Detection of Clinical Depression in Adolescents’ Speech During Family Interactions

Abstract

The properties of acoustic speech have previously been investigated as possible cues for depression in adults. However, these studies were restricted to small populations of patients and the speech recordings were made during patients’ clinical interviews or fixed-text reading sessions. Symptoms of depression often first appear during adolescence at a time when the voice is changing, in both males and females, suggesting that specific studies of these phenomena in adolescent populations are warranted. This study investigated acoustic correlates of depression in a large sample of 139 adolescents (68 clinically depressed and 71 controls). Speech recordings were made during naturalistic interactions between adolescents and their parents. Prosodic, cepstral, spectral, and glottal features, as well as features derived from the Teager energy operator (TEO), were tested within a binary classification framework. Strong gender differences in classification accuracy were observed. The TEO-based features clearly outperformed all other features and feature combinations, providing classification accuracy ranging between 81%–87% for males and 72%–79% for females. Close, but slightly less accurate, results were obtained by combining glottal features with prosodic and spectral features (67%–69% for males and 70%–75% for females). These findings indicate the importance of nonlinear mechanisms associated with the glottal flow formation as cues for clinical depression.

A comparative study of different classifiers for detecting depression from spontaneous speech

“…That might be an indication that pitch and formants features are more suitable in a speaker-dependent comparison. The voice quality group performed relatively better than pitch and formants when using SVMs, which is in line with psychological conclusions that depressed speech is characterised by irregularities in the vocal fold vibrations [6,7,8].…”

“…Jitter and shimmer voice features were analysed for depression, finding higher jitter in depression caused by the irregularity of the vocal fold vibrations [6] and lower shimmer for depressed subjects [7]. Like the jitter feature, the harmonic-to-noise (HNR) feature is higher for depressed, as the patterns of air flow in the speech production differ for depressed and control subjects [8].…”

A Cross-modal Review of Indicators for Depression Detection Systems

“…Therefore, it is important to keep these factors in mind when building a detection system. For example, in many studies systems have achieved better results using sex-dependent classifiers (Moore et al, 2008;Low et al, 2011;Yang et al, 2016; Table 2: Best performing depression detection systems. F1 score, precision, and sensitivity are reported for the depressed class.…”

Detecting depression: A comparison between spontaneous and read speech

“…On the other hand, shimmer is lower for depressed [10]. Like the jitter feature, the harmonic-to-noise (HNR) feature is higher for depressed due to the patterns of air flow in the speech production differing for depressed and healthy controls [11]. Voice energy is also a widely used distinguishing feature for depression, giving lower energy for depressed patients caused by the glottal pulses.…”