FMFilip Milisav

Filip Milisav

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Filip Milisav has a total of 66 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • netneurotools: a trainee-oriented approach to network neuroscience

    by Zhen-Qi Liu, Vincent Bazinet, Justine Y. Hansen, Filip Milisav, Andrea I. Luppi, Eric G. Ceballos, Asa Farahani, Laura E. Suarez, Golia Shafiei, Ross D. Markello, et al.

    Brain imaging is an increasingly interdisciplinary field, encompassing multiple data types and multiple analytic traditions. Projects typically involve many moving parts, such as building customized preprocessing pipelines, transforming between data formats, preparing datasets for analysis and, ultimately, displaying results. The field is conventionally built on highly specialized software packages that solve these individual challenges well but are not necessarily designed to be interoperable. Trainees new to the field are therefore often left to come up with isolated heuristics and workarounds to complete a project. Here we present a set of tools to navigate the increasingly complex informatics ecosystem of brain imaging. netneurotools is a Python toolkit that has been continuously developed and maintained by this laboratory’s trainees. The philosophy of the toolkit is that it should be the Swiss army knife of the laboratory: functions and routines that we often use but that are not part of any established pipeline or package. Since its inception, the toolkit has been open and welcomes contributions from neuroscientists across the globe. netneurotools presents a necessary counterweight to out-of-the-box software packages and highlights the importance of smaller, ad hoc functions for implementing projects. By opening a window into the inner workings of a laboratory, netneurotools also presents an opportunity to begin a new type of discourse among groups and establish tangible links within the community.

  • Open Access

    Neuromorphic hierarchical modular reservoirs

    by Filip Milisav, Andrea I. Luppi, Laura E. Suárez, Guillaume Lajoie, Bratislav Misic

    Modularity is a fundamental principle of brain organization, reflected in the presence of segregated subnetworks that enable specialized information processing. These densely connected modules are often nested within larger, higher-order modules, giving rise to a hierarchical modular architecture. Yet, how hierarchical modularity shapes network function remains unclear. Here we introduce a simple blockmodeling framework for generating multi-level hierarchical modular networks and implement them as recurrent neural network reservoirs to evaluate their computational capacity. We show that hierarchical modular networks enhance memory capacity, support multitasking, and produce a broader range of temporal dynamics compared to strictly modular and random networks. These functional advantages can be traced to topological features enriched in hierarchical modular networks, including reciprocal and cyclic network motifs. We find that these benefits extend to the heterogeneous modular organization of empirical human brain structural connectivity, where hierarchical organization enhances memory capacity and contributes to the emergence of brain-like neural timescales. Altogether, these results show that hierarchical modularity endows networks with computationally advantageous properties, providing insight into the relationship between neural network structure and function.


  • Open Access

    Homophily-informed generative models of brain maps

    by Vincent Bazinet, Zhen-Qi Liu, Filip Milisav, Andrea I. Luppi, Bratislav Misic

    The structural and functional organization of the brain can be studied across multiple scales, yielding richly detailed topographic brain maps of biological features. What are the underlying forces that shape their spatial patterning? Here we introduce a simple generative model of multiscale brain maps based on the concept of inter-regional homophily: the tendency for regions that are proximal in a given physical, molecular or functional space to display similar biological features. We evaluate the model with respect to six definitions of inter-regional homophily, including physical proximity, structural and functional connectivity, and laminar, receptor and transcriptional similarity, and across 43 empirical brain maps estimated using multiple imaging, electrophysiological and histological technologies. We show that homophilic principles are sufficient to accurately reconstruct many maps, with biological similarity and functional connectivity often contributing more than the brain's geometry. We also identify consistent patterns of unexplained variation in maps with low homophily, revealing axes of cortical organization not captured by canonical inter-regional relationships. Finally, we show that homophily-informed generative models can be used to disentangle complex relationships between brain features and make new inferences on how they fit together. Collectively, this work highlights the fundamental contribution of homophily to the topographic layout of numerous biological features of the brain.

  • Open Access

    Structure-function coupling in the human brainstem

    by Asa Farahani, Subhranil Koley, Justine Y. Hansen, Maria Guadalupe Garcia Gomar, Kavita Singh, Simone Cauzzo, Firdaus Fabrice Hannanu, Filip Milisav, Zhen-Qi Liu, Vincent Bazinet, et al.

    The nuclei of the brainstem fundamentally modulate neuronal activity throughout the central nervous system. Yet connectome reconstructions typically do not include the brainstem because it is notoriously difficult to image. As a result, the influence of the brainstem on structure-function coupling in the brain is unknown. Here we use high-resolution 7 Tesla magnetic resonance imaging (MRI) to reconstruct structural and functional brain connectomes encompassing cortex and 58 brainstem nuclei spanning the midbrain, pons, and medulla. We identify structural connectional profiles of individual brainstem nuclei to the cortex and find that they align with a spectrum of functions, spanning sensory and motor processing to higher-order cognition. Structural and functional connectivity in brainstem-augmented connectomes are positively correlated, and pairs of regions with direct anatomical projections display greater functional connectivity than pairs without. Structure-function coupling is heterogeneous across brainstem nuclei, with greatest coupling in both modulatory and relay nuclei. Collectively, this work presents an initial step towards understanding how the brainstem shapes structure-function relationships in the brain.

  • Top Cited Papers
    Open Access

    Organization of neuropeptide systems in the human brain

    by Eric G. Ceballos, Asa Farahani, Zhen-Qi Liu, Filip Milisav, Justine Y. Hansen, Alain Dagher, Bratislav Misic

    Neuropeptides are functionally diverse signaling molecules in the brain, regulating a wide range of basal bodily and cognitive processes. Despite their importance, the distribution and function of neuropeptides in the human brain remains underexplored. Here we comprehensively map the organization of human whole-brain neuropeptide receptors across multiple levels of description, including molecular and cellular embedding, mesoscale connectivity and macroscale cognitive specialization. Using gene transcription as a proxy, we reconstruct a topographical cortical and subcortical atlas of 38 neuropeptide receptors across 14 different neuropeptide families. We find that most neuropeptide receptors are highly expressed either in the cortex or subcortex, delineating an anatomical cortical–subcortical gradient. Mapping neuropeptide receptors onto hypothalamic nuclei, we demonstrate that neuropeptide receptor gene expression recapitulates fundamental anatomical divisions in the hypothalamus. Neuropeptides preferentially colocalize with metabotropic neurotransmitters, suggesting a system-wide correspondence between slow-acting molecular signaling mechanisms. To investigate the behavioral consequences of distributed neuropeptide systems, we apply meta-analytical decoding to neuropeptide maps and show a spectrum of functions, from sensory-cognitive to reward and bodily functions. Finally, using evolutionary analysis we find extended positive selection for neuropeptides in early mammals, suggesting that refinement of neuropeptides coincides with the emergence of neocortex and higher cognitive function. Collectively, these results show that neuropeptide receptors are highly organized across the human brain and closely intertwined with multiple features of brain structure and function.


  • Top Cited Papers
    Open Access

    Competitive interactions shape mammalian brain network dynamics and computation

    by Andrea I. Luppi, Yonatan Sanz Perl, Jakub Vohryzek, Hana Ali, Pedro A. M. Mediano, Fernando E. Rosas, Filip Milisav, Laura E. Suárez, Silvia Gini, Daniel Gutierrez-Barragan, et al.

    How does brain network architecture balance cooperation and competition between distributed circuits? Here we use computational whole-brain modeling to examine the dynamical and computational relevance of cooperative and competitive interactions in the mammalian connectome. Across human, macaque and mouse, we show that to faithfully reproduce brain activity, model architecture consistently combines modular cooperative interactions with diffuse, long-range competitive interactions. Across species, competitive interactions preferentially link regions characterized by opposite profiles of cytoarchitecture, gene expression and receptor expression. The model with competitive interactions provides superior subject specificity, consistently outperforming the cooperative-only model and exhibiting excellent fit to the spatiotemporal properties of the living brain. These properties were not explicitly optimized, instead emerging spontaneously. Competitive interactions in the generative connectivity produce more synergistic and hierarchical dynamics, leading to enhanced performance for neuromorphic computing. Altogether, this work provides a generative link among network architecture, dynamical properties and computational performance in the mammalian brain.


  • Open Access

    Cognitive cartography of mammalian brains using meta-analysis of AI experts

    by Andrea I Luppi, Hana Ali, Zhen-Qi I Liu, Filip Milisav, Arianna Sala, Jitka Annen, Alessandro Gozzi, Danilo Bzdok, Bratislav Misic

    The complexity of the brain is increasingly mirrored by the complexity of the neuroscientific literature, yet no individual mind can fully grasp the diversity of scales, methodologies and model organisms. Where human experts flag, the latest AI models excel: large language models can seamlessly integrate knowledge across scientific domains. Here we show how large language models can systematically and quantitatively synthesise literature-wide neuroscientific knowledge about the cognitive operations and dysfunctions associated with each brain region. Meta-analysis of AI experts reveals structure-function mappings to which existing meta-analytic frameworks are blind, demonstrated by lesions and direct intracranial stimulation. It also unlocks the possibility of extending quantitative literature meta-analysis and decoding of brain maps to other model organisms beyond human. As proof of concept, we integrate LLM meta-analysis with species-specific transcriptomics in human, macaque, and mouse, to discover an evolutionarily conserved molecular circuit for cognition. Altogether, meta-analysis of AI experts can fundamentally catalyze neuroscientific discovery by overcoming the barrier of data aggregation from heterogeneous studies, finally bringing together a scattered literature to identify emergent patterns and latent insights across disparate subfields, modalities, and species.

  • Open Access

    netneurotools: a trainee-oriented approach to network neuroscience

    by Zhen-Qi Liu, Vincent Bazinet, Justine Y Hansen, Filip Milisav, Andrea I Luppi, Eric G Ceballos, Asa Farahani, Laura E Suarez, Golia Shafiei, Ross D Markello, et al.

    Brain imaging is an increasingly inter-disciplinary field, encompassing multiple data types and multiple analytic traditions. Projects typically involve many moving parts, such as building customized preprocessing pipelines, transforming between data formats, preparing datasets for analysis, and ultimately displaying results. The field is conventionally built on highly specialized software packages that solve these individual challenges well, but are not necessarily designed to be interoperable. Trainees new to the field are therefore often left to come up with isolated heuristics and workarounds to complete a project. Here we present a way to navigate the increasingly complex informatics ecosystem of brain imaging. netneurotools is our lab's internal Python toolkit that has been continuously developed and maintained by the lab's trainees. The philosophy of the toolkit is that it should be the Swiss army knife of the lab: functions and routines that we often use but that are not part of any established pipeline or package. Since its inception, the toolkit has been open and welcomes contribution from neuroscientists across the globe. netneurotools presents a necessary counterweight to out-of-the-box software packages and highlights the importance of smaller, ad hoc functions for implementing projects. By opening a window into the inner workings of a lab, netneurotools also presents an opportunity to begin a new type of discourse among groups and establish tangible links within the community.


  • Open Access

    Neuromorphic hierarchical modular reservoirs

    by Filip Milisav, Andrea I Luppi, Laura E Suárez, Guillaume Lajoie, Bratislav Misic

    Modularity is a fundamental principle of brain organization, reflected in the presence of segregated sub-networks that enable specialized information processing. These small, densely connected modules are often nested within larger, higher-order modules, giving rise to a hierarchical modular architecture. This structure is posited to balance information segregation in specialized neuronal communities and global integration via intermodular communication. Yet, how hierarchical modularity shapes network function remains unclear. Here we introduce a simple blockmodeling framework for generating and comparing multi-level hierarchical modular networks and implement them as recurrent neural network reservoirs to evaluate their computational capacity. We show that hierarchical modular networks enhance memory capacity, support multitasking, and give rise to a broader range of temporal dynamics compared to strictly modular and random networks. These functional advantages can be traced to topological features enriched in hierarchical modular networks, which include reciprocal and cyclic network motifs. To test whether the computational advantages of hierarchical modularity subsist in empirical human brain structural connectivity patterns, we develop a novel hierarchical modularity-preserving network null model, allowing us to isolate the positive effect of empirical hierarchical modularity patterns on memory capacity. To evaluate the biomimetic validity of connectome-informed reservoir dynamics, we compare reservoir timescales to empirical brain timescales derived from MEG data and find that hierarchical modularity contributes to shaping brain-like neural timescales. Altogether, across multiple benchmarks, these results show that hierarchical modularity endows networks with computationally advantageous properties, providing insight into the relationship between neural network structure and function with potential applications for the design of neuromorphic computing architectures.


  • Open Access

    An engine for systematic discovery of cause-effect relationships between brain structure and function

    by Andrea I. Luppi, Filip Milisav, Laura E. Suarez, Golia Shafiei, Jakub Vohryzek, Yonatan Sanz Perl, Hana Ali, Fernando E. Rosas, Pedro A. M. Mediano, Bratislav Misic, et al.

    Characterising how perturbations of brain architecture influence brain function is essential to understand the origins of brain dysfunction, and devise potential avenues of treatment. Here we introduce a computational engine for systematic causal discovery of the functional consequences of altering network architecture and local biophysics in the brain. We integrate multimodal anatomical and functional neuroimaging to implement over 2, 000 in-silico brains, and provide mechanistic insight into the functional consequences of local lesions, global wiring, and empirically-derived maps of regional cytoarchitecture and chemoarchitecture. We comprehensively assess how each manipulation of brain macrostructure reshapes spatial and temporal signal coordination, information dynamics, and functional hierarchy—as well as spontaneous co-activation of meta-analytic cognitive circuits, and > 6, 000 dimensions of local neural dynamics. Our computational model systematically identifies which features of brain architecture have overlapping or antagonistic causal influence over each dimension of brain function, and how functional properties are traded off against each other across disorders and neuromodulation. We find that regions’ functional vulnerability to lesions in silico recapitulates their vulnerability to neurodevelopmental and psychiatric in vivo , along a core-periphery organisation. We provide convergent evidence that the brain’s wiring diagram is finely tuned to favour the hierarchical integration of information. Notably, our model successfully recapitulates known empirical results that have not been modelled before, including desynchronisation and flattening of the brain’s functional hierarchy induced by psychedelic 5 HT 2 A agonists. To catalyse future discoveries, we make this resource freely available to the neuroscience community through an interactive website ( https://systematic-causal-mapping.up.railway.app/ ), where users can interrogate our systematic database of simulated cause-effect relationships. Altogether, we provide a powerful computational engine to predict the functional consequences of experimental or clinical interventions, and drive neuroscientific hypothesis-generation.