The GPS data could be used to determine whether a person was in a stationary state, including at home or in the office, or in a transition state like walking in the street, the report said. It was also used to determine where participants spent most of their time.
Another factor that researchers looked at within the GPS data was to what extent a person’s movements followed a 24-hour or circadian rhythm. For example, a person who leaves for work and returns around the same time every day has a high level of circadian movement, according to the study.
Low levels of circadian movement are highly correlated with depression, the study showed.
Other features detected by Purple Robot included transition time spent traveling from one place to another, total distance moved, phone use frequency and phone use duration, the report said.
The report found that the average daily phone use duration among participants — whose average age was 28 years old — was 41 minutes, with daily frequency of use averaging 14.2 times. The average daily phone usage for depressed individuals was about 68 minutes, whereas it was about 17 minutes for non-depressed people.
"People are likely, when on their phones, to avoid thinking about things that are troubling, painful feelings or difficult relationships — it’s an avoidance behavior we see in depression,” Mohr said in the press release.
Also, people who had mild to severe symptoms of depression were found to move around less and have a less regular schedule, the report found. It added that their findings were consistent with the patterns of loss of motivation, decreased activity and social withdrawal that characterize depression.
“When people are depressed, they tend to withdraw and don’t have the motivation or energy to go out and do things,” Mohr said in the release.
The data on circadian movement, location variance, home stay, phone-use duration and phone-use frequency were “significantly different between participants with no signs of depression and the rest,” the report said.
By scoring individuals on these six features, researchers said they were able to estimate depression on unseen participants with 87 percent accuracy.