Predicting Symptoms of Depression and Anxiety Using Smartphone and Wearable Data

Moshe, Isaac; Terhorst, Yannik; Asare, Kennedy Opoku; Sander, Lasse; Ferreira, Denzil; Baumeister, Harald; Mohr, David C.; Pulkki-Råbäck, Laura

doi:10.3389/fpsyt.2021.625247

Cited by 142 publications

(132 citation statements)

References 82 publications

Supporting

Mentioning

106

Contrasting

Unclassified

Order By: Relevance

“…Nevertheless, both the correlation findings and the feature importance analysis in the prediction models clearly showed that participants’ phone screen (lock and unlock) behaviors, such as routinely and randomly locking and unlocking phone screen, and internet connectivity behaviors played the most important role in predicting their depression state. The findings in this study are also supported by prior research that investigated the relationship between screen interactions and mental health [ 3 , 27 , 28 , 33 ]. Passively sensed participants’ smartphone screen interaction (ie, on and off states) behavior was demonstrated to be an important predictor of mental health [ 27 ].…”

Section: Discussionsupporting

confidence: 88%

“…Instrumenting smartphones and wearables to capture in situ, fine-grained, and moment-by-moment data sets with sensing apps [ 20 - 24 ] has made it possible to passively collect data sets in naturalistic settings. Inherent in these data sets are behavioral patterns: routines, rhythms, activities, and interactions that are useful in complementing traditional depression assessment methods, in studying the mental health of individuals, and in developing timely mental health interventions [ 14 , 22 , 25 - 28 ].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Predicting Depression From Smartphone Behavioral Markers Using Machine Learning Methods, Hyperparameter Optimization, and Feature Importance Analysis: Exploratory Study

Asare¹,

Terhorst²,

Vega³

et al. 2021

JMIR Mhealth Uhealth

Self Cite

View full text Add to dashboard Cite

Background Depression is a prevalent mental health challenge. Current depression assessment methods using self-reported and clinician-administered questionnaires have limitations. Instrumenting smartphones to passively and continuously collect moment-by-moment data sets to quantify human behaviors has the potential to augment current depression assessment methods for early diagnosis, scalable, and longitudinal monitoring of depression. Objective The objective of this study was to investigate the feasibility of predicting depression with human behaviors quantified from smartphone data sets, and to identify behaviors that can influence depression. Methods Smartphone data sets and self-reported 8-item Patient Health Questionnaire (PHQ-8) depression assessments were collected from 629 participants in an exploratory longitudinal study over an average of 22.1 days (SD 17.90; range 8-86). We quantified 22 regularity, entropy, and SD behavioral markers from the smartphone data. We explored the relationship between the behavioral features and depression using correlation and bivariate linear mixed models (LMMs). We leveraged 5 supervised machine learning (ML) algorithms with hyperparameter optimization, nested cross-validation, and imbalanced data handling to predict depression. Finally, with the permutation importance method, we identified influential behavioral markers in predicting depression. Results Of the 629 participants from at least 56 countries, 69 (10.97%) were females, 546 (86.8%) were males, and 14 (2.2%) were nonbinary. Participants’ age distribution is as follows: 73/629 (11.6%) were aged between 18 and 24, 204/629 (32.4%) were aged between 25 and 34, 156/629 (24.8%) were aged between 35 and 44, 166/629 (26.4%) were aged between 45 and 64, and 30/629 (4.8%) were aged 65 years and over. Of the 1374 PHQ-8 assessments, 1143 (83.19%) responses were nondepressed scores (PHQ-8 score <10), while 231 (16.81%) were depressed scores (PHQ-8 score ≥10), as identified based on PHQ-8 cut-off. A significant positive Pearson correlation was found between screen status–normalized entropy and depression (r=0.14, P<.001). LMM demonstrates an intraclass correlation of 0.7584 and a significant positive association between screen status–normalized entropy and depression (β=.48, P=.03). The best ML algorithms achieved the following metrics: precision, 85.55%-92.51%; recall, 92.19%-95.56%; F1, 88.73%-94.00%; area under the curve receiver operating characteristic, 94.69%-99.06%; Cohen κ, 86.61%-92.90%; and accuracy, 96.44%-98.14%. Including age group and gender as predictors improved the ML performances. Screen and internet connectivity features were the most influential in predicting depression. Conclusions Our findings demonstrate that behavioral markers indicative of depression can be unobtrusively identified from smartphone sensors’ data. Traditional assessment of depression can be augmented with behavioral markers from smartphones for depression diagnosis and monitoring.

show abstract

Section: Discussionsupporting

confidence: 88%

Section: Introductionmentioning

confidence: 99%

Predicting Depression From Smartphone Behavioral Markers Using Machine Learning Methods, Hyperparameter Optimization, and Feature Importance Analysis: Exploratory Study

Asare¹,

Terhorst²,

Vega³

et al. 2021

JMIR Mhealth Uhealth

Self Cite

View full text Add to dashboard Cite

show abstract

“…Previous studies have documented the potential for using EWDs in continuous patient symptom monitoring, longitudinal data collection, supportive self-care, and early detection of relapse among people with mental illness ( 16 ). However, the role of EWDs in stimulating physical activity in those with mental illness remains to be proven.…”

Section: Discussionmentioning

confidence: 99%

Exploring the Association Between Electronic Wearable Device Use and Levels of Physical Activity Among Individuals With Depression and Anxiety: A Population Level Study

Onyeaka¹,

Firth²,

Enemuo³

et al. 2021

Front. Digit. Health

View full text Add to dashboard Cite

Aim: The present study aimed to investigate the cross-sectional association between self-reported use of electronic wearable devices (EWDs) and the levels of physical activity among a representative sample of adults with depression and anxiety in the United States.Methods: For this cross-sectional study, data were pooled from the Health Information National Trends Survey 2019. A sample of 1,139 adults with self-reported depression and anxiety (60.9% women; mean age of 52.5 years) was analyzed. The levels of physical activity and prevalence of EWD utilization were self-reported. The chi-square tests were used to compare individual characteristics through the use of EWDs. Multivariable logistic regression was employed to investigate the association between EWDs and physical activity levels while adjusting for sociodemographic and health-related factors.Results: From the 1,139 adults with self-reported depression and anxiety, 261 (weighted percentage 28.1%) endorsed using EWD in the last year. After adjusting for covariates, the use of EWDs was only significantly associated with a higher odds of reporting intention to lose weight (OR 2.12; 95% CI 1.04, 4.35; p = 0.04). We found no association between the use of EWDs and meeting the national weekly recommendation for physical activity or resistance/strength exercise training.Conclusion: About three in 10 adults suffering from depression and anxiety in the United States reported using EWDs in the last year. The current study findings indicate that among people living with mental illness, EWD use is associated with higher odds of weight loss intent suggesting that EWDs may serve as an opening for the clinical interactions around physical health through identifying patients primed for behavior change. Further large-scale studies using randomized trial designs are needed to examine the causal relationships between EWDs and the physical activity of people with mental health conditions.

show abstract

“…Sleep latency refers to the amount of time it takes to fall asleep, whereas sleep efficiency measures the percentage of time spent asleep while in bed [41,42]. The effect of TRE on sleep latency and efficiency was evaluated in four trials [22,23,26,27].…”

Section: Sleep Latency and Sleep Efficiencymentioning

confidence: 99%

Intermittent Fasting and Sleep: A Review of Human Trials

et al. 2021

View full text Add to dashboard Cite

This review examines the effects of two popular intermittent fasting regimens on sleep in adults with overweight and obesity. Specifically, the effects of time restricted eating (TRE; eating all food within a 4–10 h window) and alternate day fasting (ADF; 600 kcal fast day alternated with ad libitum feast day) on sleep quality, sleep duration, sleep latency, sleep efficiency, insomnia severity, and risk of obstructive sleep apnea, will be summarized. The role of weight loss will also be discussed. Results from our review reveal that the majority of these trials produced weight loss in the range of 1–6% from baseline. Sleep quality and sleep duration remained unaltered with TRE and ADF, as assessed by the Pittsburgh Sleep Quality Index (PSQI). The effects of intermittent fasting on sleep latency and sleep efficiency are mixed, with one study showing worsening of these parameters, and others showing no effect. Insomnia severity and the risk of obstructive sleep apnea remained unchanged in the trials assessing these metrics. Taken together, these preliminary findings suggest that TRE and ADF produce mild to moderate weight loss (1–6%) but their effects on sleep remain unclear. Solid conclusions are difficult to establish since participants in the studies had healthy sleep durations and no clinical insomnia at baseline, leaving little room for improvement in these metrics. Moreover, none of the trials were adequately powered to detect statistically significant changes in any measure of sleep. Future well-powered trials, conducted in individuals with diagnosed sleep disturbances, will be necessary to elucidate the effect of these popular diets on sleep.

show abstract

Predicting Symptoms of Depression and Anxiety Using Smartphone and Wearable Data

Cited by 142 publications

References 82 publications

Predicting Depression From Smartphone Behavioral Markers Using Machine Learning Methods, Hyperparameter Optimization, and Feature Importance Analysis: Exploratory Study

Predicting Depression From Smartphone Behavioral Markers Using Machine Learning Methods, Hyperparameter Optimization, and Feature Importance Analysis: Exploratory Study

Exploring the Association Between Electronic Wearable Device Use and Levels of Physical Activity Among Individuals With Depression and Anxiety: A Population Level Study

Intermittent Fasting and Sleep: A Review of Human Trials

Contact Info

Product

Resources

About