Predicting Emotional States Using Behavioral Markers Derived From Passively Sensed Data: Data-Driven Machine Learning Approach

Investigating psychological mechanisms linking pain severity to depression symptoms in women cancer survivors at a cancer center with a rural catchment area

PURPOSE

Women cancer survivors, especially those in rural areas, with high levels of depression may be acutely susceptible to pain due to the ways they think, feel, and behave. The current study seeks to elucidate the relationship between symptoms of depression and pain severity in women cancer survivors, by examining the putative mediators involved in this relationship, specifically their self-efficacy for managing their health, how overwhelmed they were from life's responsibilities, and relational burden.

METHODS

Self-report data were collected from 183 cancer survivors of breast, cervical, ovarian, or endometrial/uterine cancer, who were between 6 months and 3 years post-active therapy.

RESULTS

Women cancer survivors with higher (vs. lower) symptoms of depression had more severe pain. Individual mediation analyses revealed that survivors with higher levels of depression felt more overwhelmed by life's responsibilities and had lower self-efficacy about managing their health, which was associated with greater pain severity. When all mediators were simultaneously entered into the same model, feeling overwhelmed by life's responsibilities significantly mediated the link between survivors' symptoms of depression and their pain severity.

CONCLUSIONS

The relationship between symptoms of depression and pain severity in women cancer survivors may be attributed in part to their self-efficacy and feeling overwhelmed by life's responsibilities. Early and frequent assessment of psychosocial factors involved in pain severity for women cancer survivors may be important for managing their pain throughout the phases of cancer survivorship.

Ethical considerations for integrating multimodal computer perception and neurotechnology

Background

Artificial intelligence (AI)-based computer perception technologies (e.g., digital phenotyping and affective computing) promise to transform clinical approaches to personalized care in psychiatry and beyond by offering more objective measures of emotional states and behavior, enabling precision treatment, diagnosis, and symptom monitoring. At the same time, passive and continuous nature by which they often collect data from patients in non-clinical settings raises ethical issues related to privacy and self-determination. Little is known about how such concerns may be exacerbated by the integration of neural data, as parallel advances in computer perception, AI, and neurotechnology enable new insights into subjective states. Here, we present findings from a multi-site NCATS-funded study of ethical considerations for translating computer perception into clinical care and contextualize them within the neuroethics and neurorights literatures.

Methods

We conducted qualitative interviews with patients (n = 20), caregivers (n = 20), clinicians (n = 12), developers (n = 12), and clinician developers (n = 2) regarding their perspective toward using PC in clinical care. Transcripts were analyzed in MAXQDA using Thematic Content Analysis.

Results

Stakeholder groups voiced concerns related to (1) perceived invasiveness of passive and continuous data collection in private settings; (2) data protection and security and the potential for negative downstream/future impacts on patients of unintended disclosure; and (3) ethical issues related to patients' limited versus hyper awareness of passive and continuous data collection and monitoring. Clinicians and developers highlighted that these concerns may be exacerbated by the integration of neural data with other computer perception data.

Discussion

Our findings suggest that the integration of neurotechnologies with existing computer perception technologies raises novel concerns around dignity-related and other harms (e.g., stigma, discrimination) that stem from data security threats and the growing potential for reidentification of sensitive data. Further, our findings suggest that patients' awareness and preoccupation with feeling monitored via computer sensors ranges from hypo- to hyper-awareness, with either extreme accompanied by ethical concerns (consent vs. anxiety and preoccupation). These results highlight the need for systematic research into how best to implement these technologies into clinical care in ways that reduce disruption, maximize patient benefits, and mitigate long-term risks associated with the passive collection of sensitive emotional, behavioral and neural data.

Quantifying Digital Biomarkers for Well-Being: Stress, Anxiety, Positive and Negative Affect via Wearable Devices and Their Time-Based Predictions

Wearable devices have become ubiquitous, collecting rich temporal data that offers valuable insights into human activities, health monitoring, and behavior analysis. Leveraging these data, researchers have developed innovative approaches to classify and predict time-based patterns and events in human life. Time-based techniques allow the capture of intricate temporal dependencies, which is the nature of the data coming from wearable devices. This paper focuses on predicting well-being factors, such as stress, anxiety, and positive and negative affect, on the Tesserae dataset collected from office workers. We examine the performance of different methodologies, including deep-learning architectures, LSTM, ensemble techniques, Random Forest (RF), and XGBoost, and compare their performances for time-based and non-time-based versions. In time-based versions, we investigate the effect of previous records of well-being factors on the upcoming ones. The overall results show that time-based LSTM performs the best among conventional (non-time-based) RF, XGBoost, and LSTM. The performance even increases when we consider a more extended previous period, in this case, 3 past-days rather than 1 past-day to predict the next day. Furthermore, we explore the corresponding biomarkers for each well-being factor using feature ranking. The obtained rankings are compatible with the psychological literature. In this work, we validated them based on device measurements rather than subjective survey responses.

Ensemble Approach to Combining Episode Prediction Models Using Sequential Circadian Rhythm Sensor Data from Mental Health Patients

Managing mood disorders poses challenges in counseling and drug treatment, owing to limitations. Counseling is the most effective during hospital visits, and the side effects of drugs can be burdensome. Patient empowerment is crucial for understanding and managing these triggers. The daily monitoring of mental health and the utilization of episode prediction tools can enable self-management and provide doctors with insights into worsening lifestyle patterns. In this study, we test and validate whether the prediction of future depressive episodes in individuals with depression can be achieved by using lifelog sequence data collected from digital device sensors. Diverse models such as random forest, hidden Markov model, and recurrent neural network were used to analyze the time-series data and make predictions about the occurrence of depressive episodes in the near future. The models were then combined into a hybrid model. The prediction accuracy of the hybrid model was 0.78; especially in the prediction of rare episode events, the F1-score performance was approximately 1.88 times higher than that of the dummy model. We explored factors such as data sequence size, train-to-test data ratio, and class-labeling time slots that can affect the model performance to determine the combinations of parameters that optimize the model performance. Our findings are especially valuable because they are experimental results derived from large-scale participant data analyzed over a long period of time.

Challenges and best practices for digital unstructured data enrichment in health research: A systematic narrative review

Digital data play an increasingly important role in advancing health research and care. However, most digital data in healthcare are in an unstructured and often not readily accessible format for research. Unstructured data are often found in a format that lacks standardization and needs significant preprocessing and feature extraction efforts. This poses challenges when combining such data with other data sources to enhance the existing knowledge base, which we refer to as digital unstructured data enrichment. Overcoming these methodological challenges requires significant resources and may limit the ability to fully leverage their potential for advancing health research and, ultimately, prevention, and patient care delivery. While prevalent challenges associated with unstructured data use in health research are widely reported across literature, a comprehensive interdisciplinary summary of such challenges and possible solutions to facilitate their use in combination with structured data sources is missing. In this study, we report findings from a systematic narrative review on the seven most prevalent challenge areas connected with the digital unstructured data enrichment in the fields of cardiology, neurology and mental health, along with possible solutions to address these challenges. Based on these findings, we developed a checklist that follows the standard data flow in health research studies. This checklist aims to provide initial systematic guidance to inform early planning and feasibility assessments for health research studies aiming combining unstructured data with existing data sources. Overall, the generality of reported unstructured data enrichment methods in the studies included in this review call for more systematic reporting of such methods to achieve greater reproducibility in future studies.

Working towards a ready to implement digital literacy program

Background

As healthcare continues to expand online and digital care offerings multiply, the importance of digital inclusion and equity is now better recognized. Yet despite impressive regional grassroots efforts, today there remain few readily deployable programs designed to support patient digital literacy.

Methods

Digital Outreach for Obtaining Resources and Skills (DOORs) is one such digital literacy program that has evolved over the last 5 years to meet the rising demand. Through community partnerships, the DOORs curriculum and delivery has been updated to make the program more accessible and applicable as Coronavirus Disease 2019 (COVID-19) changes healthcare. Participants' experience in the most updated iteration of DOORS was assessed through surveys and semi-structured interviews.

Results

Improvements to DOORs include an updated DOORs curriculum, updated facilitator manual, an online platform with a learning management system, standardized training, patient-facing educational handouts, consolidation of all DOORs materials into a single package that is ready to be shared with other groups, implementation of a single-session intervention model, and Spanish translation. Participants reported improved confidence on 72% of the digital skills assessed. Thematic analysis resulted in three themes: awareness of divide, patient-centered design, and expanded skills and confidence.

Conclusions

Combined, these changes and participant outcomes better position DOORS to meet the rising need for digital literacy and offers a scalable model for teams across the world.

Identifying Current Feelings of Mild and Moderate to High Depression in Young, Healthy Individuals Using Gait and Balance: An Exploratory Study

Depressive mood states in healthy populations are prevalent but often under-reported. Biases exist in self-reporting of depression in otherwise healthy individuals. Gait and balance control can serve as objective markers for identifying those individuals, particularly in real-world settings. We utilized inertial measurement units (IMU) to measure gait and balance control. An exploratory, cross-sectional design was used to compare individuals who reported feeling depressed at the moment (n = 49) with those who did not (n = 84). The Quality Assessment Tool for Observational Cohort and Cross-sectional Studies was employed to ensure internal validity. We recruited 133 participants aged between 18-36 years from the university community. Various instruments were used to evaluate participants' present depressive symptoms, sleep, gait, and balance. Gait and balance variables were used to detect depression, and participants were categorized into three groups: not depressed, mild depression, and moderate-high depression. Participant characteristics were analyzed using ANOVA and Kruskal-Wallis tests, and no significant differences were found in age, height, weight, BMI, and prior night's sleep between the three groups. Classification models were utilized for depression detection. The most accurate model incorporated both gait and balance variables, yielding an accuracy rate of 84.91% for identifying individuals with moderate-high depression compared to non-depressed individuals.

Uncovering Emotions: A Pilot Study on Classifying Moods in the Valence-Arousal Space using In-the-Wild Passive Data

Mood classification from passive data promises to provide an unobtrusive way to track a person's emotions over time. In this exploratory study, we collected phone sensor data and physiological signals from 8 individuals, including 5 healthy participants and 3 depressed patients, for a maximum of 35 days. Participants were asked to answer a digital questionnaire three times daily, resulting in a total of 334 self-reported mood state samples. Gradient-boosting classification was applied to the collected passive data to categorize 4 mood states in the Valence-Energetic Arousal space. The cross-validation results showed better classification performance compared to a baseline model, which always predicts the majority class. The classifier using passive data had an area under the precision-recall curve of 0.39 (SD = 0.1) while the baseline had 0.26 (SD = 0.03), suggesting the presence of information in the collected features that support the classification process. The model identified the entropy of the heart rate and the average physical activity in the preceding 8 hours, along with the max normal-to-normal (NN) sinus beat interval and the NN low frequency-high frequency ratio during the questionnaire completion, as the most important features in its analysis. Additionally, the time range of data collection was considered a contextual factor.

A machine learning approach to determine resilience utilizing wearable device data: analysis of an observational cohort

Objective

To assess whether an individual's degree of psychological resilience can be determined from physiological metrics passively collected from a wearable device.

Materials and Methods

Data were analyzed in this secondary analysis of the Warrior Watch Study dataset, a prospective cohort of healthcare workers enrolled across 7 hospitals in New York City. Subjects wore an Apple Watch for the duration of their participation. Surveys were collected measuring resilience, optimism, and emotional support at baseline.

Results

We evaluated data from 329 subjects (mean age 37.4 years, 37.1% male). Across all testing sets, gradient-boosting machines (GBM) and extreme gradient-boosting models performed best for high- versus low-resilience prediction, stratified on a median Connor-Davidson Resilience Scale-2 score of 6 (interquartile range = 5-7), with an AUC of 0.60. When predicting resilience as a continuous variable, multivariate linear models had a correlation of 0.24 (P = .029) and RMSE of 1.37 in the testing data. A positive psychological construct, comprised of resilience, optimism, and emotional support was also evaluated. The oblique random forest method performed best in estimating high- versus low-composite scores stratified on a median of 32.5, with an AUC of 0.65, a sensitivity of 0.60, and a specificity of 0.70.

Discussion

In a post hoc analysis, machine learning models applied to physiological metrics collected from wearable devices had some predictive ability in identifying resilience states and a positive psychological construct.

Conclusions

These findings support the further assessment of psychological characteristics from passively collected wearable data in dedicated studies.

A Review of the Role of Artificial Intelligence in Healthcare

Artificial intelligence (AI) applications have transformed healthcare. This study is based on a general literature review uncovering the role of AI in healthcare and focuses on the following key aspects: (i) medical imaging and diagnostics, (ii) virtual patient care, (iii) medical research and drug discovery, (iv) patient engagement and compliance, (v) rehabilitation, and (vi) other administrative applications. The impact of AI is observed in detecting clinical conditions in medical imaging and diagnostic services, controlling the outbreak of coronavirus disease 2019 (COVID-19) with early diagnosis, providing virtual patient care using AI-powered tools, managing electronic health records, augmenting patient engagement and compliance with the treatment plan, reducing the administrative workload of healthcare professionals (HCPs), discovering new drugs and vaccines, spotting medical prescription errors, extensive data storage and analysis, and technology-assisted rehabilitation. Nevertheless, this science pitch meets several technical, ethical, and social challenges, including privacy, safety, the right to decide and try, costs, information and consent, access, and efficacy, while integrating AI into healthcare. The governance of AI applications is crucial for patient safety and accountability and for raising HCPs' belief in enhancing acceptance and boosting significant health consequences. Effective governance is a prerequisite to precisely address regulatory, ethical, and trust issues while advancing the acceptance and implementation of AI. Since COVID-19 hit the global health system, the concept of AI has created a revolution in healthcare, and such an uprising could be another step forward to meet future healthcare needs.

Digital Phenotyping of Mental Health using multimodal sensing of multiple situations of interest: A Systematic Literature Review

Many studies have used Digital Phenotyping of Mental Health (DPMH) to complement classic methods of mental health assessment and monitoring. This research area proposes innovative methods that perform multimodal sensing of multiple situations of interest (e.g., sleep, physical activity, mobility) to health professionals. In this paper, we present a Systematic Literature Review (SLR) to recognize, characterize and analyze the state of the art on DPMH using multimodal sensing of multiple situations of interest to professionals. We searched for studies in six digital libraries, which resulted in 1865 retrieved published papers. Next, we performed a systematic process of selecting studies based on inclusion and exclusion criteria, which selected 59 studies for the data extraction phase. First, based on the analysis of the extracted data, we describe an overview of this field, then presenting characteristics of the selected studies, the main mental health topics targeted, the physical and virtual sensors used, and the identified situations of interest. Next, we outline answers to research questions, describing the context data sources used to detect situations, the DPMH workflow used for multimodal sensing of situations, and the application of DPMH solutions in the mental health assessment and monitoring process. In addition, we recognize trends presented by DPMH studies, such as the design of solutions for high-level information recognition, association of features with mental states/disorders, classification of mental states/disorders, and prediction of mental states/disorders. We also recognize the main open issues in this research area. Based on the results of this SLR, we conclude that despite the potential and continuous evolution for using these solutions as medical decision support tools, this research area needs more work to overcome technology and methodological rigor issues to adopt proposed solutions in real clinical settings.

A Review of Converging Technologies in eHealth Pertaining to Artificial Intelligence

Over the last couple of years, in the context of the COVID-19 pandemic, many healthcare issues have been exacerbated, highlighting the paramount need to provide both reliable and affordable health services to remote locations by using the latest technologies such as video conferencing, data management, the secure transfer of patient information, and efficient data analysis tools such as machine learning algorithms. In the constant struggle to offer healthcare to everyone, many modern technologies find applicability in eHealth, mHealth, telehealth or telemedicine. Through this paper, we attempt to render an overview of what different technologies are used in certain healthcare applications, ranging from remote patient monitoring in the field of cardio-oncology to analyzing EEG signals through machine learning for the prediction of seizures, focusing on the role of artificial intelligence in eHealth.

Machine learning for passive mental health symptom prediction: Generalization across different longitudinal mobile sensing studies

Mobile sensing data processed using machine learning models can passively and remotely assess mental health symptoms from the context of patients' lives. Prior work has trained models using data from single longitudinal studies, collected from demographically homogeneous populations, over short time periods, using a single data collection platform or mobile application. The generalizability of model performance across studies has not been assessed. This study presents a first analysis to understand if models trained using combined longitudinal study data to predict mental health symptoms generalize across current publicly available data. We combined data from the CrossCheck (individuals living with schizophrenia) and StudentLife (university students) studies. In addition to assessing generalizability, we explored if personalizing models to align mobile sensing data, and oversampling less-represented severe symptoms, improved model performance. Leave-one-subject-out cross-validation (LOSO-CV) results were reported. Two symptoms (sleep quality and stress) had similar question-response structures across studies and were used as outcomes to explore cross-dataset prediction. Models trained with combined data were more likely to be predictive (significant improvement over predicting training data mean) than models trained with single-study data. Expected model performance improved if the distance between training and validation feature distributions decreased using combined versus single-study data. Personalization aligned each LOSO-CV participant with training data, but only improved predicting CrossCheck stress. Oversampling significantly improved severe symptom classification sensitivity and positive predictive value, but decreased model specificity. Taken together, these results show that machine learning models trained on combined longitudinal study data may generalize across heterogeneous datasets. We encourage researchers to disseminate collected de-identified mobile sensing and mental health symptom data, and further standardize data types collected across studies to enable better assessment of model generalizability.

Systematic Review of Functional MRI Applications for Psychiatric Disease Subtyping

Background: Psychiatric disorders have been historically classified using symptom information alone. Recently, there has been a dramatic increase in research interest not only in identifying the mechanisms underlying defined pathologies but also in redefining their etiology. This is particularly relevant for the field of personalized medicine, which searches for data-driven approaches to improve diagnosis, prognosis, and treatment selection for individual patients. Methods: This review aims to provide a high-level overview of the rapidly growing field of functional magnetic resonance imaging (fMRI) from the perspective of unsupervised machine learning applications for disease subtyping. Following the PRISMA guidelines for protocol reproducibility, we searched the PubMed database for articles describing functional MRI applications used to obtain, interpret, or validate psychiatric disease subtypes. We also employed the active learning framework ASReview to prioritize publications in a machine learning-guided way. Results: From the 20 studies that met the inclusion criteria, five used functional MRI data to interpret symptom-derived disease clusters, four used it to interpret clusters derived from biomarker data other than fMRI itself, and 11 applied clustering techniques involving fMRI directly. Major depression disorder and schizophrenia were the two most frequently studied pathologies (35% and 30% of the retrieved studies, respectively), followed by ADHD (15%), psychosis as a whole (10%), autism disorder (5%), and the consequences of early exposure to violence (5%). Conclusions: The increased interest in personalized medicine and data-driven disease subtyping also extends to psychiatric disorders. However, to date, this subfield is at an incipient exploratory stage, and all retrieved studies were mostly proofs of principle where further validation and increased sample sizes are craved for. Whereas results for all explored diseases are inconsistent, we believe this reflects the need for concerted, multisite data collection efforts with a strong focus on measuring the generalizability of results. Finally, whereas functional MRI is the best way of measuring brain function available to date, its low signal-to-noise ratio and elevated monetary cost make it a poor clinical alternative. Even with technology progressing and costs decreasing, this might incentivize the search for more accessible, clinically ready functional proxies in the future.

Understanding People's Use of and Perspectives on Mood-Tracking Apps: Interview Study

BACKGROUND

Supporting mental health and wellness is of increasing interest due to a growing recognition of the prevalence and burden of mental health issues. Mood is a central aspect of mental health, and several technologies, especially mobile apps, have helped people track and understand it. However, despite formative work on and dissemination of mood-tracking apps, it is not well understood how mood-tracking apps used in real-world contexts might benefit people and what people hope to gain from them.

OBJECTIVE

To address this gap, the purpose of this study was to understand motivations for and experiences in using mood-tracking apps from people who used them in real-world contexts.

METHODS

We interviewed 22 participants who had used mood-tracking apps using a semistructured interview and card sorting task. The interview focused on their experiences using a mood-tracking app. We then conducted a card sorting task using screenshots of various data entry and data review features from mood-tracking apps. We used thematic analysis to identify themes around why people use mood-tracking apps, what they found useful about them, and where people felt these apps fell short.

RESULTS

Users of mood-tracking apps were primarily motivated by negative life events or shifts in their own mental health that prompted them to engage in tracking and improve their situation. In general, participants felt that using a mood-tracking app facilitated self-awareness and helped them to look back on a previous emotion or mood experience to understand what was happening. Interestingly, some users reported less inclination to document their negative mood states and preferred to document their positive moods. There was a range of preferences for personalization and simplicity of tracking. Overall, users also liked features in which their previous tracked emotions and moods were visualized in figures or calendar form to understand trends. One gap in available mood-tracking apps was the lack of app-facilitated recommendations or suggestions for how to interpret their own data or improve their mood.

CONCLUSIONS

Although people find various features of mood-tracking apps helpful, the way people use mood-tracking apps, such as avoiding entering negative moods, tracking infrequently, or wanting support to understand or change their moods, demonstrate opportunities for improvement. Understanding why and how people are using current technologies can provide insights to guide future designs and implementations.