Predicting Depression Severity in Major Depressive Disorder Using Electroencephalography and Heart Rate Variability:...

Introduction Mood disorders comprise the largest burden on mental health, with major depressive disorder (MDD) being the most prevalent among them. While the diagnosis of mood disorders is generally subjective based on history and structured interviews or questionnaires, efforts are being made to identify biological markers that could provide an objective method for diagnosis, classification, prognosis, and elucidation of the underlying mechanisms of mood disorders. Electroencephalography (EEG) and heart rate variability (HRV) are two biosignals that have been studied in this regard. This study aimed to assess the associations between the severity of MDD, EEG alpha asymmetry, and HRV indices in individuals diagnosed with MDD. The EEG and HRV findings were also compared with age- and gender-matched euthymic controls. Methods In this cross-sectional study, 45 individuals (28 males and 17 females, aged 18-50 years) with MDD (as per the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition) who were either drug naive or had a drug-free interval of at least six weeks (defaulters) were included in the test group. Similarly, 45 age- and gender-matched healthy individuals were included in the control group. The study parameters, namely, the severity of depression, EEG alpha asymmetry, and HRV indices, were assessed from the Hamilton Depression Rating Scale 17 (HDRS-17) score, EEG recording, and lead II ECG data, respectively, and were compared across the test and control groups. In addition, the bivariate correlation between the severity of depression and other parameters in the MDD group and linear regression modeling with the severity of depression as the dependent variable and EEG alpha asymmetry and HRV indices as predictors were performed for the test group. Results HRV indices, namely LF n.u. and LF/HF ratio, were significantly higher, and HF n.u. was significantly lower in MDD individuals compared to healthy controls. A comparison of alpha asymmetry metrics revealed significantly high alpha asymmetry in MDD individuals compared to controls in all three frontal lead pairs (Fp1-Fp2, F3-F4, and F7-F8) and the T3-T4 temporal lead pair. While weak positive correlations were observed between HDRS-17 score and age (ρ = 0.313, p = 0.036), illness duration (ρ = 0.336, p = 0.024), F7-F8 asymmetry (r = 0.364, p = 0.014), LF n.u. (r = 0.349, p = 0.019), and LF/HF ratio (ρ = 0.416, p = 0.005), a weak negative correlation was observed between HDRS-17 and HF n.u. (r = -0.347, p = 0.019) in the test group. Similarly, age, illness duration, F7-F8 asymmetry, O1-O2 asymmetry, LF n.u., HF n.u., and LF/HF ratio were identified as correlates of HDRS-17 for multiple linear regression modeling. The resulting model, with F7-F8 asymmetry and HF n.u. as predictors, was found to be statistically significant in the test group (R² = 0.224, p = 0.005). Conclusions Evidence of a possible relationship between the severity of depression and two objective physiological measures, namely HRV HF n.u. and EEG frontal alpha asymmetry (FAA), was observed in this study. The combination of HRV HF n.u. and EEG F7-F8 FAA could explain 22.4% of the variation in severity of depression, as quantified by HDRS-17, mainly in newly diagnosed drug-naive patients of MDD at presentation.