Clinical neuroscience has long focused on identifying neurobiological differences between people with major depressive disorder (MDD) and healthy people. Recent meta-analyses have raised questions about the reproducibility and therapeutic applicability of brain changes in depression. For a study, researchers sought to compare the results with an MDD polygenic risk score (PRS) and environmental factors, as well as to quantify the upper bounds of univariate effect sizes, estimated predictive utility, and distributional dissimilarity between healthy people and people with depression on structural magnetic resonance imaging (MRI), diffusion-tensor imaging, and functional task-based as well as resting-state MRI.

A case-control, cross-sectional clinical neuroimaging research was conducted. The Marburg-Münster Affective Disorders Cohort Study provided the data. In Münster and Marburg, Germany, primary care facilities and the general community were used for enlisting depressed patients and healthy controls. Study recruiting was carried out from September 11, 2014, until September 26, 2018. The sample included 18 to 65-year-old healthy controls and patients with acute and chronic MDD. Data analysis was done between October 29, 2020, and April 7, 2022. For healthy people and those with depression, the primary analyses were univariate partial effect size (η2), classification accuracy, and distributional overlapping coefficient across neuroimaging modalities while adjusting for age, sex, and extra modality-specific confounding factors. Subgroups of acute or persistent depression patients were included in secondary analyses.

In all, 1,809 people (861 patients [47.6%] and 948 controls [52.4%]; mean [SD] age, 35.6 [13.2] years; 1,165 female patients [64.4%]) were examined. The distributions overlapped between 87% and 95%, classification accuracies varied between 54% and 56% across neuroimaging modalities, and the upper bound of the effect sizes of the single univariate variables indicated the biggest group difference ranged from partial η2 of 0.004 to 0.017. When only individuals with acute or chronic depression were considered, the trend remained essentially unaltered. Although differences were significantly less for environmental factors, they were equivalent to those identified for PRS.

According to the case-control study’s findings, even in the presence of the greatest biological univariate differences, the differences between patients with MDD and healthy controls were remarkably small, the single-participant prediction was not feasible, and similarity between study groups predominated. To boost the possibility of a clinical practice being personalized, biological psychiatry should provide meaningful outcome metrics or prediction techniques.