How Broken Brain Networks and Molecular Chaos Drive Major Depression

Major depressive disorder affects roughly 280 million people worldwide, yet its diagnosis still rests almost entirely on subjective symptom checklists. A new integrative review published in Frontiers in Integrative Neuroscience argues that this diagnostic fragility is not merely a clinical inconvenience but a direct consequence of the disorder’s staggering biological complexity. Drawing together evidence from neuroimaging, electrophysiology, genetics, and molecular studies, the review builds the case that depression is a condition of both molecular dysregulation and large-scale brain network dysfunction — while candidly acknowledging that the direct evidence connecting these two levels remains surprisingly thin.

The clinical picture of major depressive disorder extends far beyond persistent low mood. Core diagnostic symptoms include anhedonia, appetite and weight changes, sleep disturbances, fatigue, feelings of worthlessness, difficulty concentrating, and recurrent thoughts of death or suicide. But the review emphasizes that these features coexist with broader systemic effects involving the cardiovascular, immune, gastrointestinal, endocrine, and neurological systems. This multisystem impact, combined with a multifactorial etiology spanning genetic susceptibility, hypothalamic–pituitary–adrenal (HPA) axis dysregulation, chronic inflammation, oxidative stress, early-life adversity, and stressful life events, produces the marked biological and clinical heterogeneity that has frustrated attempts at a unified biological model.

At the systems level, the review highlights three large-scale brain networks whose disruption appears central to the disorder: the default mode network, the salience network, and the central executive network. Functional MRI studies have repeatedly reported abnormal connectivity within these distributed systems, alongside altered regional activity in prefrontal–limbic and reward-related circuits involved in emotional processing, cognitive control, and reward. Structural MRI has revealed abnormalities in gray matter volume, cortical thickness, and white matter integrity, particularly within prefrontal-limbic circuits, with reported changes in the hippocampus, prefrontal cortex, anterior cingulate cortex, amygdala, and thalamus. Electroencephalography adds temporal resolution, revealing alterations in oscillatory activity, neural synchronization, and microstate dynamics that complement the spatial picture painted by imaging.

Crucially, the review introduces a rigorous distinction that is often lost in popular coverage of depression neuroscience: the difference between converging evidence and direct multimodal evidence. Most molecular and neuroimaging findings originate from separate study populations, meaning that apparent correspondence between, say, reduced peripheral BDNF levels and hippocampal volume reductions should be interpreted as converging rather than demonstrating a direct molecular-to-brain relationship. Only a limited number of studies have simultaneously assessed molecular or genetic measures together with neuroimaging outcomes in the same participants, and the review’s evidence matrix makes this gap explicit across nearly every biological pathway examined.

Among the pathways with the strongest direct multimodal support is the HPA axis. The review notes that a small number of studies have directly linked serum cortisol concentrations with cortical thinning, white matter integrity changes, hippocampal shape abnormalities, and structural brain network alterations in first-episode, drug-naïve patients. One longitudinal neuroimaging study found that normalization of HPA axis dysfunction following pharmacotherapy was accompanied by changes in neural activity within prefrontal and limbic/paralimbic regions. However, HPA axis abnormalities remain heterogeneous across clinical subtypes, early-life stress exposures, and treatment statuses, and the broader correspondence between endocrine dysfunction and stress-related brain changes still rests largely on converging evidence from separate investigations.

Neurotrophic signaling, centered on brain-derived neurotrophic factor (BDNF), presents a more complicated picture. Meta-analytic evidence supports lower peripheral BDNF during acute depressive episodes and partial restoration following successful antidepressant treatment, but the review cautions that circulating BDNF cannot be treated as a direct surrogate of central BDNF activity. Direct multimodal studies are scarce: serum BDNF has been associated with regional cortical thickness and gray matter volume, BDNF promoter methylation with cortical thickness in recurrent depression, and the BDNF Val66Met polymorphism with structural covariance and resting-state connectivity during adolescence. Yet imaging-genetic findings remain inconsistent, and the biological consequences of BDNF dysregulation appear highly region-specific — reduced signaling in hippocampus and prefrontal cortex is associated with impaired plasticity, while increased activity in mesolimbic reward circuits can promote stress susceptibility in experimental models.

Neuroinflammation emerges as another major axis, but one where the evidence is even more fragmented. Meta-analyses generally report elevated pro-inflammatory cytokines such as IL-1β, IL-6, IL-18, and TNF-α in depression, yet inflammatory profiles vary substantially across patient subgroups, with elevations observed only in a subset of individuals. Direct multimodal studies linking peripheral inflammatory markers with brain network organization are limited but suggestive: blood inflammatory markers have been associated with functional network dysconnectivity, and peripheral inflammation with altered structural and functional connectivity of depression-related networks. For intracellular inflammatory cascades such as NF-κB, NLRP3 inflammasome signaling, and JAK/STAT pathways, the review finds no direct multimodal studies assessing these specific pathways together with neuroimaging measures in MDD, leaving their relationship to observed brain alterations largely inferential.

Microglial activation, the resident immune response within the brain, illustrates the interpretive challenges particularly well. PET imaging studies using translocator protein (TSPO) binding have reported increased signal in the anterior cingulate cortex, prefrontal cortex, and insula during major depressive episodes in some investigations, while others have detected no significant differences or even lower binding in patients. Postmortem findings have also been inconsistent, with several studies reporting no significant changes in microglial markers or suggesting predominantly homeostatic rather than pro-inflammatory phenotypes. The review emphasizes that TSPO binding is not specific to microglia or to any particular activation state, meaning PET findings should be read as evidence of altered glial-related neuroimmune activity rather than direct proof of a specific microglial phenotype.

Neurotransmitter systems receive perhaps the most nuanced treatment. The classical monoamine hypothesis — that depression stems from serotonin, norepinephrine, or dopamine deficiency — is acknowledged as foundational for antidepressant development but insufficient to explain clinical heterogeneity or delayed therapeutic response. Recent umbrella reviews have questioned whether serotonergic alterations alone are sufficient to explain depression’s pathophysiology. Molecular neuroimaging using PET and SERT tracers has revealed altered transporter availability, but systematic reviews report substantial heterogeneity. Dopaminergic dysfunction, linked to anhedonia and reward processing, has somewhat stronger direct multimodal support, with combined PET-fMRI studies associating altered striatal dopaminergic measures with abnormalities in striatal connectivity and reward circuitry. Glutamatergic and GABAergic abnormalities, assessed largely through proton magnetic resonance spectroscopy, show region- and stage-dependent variability, and peripheral glutaminase has been found not significantly associated with central glutamate levels — a null result the review treats as an important caution against assuming peripheral biomarkers reflect central processes.

Oxidative stress and mitochondrial dysfunction round out the molecular picture, with the Keap1-Nrf2-ARE antioxidant pathway implicated largely through preclinical evidence. One multimodal study reported that plasma antioxidant enzyme levels were associated with altered spontaneous brain activity and thalamic functional connectivity in first-episode, drug-naïve patients, though cross-sectional design precludes causal inference. The review concludes that many of the reported molecular and network alterations appear dynamic and potentially reversible, highlighting neuroplasticity as a key mechanism associated with therapeutic response. Future progress, it argues, depends on large-scale, multimodal, longitudinal studies integrating neuroimaging with molecular, genetic, and epigenetic profiling, supported by network-based models and machine learning approaches that may improve patient stratification, diagnostic accuracy, and prediction of treatment response — ultimately advancing toward a biologically informed precision psychiatry.

Subject of Research: Neurobiological mechanisms and brain network dysfunction in major depressive disorder: an integrative review

Article Title: Neurobiological mechanisms and brain network dysfunction in major depressive disorder: an integrative review

Article References: Iskakova, D., Yener, B., Kamzanova, A., Sitakanova, A., Hamoudi, R., & Kustubayeva, A. (2026). Neurobiological mechanisms and brain network dysfunction in major depressive disorder: an integrative review. Frontiers in Integrative Neuroscience, 20, Article 1904444. https://doi.org/10.3389/fnint.2026.1904444

Image Credits: AI Generated

DOI: 10.3389/fnint.2026.1904444

Keywords: major depressive disorder, brain networks, neuroinflammation, HPA axis, BDNF, precision psychiatry, Neurobiological, mechanisms, brain, network, dysfunction, major