Late last week, a group of scientists published some pre-review research indicating that people were more likely to use low-saturation Instagram filters like Inkwell when feeling depressed. Well, obviously, says Picasso. Groundbreaking stuff, guys. Even if there’s a real association between filters and feelings, the paper’s data hasn’t been vetted by other scientist—so it’s still a bit premature to act like a keyboard psychiatrist whenever you see a friend’s selfie recreating Eric Draven’s rooftop guitar solo from The Crow.

Now, that is not to say it’s impossible to learn about a person’s emotional state from their feeds: Hell, the CIA has used Facebook, Twitter, and other social platforms to take the pulse of large groups of people. But social media is too self-moderated to be accurate for single-person diagnosis. “The social media stuff has an interesting potential, but it’s definitely a cultivated picture of a person’s life,” says Stephen Schueller, behavioral psychiatrist at Northwestern University’s Feinberg School of Medicine. On the other hand, smart phones track where you go, who you talk to, how you interact with those people, and all the things you do—and do not—do. With the right permissions, soon psychiatrists might use GPS, microphones, and other sensors to get real insights into a person’s mental wellbeing.

So far, location-tracking seems to be the most useful. “People whose movement through geographic space seemed to be more rhythmic had lower anxiety and depression levels,” says Schueller. As people get depressed, the normal patterns start to slip. “Maybe they forget an appointment, decide not to go out with their friends, travel to and from work at later or earlier hours,” he says. This kind of diagnostic location tracking is still in development, but Schueller says he imagines it might one day be linked to digital calendars, to see if people are missing scheduled appointments. With enough erratic behavior, the app could send out a notification asking if the person is okay.

Other researchers have looked at using phones’ microphones to detect behaviors that indicate depression. A new study from a group of Danish researchers used mics to monitor for signs. In this study, the mic stays on continuously, listening to the ambient noises around you, your conversations with other people, your speed of speech, and tone (people who are depressed tend to have slow, flat diction). Other groups have also used smart phone mics like this, but the Danes are the first to leave leave it on continuously, which makes it harder for the patient to consciously alter their behavior.

Now, if you’re anything like me you probably spent the last four paragraphs saying “But-but-but-whattabout if the government/hackers/cyberbullies/WikiLeaks/Google’s ad sales department gets ahold of this data?” Well, that’s a valid question. It’s probably most helpful to break those worries into two separate categories: Trust and security.

Trust is making sure that people who have access to the data don’t abuse it. For instance, you’d want to make sure Google wasn’t offering up your mental health state to your insurance company, or selling you ads for Zoloft every time you showed up late for work. Security is another issue entirely, and has to do with maintaining strong cryptologic standards to protect peoples’ personal data. So yeah, the potential for a malicious hacker to access and abuse another person’s mental health data is yet another reason why anyone who is not in favor of strong crypto is—apologies for the ableist term—batshit crazy.

Passive digital diagnosis is still pretty raw. “Our group has not had success at predicting mood on the fly, and I’m pretty skeptical of any findings that claim they do,” says Schueller. This recent Instagram study, in particular, is a little suspect. Not only did it not go through peer review, it relied on a relatively small number of users. Also, it gives non-mental health specialists an excuse to offer half-baked diagnoses based on a gloomy Instagram post or tweeted Jawbreaker lyric. Which is bad. Anyway, CIA memewatching centers aside, the most useful tools for mental health aren’t those that let you lifecast your emotions, but the ones that passively watch you go about your day.

Excerpt from: 

Instagram Probably Can’t Predict Depression. GPS, Though…