Mental Health
1 item across 1 edition. First seen Wed 11 Nov, last seen Wed 11 Nov.
- The retrospective analysis covers message-level AI-derived emotional-state data from 28,870 sessions contributed by 4,463 unique users of Yuna, "an AI-powered DMHI" (digital mental health intervention), whose emotion-detection system scored 36 fine-grained emotional states at message level.
- Five trajectory types were identified: stable (7,679/28,870, 26.6%), reduced distress (7,042/28,870, 24.4%), transitional processing (6,195/28,870, 21.5%), persistent distress (4,568/28,870, 15.8%) and escalating distress (3,386/28,870, 11.7%).
- Among sessions that began in distress-associated states and had session-duration data (n=1,597), "73.3% (1,170/1,597) showed reduced-distress trajectories by session end". In mixed-effects logistic regression, more unique AI-derived emotional states (odds ratio 1.16, 95% CI 1.09–1.24; P<.001) and longer sessions (OR 1.01, 95% CI 1.00–1.02; P=.04) were positively associated with reduced-distress trajectories, while more emotional switches were negatively associated (OR 0.97, 95% CI 0.95–0.99; P=.003).
- The distress measure is the product's own classifier, not a validated self-report instrument, and all five authors are affiliated with Fit Minded, Inc., which makes Yuna. The authors state future research with self-report validation and experimental methods "is needed to determine whether these trajectory patterns carry a meaningful signal for user outcomes". Published 10 November 2026.