No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Wearable Wellness: Depression Screening via Fitbit Data Collected During COVID-19 Pandemic
ResearchGate has not been able to resolve any citations for this publication.
Background:
Disturbed heart dynamics in depression seriously increases mortality risk. Heart rate variability (HRV) is a rich source of information for studying this dynamics. This paper is a meta-analytic review with methodological commentary of the application of nonlinear analysis of HRV and its possibility to address cardiovascular diseases in depression.
Objective:
This paper aimed to appeal for the introduction of cardiological screening to patients with depression, because it is still far from established practice. The other (main) objective of the paper was to show that nonlinear methods in HRV analysis give better results than standard ones.
Methods:
We systematically searched on the web for papers on nonlinear analyses of HRV in depression, in line with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 framework recommendations. We scrutinized the chosen publications and performed random-effects meta-analysis, using the esci module in jamovi software where standardized effect sizes (ESs) are corrected to yield the proof of the practical utility of their results.
Results:
In all, 26 publications on the connection of nonlinear HRV measures and depression meeting our inclusion criteria were selected, examining a total of 1537 patients diagnosed with depression and 1041 healthy controls (N=2578). The overall ES (unbiased) was 1.03 (95% CI 0.703-1.35; diamond ratio 3.60). We performed 3 more meta-analytic comparisons, demonstrating the overall effectiveness of 3 groups of nonlinear analysis: detrended fluctuation analysis (overall ES 0.364, 95% CI 0.237-0.491), entropy-based measures (overall ES 1.05, 95% CI 0.572-1.52), and all other nonlinear measures (overall ES 0.702, 95% CI 0.422-0.982). The effectiveness of the applied methods of electrocardiogram analysis was compared and discussed in the light of detection and prevention of depression-related cardiovascular risk.
Conclusions:
We compared the ESs of nonlinear and conventional time and spectral methods (found in the literature) and demonstrated that those of the former are larger, which recommends their use for the early screening of cardiovascular abnormalities in patients with depression to prevent possible deleterious events.
Background:
Digital data on physical activity are useful for self-monitoring and preventing depression and anxiety. Although previous studies have reported machine or deep learning models that use physical activity for passive monitoring of depression and anxiety, there are no models for workers. The working population has different physical activity patterns from other populations, which is based on commuting, holiday patterns, physical demands, occupations, and industries. These working conditions are useful in optimizing the model used in predicting depression and anxiety. Further, recurrent neural networks increase predictive accuracy by using previous inputs on physical activity, depression, and anxiety.
Objective:
This study evaluated the performance of a deep learning model optimized for predicting depression and anxiety in workers. Psychological distress was considered a depression and anxiety indicator.
Methods:
A 2-week longitudinal study was conducted with workers in urban areas in Japan. Absent workers were excluded. In a daily survey, psychological distress was measured using a self-reported questionnaire. As features, activity time by intensity was determined using the Google Fit application. Additionally, we measured age, gender, occupations, employment status, work shift types, working hours, and whether the response date was a working or nonworking day. A deep learning model, using long short-term memory, was developed and validated to predict psychological distress the next day, using features of the previous day. Further, a 5-fold cross-validation method was used to evaluate the performance of the aforementioned model. As the primary indicator of performance, classification accuracy for the severity of the psychological distress (light, subthreshold, and severe) was considered.
Results:
A total of 1661 days of supervised data were obtained from 236 workers, who were aged between 20 and 69 years. The overall classification accuracy for psychological distress was 76.3% (SD 0.04%). The classification accuracy for severe-, subthreshold-, and light-level psychological distress was 51.1% (SD 0.05%), 60.6% (SD 0.05%), and 81.6% (SD 0.04%), respectively. The model predicted a light-level psychological distress the next day after the participants had been involved in 3 peaks of activity (in the morning, noon, and evening) on the previous day. Lower activity levels were predicted as subthreshold- and severe-level psychological distress. Different predictive results were observed on the basis of occupations and whether the previous day was a working or nonworking day.
Conclusions:
The developed deep learning model showed a similar performance as in previous studies and, in particular, high accuracy for light-level psychological distress. Working conditions and long short-term memory were useful in maintaining the model performance for monitoring depression and anxiety, using digitally recorded physical activity in workers. The developed model can be implemented in mobile apps and may further be practically used by workers to self-monitor and maintain their mental health state.
The use of digital tools to measure physiological and behavioural variables of potential relevance to mental health is a growing field sitting at the intersection between computer science, engineering, and clinical science. We summarised the literature on remote measuring technologies, mapping methodological challenges and threats to reproducibility, and identified leading digital signals for depression. Medical and computer science databases were searched between January 2007 and November 2019. Published studies linking depression and objective behavioural data obtained from smartphone and wearable device sensors in adults with unipolar depression and healthy subjects were included. A descriptive approach was taken to synthesise study methodologies. We included 51 studies and found threats to reproducibility and transparency arising from failure to provide comprehensive descriptions of recruitment strategies, sample information, feature construction and the determination and handling of missing data. The literature is characterised by small sample sizes, short follow-up duration and great variability in the quality of reporting, limiting the interpretability of pooled results. Bivariate analyses show consistency in statistically significant associations between depression and digital features from sleep, physical activity, location, and phone use data. Machine learning models found the predictive value of aggregated features. Given the pitfalls in the combined literature, these results should be taken purely as a starting point for hypothesis generation. Since this research is ultimately aimed at informing clinical practice, we recommend improvements in reporting standards including consideration of generalisability and reproducibility, such as wider diversity of samples, thorough reporting methodology and the reporting of potential bias in studies with numerous features.
Background
The COVID-19 pandemic and its consequences have been associated with an increase in poor population mental health. We assessed how depressive symptoms changed among U.S. adults over the course of the COVID-19 pandemic and identified the key risk factors for these symptoms.
Methods
Longitudinal panel study of a nationally representative group of U.S. adults ages 18 years and older surveyed in March-April 2020 (Time 1; N=1441) and March-April 2021 (Time 2; N=1161) in the COVID-19 and Life Stressors Impact on Mental Health and Well-being study (CLIMB). The Patient Health Questionnaire-9 (PHQ-9) was used to define elevated depressive symptoms (cut-off ≥10) and depressive symptoms score (0-27).
Findings
The prevalence of elevated depressive symptoms persisted from 27.8% in 2020 (95% CI: 24.9, 30.9) to 32.8% in 2021 (95% CI: 29.1, 36.8). Over time, the central drivers of depressive symptoms were low household income, not being married, and experiencing multiple stressors during the COVID-19 pandemic. The odds ratio of elevated depressive symptoms for low income relative to high income persons increased from 2.3 (95% CI: 1.2, 4.2) in 2020 to 7.0 (95% CI: 3.7, 13.3) in 2021. Fewer people reported experiencing 4 or more COVID-19 stressors in 2021 than in 2020 (47.5% in 2020 vs 37.1% in 2021), but the odds ratio of elevated depressive symptoms associated with 4 or more stressors relative to 1 stressor or less increased from 1.9 (95% CI: 1.2, 3.1) in 2020 to 5.4 (95% CI: 3.2, 9.2) in 2021.
Interpretation
The burden of depressive symptoms in the U.S. adult population increased over the course of the COVID-19 pandemic. Mental health gaps grew between populations with different assets and stressor experiences during the COVID-19 pandemic.
Funding
CLIMB Time 1 was sponsored by the Rockefeller Foundation-Boston University 3-D Commission. CLIMB Time 2 was sponsored by the de Beaumont Foundation.
The COVID-19 pandemic has impacted the lives of the worldwide population. Citizens suffer the social, economic, physiological, and psychological effects of this pandemic. Primary sources, scientific articles, and secondary bibliographic indexes, databases, and web pages were used for a consensus critical review. The method was a narrative review of the available literature to summarize the existing literature addressing mental health concerns and stressors related to the COVID-19 pandemic. The main search engines used in the present research were PubMed, SciELO, and Google Scholar. We found the pandemic has had a direct impact on psychopathologies such as anxiety, increasing its ratios, and depression. Other syndromes such as burnout and post-traumatic stress disorder have increased with the pandemic, showing a larger incidence among medical personnel. Moreover, eating disorders and violence have also increased. Public authorities must prepare healthcare systems for increasing incidences of mental pathologies. Mental health apps are one of the tools that can be used to reach the general population.
Emotional expressions are the behaviors that communicate our emotional state or attitude to others. They are expressed through verbal and non-verbal communication. Complex human behavior can be understood by studying physical features from multiple modalities; mainly facial, vocal and physical gestures. Recently, spontaneous multi-modal emotion recognition has been extensively studied for human behavior analysis. In this paper, we propose a new deep learning-based approach for audio-visual emotion recognition. Our approach leverages recent advances in deep learning like knowledge distillation and high-performing deep architectures. The deep feature representations of the audio and visual modalities are fused based on a model-level fusion strategy. A recurrent neural network is then used to capture the temporal dynamics. Our proposed approach substantially outperforms state-of-the-art approaches in predicting valence on the RECOLA dataset. Moreover, our proposed visual facial expression feature extraction network outperforms state-of-the-art results on the AffectNet and Google Facial Expression Comparison datasets.
Background:
The effects of endometrial cavity fluid (ECF) on in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) pregnancy outcomes following embryo transfer (ET) are still controversial. We conducted the present study to investigate whether the presence of ECF in infertile patients scheduled to undergo IVF or ICSI was associated with pregnancy outcomes.
Methods:
A retrospective cohort study design was used. Among infertile patients undergoing IVF/ICSI, those with and without ECF were matched 1:1 using propensity score matching (PSM). After ensuring that the baseline levels of the two matched groups were consistent, the pregnancy and obstetrical outcomes of the two groups were compared.
Results:
Patients with ECF had significantly lower clinical rates of pregnancy (1,061/1,862, 57% vs. 1,182/1,862, 63.5%; P<0.001), live birth (902/1,862, 48.4% vs. 1,033/1,862, 55.5%; P<0.001), biochemical pregnancy (1,182/1,862, 63.5% vs. 1,288/1,862, 69.2%; P<0.001), and embryo implantation (1,500/3,740, 40.1% vs. 1,661/3,740, 44.4%, P<0.001) than patients without ECF. Also, patients with ECF had a higher incidence of gestational diabetes (17/78, 22% vs. 8/94, 9%, P=0.014). However, there were no differences in gestational weeks at delivery or birth weight between the two groups.
Conclusions:
ECF was significantly associated with adverse pregnancy outcomes but showed no significant association with adverse obstetric outcomes (except for gestational diabetes).
Prior research has recently shown that passively collected sensor data collected within the contexts of persons daily lives via smartphones and wearable sensors can distinguish those with major depressive disorder (MDD) from controls, predict MDD severity, and predict changes in MDD severity across days and weeks. Nevertheless, very little research has examined predicting depressed mood within a day, which is essential given the large amount of variation occurring within days. The current study utilized passively collected sensor data collected from a smartphone application to future depressed mood from hour-to-hour in an ecological momentary assessment study in a sample reporting clinical levels of depression (N = 31). Using a combination of nomothetic and idiographically weighted machine learning models, the results suggest that depressed mood can be accurately from hour to hour with an average correlation between out of sample predicted depressed mood levels and observed depressed mood of 0.587, CI [0.552, 0.621]. This suggests that passively collected smartphone data can accurately predict future depressed mood among a sample reporting clinical levels of depression. If replicated in other samples, this modeling framework may allow just-in-time adaptive interventions to treat depression as it changes in the context of daily life.
Time series feature extraction is one of the preliminary steps of conventional machine learning pipelines. Quite often, this process ends being a time consuming and complex task as data scientists must consider a combination between a multitude of domain knowledge factors and coding implementation. We present in this paper a Python package entitled Time Series Feature Extraction Library (TSFEL), which computes over 60 different features extracted across temporal, statistical and spectral domains. User customisation is achieved using either an online interface or a conventional Python package for more flexibility and integration into real deployment scenarios. TSFEL is designed to support the process of fast exploratory data analysis and feature extraction on time series with computational cost evaluation.
Depression is a common mood disorder that causes severe medical problems and interferes negatively with daily life. Identifying human behavior patterns that are predictive or indicative of depressive disorder is important. Clinical diagnosis of depression relies on costly clinician assessment using survey instruments which may not objectively reflect the fluctuation of daily behavior. Self-administered surveys, such as the Quick Inventory of Depressive Symptomatology (QIDS) commonly used to monitor depression, may show disparities from clinical decision. Smartphones provide easy access to many behavioral parameters, and Fitbit wrist bands are becoming another important tool to assess variables such as heart rates and sleep efficiency that are complementary to smartphone sensors. However, data used to identify depression indicators have been limited to a single platform either iPhone, or Android, or Fitbit alone due to the variation in their methods of data collection. The present work represents a large-scale effort to collect and integrate data from mobile phones, wearable devices, and self reports in depression analysis by designing a new machine learning approach. This approach constructs sparse mappings from sensing variables collected by various tools to two separate targets: self-reported QIDS scores and clinical assessment of depression severity. We propose a so-called heterogeneous multi-task feature learning method that jointly builds inference models for related tasks but of different types including classification and regression tasks. The proposed method was evaluated using data collected from 103 college students and could predict the QIDS score with an R2 reaching 0.44 and depression severity with an F1-score as high as 0.77. By imposing appropriate regularizers, our approach identified strong depression indicators such as time staying at home and total time asleep.
Objectives:
The Center for Epidemiologic Studies Depression (CESD) scale has been useful in a broad spectrum of health research on patient and population outcomes. A brief version is used when depressive symptoms are not the primary focus. Rasch (item response) analysis previously demonstrated potential problems with positively worded items. We tested the 10-item CESD (CESD-10) scale and considered an 8-item version with both psychometric and Rasch analyses.
Methods:
This was a special sample of 2067 caregivers from three existing US databases. We describe item response patterns and internal constancy in addition to Rasch scale results.
Results:
There were few problems with missing data, and internal consistency was high (alpha = 0.86–0.88) for both CESD versions. Rasch analysis indicated that one of the positive items (“hopeful about future”) could be dropped.
Conclusions:
We partly confirmed prior work that suggested dropping positive items for the CESD-10. Among caregivers, item-level problems and scaling problems seem minimal. At present, there is not a strong rationale for dropping the CESD-10 positive items: the one poorly performing positive item might be explained by the special caregiver sample.
Resting heart rate is a determinant of cardiac output and physiological homeostasis. Although a simple, but critical, parameter, this vital sign predicts adverse outcomes, including mortality, and development of diseases in otherwise normal and healthy individuals. Temporal changes in heart rate can have valuable predictive capabilities. Heart rate can reflect disease severity in patients with various medical conditions. While heart rate represents a compilation of physiological inputs, including sympathetic and parasympathetic tone, aside from the underlying intrinsic sinus rate, how resting heart rate affects outcomes is uncertain. Mechanisms relating resting heart rate to outcomes may be disease-dependent but why resting heart rate in otherwise healthy, normal individuals affects outcomes remains obscure. For specific conditions, physiologically appropriate heart rate reductions may improve outcomes. However, to date, in the normal population, evidence that interventions aimed at reducing heart rate improves outcomes remains undefined. Emerging data suggest that reduction in heart rate via vagal activation and/or sympathetic inhibition is propitious.
Background
A considerable gap in knowledge exists around mental health trends in diverse racial and ethnic adolescent and young adult populations. The purpose of this study is to examine annual trends for mental health and help-seeking by race/ethnicity in a national sample of college students.
Methods
Survey data come from >350,000 students at 373 campuses that participated in the Healthy Minds Study between 2013 and 2021. Analyses are descriptive in nature focusing on year-by-year prevalence and help-seeking rates for each racial/ethnic group.
Results
In 2020–2021, >60% of students met criteria for one or more mental health problems, a nearly 50% increase from 2013. Mental health worsened among all groups over the study period. American Indian/Alaskan Native students experienced the largest increases in depression, anxiety, suicidal ideation, and meeting criteria for one or more mental health problem. Students of color had the lowest rates of mental health service utilization. The highest annual rate of past-year treatment for Asian, Black, and Latinx students was at or below the lowest rate for White students. Although Arab American students experienced a 22% increase in prevalence, there was an 18% decrease in treatment.
Limitations
Response rates raise the potential of nonresponse bias. Sample weights adjust along known characteristics, but there may be differences on unobserved characteristics.
Conclusions
Findings have important implications for campus mental health programming and underscore the urgency of reducing mental health inequalities in college student populations through the identification and implementation of best practices both in clinical settings and through system-level change.
Principal Component Analysis (PCA) is a multivariate exploratory analysis method, useful to separate systematic variation from noise. It allows to define a space of reduced dimensions that preserves the relevant information of the original data and allows visualization of objects (scores) and variables (loadings). PCA requires multivariate data, meaning many variables measured on many objects. Data, vectors and matrices are defined and a short summary of necessary linear algebra is given. Purely mathematical almost identical definitions of PCA and Singular Value Decomposition (SVD) are shown, but in chemometrics, PCA always has a residual and a number of meaningful components, the rank. This leads to a discussion of numerical and visual diagnostics for finding the rank and checking the residual. The visualization of scores and loadings is introduced by means of two small examples. Data preprocessing is also given consideration.
This paper presents a novel and effective audio based method on depression classification. It focuses on two important issues, \emph{i.e.} data representation and sample imbalance, which are not well addressed in literature. For the former one, in contrast to traditional shallow hand-crafted features, we propose a deep model, namely DepAudioNet, to encode the depression related characteristics in the vocal channel, combining Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) to deliver a more comprehensive audio representation. For the latter one, we introduce a random sampling strategy in the model training phase to balance the positive and negative samples, which largely alleviates the bias caused by uneven sample distribution. Evaluations are carried out on the DAIC-WOZ dataset for the Depression Classification Sub-challenge (DCC) at the 2016 Audio-Visual Emotion Challenge (AVEC), and the experimental results achieved clearly demonstrate the effectiveness of the proposed approach.
scikit-learn is an open source machine learning library written in Python.
Background
There is a need for brief instruments to screen for depression in adolescents that are valid, reliable and freely available. The aim of this study was to investigate the psychometric properties of a 10-item version of the CESD-R (CESDR-10) in two national adolescent samples.
Methods
Sample 1 consisted of N=3777 youths (mean age 15.7) and Sample 2 contained N=1150 adolescents (mean age 14.5). We performed confirmatory factor analysis, evaluated construct validity, examined differential item functioning, and assessed internal consistency reliability (α).
Results
The results suggest generally strong psychometric properties for the CESDR-10. The CFA 1-factor model showed good model fit. Construct validity was partially supported in Sample 1 and mostly supported for Sample 2 based upon the characteristics examined. The CESDR-10 showed configural and metric invariance across both samples and full measurement invariance across sex. There were no notable differences in discrimination parameters or clinically significant differential item functioning between samples or sexes.
Limitations
Criterion related validity was not assessed in this study. Further studies should evaluate the scale in comparison to a psychiatric diagnosis. In addition, this study utilized a web-based format of administration which may influence participants׳ answers. In future studies, the CESDR-10 should be administered in other settings to more thoroughly establish its generalizability.
Conclusion
In clinical and non-clinical settings alike, time pressures make the availability of brief but valid screening measures critical. Findings support future use of the CESDR-10.
Few older adults in the United States achieve the minimum recommended amount of physical activity. Lack of physical activity contributes to many chronic diseases that occur in older adults, including heart disease, stroke, diabetes mellitus, lung disease, Alzheimer disease, hypertension, and cancer. Lack of physical activity, combined with poor dietary habits, has also contributed to increased obesity in older persons. Regular exercise and increased aerobic fitness are associated with a decrease in all-cause mortality and morbidity, and are proven to reduce disease and disability, and improve quality of life in older persons. In 2008, The U.S. Department of Health and Human Services released guidelines to provide information and guidance on the amount of physical activity recommended to maintain health and fitness. For substantial health benefits, the guidelines recommend that most older adults participate in at least 150 minutes of moderate-intensity aerobic activity, 75 minutes of vigorous-intensity aerobic activity, or an equivalent combination of each per week. Older adults should also engage in strengthening activities that involve all major muscle groups at least two days a week. Those at risk of falling should add exercises that help maintain or improve balance. Generally healthy adults without chronic health conditions do not need to consult with a physician before starting an exercise regimen.
Audiface: Multimodal deep learning for depression screening
- Flores
About one-in-five americans use a smart watch or fitness tracker
- E A Vogels
Missing data imputation: focusing on single imputation
- Z Zhang