Computers Can Diagnose Depression | Mental Health Blog

Date: 08/19/2017    Written by: Hiyaguha Cohen

Your Computer Knows if You’re Depressed

Computers Can Diagnose Depression | Mental Health Blog

You know how some people post photos on social media five times a day? And how those photos always show them smiling, doing something amazing, or eating something extraordinary? The photos might excite envy in their online friends, but a new study from Harvard University and the University of Vermont indicates that computers might be telling us a different story—one that we don’t seem to be able to discern.

The researchers analyzed 43,950 photos posted on Instagram by 166 users.1 So right away, you might suspect that something was a little off in this batch of subjects, because if you do the math, that translates to about 274 photos per user, on average. Assuming that some of the users posted way more than 274 photos to create that average, we’re talking about heavy posting activity. All the users were screened for depression at the outset of the study.  Seventy one of the 166 had a history of clinical depression. Another segment tested positive for depression but had no prior history.

The study didn’t analyze whether the subjects also posted on other sites, but the heavy Instagram use might make one wonder if those users had a big hole in their schedules that allowed them to linger online too long. Not that there’s anything wrong with posting to Instagram: one of the most functional people I know posts photos daily, but you can’t ignore the fact that plenty of studies indicate a correlation between too much time online and depression.1 Given that Instagram sees over 100 million new posts a day and now enjoys more daily signups than both Facebook and Twitter, it’s possible that at least some of those users are killing time playing on their computers instead of meeting friends for coffee, going to the gym, or cooking a real meal. In other words, instead of living real life, they’ve opted for virtual life.

In any event, the researchers programmed computers to analyze the user’s photos for color, tint, brightness, whether the photo was doctored with filters that affect hue, how many faces (if any) appeared in the photos, how many likes and comments the photos received, and how often the subject posted photos.2 The computers turned out to be great diagnosticians, correctly identifying depression 70 percent of the time. In contrast, mental health clinicians averaged only 40 percent accuracy in diagnosing depression in 118 similar previous studies.

What’s up with that? Six to eight years of graduate school and the clinicians couldn’t recognize depression anywhere near as well as the computers? All that money you spent to recline on the therapy couch while dredging up memories of your miserable childhood when you could just have easily asked your computer (if it had the researchers’ program installed), “Hal, am I depressed?”

In all seriousness, it is remarkable that a computer can tell what’s up with you in seconds. But on the other hand, there’s an irony in the fact that study after study shows our computers are making us less happy. And add to that the irony that research indicates that maybe we should rely on computers to tell us whether or not we are, in fact, happy, because professionally trained humans don’t do as good a job. 

And that’s the catch-22. Never mind that a Japanese study of 25,000 workers concluded that five hours or more of screen time led employees to dramatically increased rates of depression, anxiety and insomnia.3 Never mind that brain scans show that excessive screen time (for gaming, entertainment, and social activities) leads to gray matter atrophy in the brain.4 Yes, it really does! Never mind that numerous studies indicate that kids are experiencing elevated levels of anxiety and depression in direct relation to how much time they spend online. The fact is that our computers are so efficient, so accurate, so full of shortcuts, so ripe with fascinating information and entertainment—that we can’t resist them.

In the case of this study, the computers observed that the depressed photos were more likely to be blue or gray in tone, to contain an image of someone’s face, and to have fewer people in them compared to photos posted by non-depressed individuals.5 According to study author Dr. Christopher Danforth, “Our analysis of user accounts from a popular social media app revealed that photos posted by people diagnosed with depression tended to be darker in color, received more comments from the community, were more likely to contain faces, and less likely to have a filter applied. When they did select a filter, they were more likely to use the filter that converted color images to black and white. People diagnosed with depression also posted at a higher frequency compared to non-depressed individuals."

So much for those smiley face photos indicating actual happiness! Of course, 70 percent accuracy still leaves plenty of room for error, but on the other hand, it’s a bit unnerving to know that everything you post online tells more of a story than you might have intended. Just think: the confidentiality your therapist promises you could be blown apart by the happy photos on your Instagram page, should Big Brother decide to start analyzing those photos. If you start getting lots of pop-up ads for antidepressants, it might be time to shut down your social media accounts.

In the meantime, you might want to go spend extra time with actual people instead of rushing home to post photos of them. If the depression bug bites you anyway, you’d do well to consult Jon Barron’s suggestions for overcoming the blues. And don’t forget good old fashioned exercise—which beats drugs when it comes to depression.

  • 1. a. b. Reese, Andrew G. and Danforth, Christopher M. “Instagram photos reveal predictive markers of depression.” 8 August 2017. EPJ Data Science. 9 August 2017.
  • 2. “Computer program developed to spot signs of depression from social media photos.” 7 August 2017. Eureka Alert. 9 August 2017. https://www.eurekalert.org/pub_releases/2017-08/bc-cpd080317.php
  • 3. Olinka, Koster. “Why using a computer can cause depression.” Daily Mail. 9 August 2017. http://www.dailymail.co.uk/health/article-153281/Why-using-cause-depression.html
  • 4. Dunckley, Victoria. “Gray Matters: Too Much Screen Time Damages the Brain.” 27 February 2014. Psychology Today. 9 August 2017. https://www.psychologytoday.com/blog/mental-wealth/201402/gray-matters-too-much-screen-time-damages-the-brain
  • 5. Taylor, Eryn. “Are You Depressed? Better Check Your Instagram.” 9 August 2017. News Channel 3 9 August 2017. http://wreg.com/2017/08/09/are-you-depressed-better-check-your-instagram/
Tags:

Click for Related Articles

Comments

  •  
    Submitted by Daryl on
    August 30, 2017 - 9:45am
    Westerville , Ohio

    I'm not really understanding the study. So the computer program could connect a person's Instagram activity back to a previous diagnosis of depression 70% of the time, better than a doctor who is 40% accurate?

    Isn't the original diagnosis something that was done by a doctor (who was 40% accurate)?

    I'm really not understanding the premise of the study.

Add New Comment