Abstract
BACKGROUND: Recurrent course and disruption of circadian rhythms are among the core features of bipolar disorder (BD). Thus, ongoing symptom monitoring is an essential part of good clinical management. OBJECTIVE: We conducted a study to validate the English version of the ASERT (Aktibipo questionnaire), a tool for self-assessment of mood symptoms. We also analyzed the relationship of self-assessed symptoms with clinician ratings and actigraphy measures, and investigated the possibility of predicting depressive episodes using subjective and digital measures. METHODS: This was a longitudinal study of twenty individuals with BD, followed for up to 11 months. The participants completed weekly mood self-assessments (ASERT) using a smartphone app and wore wrist actigraphs. During monthly appointments, the severity of their mood symptoms was rated by clinicians, and the participants completed questionnaires addressing overall functioning (FAST), and biological rhythms (BRIAN). RESULTS: The study confirmed the validity and reliability of the ASERT as a measure of subjective mood. Additionally, we found significant associations between ASERT responses, clinical scales, and actigraphy data (ASERT_dep vs. MADRS β = 1.42, p < 0.001, ASERT_man vs. YMRS β = 0.38, p < 0.001, mixed-effect model). In our analysis, a combination of self-assessment and actigraphy data detected depression relapse with 67% sensitivity, 90% specificity, 81% balanced accuracy, and AUC of 0.80. Furthermore, we observed a strong correlation between the actigraphy-derived interdaily stability and BRIAN scores (β=-3.86, p = 0.005) overall functioning, emphasizing the significance of circadian rhythm disruptions in BD. CONCLUSION: This study highlights the potential of digital tools, such as digitally administered self-assessments and actigraphy, to enhance the management of BD by providing valuable insights into mood states and detecting relapse. Further research is needed to refine and optimize these tools for widespread clinical application, such as informing personalized treatment plans.