【精品】以头夹肌肉肌电图分析实现重郁症之电脑辅助诊断(Part9.pptVIP

Methods and Materials An artificial neural network (ANN) Artificial intelligence Mathematical model Learning system Computer-aid diagnosis Methods and Materials Back propagation neural network (BPNN) Supervised neural network Methods and Materials Back propagation neural network (BPNN) Input-to-hidden Weight Activation function : sigmoid Methods and Materials Back propagation neural network (BPNN) Hidden-to-output Weight Activation function : pure-linear Methods and Materials Back propagation neural network (BPNN) Adjust weights Methods and Materials Back propagation neural network (BPNN) Adjust weights Methods and Materials Support vector machine (SVM) Classification Statistics Hyperplane Results T.O.V.A MDD Higher omission rates Higher mean response times More variability EMG features MDD Lower EA and RMS values Higher MF and MPF values Results T.O.V.A Compare EMG Comparison by groups during rest Results T.O.V.A Compare EMG Comparison by group during TOVA Results Accuracy Conclusion and Discussion MDD becomes more distinguishable when resting Health controls have wider range of EA and MF MDD patients have the lower capability on physiological regulation Hopefully, the system can be used to detect and to control the MDD disorder in the future. References [1] J.M. Donohue and H.A. Pincus, Reducing the societal burden of depression: a review of economic costs, quality of care and effects of treatment,Pharmacoeconomics 25 (2007) 7. [2] P. Sobocki, B. Jonsson, J. Angst, C. Rehnberg.Cost of depression in Europe. J Ment Health Policy Econ 9 (2006) 87. [3] American Psychiatric Association, Diagnostic and statistical manual of mental disorders (American Psychiatric Association, Washington, DC, 2000). [4] R.C. Kessler, P. Berglund, O. Demler, R. Jin, D.Koretz, K.R. Merikangas, A.J. Rush, E.E. Walters, P.S. Wang; National Comorbidity Survey Replication, The epidemiology of major depressive disorder: results from the National Comorbidity Survey Repli