Quantitative separation of the depressive phase of bipolar disorder and major depressive disorder using electrovestibulography

Abstract

Objectives: No electrophysiological, neuroimaging or genetic markers have been established that strongly relate to the diagnostic separation of bipolar disorder (BD) and major depressive disorder (MDD). This paper’s objective is to describe the potential of features, extracted from the recording of electrical activity from the outer ear canal, in a process called electrovestibulography (EVestG), for identifying depressed and partly remitted/remitted MDD and BD patients from each other.

Methods: From EVestG data four sensory vestibulo-acoustic features were extracted from both background (no movement) and using a single supine-vertical translation stimulus to distinguish 27 controls, 39 MDD and 43 BD patients.

Results: Using leave-one-out-cross-validation, unbiased parametric and non-parametric classification routines resulted in 78–83% (2–3 features), 80–81% (1–2 features) and 66–68% (3 features) accuracies for separation of MDD from BD, controls from depressed (BD & MDD) and the 3-way separation of BD from MDD from control groups, respectively. The main limitations of this study were the inability to fully disentangle the impact of prescribed medication from the responses and also the limited sample size.

Conclusions: EVestG features can reliably identify depressed and partly remitted/remitted MDD and BD patients from each other.