In today’s world, the researchers are focusing on your social media feeds and your search engine inputs to deduce your preferences and in this new case – your state of mind. A new algorithm can determine a person’s mental state and diagnose a patient’s depression based on his Instagram shares. Images that depict darker, sad or pessimistic moods are indicators of depression.
This algorithm can be of great help in diagnosing clinical depression. The doctors will be able to identify stressed or depressed users by running the algorithm through different Instagram posts. This will enable them to identify potential suicides and go for intervention before it’s too late.
Our first obvious question would be: How reliable is the algorithm? Scientists from the University of Vermont and Harvard have developed this algorithm after the quantum leaps made in establishing a correlation between mental state and image sharing.
“Using Instagram data from 166 individuals, we applied machine learning tools to successfully identify markers of depression. Statistical features were computationally extracted from 43,950 participant Instagram photos, using color analysis, metadata components, and algorithmic face detection. Resulting models outperformed general practitioners’ average diagnostic success rate for depression.” – Cornell University
The researchers reached out to about 500 workers in Amazon’s Mechanical Turk who were also Instagram users. The first step was to fill up questionnaires and surveys. Next, about 170 Turkers agreed to share their Instagram photos with the research team, and 70 of them were clinically depressed.
The Instagram feed gave the researchers a database of more than 40,000 photographs. These images were analyzed by crowdsourcing with a different group of Turkers. For healthy users, the team used last 100 photographs posted by each individual to be rated. For depressed users, 100 images were chosen before their diagnosis. The raters were asked to judge how happy/ likeable the photographs were on a scale of 0 – 5. Photographs were also evaluated using objective parameters like average hue, saturation, color, contrast etc.
The number of faces in each image were also counted as the study assumed that faces indicated a person’s level of social activity. The number of likes and comments on individual images were also taken into account. This data was then fed into a machine-learning algorithm to deduce a correlation between image properties and depression. It was found that depressed people shared darker, greyer images and get fewer likes than pictures posted by healthy people.
“These findings support the notion that major changes in individual psychology are transmitted in social-media use, and can be identified via computational methods,” says a researcher.
This way, depression can be detected earlier and timely intervention can be made. So what do you think of this algorithm that can diagnose depression by merely scanning your Instagram posts? Cool, right? Let us know what you think of it in the comments section
