New Algorithm Diagnoses Depression in People through their Instagram Images


A new machine has been created by researchers to diagnose depression and other mental illnesses by checking their Instagram photos.

The new study was conducted by Andrew Reece (Harvard University) and Chris Danforth (University of Vermont). The study suggests that a link exists between a person’s mental state and the colors that they use. The new research is similar to previous studies that say that insight regarding a person’s mental health and personal life can be obtained through their use of social media.

The researchers of the current study are of the belief that individuals who post grayer or images of darker color on their Instagram accounts are more likely to be depressed when compared to people who post images having bright and lively colors. The researchers were also able to create a machine that is capable of diagnosing depression in people by analyzing the images they post.


Due to the filters offered by Instagram people are able to change the colors of an image. An individual might prefer posting an already bright image; however, someone else might use a filter called ‘Inkwell’ to convert the same image into a black and white one. The results of the current study suggest that the Inkwell filter is used more by depressed people when compared to individuals who aren’t.

A total of 170 people were analyzed during the study. All of them owned Instagram accounts and were workers from Amazon’s Mechanical Turk service. All 170 individuals completed a questionnaire that also contained a survey for standard clinical depression. Participants of the study were also asked to share pictures they had posted on their Instagram accounts.

A hundred images were selected from each participant. The researchers then asked people to rate the images. The rating was on a scale of 0 to 5 and it was based on how happy, interesting or sad the images looked. The images were also categorized on the basis of the number of faces present in it, the hue and the saturation. All of the pictures were used in a machine-learning algorithm capable of seeing the correlation present between the type of image and depression.


According to the results of the algorithm, depression was predicted due to images having an increased hue and decreased brightness and saturation. The results also showed that the use of filters was different among depressed and non-depressed individuals in the sense that the use of filters was less likely in people who were depressed.

Depression was identified with a 70% success rate by the new algorithm. Keep in mind that it isn’t possible for any current technique to have a higher rate of success with regards to identifying depression. Even medical professionals and using clinical questionnaires aren’t a 100% accurate either.

“More generally, these findings support the notion that major changes in individual psychology are transmitted in social media use, and can be identified via computational methods,” said the researchers of the current study.

Both researchers, Reece and Danforth, believe that the technique they have created will help in a better understanding of mental illness. It can even detect depression in people at an early stage and lead to effective diagnosis.

Category: News

Leave a Reply

Your email address will not be published. Required fields are marked *

Article by: Douglas Norman

Douglas writes all the latest health news for BodyFatLoss. He is very diligent about finding all the facts and sources of any new health and fitness findings.