How are you feeling?: A personalized methodology for predicting mental states from temporally observable physical and behavioral information

Investigating Receptivity and Affect Using Machine Learning: Ecological Momentary Assessment and Wearable Sensing Study

BACKGROUND

As mobile health (mHealth) studies become increasingly productive owing to the advancements in wearable and mobile sensor technology, our ability to monitor and model human behavior will be constrained by participant receptivity. Many health constructs are dependent on subjective responses, and without such responses, researchers are left with little to no ground truth to accompany our ever-growing biobehavioral data. This issue can significantly impact the quality of a study, particularly for populations known to exhibit lower compliance rates. To address this challenge, researchers have proposed innovative approaches that use machine learning (ML) and sensor data to modify the timing and delivery of surveys. However, an overarching concern is the potential introduction of biases or unintended influences on participants' responses when implementing new survey delivery methods.

OBJECTIVE

This study aims to demonstrate the potential impact of an ML-based ecological momentary assessment (EMA) delivery system (using receptivity as the predictor variable) on the participants' reported emotional state. We examine the factors that affect participants' receptivity to EMAs in a 10-day wearable and EMA-based emotional state-sensing mHealth study. We study the physiological relationships indicative of receptivity and affect while also analyzing the interaction between the 2 constructs.

METHODS

We collected data from 45 healthy participants wearing 2 devices measuring electrodermal activity, accelerometer, electrocardiography, and skin temperature while answering 10 EMAs daily, containing questions about perceived mood. Owing to the nature of our constructs, we can only obtain ground truth measures for both affect and receptivity during responses. Therefore, we used unsupervised and supervised ML methods to infer affect when a participant did not respond. Our unsupervised method used k-means clustering to determine the relationship between physiology and receptivity and then inferred the emotional state during nonresponses. For the supervised learning method, we primarily used random forest and neural networks to predict the affect of unlabeled data points as well as receptivity.

RESULTS

Our findings showed that using a receptivity model to trigger EMAs decreased the reported negative affect by >3 points or 0.29 SDs in our self-reported affect measure, scored between 13 and 91. The findings also showed a bimodal distribution of our predicted affect during nonresponses. This indicates that this system initiates EMAs more commonly during states of higher positive emotions.

CONCLUSIONS

Our results showed a clear relationship between affect and receptivity. This relationship can affect the efficacy of an mHealth study, particularly those that use an ML algorithm to trigger EMAs. Therefore, we propose that future work should focus on a smart trigger that promotes EMA receptivity without influencing affect during sampled time points.

Description, prediction and causation: Methodological challenges of studying child and adolescent development

Scientific research can be categorized into: a) descriptive research, with the main goal to summarize characteristics of a group (or person); b) predictive research, with the main goal to forecast future outcomes that can be used for screening, selection, or monitoring; and c) explanatory research, with the main goal to understand the underlying causal mechanism, which can then be used to develop interventions. Since each goal requires different research methods in terms of design, operationalization, model building and evaluation, it should form an important basis for decisions on how to set up and execute a study. To determine the extent to which developmental research is motivated by each goal and how this aligns with the research designs that are used, we evaluated 100 publications from the Consortium on Individual Development (CID). This analysis shows that the match between research goal and research design is not always optimal. We discuss alternative techniques, which are not yet part of the developmental scientist's standard toolbox, but that may help bridge some of the lurking gaps that developmental scientists encounter between their research design and their research goal. These include unsupervised and supervised machine learning, directed acyclical graphs, Mendelian randomization, and target trials.

Why Loneliness Interventions Are Unsuccessful: A Call for Precision Health

Background

Loneliness has drawn increasing attention over the past few decades due to rising recognition of its close connection with serious health issues, like dementia. Yet, researchers are failing to find solutions to alleviate the globally experienced burden of loneliness.

Purpose

This review aims to shed light on possible reasons for why interventions have been ineffective. We suggest new directions for research on loneliness as it relates to precision health, emerging technologies, digital phenotyping, and machine learning.

Results

Current loneliness interventions are unsuccessful due to (i) their inconsideration of loneliness as a heterogeneous construct and (ii) not being targeted at individuals' needs and contexts. We propose a model for how loneliness interventions can move towards finding the right solution for the right person at the right time. Taking a precision health approach, we explore how transdisciplinary research can contribute to creating a more holistic picture of loneliness and shift interventions from treatment to prevention.

Conclusions

We urge the field to rethink metrics to account for diverse intra-individual experiences and trajectories of loneliness. Big data sharing and evolving technologies that emphasize human connection raise hope for realizing our model of precision health applied to loneliness. There is an urgent need for precise, integrated, and theory-driven interventions that focus on individuals' needs and the subjective burden of loneliness in the ageing context.

Prediction of stress and drug craving ninety minutes in the future with passively collected GPS data

Just-in-time adaptive interventions (JITAIs), typically smartphone apps, learn to deliver therapeutic content when users need it. The challenge is to "push" content at algorithmically chosen moments without making users trigger it with effortful input. We trained a randomForest algorithm to predict heroin craving, cocaine craving, or stress (reported via smartphone app 3x/day) 90 min into the future, using 16 weeks of field data from 189 outpatients being treated for opioid-use disorder. We used only one form of continuous input (along with person-level demographic data), collected passively: an indicator of environmental exposures along the past 5 h of movement, as assessed by GPS. Our models achieved excellent overall accuracy-as high as 0.93 by the end of 16 weeks of tailoring-but this was driven mostly by correct predictions of absence. For predictions of presence, "believability" (positive predictive value, PPV) usually peaked in the high 0.70s toward the end of the 16 weeks. When the prediction target was more rare, PPV was lower. Our findings complement those of other investigators who use machine learning with more broadly based "digital phenotyping" inputs to predict or detect mental and behavioral events. When target events are comparatively subtle, like stress or drug craving, accurate detection or prediction probably needs effortful input from users, not passive monitoring alone. We discuss ways in which accuracy is difficult to achieve or even assess, and warn that high overall accuracy (including high specificity) can mask the abundance of false alarms that low PPV reveals.

Using Machine Learning to Derive Just-In-Time and Personalized Predictors of Stress: Observational Study Bridging the Gap Between Nomothetic and Ideographic Approaches

BACKGROUND

Investigations into person-specific predictors of stress have typically taken either a population-level nomothetic approach or an individualized ideographic approach. Nomothetic approaches can quickly identify predictors but can be hindered by the heterogeneity of these predictors across individuals and time. Ideographic approaches may result in more predictive models at the individual level but require a longer period of data collection to identify robust predictors.

OBJECTIVE

Our objectives were to compare predictors of stress identified through nomothetic and ideographic models and to assess whether sequentially combining nomothetic and ideographic models could yield more accurate and actionable predictions of stress than relying on either model. At the same time, we sought to maintain the interpretability necessary to retrieve individual predictors of stress despite using nomothetic models.

METHODS

Data collected in a 1-year observational study of 79 participants performing low levels of exercise were used. Physical activity was continuously and objectively monitored by actigraphy. Perceived stress was recorded by participants via daily ecological momentary assessments on a mobile app. Environmental variables including daylight time, temperature, and precipitation were retrieved from the public archives. Using these environmental, actigraphy, and mobile assessment data, we built machine learning models to predict individual stress ratings using linear, decision tree, and neural network techniques employing nomothetic and ideographic approaches. The accuracy of the approaches for predicting individual stress ratings was compared based on classification errors.

RESULTS

Across the group of patients, an individual's recent history of stress ratings was most heavily weighted in predicting a future stress rating in the nomothetic recurrent neural network model, whereas environmental factors such as temperature and daylight, as well as duration and frequency of bouts of exercise, were more heavily weighted in the ideographic models. The nomothetic recurrent neural network model was the highest performing nomothetic model and yielded 72% accuracy for an 80%/20% train/test split. Using the same 80/20 split, the ideographic models yielded 75% accuracy. However, restricting ideographic models to participants with more than 50 valid days in the training set, with the same 80/20 split, yielded 85% accuracy.

CONCLUSIONS

We conclude that for some applications, nomothetic models may be useful for yielding higher initial performance while still surfacing personalized predictors of stress, before switching to ideographic models upon sufficient data collection.

Discovering Discontinuity in Big Financial Transaction Data

Business transactions are typically recorded in the company ledger. The primary purpose of such financial information is to accompany a monthly or quarterly report for executives to make sound business decisions and strategies for the next business period. These business strategies often result in transitions that cause underlying infrastructures and components to change, including alteration in the nomenclature system of the business components. As a result, a transaction stream of an affected component would be replaced by another stream with a different component name, resulting in discontinuity of a financial stream of the same component. Recently, advancement in large-scale data mining technologies has enabled a set of critical applications to utilize knowledge extracted from a vast amount of existing data that would otherwise have been unused or underutilized. In financial and services computing domains, recent studies have illustrated that historical financial data could be used to predict future revenues and profits, optimizing costs, among other potential applications. These prediction models rely on long-term availability of the historical data that traces back for multiple years. However, the discontinuity of the financial transaction stream associated with a business component has limited the learning capability of the prediction models. In this article, we propose a set of machine learning–based algorithms to automatically discover component name replacements, using information available in general ledger databases. The algorithms are designed to be scalable for handling massive data points, especially in large companies. Furthermore, the proposed algorithms are generalizable to other domains whose data is time series and shares the same nature as the financial data available in business ledgers. A case study of real-world IBM service delivery retrieved from four different geographical regions is used to validate the efficacy of the proposed methodology.