October 1, 2013 — Jukka-Pekka “JP” Onnela, assistant professor of biostatistics at Harvard School of Public Health, has won a prestigious Director’s New Innovator Award from the National Institutes of Health (NIH) for a proposal to collect and analyze cell phone communication and sensor data to monitor social and behavioral functioning of individuals with mood disorders. One of 41 scientists around the country receiving the award in 2013, Onnela will receive $1.5 million over the course of five years.
“I have wanted to do this particular project for many years, but it has taken time for the technology to mature,” Onnela said. “The final missing piece was funding, and I am extremely grateful and deeply honored to receive this award from the NIH.”
The New Innovator Award is part of the NIH’s High Risk-High Reward program, which provides support for exceptional innovation in biomedical research. These awards are given to early-stage investigators working on highly creative research approaches that may be at too early a stage to qualify for more traditional NIH funding, but which have the potential to produce a major impact on broad, important problems in biomedical and behavioral research. Read the NIH press release.
Given the increasing worldwide prevalence of mental health problems—and the difficulty of precisely and consistently measuring people’s mood states, behaviors, and social functioning—Onnela, along with colleagues at HSPH and Massachusetts General Hospital, plans to develop a smartphone application and the requisite data analytic tools to help monitor individuals’ well-being.
The new application will enable people with mood disorders to digitally report on how they’re faring by answering questions on their phones. At the same time, the application will unobtrusively collect data on the individuals’ behavior and communication patterns—information that could help their doctors care for them.
Onnela and his colleagues will test the application over a two-year period using two cohorts—a group of psychiatric outpatients and a control group—and collect a range of data. First, they will look at individuals’ responses to questions posed by the smartphone application about mood and well-being. Second, they will use built-in accelerometer and GPS devices on the phones to track where individuals go and how much they move—providing information, for example, on how physically active the individuals are or how much they get of the house.
Measuring social interactions
Onnela and colleagues will also track the individuals’ use of other applications on their smartphones, with a focus on social networking applications. They’ll monitor, for instance, how and when the individuals are on Facebook, if they play online games with friends, or if they’re web browsing—all activities that can suggest a level of engagement with other people and with the world at large. The researchers will also examine call and text messaging logs—which the smartphone application makes anonymous—to learn about the individuals’ social activity.
Onnela noted the importance of protecting privacy in this line of research. “All communication metadata will be anonymized such that individuals will not be identifiable,” he said. “The data will be analyzed in the aggregate because our goal is to learn about generalizable statistical trends.”
Onnela and colleagues will develop analytical tools and mathematical models to integrate the data into a set of novel behavioral metrics. They think that these new metrics could paint a picture of how individuals with mood disorders are faring—and could provide a low-cost method to help monitor the efficacy of treatment.
“Traditionally it has been very difficult to accurately measure mood states, behaviors, and social functioning of individuals affected by mood disorders,” Onnela said. “Self-reported accounts of behavior are inaccurate for a number of reasons, and this has been a longstanding barrier to progress both in clinical settings and drug trials. We hope that our work will be a step in the direction of making various behaviors and symptoms more quantifiable.”
He added, “In the long run, this could improve the accuracy of diagnoses or make it easier to monitor the responsiveness of a patient to a particular treatment strategy.”
Onnela plans to share the smartphone application, and the statistical and mathematical tools he develops to analyze the data, at no cost to the scientific community.