Understanding Ambulatory and Wearable Data for Health and Wellness

In our research, we aim (1) to recognize human internal states and behaviors (stress level, mood and sleep behaviors etc), (2) to reveal which features in which data can work as predictors and (3) to use them for intervention. We collect multi-modal (physiological, behavioral, environmental, and social) ambulatory data using wearable sensors and mobile phones, combining with standardized questionnaires and data measured in the laboratory. In this paper, we introduce our approach and some of our projects. Introduction and Our Approach Recently, we have so many devices to monitor our daily lives: pedometer, activity monitor, and sleep monitor etc. Many people wear them to quantify their personal behaviors; however, how can we use our collected data other than showing them in graphs? Our motivation to collect data is not only to visualize them but also to understand the meaning, recognize something internal behind the data (health condition or emotional states) and feedback them to users to help them to change their behaviors. Fig.1 Our approach to understand ambulatory data. In our studies, to understand human internal states and behaviors (sleep behaviors, stress, mood and performance level etc), we combine multi-modal ambulatory data from wearable sensors and mobile phones with data measured with gold standard methods in the laboratory or validated questionnaires (Fig. 1). Our interests lie in finding predictors for health and wellness and using them for treating or preventing disease or illness. In this paper, we will introduce some of our ongoing projects.