JMIR Publications

ABSTRACT

Background:

College students face significant mental health challenges, yet traditional assessments often rely on in-person self-reports and long-form self-report survey approaches. The prevalence of smartphones and longitudinal digital phenotyping—using data from devices to rapidly monitor mental health processes—offers an opportunity to assess wellbeing using a nondisruptive, prevention-focused “light touch” approach via voice journaling and emoji-based expression. While voice biomarkers are associated with clinical anxiety and depression, little research has explored voice biomarkers’ relation to psychological predictors of mental health processes in young adults.

Objective:

This study explores the utility of smartphone-captured voice biomarkers as objective indicators of psychological predictors of health outcomes in college students.

Methods:

College students (N = 289) used the Meet Pandora smartphone application for over two months. Approximately 3x/week, participants responded to voice journal prompts, such as “Describe a place you like to visit”. Voice biomarkers were extracted from responses. Voice biomarkers and their association with concurrent psychological measurements (i.e., voice journaling measured closest in time to psychology surveys) were modeled using multivariable, mixed-effects, covariate-adjusted (gender, baseline health diagnosis, prompt sentiment and time). Random intercepts were included to control for baseline individual differences, and all analyses were adjusted for multiple comparisons.

Results:

Higher word count (OR 1.27; 95% CI, 1.09, 1.49; adj P=.02) and longer recording duration (OR 1.22; 95% CI, 1.04, 1.42; adj P=.04) predicted lower self-regard. No voice features significantly predicted loneliness, acceptance, or stress. For daily emoji-based affect assessments, higher sentiment scores (OR 1.10; 95% CI, 1.04, 1.17; adj P=.002) and increased pitch (OR 1.24; 95% CI, 1.15, 1.34; adj P<.001), as well as additional acoustic features reflecting energy and voice quality, were associated with positive daily affect. Tests of valence and arousal combinations, derived from emojis, reinforced trends, revealing significant acoustic and speech differences when comparing high activation negative affect states to other states.

Conclusions:

Speech characteristics, specifically higher word count and longer duration predicted low self-regard, which correlates with internalizing symptoms. While speech and acoustic features did not predict weekly stress, loneliness, or acceptance, they were associated with daily affect fluctuations. These findings highlight the utility of high-frequency assessments via voice biomarkers as signals of affective changes in college students. This work highlights that brief vocal samples may be able to detect early, affect-specific signs of mental health risks in young adults which has important implications for psychopathology prevention and improving wellbeing.