The research focuses on cortical morphometric similarity, a technique that quantifies how similar different regions of the brain’s outer layer are to one another across multiple structural features, including cortical thickness, surface area, gray matter volume, gyrification, and sulcal depth. When two regions share a similar structural profile, they are considered linked in a similarity network. This approach builds on the well-established principle that regions with similar structural signatures often belong to the same functional or developmental systems, allowing researchers to infer large-scale brain organization from structural magnetic resonance imaging alone.

Using this framework, the researchers constructed similarity networks for individuals with anorexia nervosa and for healthy comparison participants, then examined how the networks were organized at both local and global scales. At the local scale, they looked at the connectivity of individual nodes, asking whether specific regions showed stronger or weaker similarity relationships in the patient group. At the global scale, they assessed overall network properties, including how efficiently information could in principle be integrated across the cortex and how segregated specialized communities of regions remained from one another.

The results showed a coherent pattern of disruption. Rather than a diffuse or random alteration, anorexia nervosa was associated with systematic changes concentrated in cortical systems known to support interoception, reward processing, and cognitive control, including insular, frontostriatal, and parietal regions. These are precisely the circuits that have been implicated in the distorted body image, altered appetite signaling, and rigid, perseverative behavior that characterize the illness. The overlap between structural similarity alterations and previously identified functional disturbances lends weight to the idea that the disorder reflects a deep reorganization of cortical systems rather than a localized defect.

Perhaps the most provocative finding concerns the relationship between local and global organization. In the healthy brain, local similarity relationships scale up in a predictable way to produce global network properties, so that the whole cortex functions as an integrated but differentiated system. In anorexia nervosa, this scaling appeared to break down. The coupling between local node-level features and global network measures was attenuated, suggesting that the usual rules linking microstructural or regional organization to whole-brain architecture are disrupted in the disorder. The authors describe this as a local–global breakdown, a phrase that captures both the technical observation and its potential significance for understanding symptoms.

Why should such a breakdown matter clinically? One possibility is that it reflects a failure of hierarchical integration. Eating behavior depends on the brain’s ability to combine signals from many levels of organization: the fine-grained sensing of internal bodily states, the regional circuits that assign value to food, and the global networks that coordinate decisions, self-representation, and long-term goals. If local and global levels of cortical organization no longer align, the integration of these signals may become noisy or biased, producing the characteristic mixture of body-image distortion, diminished reward from eating, and inflexible behavior seen in patients.

The study also connects to a broader shift in psychiatry toward network neuroscience, which treats mental disorders as problems of brain network organization rather than dysfunction of single regions. Under this view, symptoms emerge from patterns of interaction among distributed neural systems. The morphometric similarity approach used here is particularly well suited to this perspective because it captures structural covariance across the entire cortex in a single model, and because it can be applied to the large datasets needed to detect reliable group differences. Anorexia nervosa, which has proven difficult to localize with traditional case-control comparisons, may be exactly the kind of disorder in which network-level measures reveal what region-level analyses miss.

Methodologically, the work demonstrates the value of multi-scale analysis. Many neuroimaging studies report either local differences, such as altered cortical thickness in one region, or global differences, such as changed whole-brain connectivity, but rarely test whether the two levels remain consistent with each other. By doing so, the authors were able to show that the disorder does not simply shift individual measurements up or down; it alters the relationship between scales of organization. This kind of finding is harder to produce through artifacts of head motion, medication, or scanner differences, strengthening confidence that the effect is a genuine feature of the illness.

The clinical implications are still early, but they are worth considering. If the local–global coupling of cortical similarity networks indexes the brain’s capacity for hierarchical integration, it could eventually serve as a biomarker for illness severity, trajectory, or treatment response. Longitudinal studies could test whether the breakdown precedes symptom onset, deepens with chronicity, or reverses with weight restoration and psychotherapy. Such work would be especially valuable in anorexia nervosa, where mortality rates are among the highest of any psychiatric disorder and where current treatments help only a proportion of patients.

The study leaves open important questions. Morphometric similarity is an indirect measure, derived from structural features rather than from direct measurements of microstructure or connectivity, and the cross-sectional design cannot establish causality between network reorganization and symptoms. It remains unclear whether the local–global decoupling is a cause of disordered eating, a consequence of malnutrition and starvation, or a shared vulnerability that predates the illness. Answering these questions will require longitudinal designs, adolescent cohorts, and convergence with functional and diffusion imaging. Even so, by showing that anorexia nervova involves a measurable disruption of the brain’s multi-scale architecture, the study moves the field closer to a mechanistic account of one of psychiatry’s most stubborn disorders, and it suggests that the answers may lie not in any single brain region but in the broken dialogue between local detail and global design.

Subject of Research: Cortical morphometric similarity network organization in anorexia nervosa

Article Title: Local–global breakdown of cortical similarity networks in anorexia nervosa

Article References: Facca, M., Meregalli, V., Gentili, S., Bertoldo, A., Manara, R., Favaro, A., & Collantoni, E. (2026). Local–global breakdown of cortical similarity networks in anorexia nervosa. Nature Mental Health. https://doi.org/10.1038/s44220-026-00730-5

Image Credits: AI Generated

DOI: 10.1038/s44220-026-00730-5

Keywords: anorexia nervosa, cortical similarity networks, network neuroscience, morphometric similarity, eating disorders, brain connectivity, local–global organization, neuroimaging, psychiatry, interoception, cognitive control, Local