 patients in Relapsing-Remitting (RR) and 26 patients in Secondary Progressive (SP) and 38 patients in the control group. Measurements of 1RM quadriceps and trunk extensors, EMG leg, dynamic and static balance, the flexibility of four muscles in both legs, body composition, gait analysis via Simi Motion Analysis 3D cameras were evaluated. Twenty variables in the non-kinematic section and 200 variables in the kinematics section were analyzed. In this study, the SVM prediction method for the separation accuracy and precision of prediction.
Group I showed accurately predicted healthy with MS 99/1% resolution and accuracy was 90%. Group II: The prediction accuracy of the control with inactive MS was 95% in the kinematics section. Group III: Among the two disease models, prediction accuracy with the SP model was the most accurate (66%).
Using machine learning techniques, to achieve high accuracy in the diagnosis of healthy and important features were identified. Measuring variables and segmentation apply to costly tests can predict a person's right to have a warning for the individual and society. The effects of static balance, power, dynamic equilibrium variables were more than other variables. Using this predictive model, we can alert high-risk individuals before they have the disease to prepare for treatment before the symptoms worsen.
The purpose of this study was to predict and diagnose early MS based on movement-dependent neurophysiological variables.
This is a cross-sectional analytical study. 110 men and women with and without MS disease Mazandaran province participated in this study. Three groups were classified. Group I: 72 patients with MS and 38 healthy control group, Group II: the active control group (n = 18), non-active control (n = 20), MS active (n = 27) and non-active MS (n = 45), Group III: 46 patients in Relapsing-Remitting (RR) and 26 patients in Secondary Progressive (SP) and 38 patients in the control group. Measurements of 1RM quadriceps and trunk extensors, EMG leg, dynamic and static balance, the flexibility of four muscles in both legs, body composition, gait analysis via Simi Motion Analysis 3D cameras were evaluated. Twenty variables in the non-kinematic section and 200 variables in the kinematics section were analyzed. In this study, the SVM prediction method for the separation accuracy and precision of prediction.
Group I showed accurately predicted healthy with MS 99/1% resolution and accuracy was 90%. Group II: The prediction accuracy of the control with inactive MS was 95% in the kinematics section. Group III: Among the two disease models, prediction accuracy with the SP model was the most accurate (66%).
Using machine learning techniques, to achieve high accuracy in the diagnosis of healthy and important features were identified. Measuring variables and segmentation apply to costly tests can predict a person's right to have a warning for the individual and society. The effects of static balance, power, dynamic equilibrium variables were more than other variables. Using this predictive model, we can alert high-risk individuals before they have the disease to prepare for treatment before the symptoms worsen.
Keywords
Disease prediction; Multiple Sclerosis; Relapsing-Remitting; Secondary Progressive; machine learning; SVM
Subjects