The 'shrunken hippocampus' story of depression doesn't show up in 23,417 scans

For more than a decade, depression has been linked in both the scientific literature and the popular press to a shrunken hippocampus, the brain's main memory centre. The largest brain-scan study on the question to date, published this month in Nature Mental Health, finds that link does not hold up. The signal that does show up points to a different map of the brain: not the memory hub, but the regions that govern judgement, motivation, and how the body registers pain and emotion.

The study, led by Janine Bijsterbosch at Washington University School of Medicine in St Louis, pulled together 23,417 structural brain scans from six large-scale US and UK datasets, including UK Biobank, a long-running British health research project that has recruited more than half a million volunteers. Depression was assessed using questionnaires rather than clinician interviews, and participants skewed toward mild-to-moderate symptoms. Within those limits, the team asked a narrow question: do brain regions that prior work flagged as depression-related actually differ in volume or surface area between people with and without depression?

The effect that a 2015 study of roughly 9,000 people tied to recurrent depression did not show up in the new, larger sample. Bijsterbosch's team reports no significant link between depression and hippocampal volume across the combined dataset.

The regions that did show smaller volume or reduced cortical surface area in people with higher depression scores sit further forward and to the side of the brain: the frontal cortex, the anterior cingulate, and the insula. Each of these has a specific job the study's findings make newly legible.

The frontal cortex handles executive function: the ability to plan, inhibit impulses, and weigh long-term consequences. The anterior cingulate is involved in decision-making, motivation, and detecting when outcomes diverge from expectations. The insula processes internal body states: pain, heart rate, gut feeling, and the awareness of emotion as it is being felt. Smaller volume in these areas lines up with the symptoms patients describe most often: difficulty concentrating and making decisions, loss of motivation, and the somatic weight of mood disorders, from heaviness in the chest to dulled appetite.

The New Scientist report on the study notes that Bijsterbosch points to a unifying mechanism. The three affected regions have high neuroplasticity, the brain's capacity to remodel its connections in response to experience. That same plasticity also makes them more vulnerable to the stress and hormonal changes that accompany depression, which would explain why their volume shrinks while the more stable hippocampus does not. Stress, in this reading, leaves its footprint on the brain's most malleable tissue rather than its most-cited memory structure.

Three caveats in the study itself limit how far the conclusion travels. Depression was measured at a single time point, while the 2015 study it implicitly challenges tracked people across multiple episodes. Most participants had mild-to-moderate symptoms; no large dataset in the meta-analysis used gold-standard clinician diagnostics. Bijsterbosch and her coauthors note that hippocampal effects may still appear in recurrent or severe depression, where stress exposure accumulates over years rather than months.

John Fleming, a clinician at the Southern Colorado TMS Center who treats depression with transcranial magnetic stimulation, called the null hippocampal finding "surprising" in a separate write-up on Medical Xpress. Fleming's takeaway is that the field should not treat hippocampal shrinkage as a given. "We still need to learn more about all brain regions and not take hippocampal shrinkage as a given," he said.

The new study displaces the assumption that one region, the hippocampus, carries the whole story. For researchers, animal models of hippocampal stress may need to be paired with work on the more plastic cortical and insular regions. For clinicians and patients, the new map offers a more useful mental model. The symptoms that feel most like depression, brain fog, lost motivation, a body that registers pain and emotion differently, sit in the regions this study actually flags.

Bijsterbosch's team stops short of clinical advice. The next step, they write, is to repeat the analysis on samples with clinician-confirmed diagnoses and to track brain changes over time rather than at a single scan.