Previously submitted to: JMIR Mental Health (no longer under consideration since Feb 12, 2026)
Date Submitted: Feb 7, 2026
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
When Psychological Change Becomes Textually Latent: A Viewpoint on Retrospective Longitudinal Analysis of Reflective Mental Health Text
ABSTRACT
Background:
Digital mental health applications increasingly deploy sentiment analysis tools to monitor psychological change. While these tools effectively detect affective transformations (emotional valence changes), their performance on cognitive-structural transformations (identity integration, meaning-making, acceptance development) remains underexplored.
Objective:
This Viewpoint evaluates whether standard affect-detection tools can capture cognitive-structural transformations, using a longitudinal single-subject corpus exhibiting both transformation types.
Methods:
Using a single-subject, multi-environment corpus spanning both introspective therapeutic dialogue and emotionally expressive casual chat, we compared three widely used paradigms: deterministic lexicon accumulation, supervised emotion classification, and LLM-based message-level affect inference. Lexicon-based (Phase A) and supervised models (Phase B) marginally tracked affective change across time in both expressive and reflective contexts. LLM-based inference (Phase C), however, failed to detect drift in introspective text while succeeding in high-affect, stylistically foregrounded language, the casual communications corpus.
Results:
While these methods detected modest affective changes in casual conversation, they showed no systematic differentiation within the therapeutic corpus once message length and aggregation effects were considered. In contrast, construct-specific analyses focused on identity integration and acceptance language, together with longitudinal contextual interpretation, identified a coherent psychological transformation corroborated by self-report and behavioral change. Without message length normalization, therapeutic progress could be confounded with changes in verbosity, elaboration, or reflective depth -all of which may increase during productive therapeutic engagement independent of emotional state changes.
Conclusions:
Standard sentiment tools remain effective for detecting affective change but as illustrated by this case, might be insufficient for identifying meaning-oriented psychological transformation. Future digital mental health research should distinguish between affective and cognitive change phenotypes and develop construct-appropriate analytic frameworks to avoid systematic blind spots
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.