StudentLife app, sensing, and analytics system architecture (credit: Rui Wang et al.)
Northwestern scientists believe an open-access android cell phone app called Purple Robot can detect depression simply by tracking the number of minutes you use the phone and your daily geographical locations.
The more time you spend using your phone, the more likely you are depressed, they found in a small Northwestern Medicine study published yesterday (July 15) in the Journal of Medical Internet Research. The average daily usage for depressed individuals was about 68 minutes, while for non-depressed individuals it was about 17 minutes.
Another factor was your location. Spending most of your time at home and most of your time in fewer locations — as measured by GPS tracking — also are linked to depression.
In addition, having a less regular day-to-day schedule, leaving your house and going to work at different times each day, for example, also is linked to depression.
Based on those three factors, they claim they could identify which of 28 individuals they recruited from Craig’s List had depressive symptoms — based on a standardized questionnaire measuring depression called the PHQ-9 — 87 percent accuracy.
Example phone usage data from a participant. Each row is a day, and the black bars show the extent of time during which the phone has been is use. The bars on the right side show the overall phone usage duration for each day. (credit: Sohrab Saeb et al./Journal of Medical Internet Research)
“The significance of this is we can detect if a person has depressive symptoms and the severity of those symptoms without asking them any questions,” said senior author David Mohr, director of the Center for Behavioral Intervention Technologies at Northwestern University Feinberg School of Medicine. “We now have an objective measure of behavior related to depression. And we’re detecting it passively. Phones can provide data unobtrusively and with no effort on the part of the user.”
Better than questionnaires
The smartphone data was more reliable in detecting depression than daily questions participants answered about how sad they were feeling on a scale of 1 to 10. Those answers may be rote and often not reliable, said lead author Sohrob Saeb, a postdoctoral fellow and computer scientist in preventive medicine at Feinberg.
“The data showing depressed people tended not to go many places reflects the loss of motivation seen in depression,” said Mohr, who is a clinical psychologist and professor of preventive medicine at Feinberg. “When people are depressed, they tend to withdraw and don’t have the motivation or energy to go out and do things.”
The research could ultimately lead to monitoring people at risk of depression and enabling health care providers to intervene more quickly, they suggest.
While the phone usage data didn’t identify how people were using their phones, Mohr suspects people who spent the most time on them were surfing the web or playing games, rather than talking to friends. “People are likely, when on their phones, to avoid thinking about things that are troubling, painful feelings or difficult relationships,” Mohr said. “It’s an avoidance behavior we see in depression.”
That assumption seems questionable; non-depressed people often spend time on phones texting, checking Facebook, reading, emails, etc.
But Saeb also analyzed the GPS locations and phone usage for 28 individuals (20 females and eight males, average age of 29) over two weeks. The sensor tracked GPS locations every five minutes.
To determine the relationship between phone usage and geographical location and depression, the subjects took a widely used standardized questionnaire measuring depression, the PHQ-9, at the beginning of the two-week study. The PHQ-9 asks about symptoms used to diagnose depression such as sadness, loss of pleasure, hopelessness, disturbances in sleep and appetite, and difficulty concentrating. Then, Saeb developed algorithms using the GPS and phone usage data collected from the phone, and correlated the results of those GPS and phone usage algorithms with the subjects’ depression test results.
Of the participants, 14 did not have any signs of depression and 14 had symptoms ranging from mild to severe depression.
The goal of the research is to passively detect depression and different levels of emotional states related to depression, Saeb said. The information ultimately could be used to monitor people who are at risk of depression to, perhaps, offer them interventions if the sensor detected depression or to deliver the information to their clinicians. Future Northwestern research will look at whether getting people to change those behaviors linked to depression improves their mood.
“We will see if we can reduce symptoms of depression by encouraging people to visit more locations throughout the day, have a more regular routine, spend more time in a variety of places or reduce mobile phone use,” Saeb said.
In addition to studies that use mobile phone sensor data to better understand depression, Mohr’s team also is running clinical trials to treat depression and anxiety using evidence-based interventions.
Contact ehealth@northwestern.edu or 855-682-2487 to learn how to join one of their paid research studies, or visit http://cbitshealth.northwestern.edu/.
This research was funded by research grants from the National Institute of Mental Health of the National Institutes of Health.
Abstract of Mobile Phone Sensor Correlates of Depressive Symptom Severity in Daily-Life Behavior: An Exploratory Study
Background: Depression is a common, burdensome, often recurring mental health disorder that frequently goes undetected and untreated. Mobile phones are ubiquitous and have an increasingly large complement of sensors that can potentially be useful in monitoring behavioral patterns that might be indicative of depressive symptoms.
Objective: The objective of this study was to explore the detection of daily-life behavioral markers using mobile phone global positioning systems (GPS) and usage sensors, and their use in identifying depressive symptom severity.
Methods: A total of 40 adult participants were recruited from the general community to carry a mobile phone with a sensor data acquisition app (Purple Robot) for 2 weeks. Of these participants, 28 had sufficient sensor data received to conduct analysis. At the beginning of the 2-week period, participants completed a self-reported depression survey (PHQ-9). Behavioral features were developed and extracted from GPS location and phone usage data.
Results: A number of features from GPS data were related to depressive symptom severity, including circadian movement (regularity in 24-hour rhythm;r=-.63, P=.005), normalized entropy (mobility between favorite locations; r=-.58,P=.012), and location variance (GPS mobility independent of location; r=-.58,P=.012). Phone usage features, usage duration, and usage frequency were also correlated (r=.54, P=.011, and r=.52, P=.015, respectively). Using the normalized entropy feature and a classifier that distinguished participants with depressive symptoms (PHQ-9 score ≥5) from those without (PHQ-9 score <5), we achieved an accuracy of 86.5%. Furthermore, a regression model that used the same feature to estimate the participants’ PHQ-9 scores obtained an average error of 23.5%.
Conclusions: Features extracted from mobile phone sensor data, including GPS and phone usage, provided behavioral markers that were strongly related to depressive symptom severity. While these findings must be replicated in a larger study among participants with confirmed clinical symptoms, they suggest that phone sensors offer numerous clinical opportunities, including continuous monitoring of at-risk populations with little patient burden and interventions that can provide just-in-time outreach.
