Bipolar disorder (BD) is associated with significant mortality and morbidity. It typically begins in adolescence or early adulthood, an important developmental period during which higher education, first jobs, and relationships are pursued. Recurrent mood episodes during this period can have a devastating impact on a young person's ability to achieve a high quality of life as an adult. A method by which to predict the onset of mood symptoms in adolescence would create an opportunity to intervene and reduce exposure to the harmful effects of recurrent episodes. A new approach ? digital phenotyping ? may make this possible. Digital phenotyping is defined as the ?moment-by-moment quantification of the human phenotype in situ? using data collected from smartphone sensors (accelerometer, texts, calls, GPS). Digital phenotyping has been used to identify mood changes and potential signs of relapse in adults with BD, but has not yet been applied to adolescents. We will use Beiwe, a digital phenotyping application for iOS and Android phones, to collect digital phenotypes from participants (aged 14-19) over 18-months (N=120; n=70 with BD [I, II, Other Specified], n=50 typically-developing). Over the follow-up period, participants will complete biweekly mood assessments, and both participants and caregivers will be interviewed monthly to track changes in mood/behavior. This will allow the phone sensor data collected with Beiwe to be closely linked to symptom changes.
The specific aims of this project are (1) to characterize the digital phenotype of BD symptoms in adolescents, (2) to describe differences in the digital phenotypes of the BD and typically developing groups, and (3) to develop a model for predicting mood symptoms prospectively. The proposed study is consistent with all four NIMH strategic objectives for the future of mental health research. This K23 Award will provide Anna Van Meter, PhD with the necessary training and mentorship to (1) gain proficiency in computational psychiatry by learning to analyze longitudinal data using statistical and machine learning techniques, (2) build expertise in patient-oriented translational research by designing and conducting a longitudinal study with youth participants; (3) learn to employ state-of-the-art mobile technology to personalize assessment and intervention using patient data. To accomplish these training goals, Dr. Van Meter has organized an outstanding mentorship team (Anil Malhotra, MD, Jukka-Pekka Onnela, DSc, John Kane, MD, Christoph Correll, MD, and Deborah Estrin, PhD), with expertise in patient-oriented research, technology-based mental health research, computational psychiatry, bipolar disorder in youth, and computer science. The proposed study will be the first to describe the digital phenotype of BD in adolescents, a population at great risk for the onset of BD as well as the damaging effects of repeated episodes. The completion of the proposed K23 Mentored Career Award will support an innovative program of patient-oriented research, and will provide Dr. Van Meter with the skills necessary to become an independent investigator pursuing novel technological solutions to improve patients' quality of life.
Bipolar disorder typically begins in adolescence; recurrent mood symptoms during this period can have a devastating impact on a person's ability to achieve a high quality of life. Digital phenotyping, defined as the ?moment-by-moment quantification of the human phenotype in situ? using data collected from smartphone sensors, may make it possible to better understand and predict the onset of mood symptoms with minimal burden to the individual. The goal of this K23 proposal is to characterize the digital phenotype of bipolar disorder in adolescents in order to prospectively predict mood symptoms, creating an opportunity to reduce the harmful effects of recurrent episodes.
