In a groundbreaking study published in Schizophrenia (2025), Janssen, Guil Gallego, Díaz-Caneja, and colleagues have unveiled profound insights into the heterogeneity of morphometric similarity networks (MSNs) within both healthy individuals and those diagnosed with schizophrenia. This research marks a pivotal stride in understanding the intricate architecture of brain structure variability and its implications for mental health disorders. Morphometric similarity networks, which map correlations between anatomical brain features, have become a frontier for neuroscientists aiming to decode the complex organization of the brain’s structure-function relationships. The findings from this study compel us to reconsider the interplay between brain morphology and psychiatric conditions, especially schizophrenia, illuminating both typical developmental trajectories and pathological deviations.
The concept of morphometric similarity networks revolves around measuring the resemblance of morphometric features such as cortical thickness, surface area, and gray matter volume across different regions of the brain. This methodology transcends traditional volumetric analysis by providing a network-based perspective on brain morphology. The researchers utilized high-resolution magnetic resonance imaging (MRI) data to construct MSNs, quantifying the degree of similarity between brain regions and thereby creating individualized brain network maps. These networks offer a refined understanding of how structural variability correlates with cognitive and clinical phenotypes, reflecting underlying neurobiological mechanisms that might otherwise remain concealed.
Importantly, the study delves into the heterogeneity—a term denoting variability and diversity—in the organization of these MSNs in both healthy and schizophrenic populations. Schizophrenia, a complex psychiatric disorder characterized by disrupted thought processes, emotional dysregulation, and cognitive deficits, has long eluded definitive neuroanatomical characterization due to its highly heterogeneous presentation. The authors hypothesize that rather than a uniform disruption, schizophrenia features multiple patterns of altered morphometric similarity, each potentially corresponding to different symptom profiles or disease trajectories. This heterogeneity challenges the prevailing one-size-fits-all model of brain dysfunction in schizophrenia and underscores the urgency of personalized approaches to diagnosis and treatment.
The methodology employed by Janssen et al. was meticulously designed to capture this nuanced heterogeneity. Participants included both healthy controls and individuals diagnosed with schizophrenia, whose structural MRI scans were processed using advanced morphometric techniques. Each participant’s brain was parcelled into anatomically defined regions, and multiple morphometric features were extracted. These features were correlated pairwise across regions to generate the MSNs, which were then subjected to rigorous statistical analysis, including measures of network topology such as modularity, hierarchy, and small-worldness. This approach allowed the authors to map the distinctive patterns of morphometric similarity that delineate healthy from schizophrenic brain networks.
One of the most striking findings in the study is the degree of inter-individual variability found in the morphometric similarity networks within the healthy population itself. Contrary to earlier assumptions of relative uniformity, healthy brains exhibited considerable differences in network organization, reflecting the diverse cognitive and behavioral phenotypes manifest in the general population. This intrinsic heterogeneity in healthy MSNs suggests a dynamic developmental process influenced by genetic, epigenetic, and environmental factors, all contributing to a unique neuroanatomical fingerprint in every individual. Thus, the brain’s structural network is not static but a reflection of lifelong adaptations.
When comparing these patterns to those observed in schizophrenia, Janssen and colleagues identified both overlapping and distinct alterations in morphometric similarity networks. While some schizophrenic patients exhibited globally reduced similarity suggestive of widespread cortical disruptions, others demonstrated focal alterations concentrated in regions implicated in cognitive control, sensory processing, and emotional regulation. These region-specific disruptions resonated with clinical symptom clusters, such as negative symptoms or hallucinations, highlighting how MSN profiles could serve as biomarkers for disease subtypes. The multidimensional nature of these network disruptions supports a reconceptualization of schizophrenia as a constellation of neuroanatomical phenotypes rather than a monolithic disease entity.
The study also explores the potential neurodevelopmental origins of these heterogeneous morphometric similarity patterns. Given that schizophrenia often emerges in late adolescence or early adulthood, a period marked by significant synaptic pruning and brain maturation, changes in MSNs likely reflect both aberrant developmental processes and subsequent neurodegeneration. The authors discuss how disruptions in early cortical development may set the stage for maladaptive network configurations, which are later exacerbated by environmental stressors or neuroinflammatory responses. Longitudinal data integration within the study hints at a progressive destabilization of MSNs in schizophrenia over time, offering a window into the temporal dynamics of structural brain changes.
Technically, the study leverages cutting-edge neuroimaging analysis pipelines combined with machine learning algorithms to classify individual connectivity profiles. This combination enhances the sensitivity and specificity of brain network characterization, enabling detection of subtle but clinically relevant variations. Notably, the authors applied community detection and network-based predictive modeling to identify subnetworks that most strongly differentiate schizophrenia from healthy controls. Such methodological innovations not only refine biomarker discovery but also pave the way for stratified medicine in psychiatry, where treatments can be aligned with specific neuroanatomical profiles.
The implications of these findings extend beyond schizophrenia alone. Morphometric similarity networks, by capturing multivariate relationships in brain structure, may hold promise for understanding other neuropsychiatric disorders characterized by heterogeneous phenotypes, such as bipolar disorder, autism spectrum disorder, or major depressive disorder. The heterogeneity principle highlighted by Janssen’s research emphasizes the importance of adopting network neuroscience frameworks that move away from univariate, region-centric approaches. This paradigm shift enhances our ability to unravel the pathological mechanisms underpinning complex brain disorders, encouraging integrative models that link structure, function, and behavior.
Moreover, the study illuminates the intersection between genetics and morphometric similarity. The researchers discuss how polygenic risk scores for schizophrenia correlate with MSN disruption patterns, suggesting that genetic vulnerability exerts measurable effects on brain network architecture. This genetic imaging-genetics nexus offers exciting avenues for early identification of individuals at risk for schizophrenia before clinical onset. By combining genetic data with MSN profiles, future research could potentially identify neurodevelopmental trajectories predictive of disease, opening the way for preemptive interventions.
Another noteworthy aspect of the research is its emphasis on reproducibility and data sharing. Janssen and colleagues employed open-source analytic tools and have made their code publicly available, encouraging the scientific community to validate and extend their findings. Such transparency addresses longstanding challenges in psychiatric neuroimaging studies related to sample heterogeneity, imaging protocols, and analysis pipelines. The availability of standardized workflows facilitates large-scale meta-analyses and cross-cohort comparisons, which are essential for establishing robust biomarkers that generalize across populations.
The clinical translation potential of MSN heterogeneity is profound. Personalized medicine approaches in psychiatry have lagged behind those in oncology or cardiology due in part to the paucity of reliable biomarkers. By identifying distinct morphometric similarity profiles associated with different clinical presentations and functional outcomes, this study contributes foundational knowledge toward individualized prognosis and treatment planning. For instance, patients exhibiting certain MSN patterns might benefit more from cognitive remediation therapies, while others could respond better to pharmacological interventions targeting specific neural circuits.
Importantly, the authors caution that while MSN analyses are powerful, they must be integrated with complementary modalities such as functional MRI, diffusion tensor imaging, and electrophysiology to achieve a comprehensive understanding of schizophrenia’s neurobiology. Each modality captures different facets of brain organization—structural covariance, connectivity dynamics, and electrophysiological synchrony—which together form a holistic picture. Future multimodal studies incorporating MSN heterogeneity will likely unravel complex genotype-phenotype relationships in psychiatric disorders with unprecedented granularity.
In framing their findings, Janssen and colleagues emphasize the broader implications for conceptualizing mental illnesses beyond categorical diagnoses. The heterogeneity observed in morphometric similarity networks reflects the nuanced variations in brain structure underlying diverse symptom clusters, advocating for dimensional and network-based models in psychiatry. This approach aligns with initiatives like the Research Domain Criteria (RDoC) framework, which prioritizes neural circuits and behavioral dimensions over traditional diagnostic boundaries, ultimately aiming to improve treatment outcomes through precision neuroscience.
The research also touches upon the challenges inherent in interpreting morphometric similarity metrics. While such measures provide valuable information about anatomical covariance, they do not directly reflect causal mechanisms. Structural similarities may arise from shared developmental origins, genetic regulation, or coordinated activity-dependent plasticity. Distinguishing among these requires multimodal longitudinal investigations and integration with molecular and cellular data. Nonetheless, the findings by Janssen et al. offer a compelling step toward bridging macrostructural network alterations with microscale neuropathological processes in schizophrenia.
In summary, this seminal study on the heterogeneity of morphometric similarity networks in health and schizophrenia elevates our understanding of brain structural variability’s role in psychiatric disorders. By revealing distinct neuroanatomical network profiles linked to symptomatology and genetic risk, Janssen and colleagues provide a robust framework for advancing personalized medicine in schizophrenia. Their work exemplifies the power of network neuroscience methodologies and sets a high standard for future investigations aimed at dissecting the complex neurobiology of mental illness. As the field progresses, integrating these insights with functional and molecular data promises to transform psychiatric diagnosis and therapeutics in the coming decade.
Subject of Research: Brain morphometric similarity networks; structural brain variability; schizophrenia neuroanatomy
Article Title: Heterogeneity of morphometric similarity networks in health and schizophrenia
Article References:
Janssen, J., Guil Gallego, A., Díaz-Caneja, C.M. et al. Heterogeneity of morphometric similarity networks in health and schizophrenia. Schizophr 11, 70 (2025). https://doi.org/10.1038/s41537-025-00612-2
Image Credits: AI Generated
