Objective Prediction of Next-Day's Affect Using Multimodal Physiological and Behavioral Data: Algorithm Development and Validation Study.

BACKGROUND: Affective states are important aspects of healthy functioning; as such, monitoring and understanding affect is necessary for the assessment and treatment of mood-based disorders. Recent advancements in wearable technologies have increased the use of such tools in detecting and accurately estimating mental states (eg, affect, mood, and stress), offering comprehensive and continuous monitoring of individuals over time. OBJECTIVE: Previous attempts to model an individual's mental state relied on subjective measurements or the inclusion of only a few objective monitoring modalities (eg, smartphones). This study aims to investigate the capacity of monitoring affect using fully objective measurements. We conducted a comparatively long-term (12-month) study with a holistic sampling of participants' moods, including 20 affective states. METHODS: Longitudinal physiological data (eg, sleep and heart rate), as well as daily assessments of affect, were collected using 3 modalities (ie, smartphone, watch, and ring) from 20 college students over a year. We examined the difference between the distributions of data collected from each modality along with the differences between their rates of missingness. Out of the 20 participants, 7 provided us with 200 or more days' worth of data, and we used this for our predictive modeling setup. Distributions of positive affect (PA) and negative affect (NA) among the 7 selected participants were observed. For predictive modeling, we assessed the performance of different machine learning models, including random forests (RFs), support vector machines (SVMs), multilayer perceptron (MLP), and K-nearest neighbor (KNN). We also investigated the capability of each modality in predicting mood and the most important features of PA and NA RF models. RESULTS: RF was the best-performing model in our analysis and performed mood and stress (nervousness) prediction with ~81% and ~72% accuracy, respectively. PA models resulted in better performance compared to NA. The order of the most important modalities in predicting PA and NA was the smart ring, phone, and watch, respectively. SHAP (Shapley Additive Explanations) analysis showed that sleep and activity-related features were the most impactful in predicting PA and NA. CONCLUSIONS: Generic machine learning-based affect prediction models, trained with population data, outperform existing methods, which use the individual's historical information. Our findings indicated that our mood prediction method outperformed the existing methods. Additionally, we found that sleep and activity level were the most important features for predicting next-day PA and NA, respectively.

Personal mental health navigator: Harnessing the power of data, personal models, and health cybernetics to promote psychological well-being.

Current digital mental healthcare solutions conventionally take on a reactive approach, requiring individuals to self-monitor and document existing symptoms. These solutions are unable to provide comprehensive, wrap-around, customized treatments that capture an individual's holistic mental health model as it unfolds over time. Recognizing that each individual requires personally tailored mental health treatment, we introduce the notion of Personalized Mental Health Navigation (MHN): a cybernetic goal-based system that deploys a continuous loop of monitoring, estimation, and guidance to steer the individual towards mental flourishing. We present the core components of MHN that are premised on the importance of addressing an individual's personal mental health state. Moreover, we provide an overview of the existing physical health navigation systems and highlight the requirements and challenges of deploying the navigational approach to the mental health domain.

ECG Biosignal Deidentification Using Conditional Generative Adversarial Networks.

Electrocardiogram (ECG) signals provide rich information on individuals' potential cardiovascular conditions and disease, ranging from coronary artery disease to the risk of a heart attack. While health providers store and share these information for medical and research purposes, such data is highly vulnerable to privacy concerns, similar to many other types of healthcare data. Recent works have shown the feasibility of identifying and authenticating individuals by using ECG as a biometric due to the highly individualized nature of ECG signals. However, to the best of our knowledge, there does not exist a method in the literature attempting to de-identify ECG signals. In this paper, to address this privacy protection gap, we propose a Generative Adversarial Network (GAN)-based framework for de-identification of ECG signals. We leverage a combination of a standard GAN loss, an Ordinary Differential Equations (ODE)-based, and identity-based loss values to train a generator that de-identifies a ECG signal while preserving structure the ECG signal and information regarding the target cardio vascular condition. We evaluate our framework in terms of both qualitative and quantitative metrics considering different weightings over the above-mentioned losses. Our experiments demonstrate the efficiency of our framework in terms of privacy protection and ECG signal structural preservation.

Sleep Patterns and Affect Dynamics Among College Students During the COVID-19 Pandemic: Intensive Longitudinal Study.

BACKGROUND: Sleep disturbance is a transdiagnostic risk factor that is so prevalent among young adults that it is considered a public health epidemic, which has been exacerbated by the COVID-19 pandemic. Sleep may contribute to mental health via affect dynamics. Prior literature on the contribution of sleep to affect is largely based on correlational studies or experiments that do not generalize to the daily lives of young adults. Furthermore, the literature examining the associations between sleep variability and affect dynamics remains scant. OBJECTIVE: In an ecologically valid context, using an intensive longitudinal design, we aimed to assess the daily and long-term associations between sleep patterns and affect dynamics among young adults during the COVID-19 pandemic. METHODS: College student participants (N=20; female: 13/20, 65%) wore an Oura ring (Oura Health Ltd) continuously for 3 months to measure sleep patterns, such as average and variability in total sleep time (TST), wake after sleep onset (WASO), sleep efficiency, and sleep onset latency (SOL), resulting in 1173 unique observations. We administered a daily ecological momentary assessment by using a mobile health app to evaluate positive affect (PA), negative affect (NA), and COVID-19 worry once per day. RESULTS: Participants with a higher sleep onset latency (b=-1.09, SE 0.36; P=.006) and TST (b=-0.15, SE 0.05; P=.008) on the prior day had lower PA on the next day. Further, higher average TST across the 3-month period predicted lower average PA (b=-0.36, SE 0.12; P=.009). TST variability predicted higher affect variability across all affect domains. Specifically, higher variability in TST was associated higher PA variability (b=0.09, SE 0.03; P=.007), higher negative affect variability (b=0.12, SE 0.05; P=.03), and higher COVID-19 worry variability (b=0.16, SE 0.07; P=.04). CONCLUSIONS: Fluctuating sleep patterns are associated with affect dynamics at the daily and long-term scales. Low PA and affect variability may be potential pathways through which sleep has implications for mental health.

Investigation of Machine Learning and Deep Learning Approaches for Detection of Mild Traumatic Brain Injury from Human Sleep Electroencephalogram.

Traumatic Brain Injury (TBI) is a highly prevalent and serious public health concern. Most cases of TBI are mild in nature, yet some individuals may develop following-up persistent disability. The pathophysiologic causes for those with persistent postconcussive symptoms are most likely multifactorial and the underlying mechanism is not well understood, although it is clear that sleep disturbances feature prominently in those with persistent disability. The sleep electroencephalogram (EEG) provides a direct window into neuronal activity during an otherwise highly stereotyped behavioral state, and represents a promising quantitative measure for TBI diagnosis and prognosis. With the ever-evolving domain of machine learning, deep convolutional neural networks, and the development of better architectures, these approaches hold promise to solve some of the long entrenched challenges of personalized medicine for uses in recommendation systems and/or in health monitoring systems. In particular, advanced EEG analysis to identify putative EEG biomarkers of neurological disease could be highly relevant in the prognostication of mild TBI, an otherwise heterogeneous disorder with a wide range of affected phenotypes and disability levels. In this work, we investigate the use of various machine learning techniques and deep neural network architectures on a cohort of human subjects with sleep EEG recordings from overnight, in-lab, diagnostic polysomnography (PSG). An optimal scheme is explored for the classification of TBI versus non-TBI control subjects. The results were promising with an accuracy of ∼95% in random sampling arrangement and ∼70% in independent validation arrangement when appropriate parameters were used using a small number of subjects (10 mTBI subjects and 9 age- and sex-matched controls). We are thus confident that, with additional data and further studies, we would be able to build a generalized model to detect TBI accurately, not only via attended, in-lab PSG recordings, but also in practical scenarios such as EEG data obtained from simple wearables in daily life.

Classification of Electroencephalogram in a Mouse Model of Traumatic Brain Injury Using Machine Learning Approaches.

Traumatic Brain Injury (TBI) is highly prevalent, affecting ~1% of the U.S. population, with lifetime economic costs estimated to be over $75 billion. In the U.S., there are about 50,000 deaths annually related to TBI, and many others are permanently disabled. However, it is currently unknown which individuals will develop persistent disability following TBI and what brain mechanisms underlie these distinct populations. The pathophysiologic causes for those are most likely multifactorial. Electroencephalogram (EEG) has been used as a promising quantitative measure for TBI diagnosis and prognosis. The recent rise of advanced data science approaches such as machine learning and deep learning holds promise to further analyze EEG data, looking for EEG biomarkers of neurological disease, including TBI. In this work, we investigated various machine learning approaches on our unique 24-hour recording dataset of a mouse TBI model, in order to look for an optimal scheme in classification of TBI and control subjects. The epoch lengths were 1 and 2 minutes. The results were promising with accuracy of ~80-90% when appropriate features and parameters were used using a small number of subjects (5 shams and 4 TBIs). We are thus confident that, with more data and studies, we would be able to detect TBI accurately, not only via long-term recordings but also in practical scenarios, with EEG data obtained from simple wearables in the daily life.

Investigation of Machine Learning Approaches for Traumatic Brain Injury Classification via EEG Assessment in Mice.

Due to the difficulties and complications in the quantitative assessment of traumatic brain injury (TBI) and its increasing relevance in today's world, robust detection of TBI has become more significant than ever. In this work, we investigate several machine learning approaches to assess their performance in classifying electroencephalogram (EEG) data of TBI in a mouse model. Algorithms such as decision trees (DT), random forest (RF), neural network (NN), support vector machine (SVM), K-nearest neighbors (KNN) and convolutional neural network (CNN) were analyzed based on their performance to classify mild TBI (mTBI) data from those of the control group in wake stages for different epoch lengths. Average power in different frequency sub-bands and alpha:theta power ratio in EEG were used as input features for machine learning approaches. Results in this mouse model were promising, suggesting similar approaches may be applicable to detect TBI in humans in practical scenarios.