Development of a Cohesive Predictive Model for Substance Use Disorder Rehabilitation Using Passive Digital Biomarkers, Psychological Assessments, and Automated Facial Emotion Recognition (Preprint)

Опубликована: Янв. 16, 2025

Substance Use Disorder (SUD) involves excessive substance consumption and persistent reward-seeking behaviors, leading to serious physical, cognitive, social consequences. This disorder is a global health crisis tied increased mortality, unemployment, reduced quality of life. Altered brain connectivity, circadian rhythms, dopaminergic pathways contribute sleep disorders, anxiety, stress, which worsen SUD severity relapse. Factors like trauma socioeconomic disadvantages heighten risk. Digital technologies, including wearables machine learning, show promise for diagnosis, monitoring, intervention, from relapse prediction early detection comorbidities. With high rates younger patient cases, these innovations could enhance treatment outcomes SUD. Develop validate predictive model with Machine Learning the duration therapy rehabilitation or in patients SUD, using digital physiological measurements, psychological profile, automatic facial emotion recognition emotional state during craving. study will be conducted adult male at center control volunteers. Participants undergo demographic, craving assessment, also monitored smartwatch eighteen six months respectively. All participants reassessed sixth month monitoring. The collected data then used train models neural network, validated against other compared algorithms. Demographic, psychological, biomarkers profiles created, correlations analyzed, they controls, generate phenotype When achieves an adequate validity (AUC=≥0.80) graphic user interface designed clinical use. integration accessible wearables, routine recovery data, assessments by enabling personalized reducing risks. approach, leveraging affordable technology, addresses public challenges supports reintegration, particularly economically vulnerable populations. Accessible commercial smartwatches, combined psychologic, demographic learning model, may able as tools preventing

Язык: Английский

---

Large-scale digital phenotyping: Identifying depression and anxiety indicators in a general UK population with over 10,000 participants

Journal of Affective Disorders, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

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

---

Digital phenotyping for mental health based on data analytics: A systematic literature review

Artificial Intelligence in Medicine, Год журнала: 2025, Номер 163, С. 103094 - 103094

Опубликована: Март 1, 2025

Язык: Английский

---

AI-based personalized real-time risk prediction for behavioral management in psychiatric wards using multimodal data

International Journal of Medical Informatics, Год журнала: 2025, Номер unknown, С. 105870 - 105870

Опубликована: Март 1, 2025

Язык: Английский

---

Comprehensive Symptom Prediction in Acute Psychiatric Inpatients Using Wearable-Based Deep Learning Models: Development and Validation Study (Preprint)

Journal of Medical Internet Research, Год журнала: 2024, Номер 26, С. e65994 - e65994

Опубликована: Окт. 20, 2024

Background Assessing the complex and multifaceted symptoms of patients with acute psychiatric disorders proves to be significantly challenging for clinicians. Moreover, staff in wards face high work intensity risk burnout, yet research on introduction digital technologies this field remains limited. The combination continuous objective wearable sensor data acquired from deep learning techniques holds potential overcome limitations traditional assessments support clinical decision-making. Objective This study aimed develop validate wearable-based models comprehensively predict patient across various South Korea. Methods Participants diagnosed schizophrenia mood were recruited 4 3 hospitals prospectively observed using wrist-worn devices during their admission period. Trained raters conducted periodic Brief Psychiatric Rating Scale, Hamilton Anxiety Montgomery-Asberg Depression Young Mania Scale. Wearable collected patients’ heart rate, accelerometer, location data. Deep developed 2 distinct approaches: single individually (Single) multiple simultaneously via multitask (Multi). These further addressed problems: within-subject relative changes (Deterioration) between-subject absolute severity (Score). Four configurations consequently each scale: Single-Deterioration, Single-Score, Multi-Deterioration, Multi-Score. Data participants before May 1, 2024, underwent cross-validation, resulting fine-tuned then externally validated remaining participants. Results Of 244 enrolled participants, 191 (78.3%; 3954 person-days) included final analysis after applying exclusion criteria. demographic characteristics as well distribution data, showed considerable variations hospitals. 139 used while 52 external validation. Single-Deterioration Multi-Deterioration achieved similar overall accuracy values 0.75 cross-validation 0.73 Single-Score Multi-Score attained R² 0.78 0.83 0.66 0.74 validation, respectively, model demonstrating superior performance. Conclusions based effectively classified symptom deterioration predicted wards. Despite lower computational costs, Multi demonstrated equivalent or performance than Single models, suggesting that is a promising approach comprehensive prediction. However, significant wards, which presents key challenge developing decision systems Future studies may benefit recurring local validation federated address generalizability issues.

Язык: Английский

---

Detection of Symptoms of Depression Using Data From the iPhone and Apple Watch

2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Год журнала: 2023, Номер unknown, С. 1818 - 1823

Опубликована: Дек. 5, 2023

Digital health data from consumer wearable devices and smartphones have the potential to improve our understanding of mental illness. However, in conditions like depression, there is not yet a consistent uniform measurement tool whose result can be reliably used as gold standard measure depression severity. This work seeks specify what symptoms dimensions detected using vitals, activity, sleep monitored by devices. Machine learning models are fit digital detect responses individual questions surveys (self-reports) well summary scores these self-reports. For high performing models, feature importance investigated. Analysis conducted on preliminary 99 participants an ongoing study with Apple Watch iPhone along validated self-reports relevant severity, anhedonia quality. Receiver operator characteristic area under curve (ROC AUC) average precision assess model performance. The sensor investigated was found significantly five 74 measures, including overall severity specific poor appetite, aspects anhedonia, timings AUC between 0.63 0.72). features use detection vary per task suggest further areas for investigation right look at symptom.

Язык: Английский

---

Comparison of self-reported and physiological sleep quality from consumer devices to depression and neurocognitive performance

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Авг. 15, 2024

This study examines the relationship between self-reported and physiologically measured sleep quality in individuals with depression its impact on neurocognitive performance. Using data from 249 participants medium to high 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 Patient Health Questionnaire-14, whereas only detected “sleeping too much” low libido. Notably, disturbances correlated significantly Physiological able changes domains of medication use latency. These findings emphasize are not measuring same construct, both important monitor when studying relation depression.

Язык: Английский

---

Comprehensive Symptom Prediction for Acute Psychiatric Inpatients: Development and Validation of Wearable–Based Deep Learning Models (Preprint)

Опубликована: Авг. 31, 2024

Assessing complex and multifaceted symptoms of patients with acute psychiatric disorders proves significantly challenging for clinicians. Moreover, the staff in wards face high work intensity risk burnout, yet research on introduction digital technologies this field remains limited. The combination continuous objective wearable sensor data acquired from deep learning techniques holds potential to overcome limitations traditional assessments support clinical decision-making. We aimed develop validate wearable–based models comprehensively predict patient across various South Korea. Participants diagnosed schizophrenia mood were recruited four three hospitals prospectively observed using wrist-worn devices during their admission period. Trained raters conducted periodic Brief Psychiatric Rating Scale, Hamilton Anxiety Montgomery–Asberg Depression Young Mania Scale. Wearable collected patients’ heart rate, accelerometer, location data. Deep developed two distinct approaches: single individually (Single) multiple simultaneously via multitask (Multi). These further addressed problems: within-subject relative changes (Deterioration) between-subject absolute severity (Score). Four configurations consequently each scale: Single-Deterioration, Single-Score, Multi-Deterioration, Multi-Score. Data participants before May 1, 2024, underwent cross-validation, resulting fine-tuned then externally validated remaining participants. Of 244 enrolled participants, 191 (3,954 person-days) included final analysis after applying exclusion criteria. demographic characteristics as well distribution data, showed considerable variations hospitals. 139 used while 52 external validation. Single-Deterioration Multi-Deterioration achieved similar overall accuracy values 0.75 cross-validation 0.73 Single-Score Multi-Score attained R² 0.78 0.83 0.66 0.74 validation, respectively, Multi model demonstrating superior performance. based effectively classified symptom deterioration predicted wards. Despite lower computational costs, demonstrated equivalent or performance than Single models, suggesting that is a promising approach comprehensive prediction. However, significant wards, which presents key challenge developing decision systems Future studies may benefit recurring local validation federated address generalizability issues.

Язык: Английский

---

No prediction without prevention: A global qualitative study of attitudes toward using a prediction tool for risk of developing depression during adolescence

Cambridge Prisms Global Mental Health, Год журнала: 2024, Номер 11

Опубликована: Янв. 1, 2024

Abstract Given the rate of advancement in predictive psychiatry, there is a threat that it outpaces public and professional willingness for use clinical care health. Prediction tools psychiatry estimate risk future development mental health conditions. used with young populations have potential to reduce worldwide burden depression. However, little known globally about adolescents’ other stakeholders’ attitudes toward depression prediction tools. To address this, key informant interviews focus group discussions were conducted Brazil, Nepal, Nigeria United Kingdom 23 adolescents, 45 parents, 47 teachers, 48 health-care practitioners 78 stakeholders (total sample = 241) assess using calculator based on Identifying Depression Early Adolescence Risk Score. Three attributes identified an acceptable tool: should be understandable, confidential actionable. Understandability includes literacy differentiating between having condition versus condition. Confidentiality concerns are disclosing impeding educational occupational opportunities. results must also actionable through prevention services high-risk adolescents. Six recommendations provided guide research preparedness implementing

Язык: Английский

---