A new study published in npj Mental Health Research reports that a specific brain-network signal may reliably predict whether a person with major depression will respond to antidepressant treatment.
Major depressive disorder affects millions worldwide, yet doctors still lack tools to determine which patients will benefit from antidepressants. Current treatment is largely trial-and-error, often requiring months before knowing whether a medication will work.
Scientists have long suspected that the brain’s “default mode network”—a system active during self-reflection and rumination—plays a central role in depression. But until now, no study had convincingly shown that patterns within this network could predict treatment outcomes.
The research team, led by Kaizhong Zheng and Liangjun Chen, set out to test whether communication between the medial prefrontal cortex (mPFC) and the posterior cingulate cortex (PCC)—two hubs of the default mode network—could serve as such a predictor. These regions are known to be involved in self-focused thinking and emotional regulation, both of which are disrupted in depression.
To investigate this, the research team analyzed resting-state brain scans from a total of 4,271 participants across four datasets. The largest of these cohorts included 2,142 people diagnosed with major depression and 1,991 healthy individuals.
The sample included both first-episode patients who had never taken antidepressants and those with recurrent depression. Additional datasets followed patients undergoing antidepressant medication or repetitive transcranial magnetic stimulation (rTMS), allowing the team to examine how brain connectivity correlated with treatment.
Using a technique called Granger causality analysis, the team measured the directional flow of information from the mPFC to the PCC. They found that people with recurrent depression had significantly reduced connectivity compared with both healthy participants and those experiencing their first depressive episode who had not taken any antidepressant medication. This reduction also correlated with longer illness duration and prior antidepressant use.
Most strikingly, the pre-treatment baseline signal predicted who would improve with therapy. The researchers noted that successful antidepressant treatment actually decreased mPFC-to-PCC connectivity. More importantly, machine-learning models trained on a patient’s baseline connectivity measure were able to distinguish future responders from non-responders with high accuracy before treatment even began.
The baseline connectivity measure was also linked to eventual treatment improvement rather than to the initial severity of core depressive symptoms, such as anhedonia or suicidal thinking, suggesting it reflects a treatment-specific mechanism rather than general illness severity.
Zheng and Chen concluded: “Despite the known pivotal role of the DMN in various cognitive and emotional processes, it has not yet been targeted for therapeutic intervention. Our study reveals a significant association between the DMN and treatment outcomes, providing strong evidence for the feasibility of DMN-targeted interventions.”
Despite its promise, the study has limitations. For instance, the study examined only antidepressant medication and rTMS, while other treatments with distinct mechanisms—such as electroconvulsive therapy (ECT) or psychotherapy—were not included and may show different patterns of brain connectivity.
The study, “Beyond depression symptoms: the default mode network as a predictor of antidepressant response,” was authored by Kaizhong Zheng, Liangjun Chen, Huaning Wang, the DIRECT consortium, Baojuan Li, and Badong Chen.