Acoustic and language analysis of speech for suicidal ideation among US veterans

Abstract

Background Screening for suicidal ideation in high-risk groups such as U.S. veterans is crucial for early detection and suicide prevention. Currently, screening is based on clinical interviews or self-report measures. Both approaches rely on subjects to disclose their suicidal thoughts. Innovative approaches are necessary to develop objective and clinically applicable assessments. Speech has been investigated as an objective marker to understand various mental states including suicidal ideation. In this work, we developed a machine learning and natural language processing classifier based on speech markers to screen for suicidal ideation in US veterans. Methodology Veterans submitted 588 narrative audio recordings via a mobile app in a real-life setting. In addition, participants completed self-report psychiatric scales and questionnaires. Recordings were analyzed to extract voice characteristics including prosodic, phonation, and glottal. The audios were also transcribed to extract textual features for linguistic analysis. We evaluated the acoustic and linguistic features using both statistical significance and ensemble feature selection. We also examined the performance of different machine learning algorithms on multiple combinations of features to classify suicidal and non-suicidal audios. Results A combined set of 15 acoustic and linguistic features of speech were identified by the ensemble feature selection. Random Forest classifier, using the selected set of features, correctly identified suicidal ideation in veterans with 86% sensitivity, 70% specificity, and an area under the receiver operating characteristic curve (AUC) of 80%. Conclusions Speech analysis of audios collected from veterans in everyday life settings using smartphones offers a promising approach for suicidal ideation detection. A machine learning classifier may eventually help clinicians identify and monitor high-risk veterans.

Utilizing digital predictive biomarkers to identify Veteran suicide risk

“…Belouali et al underscored how speech pattern recognition can be a digital biomarker for insight into suicidal ideation among Veterans ( 17 ). In this study, 124 Veteran participants provided 588 audio recordings of themselves answering opened ended questionnaires inquiring about their general health via a smartphone application that researchers designed.…”

“…The researchers also noted that demographics and mental states such as anxiety, depression, and PTSD might have confounded results. Further studies in this novel arena were recommended ( 17 ).…”

“…Screening refers to the systematic application of a test or enquiry to identify individuals at sufficient risk of a specific diagnosis to warrant further investigation ( 16 ). Currently, the screening of Veterans for suicide risk is limited to those who are willing to divulge their mental health status on self-report assessments, patient interviews in a clinical setting, suicide hotlines, and self, family, or friend reporting to a healthcare professional ( 17 ). In 2019, utilizing the VA's Suicide Risk Identification Strategy (Risk ID), the prevalence of suicidal ideation within the Veteran population was 3.5% ( 18 ).…”

Detection of Suicide Risk Using Vocal Characteristics: Systematic Review

“…The risk of bias was assessed as low, moderate, or high. The risk-of-bias assessments are available in Multimedia Appendix 3 [15,[22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41].…”

“…None of the included studies were at low risk of bias. In 43% (9/21) of the studies, the risk of bias was moderate [15,22,23,26,[28][29][30]32,[36][37][38][39][40][41], whereas in the remaining 57% (12/21), the risk of bias was assessed as high [24,25,27,31,[33][34][35]. The main sources of bias were confounding factors, selection of participants, and selective reporting of results.…”