AI-driven dynamic psychological measurement: correcting university student mental health scales using daily behavioral and cognitive data.
Source check: Source checked. Not independent scientific confirmation.
Finding: LLM/RAG-based score adjustment used daily behavioral and cognitive data from 177 university students. The paper reports higher AUCs than standard scales for anxiety and depression classifications and longitudinal decreases on clinician-rated scales, while SCL-90 specificity worsened; these observational measurements do not show intervention efficacy or diagnostic validity.
Technology: Large language models with retrieval-augmented generation dynamically adjust questionnaire scores using daily behavioral and cognitive data.
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
- AI contribution
- LLM/RAG score adjustment is linked to differences in scale-based assessment and longitudinal score patterns.
- Test and result
- Among 177 students, reported AUC was 0.95 versus 0.86 for anxiety and 0.93 versus 0.82 for depression; HAM-A/HAM-D scores declined while SCL-90 specificity worsened.
- Limitation / still unknown
- Longitudinal observational comparison; score changes do not demonstrate an intervention effect, clinical diagnosis, or generalizable measurement validity.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.frontiersin.org/journals/digital-health/articles/10.3389/fdgth.2025.1615250/full
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary source full-text Methods and Results; version date and reported validation scope checked..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.