This article originally appeared on YourTango, by Tom Miller
Well, this is kind of shocking...
I suppose periodically we have to hand it to Scientology: mental health science is really hard. So hard, in fact, that flipping a coin sometimes as steady-handed in diagnosis and treatment as $500K worth of student loans for a stay at BMS and The House Of God. But now that we have Instagram, we should be able to let people know if they're sad for once and all.
A pair of researchers from Harvard and the University of Vermont studied 166 individuals of whom some 71 had previously been diagnosed with depression. Analyzing the Instagram accounts of these people, the teams of researchers (Chris Danforth and Andrew Reece of University of Vermont and a little school in Boston, respectively) were able to find pinpoint the likelihood of which were depressed at a 70 percent clip compared to a much lower likelihood from a traditional practitioner (42 percent).
As this is the internet, we're supposed to be OK with merely reporting on something someone else is reporting and adding snarky comments and Game Of Thrones GIFs. However, we're going the extra mile today with how Instagram diagnoses depression.
The researchers looked for several "tells" when studying the nearly 44,000 photos. One of the first things they analyzed was filter hue. As you'd imagine, the hypothesis that brighter colors (yay Valencia filter!) indicated a lower likelihood of depression than darker ones.
Valencia was used by Instagram users who were least likely to be depressed:
Whereas people who used the Inkwell filter were MORE likely to suffer from depression:
And right in between least and most likely to be depressed were those who used the Maven filter:
Black & white images coincided with a greater likelihood of diagnosis, and why not? All of our pictures from The Great Depression are in black & white (or colorized in a painfully amateurish way). Plus, it certainly gives more of a tortured artist vibe than 80 degrees and sunny.
Photo composition was another correlative factor. Using a face-detection algorithm (much more useful than a hotdog-not-a-hotdog app), the researchers were able to analyze the IG posts for the presence of faces though not necessarily selfies. And this is where it gets interesting. Depressed individuals were MORE likely to have photos with faces but LESS likely to have photos with a lot of faces.
That is to say, the less sanguine group may have had one or two people in every single photo while the other group had some photos of funny road signs and some photos with a dozen smiling maniacs.
And the all important social vindication. Photos posted by diagnosed as depressed individuals generally had fewer "Likes" than those of individuals not diagnosed as depressed. You'd guess that the chicken-or-egg of this may be that so-called depressed participants may have a smaller social group even if it's on a digital platform. Or that people are happier because they're validated by friends and acquaintances. Or people just don't like sad bastard pictures. Either way, Instagram diagnoses depression according to this study.
Whew. OK, now to the comparative methodology. The researchers arrived at the 42 percent control number by using an aggregation of 118 studies of 50,371 patients.
Per the methodology, practitioners were able to correctly diagnose depression in 42% of depressed patients "without assistance from scales, questionnaires, or other measurement instruments." (The read is well, well worth it if you want to dig into the methodology even if the experimental group seems a little bit small.)
And now to make it somewhat personal, as is procedure with internet writing. I've fought with depression and am, publicly, a relatively optimistic character even if my girlfriend(s) have to hear from me how misunderstood my brilliance is. Yes, attempts at humor are likely sometimes an attempt to mask pain points and every other cliché we want to pitch at emotional health diagnoses.
But it's possibly slightly premature to use social media footprint as a guidepost for mental health, or else we'd seriously need to consider institutionalizing all YouTube commenters and anyone on Twitter that's not named Kevin Hart.