Depression Recognition Using Daily Wearable-Derived Physiological Data

Sensors, Год журнала: 2025, Номер 25(2), С. 567 - 567

Опубликована: Янв. 19, 2025

The objective identification of depression using physiological data has emerged as a significant research focus within the field psychiatry. advancement wearable measurement devices opened new avenues for individuals with in everyday-life contexts. Compared to other methods, wearables offer potential continuous, unobtrusive monitoring, which can capture subtle changes indicative depressive states. present study leverages multimodal wristband collect from fifty-eight participants clinically diagnosed during their normal daytime activities over six hours. Data collected include pulse wave, skin conductance, and triaxial acceleration. For comparison, we also utilized matched healthy controls publicly available dataset, same equivalent durations. Our aim was identify through analysis measurements derived daily life scenarios. We extracted static features such mean, variance, skewness, kurtosis indicators like heart rate, acceleration, well autoregressive coefficients these signals reflecting temporal dynamics. Utilizing Random Forest algorithm, distinguished non-depressive varying classification accuracies on aggregated 6 h, 2 30 min, 5 min segments, 90.0%, 84.7%, 80.1%, 76.0%, respectively. results demonstrate feasibility wearable-derived recognition. achieved suggest that this approach could be integrated into clinical settings early detection monitoring symptoms. Future work will explore methods personalized interventions real-time offering promising avenue enhancing mental health care integration technology.

Язык: Английский

---

Uncovering Social States in Healthy and Clinical Populations Using Digital Phenotyping and Hidden Markov Models: Observational Study

Journal of Medical Internet Research, Год журнала: 2025, Номер 27, С. e64007 - e64007

Опубликована: Апрель 28, 2025

Background Brain-related disorders are characterized by observable behavioral symptoms, for example, social withdrawal. Smartphones can passively collect data reflecting digital activities such as communication app usage and calls. These collected objectively in real time, avoiding recall bias, may, therefore, be a useful tool measuring behaviors related to functioning. Despite promising clinical utility, analyzing smartphone is challenging datasets often include range of temporal features prone missingness. Objective Hidden Markov models (HMMs) provide interpretable, lower-dimensional representations data, allowing This study aimed investigate the HMM method modeling time series data. Methods We applied an aggregate dataset measures designed assess phone-related functioning healthy controls (HCs) participants with schizophrenia, Alzheimer disease (AD), memory complaints. trained on subset HCs (91/348, 26.1%) selected model socially active inactive states. Then, we generated hidden state sequences per participant calculated their “total dwell time,” that is, percentage spent state. Linear regression were used compare total available measures, logistic was times between diagnostic groups HCs. primarily reported results from 2-state but also verified HMMs more states whole dataset. Results identified lower AD (26/257, 10.1%) versus withheld (156/257, 60.7%; odds ratio 0.95, 95% CI 0.92-0.97; false discovery rate [FDR]–corrected P<.001), well complaints (57/257, 22.2%; 0.97, 0.96-0.99; FDR-corrected P=.004). The result group very robust across variations, whereas less robust. observed interaction when predicting (FDR-corrected P=.02). No significant relationships regarding schizophrenia (18/257, 7%; P>.99). Conclusions found practical, interpretable phenotyping analysis, providing objective phenotype possible indicator

Язык: Английский

---

Mobility-Based Smartphone Digital Phenotypes Unobtrusively Capture Everyday Cognition, Mood, and Community Life-Space in Older Adults: A Pilot Study (Preprint)

JMIR Human Factors, Год журнала: 2024, Номер 11, С. e59974 - e59974

Опубликована: Сен. 30, 2024

Background Current methods of monitoring cognition in older adults are insufficient to address the growing burden Alzheimer disease and related dementias (AD/ADRD). New approaches that sensitive, scalable, objective, reflective meaningful functional outcomes direly needed. Mobility trajectories geospatial life space patterns reflect many aspects cognitive integrity may be useful proxies age-related decline. Objective We investigated feasibility, acceptability, preliminary validity a 1-month smartphone digital phenotyping protocol infer everyday cognition, function, mood from passively obtained GPS data. also sought clarify intrinsic extrinsic factors associated with mobility phenotypes for consideration future studies. Methods Overall, 37 aged between 63 85 years healthy (n=31, 84%), mild impairment (n=5, 13%), dementia (n=1, 3%) used an open-source app (mindLAMP) unobtrusively capture 4 weeks. data were processed into interpretable features across categories activity, inactivity, routine, location diversity. Monthly average day-to-day intraindividual variability (IIV) metrics calculated each feature test priori hypotheses neuropsychological framework. Validation measures collected at baseline compared against monthly examine construct validity. Feasibility acceptability included retention, comprehension study procedures, technical difficulties, satisfaction ratings debriefing. Results All (37/37, 100%) participants completed 4-week period without major adverse events, 100% (37/37) reported explanation 97% (36/37) no feelings discomfort. Participants’ scores on consent quiz education race. Technical issues requiring troubleshooting infrequent, though 41% (15/37) battery drain. Moderate strong correlations (r≥0.3) identified validators. Specifically, individuals greater activity more diversity demonstrated better less impairment, depression, community participation, objective subjective validation measures. Contrary predictions, IIV routine habits positive outcomes. Many demographic technology-related not features; however, income, being native English speaker, season occupational status features. Conclusions Theoretically informed feasibly captured adults’ personal smartphones relate clinically including performance, decline, mood, activity. Future studies should consider impact when interpreting phenotypes. is promising method relevant risk resilience context aging AD/ADRD continue large, diverse samples.

Язык: Английский

---

Digital biomarkers in depression: A systematic review and call for standardization and harmonization of feature engineering

Journal of Affective Disorders, Год журнала: 2024, Номер 356, С. 438 - 449

Опубликована: Апрель 5, 2024

Язык: Английский

---

Predicting and Monitoring Symptoms in Diagnosed Depression Using Smartphone Data: An Observational Study (Preprint)

Journal of Medical Internet Research, Год журнала: 2024, Номер 26, С. e56874 - e56874

Опубликована: Сен. 24, 2024

Background Clinical diagnostic assessments and the outcome monitoring of patients with depression rely predominantly on interviews by professionals use self-report questionnaires. The ubiquity smartphones other personal consumer devices has prompted research into potential data collected via these to serve as digital behavioral markers for indicating presence depression. Objective This paper explores using detect monitor symptoms in diagnosed Specifically, it investigates whether this can accurately classify depression, well changes depressive states over time. Methods In a prospective cohort study, we smartphone up 1 year. study consists observations from 164 participants, including healthy controls (n=31) various disorders: major disorder (MDD; n=85), MDD comorbid borderline personality (n=27), episodes bipolar (n=21). Data were labeled based severity 9-item Patient Health Questionnaire (PHQ-9) scores. We performed statistical analysis used supervised machine learning observe state Results Our correlation revealed 32 associated state. classified who are depressed an accuracy 82% (95% CI 80%-84%) change 75% 72%-76%). Notably, most important features classifying screen-off events, battery charge levels, communication patterns, app usage, location data. Similarly, predicting state, related location, level, screen, accelerometer patterns. Conclusions supplement clinical evaluations may aid detecting particularly if combined intermittent symptoms.

Язык: Английский

---

Comparing self reported and physiological sleep quality from consumer devices to depression and neurocognitive performance

npj Digital Medicine, Год журнала: 2025, Номер 8(1)

Опубликована: Фев. 9, 2025

Abstract This study examines the relationship between self-reported and physiologically measured sleep quality their impact on neurocognitive performance in individuals with depression. Using data from 249 participants medium to severe depression monitored over 13 weeks, was assessed via retrospective self-report physiological measures consumer smartphones smartwatches. Correlations were generally weak. Machine learning models revealed that could detect all symptoms using Patient Health Questionnaire-14, whereas detected “sleeping too much” low libido. Notably, only disturbances correlated significantly performance, specifically processing speed. Physiological able changes sleep, medication use, latency. These findings emphasize are not measuring same construct, both important monitor when studying relation

Язык: Английский

---

Associations between smartphone GPS data and changes in psychological health and burden outcomes among family caregivers and patients with advanced cancer: an exploratory longitudinal cohort study

BMC Cancer, Год журнала: 2025, Номер 25(1)

Опубликована: Апрель 4, 2025

Managing advanced cancer can be psychologically distressing and burdensome for family caregivers their care recipients. Innovations in the collection modelling of passive data from personally-owned smartphones (e.g., GPS), called digital phenotyping, may afford possibility remotely monitoring detecting distress burden. We explored potential using passively-collected GPS to assess predict caregiver patient This exploratory longitudinal cohort study enrolled smartphone-owning participants with (August 2021-July 2023) recruited via an oncology clinic or self-referral through Facebook. Participants downloaded a phenotyping research app, Beiwe, that passively collected 24 weeks. completed self-report measures (PROs) anxiety depressive symptoms (Hospital Anxiety Depression Scale [HADS]), mental health (PROMIS Mental Health), burden (Montgomery-Borgatta Caregiver Burden scale) at baseline every 6 weeks After pre-processing raw into daily features time spent home, distance traveled/day), computing biweekly moving averages standard deviations, conducting principal components analysis (PCA) resulting variables, within-person regression models were used associations between changes PRO PCA scores, adjusted-R2 as measure effect size (small = 0.02, medium 0.13, large 0.26). Evaluable 48 (family 32; patients 16). smartphone explained small-to-medium variance (0.06), depression (0.15), (0.07). Patient predicted small (0.12) (0.05). Combined (0.02) (0.10) PROMIS-mental (0.36) (0.50). For outcomes, accounted (0.07); (0.24). (0.18). The demonstrates predictive utility detect psychological A larger is needed validate these findings further explore clinical application cancer.

Язык: Английский

---

Did you miss me? Making the most of digital phenotyping data by imputing missingness with point process models

medRxiv (Cold Spring Harbor Laboratory), Год журнала: 2025, Номер unknown

Опубликована: Май 13, 2025

Abstract Digital phenotyping has broad clinical potential, providing low-burden objective measures of behaviour as individuals go about their day. However, progress in making inferences from these data is severely challenged by the common occurrence missing data. We investigated non-homogeneous Poisson Point Process Models (PPPMs) a method for imputing temporal digital data, considering smartphone-based activities ‘points’. assessed inclusion time-varying covariates (‘hour’ and ‘day’) personalised PPPMs. used participants SMARD (n=26) ground truth evaluation, PRISM (n=65) Hersenonderzoek (n=283) replication analysis involving Hidden Markov Models, evaluating effectiveness PPPMs imputation influence on downstream analysis. In using one-hot encoded hour provided highest out-of-sample likelihood. successfully replicated findings our prior work this method, demonstrating that are promising tool phenotyping.

Язык: Английский

---

Detecting your depression with your smartphone? – An ethical analysis of epistemic injustice in passive self-tracking apps

Ethics and Information Technology, Год журнала: 2024, Номер 26(2)

Опубликована: Апрель 15, 2024

Abstract Smartphone apps might offer a low-threshold approach to the detection of mental health conditions, such as depression. Based on gathering ‘passive data,’ some generate user’s ‘digital phenotype,’ compare it those users with clinically confirmed depression and issue warning if depressive episode is likely. These can, thus, serve epistemic tools for affected users. From an ethical perspective, crucial consider injustice promote socially responsible innovations within digital healthcare. In cases injustice, people are wronged specifically agents, i.e., agents production distribution knowledge. We suggest that agency relies different resource- uptake-related preconditions which can be impacted by functionality passive self-tracking apps. how this lead forms (testimonial, hermeneutical, contributory injustice) analyze influence apps’ use practices individual level, in healthcare settings, structural level.

Язык: Английский

---

Predicting and Monitoring Symptoms in Diagnosed Depression Using Mobile Phone Data: An Observational Study

medRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Июнь 17, 2024

Abstract Background Clinical diagnostic assessments and outcome monitoring of patients with depression rely predominantly on interviews by professionals the use self-report questionnaires. The ubiquity smartphones other personal consumer devices has prompted research into potential data collected via these to serve as digital behavioral markers for indicating presence depression. Objective This paper explores using mobile phones detect monitor symptoms in diagnosed Methods In a prospective cohort study, we smartphone up one year. study consists observations from 99 subjects, including healthy controls (n=25) various depressive disorders: major disorder (MDD) (n=46), comorbid borderline personality (MDD|BPD) (n=16), bipolar episodes (MDE|BD) (n=12). Data were labeled based severity, 9-item Patient Health Questionnaire (PHQ-9) scores. We performed statistical analysis employed supervised machine learning classify severity observe changes state over time. Results identified 32 associated state. Our classified depressed subjects an accuracy 82% transitions 75%. Conclusions phone supplement clinical evaluations may aid detecting relapse its outcome, particularly if combined intermittent symptoms.

Язык: Английский

---