A new study reports that computer programs are able to use brain scans to distinguish between healthy individuals and those at risk of developing anxiety and depression. Furthermore, the research suggests that software may someday be able to predict exactly which at-risk adolescents will develop one of those diseases.
The research program is focused on early detection of psychiatric maladies because most such disorders first begin to manifest during adolescence, and it is then that treatments may be most effective at stopping or preventing them.
Currently, although a computer may be able to tell that a person with certain risk factors has a 15 percent chance of developing major depression, it cannot tell exactly which patient will finally manifest the disease. This largely stems from the lack of a single biomarker for each disease.
To judge the computer's ability to gauge mental illness risk, the computer analyzed 32 healthy adolescents—16 who had one bipolar parent, and 16 who had parents without mental illnesses. Each participant was required to label a series of faces as being either happy, neutral, or scared while being monitored by a functionalmagnetic resonance imaging device.
Using the MRI results, the computer was task with sorting the participants into groups of "healthy" versus "at-risk" individuals.
The computer was correct in its estimation of each person's individual risk approximately three fourths of the time. The highest rate of success was achieved using the scans taken when looking at neutral faces, confirming earlier research which suggested that people with mooddisorders tend to be ambivalent or fearful toward neutral faces.