A researcher from Stanford University in California is developing artificial intelligence (AI) that can assist medical personnel in selecting the appropriate antidepressant for treating depression.

To determine the appropriate prescription for treating depression, researchers outfitted roughly 300 patients with electroencephalograms (EEGs) to measure electrical activity within the patients’ brains, building upon related research that suggests EEGs might be able to distinguish among different types of depression.

Once captured, the team fed that information to an advanced computer program that can discover how each type of depression would respond to specific medications.

The AI assigned the 300 patients either the anti-depressant Zoloft or a placebo and discovered that the AI effectively predicted which patients would best respond to the antidepressant medication.

Largely considered a guessing game, prescribing antidepressants is an imperfect process built on the assumption that there is just one type of depression. Yet, researcher have found that there are different types of depression. As such, researchers are learning that there may be more than one kind of medication appropriate for treating the different types of depression.

The research appears in the journal Nature Biotechnology.

In addition to making predictions about how prescribed antidepressants will treat certain types of depression, AI is being used by virtually every industry imaginable to make predictions about everything from premature death to crime.