Assisted Diagnosis Technology Of Depression Using Heart Rate Variability

Depression is a typical mental illness and mainly characterized by persistent depressed mood,insomnia,fatigue and so on.It damages the physiology and psychology of people.The diagnosis and treatment of depression is very important.The current clinical diagnosis relies on the subjective judgment of doctors and patients,but an objective diagnosis is absence.The autonomic nervous system(ANS)of depressive patients is abnormal.Many researchers studied the ANS of depressive patients and healthy people using heart rate variability(HRV).They found that the HRV of depressive patients is significantly differ from that of heathy people.This phenomenon indicates that HRV be used to diagnose depression.In this study,we establish classier to recognize depressive patients using HRV.Depression is about twice as common in females as it is in males.Researches demonstrated the differences of ANS and HRV between depressive patients and heathy people,thus the difference between the depression effect of males and females may exist.Such differences may affect the effect of the classifier when using HRV to establish classifier to recognize depressive patients.Therefore,we make some works to establish more precise classifier.(1)Acquire HRV signals and calculate features to quantify the HRV signals.We collect the electrocardiogram signals of depressive patients and healthy people under multistate tests.Abnormal signals are detected using Shannon entropy of the normalized histogram of signals.Then we calculate HRV signals and their time and frequency domains and nonlinear features.(2)Survey the difference between the depression effect of males and females.This difference is surveyed according to the interaction of factors(Depression and Gender)on HRV features using ANOVA analysis.The results indicate that the depression effect of males differs from that of females.(3)Establish Bayesian Networks classifier.According to above difference and to obtain more precise classifier,only the HRV of females is employed to establish classifier using Bayesian Networks.In conclusion,we found the difference between depression effect of males and females using ANOVA analysis.The difference may affect the performance of the classier established using HRV.Therefore,only the HRV of females is employed to establish the classifier using Bayesian Networks.The sensitivity,specificity,and accuracy of the classifier is 89.5%,84.2%,and 86.8%,respectively.It indicates that it is feasible to establish the classifier using HRV for the assisted diagnosis of depression.