• NSF-IIS-1816687: Computational Modeling of Human Rhythms to Improve Health and Quality of Life

    Recent research indicates balance is vital to health and well-being, however humans experience stress and imbalances through lack of attention to mental and physical states when meeting work and life demands. This research proposes to improve health and wellbeing in individuals through awareness of their personal rhythms, which are repeated cycles of internal and external events including biological, mental, social, and environmental. The investigators seek to design and evaluate a data analytic and modeling method to make users aware of potential activities at any given time that align with their biological clock, achieving higher performance in life and work without invoking stress-related disorders. The work leverages advances in wearable devices, sensing technology, and online sources to proactively collect and analyze physiological, psychological, behavioral, social, and environmental data to identify personal rhythms, to explore their relationship to positive physical, mental, and behavioral outcomes, and to provide people with tools to reason about these relationships and to improve outcomes. This research highlights the positive performance and outcomes resulting from integrating personal rhythms into individuals’ daily lives.

    We are currently recruiting participants interested in improving their sleep, productivity, and overall wellness by participating in our ongoing study on Human Rhythms. Our study focuses on measuring human rhythms through data collected by participants’ smartphones as well as through a FitBit fitness tracker. If interested, see below recruitment flyer for details, qualifications and how to join. 

    Related Publications: 

    "Modeling Biobehavioral Rhythms with Passive Sensing in the Wild: A Case Study to Predict Readmission Risk after Pancreatic Surgery"
    "Identifying Behavioral Phenotypes of Loneliness and Social Isolation with Passive Sensing: Statistical Analysis, Data Mining and Machine Learning of Smartphone and Fitbit Data”

  • MoodRing: Mobile Application Detecting Changes in Adolescent Depression

    Adolescent depression and suicide are increasingly prominent health issues that require thorough diagnosis, assessment and treatment initiatives by pediatric health care. Minimal reassessment and follow up care of adolescents’ depression can lead to worsening symptoms, increased costs to the child’s health, and increased healthcare utilization. The aim of this study is to use passive sensing and machine learning to monitor adolescent’s daily behavior patterns and provide constant indicators for changes in depressive symptoms. This study will develop a machine learning pipeline from adolescents’ smartphone sensor and activity tracker data and determine the reliability of using passive sensing to predict self-reported depressive symptoms. A mobile application, MoodRing, will be developed from the machine learning pipeline to display health feedback to adolescents and parents, provide just-in-time self-management interventions, and update healthcare professionals for more efficient clinical decision-making and patient care. 

  • Predicting the Probability of Readmission After Cancer Surgery with LSTM

    Hospital readmissions cost the US healthcare system billions of dollars annually, are associated with high mortality rates and are a source of stress and suffering for patients and family members. Traditional approaches to readmission risk stratification rely on static administrative and medical record data and generally classify all surgical oncology patients at high-risk. However, different factors related to daily behavior and activities may contribute to or signal increased or decreased risk of readmission. Our research utilizes mobile sensing and deep learning to measure daily readmission risk in cancer patients after discharge. Using data from mobile and Fitbit devices of 49 patients collected over 90 days after discharge from the hospital, we build a probabilistic model in an LSTM structure to infer the risk progression trajectory in each patient. Our results show that using only sensor data, the model can predict the risk progression trajectory aligned with the ground truth data.

  • Connected Steps

    Physical inactivity and social isolation are growing epidemics linked to increased morbidity and mortality, particularly among aging Americans. This research aims to address both problems by encouraging co-productive physical activities, specifically walking together. We aim to build an intelligent technological system that 1) connects people in the neighborhood at the right time and place by inferring daily activities and routines of individuals from smartphone sensors and recommending just-in-time co-productive physical activities via smartphones; 2) uses social incentives to motivate people for co-productive activity through recognizing and rewarding real-time activities of a group of people. We hypothesize that this system will increase participant physical activity and social connectedness and improve health.

  • Adaptive Humanoid Robots for Pain Management in Children

    EIM funded project

    Accurate pain assessment and management is particularly important in children exposed to prolonged or repeated acute pain including procedural pain because of elevated risk for adverse outcomes such as traumatic medical stress, intense pain response for subsequent pain and also developing chronic pain. Our current work in progress tries to help pain management in children through developing intelligent adaptive humanoid robots as a multi-modal non-pharmacological intervention. Our current work increases the interactive capabilities of Nao humanoid robots by using the camera and microphone to assess pain and emotion in children undergoing procedural treatment through combining detection models for facial expression, voice quality, and adapt the robot’s verbal and non-verbal interactive responses accordingly for optimal distraction through adaptive behavioral models. Using recognized emotion as an environment input to a reinforcement learning model, the robot acts as the agent to choose the best action out of a set of entertaining and distracting verbal and non-verbal actions to cheer up the child and distract them from the pain and fear of the medical procedure.