Background Feelings of loneliness are associated with poor physical and mental health. Detection of loneliness through passive sensing on personal devices can lead to the development of interventions aimed at decreasing rates of loneliness. Objective The aim of this study was to explore the potential of using passive sensing to infer levels of loneliness and to identify the corresponding behavioral patterns. Methods Data were collected from smartphones and Fitbits (Flex 2) of 160 college students over a semester. The participants completed the University of California, Los Angeles (UCLA) loneliness questionnaire at the beginning and end of the semester. For a classification purpose, the scores were categorized into high (questionnaire score>40) and low (≤40) levels of loneliness. Daily features were extracted from both devices to capture activity and mobility, communication and phone usage, and sleep behaviors. The features were then averaged to generate semester-level features. We used 3 analytic methods: (1) statistical analysis to provide an overview of loneliness in college students, (2) data mining using the Apriori algorithm to extract behavior patterns associated with loneliness, and (3) machine learning classification to infer the level of loneliness and the change in levels of loneliness using an ensemble of gradient boosting and logistic regression algorithms with feature selection in a leave-one-student-out cross-validation manner. Results The average loneliness score from the presurveys and postsurveys was above 43 (presurvey SD 9.4 and postsurvey SD 10.4), and the majority of participants fell into the high loneliness category (scores above 40) with 63.8% (102/160) in the presurvey and 58.8% (94/160) in the postsurvey. Scores greater than 1 standard deviation above the mean were observed in 12.5% (20/160) of the participants in both pre- and postsurvey scores. The majority of scores, however, fell between 1 standard deviation below and above the mean (pre=66.9% [107/160] and post=73.1% [117/160]). Our machine learning pipeline achieved an accuracy of 80.2% in detecting the binary level of loneliness and an 88.4% accuracy in detecting change in the loneliness level. The mining of associations between classifier-selected behavioral features and loneliness indicated that compared with students with low loneliness, students with high levels of loneliness were spending less time outside of campus during evening hours on weekends and spending less time in places for social events in the evening on weekdays (support=17% and confidence=92%). The analysis also indicated that more activity and less sedentary behavior, especially in the evening, was associated with a decrease in levels of loneliness from the beginning of the semester to the end of it (support=31% and confidence=92%). Conclusions Passive sensing has the potential for detecting loneliness in college stu...

Objective: There has been substantial research and public interest in mindfulness interventions, biological pathways, and health over the past two decades. This article reviews recent developments in understanding relationships between mindfulness interventions and physical health. Methods: A selective review was conducted with the goal of synthesizing conceptual and empirical relationships between mindfulness interventions and physical health outcomes. Results: Initial randomized controlled trials (RCTs) in this area suggest that mindfulness interventions can improve pain management outcomes among chronic pain populations, and there is preliminary evidence for mindfulness interventions improving specific stress-related disease outcomes in some patient populations (i.e., clinical colds, psoriasis, IBS, PTSD, diabetes, HIV). We offer a stress buffering framework for the observed beneficial effects of mindfulness interventions and summarize supporting biobehavioral and neuroimaging studies that provide plausible mechanistic pathways linking mindfulness interventions with positive physical health outcomes. Conclusion: We conclude with new opportunities for research and clinical implementations to consider in the next two decades.

Alcohol typically has a detrimental impact on memory across a variety of encoding and retrieval conditions (e.g., Mintzer, 2007; Ray & Bates, 2006). No research has addressed alcohol's effect on memory for lengthy and interactive events and little has tested alcohol's effect on free recall. In this study 94 participants were randomly assigned to alcohol, placebo, or control groups and consumed drinks in a bar-lab setting while interacting with a "bartender". Immediately afterwards all participants freely recalled the bar interaction. Consistent with alcohol myopia theory, intoxicated participants only differed from placebo and control groups when recalling peripheral information. Expanding on the original hypervigilance hypothesis, placebo participants showed more conservative reporting behaviour than the alcohol or control groups by providing more uncertain and "don't know" responses. Thus, alcohol intoxication had confined effects on memory for events, supporting and extending current theories.

The rate of depression in college students is rising, which is known to increase suicide risk, lower academic performance and double the likelihood of dropping out of school. Existing work on fnding relationships between passively sensed behavior and depression, as well as detecting depression, mainly derives relevant unimodal features from a single sensor. However, co-occurrence of values in multiple sensors may provide better features, because such features can describe behavior in context. We present a new method to extract contextually fltered features from passively collected, time-series mobile data via association rule mining. After calculating traditional unimodal features from the data, we extract rules that relate unimodal features to each other using association rule mining. We extract rules from each class separately (e.g., depression vs. nondepression). We introduce a new metric to select a subset of rules that distinguish between the two classes. From these rules, which capture the relationship between multiple unimodal features, we automatically extract contextually fltered features. These features are then fed into a traditional machine learning pipeline to detect the class of interest (in our case, depression), defned by whether a student has a high BDI-II score at the end of the semester. The behavior rules generated by our methods are highly interpretable representations of diferences between classes. Our best model uses contextually-fltered features to signifcantly outperform a standard model that uses only unimodal features, by an average of 9.7% across a variety of metrics. We further verifed the generalizability of our approach on a second dataset, and achieved very similar results. CCS Concepts: • Human-centered computing Ubiquitous and mobile computing; • Applied computing Life and medical sciences.

We present a machine learning approach that uses data from smartphones and fitness trackers of 138 college students to identify students that experienced depressive symptoms at the end of the semester and students whose depressive symptoms worsened over the semester. Our novel approach is a feature extraction technique that allows us to select meaningful features indicative of depressive symptoms from longitudinal data. It allows us to detect the presence of post-semester depressive symptoms with an accuracy of 85.7% and change in symptom severity with an accuracy of 85.4%. It also predicts these outcomes with an accuracy of >80%, 11–15 weeks before the end of the semester, allowing ample time for pre-emptive interventions. Our work has significant implications for the detection of health outcomes using longitudinal behavioral data and limited ground truth. By detecting change and predicting symptoms several weeks before their onset, our work also has implications for preventing depression.

According to law enforcement, many witnesses are intoxicated either at the time of the crime, the interview, or both (Evans et al., Public Policy Law 15(3):194-221, 2009). However, no study to date has examined whether intoxicated witnesses' recall is different from sober witnesses' and whether they are more vulnerable to misinformation using an ecologically valid experimental design. Intoxicated, placebo, and sober witnesses observed a live, staged theft, overheard subsequent misinformation about the theft, and took part in an investigative interview. Participants generally believed they witnessed a real crime and experienced a real interview. Intoxicated witnesses were not different from placebo or sober witnesses in the number of accurate details, inaccurate details, or "don't know" answers reported. All the participants demonstrated a misinformation effect, but there were no differences between intoxication levels: Intoxicated participants were not more susceptible to misinformation than sober or placebo participants. Results are discussed in the light of their theoretical and applied relevance.

Legal decision-making studies often demonstrate context effects: People's initial beliefs about a suspect's guilt influence their evaluation of subsequent evidence. We examine three potential moderators of these context effects: Order of evidence presentation, ability to ruminate, and valence of the initial belief (innocence or guilt). College students (n = 382) were presented with DNA evidence (incriminating or exonerating) and an ambiguous alibi in one of two orders (or just the alibi), and then evaluated how strongly the alibi incriminated the suspect and the suspect's likelihood of guilt. Results indicated that alibi evaluation exhibited context effects when (i) initial beliefs were of guilt (but not of innocence) and when (ii) evaluating subsequent evidence (but not when retroactively evaluating prior evidence). Rumination failed to moderate any effects. The integration of evidence exhibited recency effects: DNA had a greater impact on participants' beliefs in the suspect's guilt when presented last rather than first.

After controlling for initial confidence, inaccurate memories were shown to be more easily distorted than accurate memories. In two experiments groups of participants viewed 50 stimuli and were then presented with these stimuli plus 50 fillers. During this test phase participants reported their confidence that each stimulus was originally shown. This was followed by computer-generated responses from a bogus participant. After being exposed to this response participants again rated the confidence of their memory. The computer-generated responses systematically distorted participants' responses. Memory distortion depended on initial memory confidence, with uncertain memories being more malleable than confident memories. This effect was moderated by whether the participant's memory was initially accurate or inaccurate. Inaccurate memories were more malleable than accurate memories. The data were consistent with a model describing two types of memory (i.e., recollective and non-recollective memories), which differ in how susceptible these memories are to memory distortion.