Multivariate Assessment of Voice Patterns as Biomarkers of Mental Health in Pre-College Students

• Grant number: 24/09587-9
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: March 01, 2025
• End date: December 31, 2025
• Field of knowledge: Health Sciences - Medicine - Psychiatry
• Principal Investigator: Lucas Murrins Marques
• Grantee: Ana Luisa Abutara Novaes
• Host Institution: Faculdade de Ciências Médicas da Santa Casa de São Paulo (FCMSCSP). Fundação Arnaldo Vieira de Carvalho. São Paulo , SP, Brazil

Abstract

Early detection of psychological disorders in young people, especially during periods of high academic pressure, is critical for effective interventions and adequate support. Recent studies have highlighted the human voice as a promising channel for non-invasive identification of signs of anxiety and depression. Vocal characteristic analysis offers a unique means of objectively inferring an individual's mental state using signal processing and machine learning technologies. This study represents a detailed exploratory analysis of the data collected in the previous project "Acoustic Assessment of the Voice and Anxiety of University Students during Exam Periods: A Test-Retest Study". The current research aims to significantly expand the scope of the initial study by applying advanced multivariate analysis techniques to explore the associations between various vocal characteristics and levels of anxiety and depression in students preparing for university entrance exams. Using a combination of multivariate regression methods, this study aims to identify vocal patterns that may serve as reliable indicators of adverse psychological states. The acoustic characteristics of the voice, such as zero-crossing rate, energy, energy entropy, spectral centroids, among others, will be analyzed in relation to their correlation with the scores obtained on the 7-item Generalized Anxiety Disorder scales (GAD- 7) and the Patient Health Questionnaire-9 (PHQ-9). By focusing on a quantitative and exploratory approach, the project seeks to offer substantial insights into how subtle variations in speech may reflect significant psychological changes, thus providing a promising avenue for early and personalized interventions in educational and clinical settings. This study not only expands the understanding of vocal biomarkers in mental health contexts, but also reinforces the applicability of voice analysis as a complementary diagnostic tool in psychological assessment.