Predicting Momentary Suicidal Ideation From Smartphone Screenshots Using Vision-Language Models: Prospective Machine Learning Study
JMIR Mental Health · 2026
DOI: 10.2196/90581 ↗At a glance
Journal article in JMIR Mental Health, 2026. Written by Ross Jacobucci and 3 co-authors. Institution: Center for Healthy Minds, University of Wisconsin-Madison, 625 W Washington Ave, Madison, WI, 53703, United States, 1 (608) 263-6321. Major MEDLINE topics: Smartphone, Suicidal Ideation, Machine Learning. Cited 1 times so far (Crossref); 76 references. The full text is freely available in PubMed Central. Funded by: NIMH NIH HHS.
Abstract (original language)
Background
Passive smartphone sensing shows promise for suicide prevention, but behavioral metadata (GPS, screen time, and accelerometry) often lacks the contextual information needed to detect acute psychological distress. Analyzing what people actually see, read, and type on their phones—rather than just usage patterns—may provide more proximal signals of risk.
Objective
This study aimed to test whether vision-language models (VLMs) applied to passively captured smartphone screenshots can predict momentary suicidal ideation (SI).
Methods
Seventy-nine adults with past month suicidal thoughts or behaviors completed ecological momentary assessments (EMA) over 28 days while screenshots were captured every 5 seconds during active phone use. We fine-tuned open-source VLMs (Qwen2.5-VL [Alibaba Cloud], LFM2-VL [Liquid AI]), and text-only models (Qwen3 [Alibaba Cloud]) to predict SI from screenshots captured in the 2 hours preceding each EMA. We evaluated performance with temporal and subject holdouts.
Results
The analytic sample comprised 2.5 million screenshots from 70 participants. Temporal holdout models achieved strong discrimination at the EMA level (AUC=0.83; AUPRC=0.77), with image-based models outperforming text-only models (AUC=0.83 vs 0.79; 95% CI 0.003-0.07). Subject holdout generalization was near chance (AUC≈0.50), though a simple lexical screening method retained modest discrimination (AUC=0.62). Smaller models performed comparably to larger models, supporting feasible on-device deployment.
Conclusions
Screen content predicts short-term SI with clinically meaningful accuracy when models are personalized but does not generalize across individuals. These findings support a 2-stage clinical architecture, coarse lexical screening for new patients, with personalized VLM-based monitoring after a calibration period. On-device inference may enable privacy-preserving deployment.
Subjects
MeSH terms (MEDLINE)
Author keywords
Authors and institutions
- Center for Healthy Minds, University of Wisconsin-Madison, 625 W Washington Ave, Madison, WI, 53703, United States, 1 (608) 263-6321.
- Department of Psychology, University of Wisconsin-Madison, Madison, WI, United States.
Funding
- NIMH NIH HHS
Cite
APA
Jacobucci, R., Shao, W., Kobrinsky, V., & Ammerman, B. (2026). Predicting Momentary Suicidal Ideation From Smartphone Screenshots Using Vision-Language Models: Prospective Machine Learning Study. JMIR Mental Health. https://doi.org/10.2196/90581
Vancouver
Jacobucci R, Shao W, Kobrinsky V, Ammerman B. Predicting Momentary Suicidal Ideation From Smartphone Screenshots Using Vision-Language Models: Prospective Machine Learning Study. JMIR Mental Health. 2026. doi:10.2196/90581
BibTeX
@article{Jacobucci2026,
author = {Jacobucci, Ross and Shao, Wenpei and Kobrinsky, Veronika and Ammerman, Brooke},
title = {Predicting Momentary Suicidal Ideation From Smartphone Screenshots Using Vision-Language Models: Prospective Machine Learning Study},
journal = {JMIR Mental Health},
year = {2026},
doi = {10.2196/90581}
}
EUVIMED