For people with depression, finding a suitable antidepressant medication can be difficult and involve a lot of trial and error before finding one that works for them. Now a new study led by scientists at UT Southwestern has come up with a new imaging technique which the researchers claim will allow them to predict a person’s response to different types of antidepressant medication without sometimes having to spend months trying to find one that works.

The research first looked at the common antidepressant drug sertraline, one of a class of drugs called selective serotonin reuptake inhibitors (SSRIs), comparing people taking the drug to those taking a placebo. Before they started the medication, participants had their brains scanned in an MRI machine, both while they were resting and performing a reward task. This was repeated again after they had been on the drug or placebo for 8 weeks as well as measuring how their depression had changed, if at all. People who had not responded to sertraline after this time were switched to another antidepressant called bupropion and underwent the MRI tests and evaluation of their depression symptoms again after 8 weeks.

With this data from over 300 people, the researchers used machine learning techniques to map which brain regions and circuits where associated with a response to each drug, allowing them to predict how other people might respond in the future.