AMAgoston Mihalik

Agoston Mihalik

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Agoston Mihalik has a total of 138 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A transcriptomic dimension of neuronal and immune gene programs within the subgenual anterior cingulate cortex in schizophrenia

    by Rachel L. Smith, Agoston Mihalik, Nirmala Akula, Pavan K. Auluck, Stefano Marenco, Armin Raznahan, Petra E. Vértes, Francis J. McMahon

    Many psychiatric disorders are heritable, but the molecular consequences of genetic risk remain difficult to resolve, in part due to environmental confounds and the complexity of transcriptomic data. This challenge impedes therapeutic development, which relies on integrating genetic and genomic insights. Here, we integrate diagnosis, toxicological exposure, and gene expression to clarify disease-associated transcriptomic patterns in the subgenual anterior cingulate cortex (sgACC), a brain region implicated in affective regulation and psychiatric illness. We applied group regularized canonical correlation analysis (GRCCA)—a multivariate regression method that models interdependent features—to deeply sequenced bulk RNA-seq data from individuals with bipolar disorder (BD; N = 35), major depression (MDD; N = 51), schizophrenia (SCZ; N = 44), and controls (N = 55). Toxicology data from 17 known compounds were included to assess the relative contribution of known environmental exposures. Case-control expression changes were also analyzed using traditional differential gene expression (DGE) analysis to compare biological interpretability across methods. Gene set enrichment analyses evaluated enrichments for neuropsychiatric risk genes, gene ontology pathways, and cell type markers. GRCCA identified a latent variable significantly associated with schizophrenia (pperm = 0.001). This expression pattern was enriched for upregulated neuronal pathways, downregulated immune processes, and genes within loci associated with schizophrenia by GWAS. While DGE results were correlated (r = 0.43; pperm = 1.0 × 10−4) and enriched for similar functional pathways, GRCCA showed stronger alignment with schizophrenia risk genes implicated by genome-wide association studies. Together, these findings define a schizophrenia-associated expression gradient in the sgACC and illustrate how multivariate integration can refine transcriptomic signals in the context of complex psychiatric disease.


  • Open Access

    Cardiometabolic health and physical robustness map onto distinct patterns of brain structure and neurotransmitter systems

    by Eliana Nicolaisen-Sobesky, Somayeh Maleki Balajoo, Mostafa Mahdipour, Agoston Mihalik, Mahnaz Olfati, Felix Hoffstaedter, Janaina Mourão-Miranda, Masoud Tahmasian, Simon B. Eickhoff, Sarah Genon

    The link between brain health and risk/protective factors for non-communicable diseases (such as high blood pressure, high body mass index, diet, smoking, physical activity, etc.) is increasingly acknowledged. However, the specific effects that these factors have on brain health are still poorly understood, delaying their implementation in precision brain health. Here, we studied the multivariate relationships between risk factors for non-communicable diseases and brain structure, including cortical thickness (CT) and gray matter volume (GMV). Furthermore, we adopted a systems-level perspective to understand such relationships, by characterizing the cortical patterns (yielded in association to risk factors) with regards to brain morphological and functional features, as well as with neurotransmitter systems. Similarly, we related the pattern of risk/protective factors dimensions with a peripheral marker of inflammation. First, we identified latent dimensions linking a broad set of risk factors for non-communicable diseases to parcel-wise CT and GMV across the whole cortex. Data was obtained from the UK Biobank ( n = 7,370, age range = 46–81 years). We used regularized canonical correlation analysis (RCCA) embedded in a machine learning framework. This approach allows us to capture inter-individual variability in a multivariate association and to assess the generalizability of the model. The brain patterns (captured in association with risk/protective factors) were characterized from a multi-level perspective, by performing correlations (spin tests) between them and different brain patterns of structure, function, and neurotransmitter systems. The association between the risk/protective factors pattern and C-reactive protein (CRP, a marker of inflammation) was examined using Spearman correlation. We found two significant and partly replicable latent dimensions. One latent dimension linked cardiometabolic health to brain patterns of CT and GMV and was consistent across sexes. The other latent dimension linked physical robustness (including non-fat mass and strength) to patterns of CT and GMV, with the association to GMV being consistent across sexes and the association to CT appearing only in men. The CT and GMV patterns of both latent dimensions were associated to the binding potentials of several neurotransmitter systems. Finally, the cardiometabolic health dimension was correlated to CRP, while physical robustness was only very weakly associated to it. We observed robust, multi-level and multivariate links between both cardiometabolic health and physical robustness with respect to CT, GMV, and neurotransmitter systems. Interestingly, we found that cardiometabolic health and physical robustness are associated with not only increases in CT or GMV, but also with decreases of CT or GMV in some brain regions. Our results also suggested a role for low-grade chronic inflammation in the association between cardiometabolic health and brain structural health. These findings support the relevance of adopting a holistic perspective in health, by integrating neurocognitive and physical health. Moreover, our findings contribute to the challenge to the classical conceptualization of neuropsychiatric and physical illnesses as categorical entities. In this perspective, future studies should further examine the effects of risk/protective factors on different brain regions in order to deepen our understanding of the clinical significance of such increased and decreased CT and GMV.


  • Open Access

    Reduced brain structural similarity is associated with maturation, neurobiological features, and clinical status in schizophrenia

    by Natalia García-San-Martín, Richard Ai Bethlehem, Patricia Segura, Agoston Mihalik, Jakob Seidlitz, Isaac Sebenius, Claudio Alemán-Morillo, Lena Dorfschmidt, Golia Shafiei, Sarah E. Morgan, et al.

    Schizophrenia spectrum disorders (SSD) are characterized by atypical brain maturation, including alterations in structural similarity between regions. Using structural MRI data from 195 healthy controls (HC) and 352 individuals with SSD, we construct individual Morphometric INverse Divergence (MIND) networks. Compared to HC, individuals with SSD mainly exhibit reduced structural similarity in the temporal, cingulate, and insular lobes, being more pronounced in individuals exhibiting a ‘poor’ clinical status (more impaired cognitive functioning and more severe symptomatology). These alterations are associated with cortical hierarchy and maturational events, locating MIND reductions in higher-order association areas that mature later. Finally, we map 46 neurobiological features onto MIND networks, revealing a high presence of neurotransmitters and astrocytes, along with decreased metabolism and microstructure, in regions with reduced similarity in SSD. These findings provide evidence on the complex interplay between structural similarity, maturational events, and the underlying neurobiology in determining clinical status of individuals with SSD.


  • Open Access

    Cardiometabolic health and physical robustness map onto distinct patterns of brain structure and neurotransmitter systems

    by Eliana Nicolaisen-Sobesky, Somayeh Maleki Balajoo, Mostafa Mahdipour, Agoston Mihalik, Mahnaz Olfati, Felix Hoffstaedter, Janaina Mourao-Miranda, Masoud Tahmasian, Simon B. Eickhoff, Sarah Genon

    There is a recognized link between risk factors for non-communicable diseases and brain health. However, the specific effects that they have on brain health are still poorly understood, preventing its implementation in clinical practice. For instance, the association between such risk factors and cortical thickness (CT) has been primarily explored using univariate/bivariate methods and global/lobar measures of CT and has yielded inconsistent results. In this work, we aim to study the relationship between risk factors for non-communicable diseases and CT. In addition, we adopt a systems-level perspective to understand such relationship, by integrating several brain features including brain structure and function as well as neurotransmitter systems. Here, we analyzed latent dimensions linking a broad set of risk factors for non-communicable diseases to parcel-wise CT across the whole cortex (including raw, proportional, and brain size- corrected measures). We used a multivariate approach (regularized canonical correlation analysis (RCCA)) embedded in a machine learning framework that allows to capture inter- individual variability and to assess the generalizability of the model. The brain patterns (captured in association with risk factors) were characterized from a multi-level perspective, by comparing them with patterns of brain structure, function, and neurotransmitter systems. Analyses were performed separately in women (n=3685, 46-81 years) and in age-matched men (n=3685, 46-81 years) to avoid sex-bias on the results. We found one significant latent dimension (women: rrange=0.25-0.30, p=0.005-0.005; men: rrange=0.31-0.34, p=0.005-0.005), capturing variability in cardiometabolic health, including physical activity, body morphology/composition, basal metabolic rate, and blood pressure. This cardiometabolic health dimension was linked to a CT axis of inter-individual variability from the insula and cingulate cortex to occipital and parietal areas. Interestingly, this brain pattern was associated with the binding potentials of several neurotransmitter systems, including serotoninergic, dopaminergic, cholinergic, and GABAergic systems. Of note, this latent dimension was similar across sexes and across CT measures (raw, proportional, and brain-size corrected). We observed a robust, multi-level and multivariate link between cardiometabolic health, CT, and neurotransmitter systems. These findings support the urgency of further investigation into the interaction between brain health and physical health and contributes to the challenge to the classical conceptualization of neuropsychiatric and physical illnesses as categorical entities. Therefore, regular monitoring of cardiometabolic risk factors may reduce their adverse effects on brain health and prevent the development of brain diseases.


  • Open Access

    Connectome-based reservoir computing with the conn2res toolbox

    by Laura E. Suárez, Agoston Mihalik, Filip Milisav, Kenji Marshall, Mingze Li, Petra E. Vértes, Guillaume Lajoie, Bratislav Misic

    The connection patterns of neural circuits form a complex network. How signaling in these circuits manifests as complex cognition and adaptive behaviour remains the central question in neuroscience. Concomitant advances in connectomics and artificial intelligence open fundamentally new opportunities to understand how connection patterns shape computational capacity in biological brain networks. Reservoir computing is a versatile paradigm that uses high-dimensional, nonlinear dynamical systems to perform computations and approximate cognitive functions. Here we present : an open-source Python toolbox for implementing biological neural networks as artificial neural networks. is modular, allowing arbitrary network architecture and dynamics to be imposed. The toolbox allows researchers to input connectomes reconstructed using multiple techniques, from tract tracing to noninvasive diffusion imaging, and to impose multiple dynamical systems, from spiking neurons to memristive dynamics. The versatility of the toolbox allows us to ask new questions at the confluence of neuroscience and artificial intelligence. By reconceptualizing function as computation, sets the stage for a more mechanistic understanding of structure-function relationships in brain networks.


  • Open Access

    conn2res: A toolbox for connectome-based reservoir computing

    by Bratislav Misic, Laura Suarez, Agoston Mihalik, Filip Milisav, Kenji Marshall, Mingze Li, Petra Vertes, Guillaume Lajoie

    The connection patterns of neural circuits form a complex network. How signaling in these circuits manifests as complex cognition and adaptive behaviour remains the central question in neuroscience. Concomitant advances in connectomics and artificial intelligence open fundamentally new opportunities to understand how connection patterns shape computational capacity in biological brain networks. Reservoir computing is a versatile paradigm that uses nonlinear dynamics of high-dimensional dynamical systems to perform computations and approximate cognitive functions. Here we present conn2res: an open-source Python toolbox for implementing biological neural networks as artificial neural networks. conn2res is modular, allowing arbitrary architectures and arbitrary dynamics to be imposed. The toolbox allows researchers to input connectomes reconstructed using multiple techniques, from tract tracing to noninvasive diffusion imaging, and to impose multiple dynamical systems, from simple spiking neurons to memristive dynamics. The versatility of the conn2res toolbox allows us to ask new questions at the confluence of neuroscience and artificial intelligence. By reconceptualizing function as computation, conn2res sets the stage for a more mechanistic understanding of structure-function relationships in brain networks.

  • Open Access

    conn2res : A toolbox for connectome-based reservoir computing

    by Laura E. Suárez, Agoston Mihalik, Filip Milisav, Kenji Marshall, Mingze Li, Petra E. Vértes, Guillaume Lajoie, Bratislav Misic

    The connection patterns of neural circuits form a complex network. How signaling in these circuits manifests as complex cognition and adaptive behaviour remains the central question in neuroscience. Concomitant advances in connectomics and artificial intelligence open fundamentally new opportunities to understand how connection patterns shape computational capacity in biological brain networks. Reservoir computing is a versatile paradigm that uses nonlinear dynamics of high-dimensional dynamical systems to perform computations and approximate cognitive functions. Here we present conn2res : an open-source Python toolbox for implementing biological neural networks as artificial neural networks. conn2res is modular, allowing arbitrary architectures and arbitrary dynamics to be imposed. The toolbox allows researchers to input connectomes reconstructed using multiple techniques, from tract tracing to noninvasive diffusion imaging, and to impose multiple dynamical systems, from simple spiking neurons to memristive dynamics. The versatility of the conn2res toolbox allows us to ask new questions at the confluence of neuroscience and artificial intelligence. By reconceptualizing function as computation, conn2res sets the stage for a more mechanistic understanding of structure-function relationships in brain networks.


  • Open Access

    Author Correction: A cross-cohort replicable and heritable latent dimension linking behaviour to multi-featured brain structure

    by Eliana Nicolaisen-Sobesky, Agoston Mihalik, Shahrzad Kharabian-Masouleh, Fabio S. Ferreira, Felix Hoffstaedter, Holger Schwender, Somayeh Maleki Balajoo, Sofie L. Valk, Simon B. Eickhoff, B. T. Thomas Yeo, et al.

  • Open Access

    A cross-cohort replicable and heritable latent dimension linking behaviour to multi-featured brain structure

    by Eliana Nicolaisen-Sobesky, Agoston Mihalik, Shahrzad Kharabian-Masouleh, Fabio S. Ferreira, Felix Hoffstaedter, Holger Schwender, Somayeh Maleki Balajoo, Sofie L. Valk, Simon B. Eickhoff, B. T. Thomas Yeo, et al.

    Identifying associations between interindividual variability in brain structure and behaviour requires large cohorts, multivariate methods, out-of-sample validation and, ideally, out-of-cohort replication. Moreover, the influence of nature vs nurture on brain-behaviour associations should be analysed. We analysed associations between brain structure (grey matter volume, cortical thickness, and surface area) and behaviour (spanning cognition, emotion, and alertness) using regularized canonical correlation analysis and a machine learning framework that tests the generalisability and stability of such associations. The replicability of brain-behaviour associations was assessed in two large, independent cohorts. The load of genetic factors on these associations was analysed with heritability and genetic correlation. We found one heritable and replicable latent dimension linking cognitive-control/executive-functions and positive affect to brain structural variability in areas typically associated with higher cognitive functions, and with areas typically associated with sensorimotor functions. These results revealed a major axis of interindividual behavioural variability linking to a whole-brain structural pattern.


  • Top Cited Papers
    Open Access

    Canonical Correlation Analysis and Partial Least Squares for Identifying Brain–Behavior Associations: A Tutorial and a Comparative Study

    by Agoston Mihalik, James Chapman, Rick A. Adams, Nils R. Winter, Fabio S. Ferreira, John Shawe-Taylor, Janaina Mourão-Miranda