Within-Person Symptom Dynamics in Serious Mental Illness: Insights from Daily Self-reports and Audio Journals
Abstract
Understanding how clinical symptoms fluctuate within individuals is crucial for developing personalized treatment, yet these dynamics remain understudied. We examined whether daily self-reports and speech features from audio journals predicted monthly fluctuations in clinician-rated symptoms among 54 outpatients with affective and psychotic disorders. We identified seven within-person symptom dimensions: disorganized thinking, mania, social withdrawal, inexpressivity, affective depression, cognitive depression, and anxiety. Daily self-reports and speech features were summarized on a monthly level across different temporal windows highlighting 3, 7, 14, and 28 days of daily data prior to the clinical interview to assess the predictive value of emphasizing more or less recent daily histories. Bayesian multilevel models indicated that both the intensity and day-to-day variability of daily experiences and speech features tracked month-to-month symptom fluctuations. Daily Positive Affect and speech sentiment were the most broadly informative indicators of fluctuations across symptom domains. Greater Daily Positive Affect or more positive sentiment predicted increased mania, whereas lower levels tracked increased other symptoms except for disorganized thinking. Model comparisons across temporal windows showed that models using daily self-reports were generally robust to different windows, although social withdrawal showed improved performance highlighting longer daily histories. Models using speech features generally performed better with longer temporal windows, whereas mania and anxiety showed no clear temporal window preference. Overall, these findings demonstrate the utility of daily self-reports and audio journals for dynamic symptom tracking and showcase a data-driven approach to investigating the optimal length of daily history for developing personalized monitoring tools in psychiatry.
Related
- Demographic Moderators of the Relationship Between Psychiatric Diagnosis and Functional Capacity in People with Serious Mental Illness
- The Paisa Kids Study: Rationale, Methods, and Initial Characterization of a Youth Cohort at Transdiagnostic High Risk for Serious Mental Illness
- Algorithmic Circuit Psychiatry: A Roadmap for Precision Treatment of Serious Mental Illness
- Acceptability of blinded continuous glucose monitoring for people with serious mental illness and type 2 diabetes: findings from an exploratory mixed-methods study.
- Glucagon‐Like Peptide‐1 Receptor Agonists and Incident Cardiovascular Events and Death in Treated Serious Mental Illness, Type 2 Diabetes, and Obesity
BibTeX
@misc{jun2026withinperson,
title = {{Within-Person Symptom Dynamics in Serious Mental Illness: Insights from Daily Self-reports and Audio Journals}},
author = {Daiil Jun and Jeffrey M. Girard and Yoonho Chung and Scott L. Rauch and Einat Liebenthal and Justin T. Baker},
howpublished = {PsyArXiv},
year = {2026},
month = sep,
doi = {10.31234/osf.io/aqnmw_v2},
url = {https://osf.io/preprints/psyarxiv/aqnmw_v2/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Within-Person Symptom Dynamics in Serious Mental Illness: Insights from Daily Self-reports and Audio Journals
AU - Daiil Jun
AU - Jeffrey M. Girard
AU - Yoonho Chung
AU - Scott L. Rauch
AU - Einat Liebenthal
AU - Justin T. Baker
PY - 2026
DA - 2026/09/16
PB - PsyArXiv
DO - 10.31234/osf.io/aqnmw_v2
UR - https://osf.io/preprints/psyarxiv/aqnmw_v2/
AB - Understanding how clinical symptoms fluctuate within individuals is crucial for developing personalized treatment, yet these dynamics remain understudied. We examined whether daily self-reports and speech features from audio journals predicted monthly fluctuations in clinician-rated symptoms among 54 outpatients with affective and psychotic disorders. We identified seven within-person symptom dimensions: disorganized thinking, mania, social withdrawal, inexpressivity, affective depression, cognitive depression, and anxiety. Daily self-reports and speech features were summarized on a monthly level across different temporal windows highlighting 3, 7, 14, and 28 days of daily data prior to the clinical interview to assess the predictive value of emphasizing more or less recent daily histories. Bayesian multilevel models indicated that both the intensity and day-to-day variability of daily experiences and speech features tracked month-to-month symptom fluctuations. Daily Positive Affect and speech sentiment were the most broadly informative indicators of fluctuations across symptom domains. Greater Daily Positive Affect or more positive sentiment predicted increased mania, whereas lower levels tracked increased other symptoms except for disorganized thinking. Model comparisons across temporal windows showed that models using daily self-reports were generally robust to different windows, although social withdrawal showed improved performance highlighting longer daily histories. Models using speech features generally performed better with longer temporal windows, whereas mania and anxiety showed no clear temporal window preference. Overall, these findings demonstrate the utility of daily self-reports and audio journals for dynamic symptom tracking and showcase a data-driven approach to investigating the optimal length of daily history for developing personalized monitoring tools in psychiatry.
ER -