For once, social media could impact your mental health for the better.
Feeling gloomy? Instagram feels you.
It’s no secret that excessive social media use can leave you feeling blue, but picture posts can now be used to predict if you’re entering a wave of depression, American researchers have found.
Machine learning can today analyse the colours, faces, and metadata (time, location, or ‘likes’, etc) of your social media posts, to correctly pinpoint markers of low mood.
These computer results are so accurate, they can even outperform GPs in reaching the correct diagnosis.
Numerous studies have previously shown that text posts (on Facebook and Twitter, for example) can reveal information about users’ mental health. (Facebook even has it’s own suicide alert system).
But with almost 100 million new posts per day – and surging growth rates – scientists from Harvard and Vermont universities decided that its Instagram and image analysis that will be key in improving our future health.
Psychologist Andrew Reece and computer scientist Christopher Danforth programmed a robot to trawl through a whopping 43,950 images from 166 social media users looking for clues that could indicate poor mental wellbeing.
“Are there people present? Is the setting in nature or indoors? Is it night or day?” the researchers ask in the paper.
“Did the photo receive any comments? How many ‘Likes’ did it get?”
Markers of depression analysed included how frequently you post (depressed people post more), and how many different faces crop up in your images (depressed people have a lower average face count per photo).
Depressed posters tended to get fewer likes, but more comments, say the researchers.
They are also less likely to apply Instagram filters to their posted photos (but if they did could be spotted by a preference toward black and white images, and darker blue-toned filters).
The computer was able of noticing these markers, prior to formal medical diagnosis.
In the study, more than half of GPs depression diagnoses were false positives, while the majority of the computerised diagnoses were correct.
“False diagnoses are costly for both healthcare programs and individuals,” the researchers observe. With this in mind, computer analysis could better identify depressed individuals and improve quality of care.
“Given that mental health services are unavailable or underfunded in many countries, this computational approach, requiring only patients’ digital consent to share their social media histories, may open avenues to care which are currently difficult or impossible to provide” they write.
“These results suggest new avenues for early screening and detection of mental illness.”
Already the researchers are hoping to improve their robot diagnoses further by including textual analysis of Instagram posts’ comments, captions, and tags.
You might appreciate seeing a human face when you visit the doctors, but it’s computer wizardry that could help improve the world’s mental health.
Finally, social media being put to good use.