Biomarkers have transformed modern medicine but remain largely elusive in psychiatry, partly because there is a weak correspondence between diagnostic labels and their neurobiological substrates.

Like other neuropsychiatric disorders, depression is not a single disease but rather a syndrome encompassing co-occurring symptoms that have different responses to treatment.

Using brain imaging (fMRI specifically) in a large multisite sample, we showed that people with depression can be subdivided into four subtypes defined by specific patterns of brain connectivity in depression-related brain networks.

Using machine learning methods, we identified neuroimaging biomarkers for diagnosing depression subtypes with high accuracy rates, based solely on patient brain scans. Importantly, these subtypes also predict how depressed people will respond to a brain stimulation-based antidepressant treatment (transcranial magnetic stimulation), allowing us to identify which individuals are most likely to benefit from this treatment.

Our results defined novel subtypes of depression that go beyond current diagnostic boundaries and may be useful for identifying individuals who would most likely benefit from targeted neurostimulation therapies.