Depression Detection from Short Utterances via Diverse Smartphones in Natural Environmental Conditions

Abstract

Depression is a leading cause of disease burden worldwide, however there is an unmet need for screening and diagnostic measures that can be widely deployed in real-world environments. Voice-based diagnostic methods are convenient, non-invasive to elicit, and can be collected and processed in near real-time using modern smartphones, smart speakers, and other devices. Studies in voice-based depression detection to date have primarily focused on laboratory-collected voice samples, which are not representative of typical user environments or devices. This paper conducts the first investigation of voice-based depression assessment techniques on real-world data from 887 speakers, recorded using a variety of different smartphones. Evaluations on 16 hours of speech show that conservative segment selection strategies using highly thresholded voice activity detection, coupled with tailored normalization approaches are effective for mitigating smartphone channel variability and background environmental noise. Together, these strategies can achieve F1 scores comparable with or better than those from a combination of clean recordings, a single recording environment and long utterances. The scalability of speech elicitation via smartphone allows detailed models dependent on gender, smartphone manufacturer and/or elicitation task. Interestingly, results herein suggest that normalization based on these criteria may be more effective than tailored models for detecting depressed speech.

How to Design a Relevant Corpus for Sleepiness Detection Through Voice?

“…The other aspect impacting the quality of the recordings is the chosen recording device, which takes over the previous dichotomy between realism and control over the collected data. While an increasing number of studies are based on smartphones recordings [e.g., Huang et al ( 66 )], it could be feared that acoustic features extracted from these recordings may suffer from the low recording quality of these devices.…”

Automatic Elicitation Compliance for Short-Duration Speech Based Depression Detection

“…This study adopted four datasets that included only 'pataka' task utterances. They were derived from subsets of the Sonde Health 1 (SH1) [21], Sonde Health 2 (SH2) [10,18], Sonde Health 3 (SH3), and Yale depression datasets. Similarly to the SH1 and SH2, the SH3 was privately collected via personal Android and iOS smart devices (e.g.…”

“…Recently, automatic speech-based depression studies found among a review of dozens of studies [4] that the DDK 'pataka' task has been used due to its clinical history as an evaluative tool and restriction of speakers' phonetic variability unlike conversational speech activities. For example, [10] utilized acoustic speech features from 'pataka' recordings to automatically detect individuals with depression with nearly 70% accuracy. However, still little is known about what kind of influence the number of 'pataka' utterance or rate of speech have on acoustic-based features, and further, the effects these attributes have on automatic speech-based depression classification.…”

“…Speech landmark features provide information about abrupt articulatory-event changes that are effective for detecting depression in naturalistic environments [10]. Moreover, speech landmarks were initially designed to capture consonant-vowel transitions; and hence, are a logical fit for potentially detecting counts of 'pataka' [19].…”

Privacy Implications of Voice and Speech Analysis – Information Disclosure by Inference

“…Depressive speech can be detected automatically with high accuracy based on voice cues, even under adverse recording conditions, such as low microphone quality, short utterances, and background environmental noise [19,41]. Not only the detection, but also a severity assessment of depression is possible using a speech sample: In men and women, certain voice features were found to be highly predictive of their HAMD (Hamilton Depression Rating Scale) score, which is the most widely used diagnostic tool to measure a patient's degree of depression and suicide risk [36].…”