Perception Of Musical Emotions In OCD:a Behavioral And SVM Study

This study explores the characteristics of the emotion perception of obsessivecompulsive disorder patients through behavioral and machine learning methods.Based on the experimental results,we try to classify and predict the severity of obsessive-compulsive disorder symptoms by machine learning.Nineteen mild obsessive-compulsive disorder patients were recruited from the psychological consultation clinic and inpatients of Guangji Hospital in Suzhou.In this study,a standardized music fragment of five emotions: anger,fear,sadness,tenderness,and happy was selected as the experimental material for the behavioral experiment.The subjects were tested in anger,fear,sadness,tenderness and happiness..The results of repeated measures analysis of variance of matched obsessive-compulsive disorder group and control group showed significant difference in many items,and the data showing these differences were input into support vector regression machine to learn and classify.The results of music emotion perception predicted the symptoms of obsessive-compulsive disorder.The conclusions are as follows :(1)The sensitivity of the patients with mild OCD to the negative emotion(anger,fear,sadness)was higher than that of the control group,while the sensitivities of the positive mood stimulation(tenderness and happy)were differentiated,And the sensitivity of happy was not significantly different between the two groups.It indicated that OCD mainly changed the process of negative emotion recognition,and the positive emotion recognition of different awakening degree was different.(2)Anger,fear and sadness of the emotional material of the target emotional and non-target emotions were found significant differences,indicating anger and fear,sadness and peace may have a closer relationship in the production mechanism,such as by the patient personality factors,Regulating ability;anger and fear,sadness and peace may have homology to the patient's meaning.(3)To establish a input plane,and to input the target emotion score of anger,fear,sadness and calm material,the output of support vector regression is the best.(4)support vector regression machine in cognitive modification,judgments,combined with cognitive therapy,there is a larger application space.This study may give reference to clinical diagnosis and treatment of obsessivecompulsive disorder and other mental disorders.