Digital phenotyping in bipolar disorder: Using longitudinal Fitbit data and personalized machine learning to predict mood symptomatology DOI Open Access
Jessica M. Lipschitz, Shuwen Lin, Soroush Saghafian

et al.

Acta Psychiatrica Scandinavica, Journal Year: 2024, Volume and Issue: 151(3), P. 434 - 447

Published: Oct. 13, 2024

Effective treatment of bipolar disorder (BD) requires prompt response to mood episodes. Preliminary studies suggest that predictions based on passive sensor data from personal digital devices can accurately detect episodes (e.g., between routine care appointments), but date do not use methods designed for broad application. This study evaluated whether a novel, personalized machine learning approach, trained entirely Fitbit data, with limited filtering could symptomatology in BD patients. We analyzed 54 adults BD, who wore Fitbits and completed bi-weekly self-report measures 9 months. applied (ML) models aggregated over two-week observation windows occurrences depressive (hypo)manic symptomatology, which were defined as scores above established clinical cutoffs the Patient Health Questionnaire-8 (PHQ-8) Altman Self-Rating Mania Scale (ASRM) respectively. As hypothesized, among several ML algorithms, Binary Mixed Model (BiMM) forest achieved highest area under receiver operating curve (ROC-AUC) validation process. In testing set, ROC-AUC was 86.0% depression 85.2% (hypo)mania. Using optimized thresholds calculated Youden's J statistic, predictive accuracy 80.1% (sensitivity 71.2% specificity 85.6%) 89.1% (hypo)mania 80.0% 90.1%). sound performance detecting patients using Findings expand upon evidence produce accurate predictions. Additionally, best our knowledge, this represents first application BiMM prediction. Overall, results move field step toward algorithms suitable full population patients, rather than only those high compliance, access specialized devices, or willingness share invasive data.

Language: Английский

A translationally informed approach to vital signs for psychiatry: a preliminary proof of concept DOI Creative Commons
Meredith L. Wallace, Ellen Frank, Colleen A. McClung

et al.

NPP—Digital Psychiatry and Neuroscience, Journal Year: 2024, Volume and Issue: 2(1)

Published: Aug. 26, 2024

Abstract The nature of data obtainable from the commercial smartphone – bolstered by a translational model emphasizing impact social and physical zeitgebers on circadian rhythms mood offers possibility scalable objective vital signs for major depression. Our was to explore associations between passively sensed behavioral repeatedly measured depressive symptoms suggest which features could eventually lead towards We collected continuous bi-weekly (PHQ-8) 131 psychiatric outpatients with lifetime DSM-5 diagnosis depression and/or anxiety over 16-week period. Using linear mixed-effects models, we related concurrent summary (mean variability sleep, activity, engagement metrics), considering both between- within-person associations. Individuals more variable wake-up times across study reported higher relative individuals less (B [95% CI] = 1.53 [0.13, 2.93]). On given week, having lower step count (−0.16 [−0.32, −0.01]), slower walking rate (−1.46 [−2.60, −0.32]), normalized location entropy (−3.01 [−5.51, −0.52]), time at home (0.05 [0.00, 0.10]), distances traveled (−0.97 [−1.72, −0.22]), one’s own typical levels, were each associated symptoms. With replication in larger samples clear understanding how these components are best combined, composite measure potentially offer kinds medicine that have proven invaluable assessment decision-making medicine. Clinical Trials Registration: form basis this report as part clinical trial number NCT03152864.

Language: Английский

Citations

0
Using Wearable Device and Artificial Intelligence to Predict Mood Symptoms in Bipolar Disorder (Preprint) DOI Creative Commons
Yi‐Ling Chien

Published: Aug. 28, 2024

UNSTRUCTURED Bipolar disorder (BD) is a highly recurrent disorder. Early detection, early intervention, and prevention of bipolar mood symptoms are key for better prognosis. In this study, we build prediction models with machine learning algorithms. This study recruited 24 participants BD. The Beck Depression Inventory (BDI) Young Mania Rating Scale (YMRS) were used to evaluate depressive manic episodes respectively. Using digital biomarkers collected from wearable devices as input, six algorithms (Logistic Regression, Decision Tree, K-Nearest Neighbors, Random Forest, Adaptive Boosting, Extreme Gradient Boosting) predictive models. model achieved 83% accuracy, 0.89 Area Under the Receiver Operating Characteristic curve (AUROC), 0.65 F1 score on testing data. 91% 0.88 AUROC, 0.25 With interpretable Shapely Additive exPlanations (SHAP), found that relatively high resting heart rate, low activity, lack sleep may predict symptoms. demonstrated could be Moreover, based findings model, provide clinical assessment treatment earlier prevent recurrence.

Language: Английский

Citations

0
Using Wearable Device and Artificial Intelligence to Predict Mood Symptoms in Bipolar Disorder (Preprint) DOI
Chia‐Tung Wu,

Lian-Yin Jhao,

Ding-Shan Liu

et al.

Published: Sept. 9, 2024

UNSTRUCTURED Bipolar disorder (BD) is a highly recurrent disorder. Early detection, early intervention, and prevention of bipolar mood symptoms are key for better prognosis. In this study, we build prediction models with machine learning algorithms. This study recruited 24 participants BD. The Beck Depression Inventory (BDI) Young Mania Rating Scale (YMRS) were used to evaluate depressive manic episodes respectively. Using digital biomarkers collected from wearable devices as input, six algorithms (Logistic Regression, Decision Tree, K-Nearest Neighbors, Random Forest, Adaptive Boosting, Extreme Gradient Boosting) predictive models. model achieved 83% accuracy, 0.89 Area Under the Receiver Operating Characteristic curve (AUROC), 0.65 F1 score on testing data. 91% 0.88 AUROC, 0.25 With interpretable Shapely Additive exPlanations (SHAP), found that relatively high resting heart rate, low activity, lack sleep may predict symptoms. demonstrated could be Moreover, based findings model, provide clinical assessment treatment earlier prevent recurrence.

Language: Английский

Citations

0
Adolescent depression treatment pathways in Primary Care – protocol for a longitudinal Cohort Study Describing Naturalistic Flow of Treatment and Evaluating Effectiveness and Cost- effectiveness of Interpersonal Counseling Compared to Treatment as Usual DOI
Outi Mantere, Aija Myllyniemi, Emma Salusjärvi

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 23, 2024

Abstract Background:Implementation of evidence-based interventions is one the proposed responses to increased demand for treatment adolescent depression. While efficacy interpersonal psychotherapy treat depression adolescents (IPT-A) well established, effectiveness and cost-effectiveness shorter counseling (IPC-A) remains open. Objective: We present a protocol prospective evaluation naturalistic flow with sustained depression, IPC-A, as compared usual or no Methods: will collect cohort grade 7 9 (13–16-year-olds) in selected Finnish schools using convenience sampling (n=9000). We compare three groups defined at 6 months (targeting n=100; (TAU), n=200; treatment, n=100). The primary outcome measure will be proportion who received specialized psychiatric services by 2 years after baseline. Secondary measures include longitudinal changes PHQ-9-A scores 12 months, positive mental health, social inclusion, quality life. Cost-effectiveness evaluated survey data an economic evaluation register information on service use before up 10 A universal all adolescents, independent mood, provide description a) sustained depression over follow-up period (Patient Health Questionnaire items, version, ≥ two measurements months), b) self-reported need motivation support, c) therapeutic intervention, d) benefits harms treatment. describe treatment received predictors based reports from caretakers, therapists, electronic patient records. Impact training IPC-A competence access evaluated. Conclusions: The study will describe for, pathways to, content health depressed adolescents. The results can improve detection equal care, inform decision -makers about best practices prevention, including utility implementation IPC-A. Trial registration: ClinicalTrials.com NCT06390462 registered 2024-03-19

Language: Английский

Citations

0
Digital phenotyping in bipolar disorder: Using longitudinal Fitbit data and personalized machine learning to predict mood symptomatology DOI Open Access
Jessica M. Lipschitz, Shuwen Lin, Soroush Saghafian

et al.

Acta Psychiatrica Scandinavica, Journal Year: 2024, Volume and Issue: 151(3), P. 434 - 447

Published: Oct. 13, 2024

Effective treatment of bipolar disorder (BD) requires prompt response to mood episodes. Preliminary studies suggest that predictions based on passive sensor data from personal digital devices can accurately detect episodes (e.g., between routine care appointments), but date do not use methods designed for broad application. This study evaluated whether a novel, personalized machine learning approach, trained entirely Fitbit data, with limited filtering could symptomatology in BD patients. We analyzed 54 adults BD, who wore Fitbits and completed bi-weekly self-report measures 9 months. applied (ML) models aggregated over two-week observation windows occurrences depressive (hypo)manic symptomatology, which were defined as scores above established clinical cutoffs the Patient Health Questionnaire-8 (PHQ-8) Altman Self-Rating Mania Scale (ASRM) respectively. As hypothesized, among several ML algorithms, Binary Mixed Model (BiMM) forest achieved highest area under receiver operating curve (ROC-AUC) validation process. In testing set, ROC-AUC was 86.0% depression 85.2% (hypo)mania. Using optimized thresholds calculated Youden's J statistic, predictive accuracy 80.1% (sensitivity 71.2% specificity 85.6%) 89.1% (hypo)mania 80.0% 90.1%). sound performance detecting patients using Findings expand upon evidence produce accurate predictions. Additionally, best our knowledge, this represents first application BiMM prediction. Overall, results move field step toward algorithms suitable full population patients, rather than only those high compliance, access specialized devices, or willingness share invasive data.

Language: Английский

Citations

0