Development of a Mental Disorder Screening System Using Support Vector Machine for Classification of Heart Rate Variability Measured from Single-lead Electrocardiography

The diagnosis of psychiatric disorders, such as major depressive disorder (MDD), depends on clinical interviews and assessment of symptoms. However, due to the fact that mental states cannot be objectively assessed, diagnosis procedures are often influenced by clinical experience of psychiatrist. Hence, the aim of this study is to develop a simple, objective, highly accurate self-check system for screening of psychiatric disorders based on heart rate variability (HRV) measured from single- lead electrocardiography (ECG) or photoplethysmogram (PPG) for home healthcare monitoring. HRV is widely used as objective biomarker for assessment of autonomic nerve system. The low frequency (LF) of HRV originates from the sympathetic and parasympathetic nerves. The high frequency (HF) originates from the parasympathetic nerves. In our previous clinical trial, we confirmed that HRV of MDD patients is less reactive than healthy subjects during a mental task (generate random numbers) condition. However, mental task alone is difficult to assess HRV accurately owing to influence of measurement condition and individual differences. In this study, we implemented a single-lead ECG system based on reactivity of HRV, and combined mental task and paced deep breathing thereby improving the screening accuracy. Moreover, support vector machine (SVM) model was applied for classification of HRV indices. We tested the system on 16 healthy subjects and 7 psychiatric patients with depression or somatoform disorder. A significant difference was found between the healthy group and the patient group for the response of the HRV indices on dual mental tasks. The SVM non-linear classification model achieved a sensitivity of 71.4% and specificity of 93.8%.