Data-driven mean-field within whole-brain models

Marmaduke Woodman; Martin Breyton; Meysam Hashemi; Spase Petkoski; Viktor Jirsa; Viktor Sip

REVIEW 3 major objections 5 minor 1 cited by

Data-driven mean-field within whole-brain models

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A multilayer perceptron trained directly on simulations of a quadratic integrate-and-fire network can learn its macroscopic mean-field dynamics, including sparse connectivity as a new parameter, and when embedded in a whole-brain model it r

desk verdict Promising MLP-learned mean-field for QIF networks, but the whole-brain validation rests on a violated linear-coupling approximation and the new cusp is not independently verified. read the letter →

arxiv 2509.02799 v2 pith:4NSW6DE2 submitted 2025-09-02 q-bio.NC

Martin Breyton , Viktor Sip , Marmaduke Woodman , Meysam Hashemi , Spase Petkoski , Viktor Jirsa This is my paper · ORCID

classification q-bio.NC

keywords data-drivenmean-fieldQIFnetworkwhole-brainmodelmultilayerperceptronbifurcationanalysissimulation-basedinferenceconnectionprobabilityneuralmass

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that a multilayer perceptron can learn the macroscopic dynamics of a network of quadratic integrate-and-fire (QIF) neurons directly from simulations, bypassing the restrictive all-to-all connectivity assumption of analytic mean-field theory. It introduces the connection probability p as a new parameter, performs bifurcation analysis on the trained network, and finds a cusp bifurcation in the (p, η) plane that reshapes the phase diagram in a way degenerate with synaptic coupling J. Embedding this learned mean-field in a whole-brain model, the authors use simulation-based inference on synthetic fMRI data and show accurate parameter recovery, while the analytic MPR model yields estimates that are confidently wrong. If correct, this offers a flexible data-driven bridge from micro-scale spiking rules to macro-scale brain dynamics, with the potential to build mean-field models for neuron models where analytic closure is intractable.

What carries the argument

The central object is the trained MLP that represents the vector field ẋ = MLPΘ(x, η, J, p, I(t)) with state x = (r, v). It is trained by unrolling the MLP through a Heun integrator and minimizing the average Euclidean distance between predicted and observed population trajectories from QIF simulations. Once trained, the frozen weights define a deterministic dynamical system on which numerical continuation can be run to locate fixed points and bifurcation branches; the derivative with respect to the input current is computed via automatic differentiation to test coupling linearity. Adding p as an input parameter is the mechanism that bypasses the analytic all-to-all connectivity assumption,

What would settle it

Measure the partial derivatives ∂ṙ/∂I_ext and ∂v̇/∂I_ext across the (r, v) phase space for the trained MLP; if their non-constant variation shifts the cusp location in the (p, η) plane or changes whole-brain functional connectivity by a detectable amount, the claimed phase diagram and inference results are not robust. Alternatively, simulate a two-node spiking network with true synaptic coupling and compare its joint activity to a two-node MLP whole-brain model under the same input; a clear mismatch in correlation structure would falsify the linear-coupling approximation.

Watch

Extended reading notes

Core claim

The central claim is that the phase flow of a network of QIF neurons—its mean firing rate r and mean membrane potential v—can be learned by an MLP from simulated time series without needing an analytic closure. The trained MLP reproduces the analytical MPR model almost exactly at p=1, and extends it to sparse connectivity p<1, which the analytic derivation cannot handle. Numerical continuation on the frozen MLP reveals a cusp bifurcation in the (η, p) plane with the same shape as the known cusp in (η, J), implying p and J are nearly degenerate: decreasing connection probability acts like decreasing synaptic coupling. The learned mean-field is then used as a neural mass inside a whole-brain m

Load-bearing premise

The whole-brain results assume that coupling between brain regions can be treated as a linear input to the local neural mass, even though the learned MLP is not linear in that input; the paper forces the linear coupling anyway and asserts the behavior stays consistent.

Editorial extensions

If this is right

A mean-field can be derived for any spiking network simulation with a fully observed macroscopic state, even when no analytic closure exists, such as with sparse connectivity or non-Lorentzian heterogeneity.
Connection probability becomes a first-class parameter in whole-brain models; because it is degenerate with synaptic coupling, inference must estimate both to avoid biased parameter recovery.
Simulation-based inference using the learned mean-field recovers ground-truth parameters from synthetic fMRI better than the analytical model, suggesting that data-driven mean-fields reduce model misspecification bias in whole-brain inversion.
The learned dynamics extrapolate beyond the training data: cusp branches and the chaotic regime under sinusoidal input emerge without explicit sampling, indicating the MLP captures the underlying vector field rather than memorizing examples.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The linear-coupling assumption (Appendix B) is violated for the MLP, yet the whole-brain simulations rely on it; a direct test comparing the MLP whole-brain model with the original spiking network under identical structural connectivity would show whether the parameter-recovery results survive without this approximation.
Because p and J are degenerate, the posterior in (p, J) is likely non-identifiable even in the MLP model; the improved inference may reflect a better match to finite-size effects rather than a true separation of p and J.
Extending the same pipeline to partially observed states (e.g., only firing rate) would require latent-state reconstruction or additional observables; the paper explicitly restricts to full observability, which limits direct application to real neural recordings.
The method naturally extends to detailed neuronal simulators (multi-compartment, conductance-based) where analytic mean-field is intractable; a concrete next step is to train on such a simulator and compare its macroscopic activity against the learned model.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this paper has a good idea and some honest internal checks, but its main whole-brain validation leans on an approximation the authors admit is violated, and the flagship 'new cusp' is never checked against direct spiking simulations. I'd send it to review, but the revision needs to be substantial.

What's new and valuable: using an MLP to learn the phase flow of a QIF network, with connection probability p as an input, is a genuinely useful move for mean-field approaches that are analytically intractable. For p=1 the trained MLP matches the MPR phase diagram and recovers the known chaotic regime even though that regime wasn't in the training data—real evidence the surrogate learned the dynamics, not just fitted a surface. The continuation analysis on the trained MLP is a sensible way to extract bifurcation structure. The paper is also honest enough to note that p and J are degenerate, and the point that inference with an incomplete model can be 'confidently wrong' is fair.

Where it gets shaky: the whole-brain model and the SBI comparison are built on Equation (5)'s linear coupling, and Appendix B plainly states the MLP does not satisfy the required affine-in-I condition. The authors say the behavior 'still remains consistent' but give no quantitative error check. That matters because the SBI results in Figure 5 are supposed to show the learned model recovers parameters better; if the coupling approximation distorts TVB_QIF and TVB_MPR differently, the comparison is compromised. The cusp in the (p, eta) plane is claimed as a new result, but it is derived from the fitted MLP itself, and with p and J effectively interchangeable, it may be little more than a re-expression of the known (J, eta) cusp. There is no independent verification against spiking simulations for p<1, no uncertainty quantification on the MLP, and no released code or data. The training benchmark in Appendix E is nice, but it doesn't address these issues.

Proportionally: none of this kills the core idea. The MLP surrogate for p=1 is well validated, and the framework is clearly transferable to other micro-scale parameters. The deficiencies are in validation of the novel parts, and they are fixable: quantify the linear-coupling error in the operating regime, test the predicted cusp against direct spiking simulations for p<1, and release the code.

Who is this for? Computational neuroscientists working on whole-brain models and mean-field reductions. A serious referee should engage, but I would not cite it for the whole-brain claims until the linear-coupling issue is addressed.

Referee Report

3 major / 5 minor

Summary. The paper introduces a data-driven framework in which a multi-layer perceptron (MLP) is trained to reproduce the macroscopic phase flow of a network of Quadratic-Integrate-and-Fire (QIF) neurons, using the population firing rate r, mean membrane potential v, the excitability parameter η, synaptic weight J, connection probability p, and an external input I(t) as inputs. The trained MLP is validated against the analytical Montbrió-Pazó-Roxin (MPR) mean-field model for p=1, including recovery of the phase diagram and a chaotic regime. Bifurcation analysis on the MLP is used to claim a cusp bifurcation in the (p,η) plane that is degenerate with J. The MLP is then embedded in a whole-brain model with linear structural coupling (TVB_QIF), compared to the standard TVB_MPR model, and used in simulation-based inference (SBI) to recover η, J, and p from synthetic and real fMRI data. The authors report that TVB_QIF recovers ground-truth parameters with lower z-scores than TVB_MPR, while TVB_MPR produces confidently wrong estimates.

Significance. If the claims hold, the paper offers a flexible and generic route from spiking-network simulations to mean-field models that are not constrained by analytical closure, and demonstrates a concrete use-case in whole-brain modeling and Bayesian parameter inference. The p=1 validation against the exact MPR solution and the reproduction of the chaotic regime are genuine strengths and give independent support for the MLP reconstruction method. The paper is also notably transparent in Appendix B about the failure of the linear-coupling assumption. However, the whole-brain and SBI results rest on an unquantified approximation, and the claim of a new cusp bifurcation in p has no independent ground truth. These issues are load-bearing for the paper's central claims and require additional validation before the results can be accepted.

major comments (3)

[§III C and Appendix B, Eq. (5), Eq. (B1)] The whole-brain model and all downstream SBI results rely on the linear-coupling form N(ψ,I) ≈ N(ψ,0)+I. Appendix B explicitly shows ∂ṙ/∂Iext and ∂v̇/∂Iext are not constant in (r,v), so this approximation is violated. The text only says that 'when enforcing it, the behavior of the model still remains consistent,' but no quantitative error measure is provided. Since the coupling term gΣW_ijS(ψ_j) is not small in the operating regime (Fig. 4 sweeps g around 0.55–0.66), the nonlinearity in Iext could materially change the whole-brain dynamics. The claimed SBI accuracy may therefore be an artifact of the linearized model rather than a property of the learned mean-field. Please provide a concrete error quantification over the operating range of (r,v) and coupling strengths, and ideally validate a coupled spiking-network simulation against TVB_QIF; alternatively, implement the coupling direct
[§III B and Fig. 3] The paper claims the trained MLP reveals a new cusp bifurcation in the (p,η) plane. However, for p<1 there is no analytical ground truth, and the MLP itself is the object under study, so the cusp is a property of the fitted function rather than of the spiking network. The text also acknowledges that p and J are 'highly degenerate' and that decreasing p has nearly the same effect as decreasing J. In the large-N limit the mean-field input should scale as pJ r, suggesting that the 'new' cusp may simply be the known (J,η) cusp reparameterized by the product pJ. Please test this directly: compare the cusp curves from Fig. 3(a) with those obtained by rescaling J by p, or train an MLP on the product pJ and show whether the two-parameter dependence is irreducible. Additionally, validate the cusp location at p<1 using direct spiking simulations at representative parameter points, and report varia
[§III D and Fig. 5] The SBI ground-truth observations are generated with TVB_QIF, which is the same model family used for inference. This is a standard self-consistency test, but it does not validate the model against the actual spiking network. Since TVB_QIF is built on the linear-coupling approximation of Appendix B, the recovered posteriors for η, J, and p characterize the linearized model, not necessarily the QIF network. The comparison with TVB_MPR is informative as a model-misspecification example, but the conclusion that the data-driven model is more accurate for fMRI parameter recovery requires an end-to-end validation with data generated from a true coupled QIF network (or at least from a coupling scheme that avoids the linearization error). Without this, the reported improvement may be an artifact of the violated coupling assumption.

minor comments (5)

[Abstract] Grammar: 'existence of new cusp bifurcation' should be 'existence of a new cusp bifurcation'.
[Figure 4 caption] The caption contains a typo: 'p = 0.92p = 0.88 simulated with simulated with and TVB QIF' appears garbled. Please correct.
[Figure 5 caption] The caption says 'In (d),(h) and (e),(f),(h)'—the enumeration of panels is inconsistent and should be corrected.
[§III A] The text states 'a total of N = 220000 simulations.' Given the stated grid: η ∈ [−8,−1] with step 0.05, p ∈ [0.6,1] with step 0.1, J ∈ [5,15] with step 0.5, and 5 seeds, the number of simulations is approximately 74,000. The discrepancy should be clarified (e.g., number of training segments rather than simulations).
[Fig. 3] The cusp curves and fixed-point locations are shown without any measure of variability across MLP training runs or across noise seeds. A brief quantitative robustness statement would strengthen the presentation.

Circularity Check

1 steps flagged · score 6.0 of 10

SBI validation is a closed loop: ground-truth data generated from TVB_QIF, the same model used for inference, compounded by the acknowledged linear-coupling violation.

fitted input called prediction [Section II.D, Model inversion (Figure 5 and surrounding text)]
"We then generated 500 ground truth observations using TVB QIF with different combinations of {¯η, J, p} drawn from the prior distribution, and in Figure 5, we display the posterior z-scores and shrinkage PM P R(red points) and PQIF (black points) for all of them. All posterior distributions PQIF have a low z-score value which indicates that the estimation is always close to, or contains the ground truth."

The ground-truth observations are not independent spiking-network data: they are generated by TVB_QIF, the same whole-brain model whose node dynamics are the fitted MLP_QIF, and the same simulator is used to train the amortized posterior P_QIF. Low z-scores for P_QIF are therefore a self-consistency property of a posterior on its own generative model, not evidence that the learned mean-field predicts independent dynamics. The red P_MPR points come from a deliberately misspecified comparison in which the data are generated from the favored model, so the conclusion that MPR is 'confidently wrong' is forced by the experimental design. This is further compounded by Appendix B, which states that the linear coupling assumption, Eq. (B1), does not hold for the MLP, yet TVB_QIF is built by enforci

full rationale

The paper has genuine external anchoring for p=1 (the MLP reproduces the MPR bifurcation diagram and the chaotic regime), and the cusp bifurcation is a real emergent property of the trained regressor rather than a direct copy of the training labels. However, the central whole-brain and SBI validation is circular in a specific sense: the 500 'ground truth' observations used to claim accurate parameter recovery for TVB_QIF are generated from TVB_QIF itself, the same model whose posterior is being computed. Thus the low z-scores are a calibration check, not an external validation of the MLP_QIF mean-field. The comparison to TVB_MPR is a misspecification test with data generated under the favored model, making the claimed superiority of TVB_QIF over TVB_MPR a consequence of the experiment design. This is compounded by the manuscript's own Appendix B: the authors explicitly show that MLP_QIF is not affine in Iext, contradicting Eq. (B1), and yet they 'enforce' linear coupling anyway to build TVB_QIF. The whole-brain simulations and SBI therefore test a linearized surrogate, not the MLP as trained. The p-versus-J degeneracy is acknowledged ('highly degenerate with J') and further weakens the novelty of the 'new' cusp, though the cusp itself is not circular because it is derived from the fitted MLP and not assumed a priori. Overall, the main 'accurate parameter recovery' claim reduces to a closed-loop self-consistency check, so a score of 6 is warranted.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The core of the paper is a fitted MLP used as a surrogate for microscopic dynamics. The few explicit axioms are the analytical ground truth, the sufficiency of the (r,v,p) state description, the surrogate's accuracy, the enforced linear coupling, and the informativeness of the SBI features. No new physical entities are postulated.

free parameters (2)

MLP weights and biases Θ = Not reported; weights not released
The surrogate phase-flow model is trained on QIF/MPR time series; all downstream bifurcation and whole-brain results depend on this fit.
MLP architecture and training hyperparameters = Not reported
Number of layers/units, activation schedule, learning rate, batch size, and regularization are not specified; these choices affect surrogate accuracy and derived bifurcations.

assumptions (5)

standard math The MPR equations are the exact ground-truth macroscopic model for the all-to-all QIF network in the thermodynamic limit with Lorentzian heterogeneity
Borrowed from Montbrió et al. [23]; used as ground truth to validate the MLP and as the comparator whole-brain model.
domain assumption A finite QIF network with N=10^4 and connection probability p has macroscopic dynamics adequately captured by the pair (r,v) with homogeneous p
Central modeling assumption: the MLP is trained on these two summary variables and ignores higher-order statistics; no evidence that (r,v,p) is a sufficient statistic when p<1.
domain assumption The MLP is an accurate surrogate of the true phase flow over the sampled parameter range and extrapolates outside it
Required for the bifurcation analysis and whole-brain integration; only partially tested (p=1 against MPR, one chaotic regime).
ad hoc to paper Linear coupling in the whole-brain model is a valid approximation even though the MLP derivative with respect to Iext is not constant
Appendix B shows the partial derivatives are not constant; linear coupling is enforced anyway, which affects the fidelity of TVB_QIF simulations.
domain assumption SBI summary features from FC/FCD are sufficient to identify eta, J, and p; the priors and fixed g, sigma are appropriate
No formal identifiability analysis; posterior shrinkage variability in Figure 5 suggests partial identifiability issues even for the matched model.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Data-driven mean-field within whole-brain models." pith.science (2026). https://pith.science/paper/4NSW6DE2

@misc{pith2026250902799,
  author       = {Pith},
  title        = {Pith review of: Data-driven mean-field within whole-brain models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4NSW6DE2}},
  note         = {Machine review of arXiv:2509.02799}
}

read the original abstract

Mean-field models provide a link between microscopic neuronal activity and macroscopic brain dynamics. Their derivation depends on simplifying assumptions, such as all-to-all connectivity, limiting their biological realism. To overcome this, we introduce a data-driven framework in which a multi-layer perceptron (MLP) learns the macroscopic dynamics directly from simulations of a network of spiking neurons. The network connection probability serves here as a new parameter, inaccessible to purely analytical treatment, which is validated against ground truth analytical solutions. Through bifurcation analysis on the trained MLP, we demonstrate the existence of new cusp bifurcation that systematically reshapes the system's phase diagram in a degenerate manner with synaptic coupling. By integrating this data-driven mean-field model into a whole-brain computational framework, we show that it extends beyond the macroscopic emergent dynamics generated by the analytical model. For validation, we use simulation-based inference on synthetic functional magnetic resonance imaging (fMRI) data and demonstrate accurate parameter recovery for the novel mean-field model, while the current state-of-the-art models lead to biased estimates. This work presents a flexible and generic framework for building more realistic whole-brain models, bridging the gap between microscale mechanisms and macroscopic brain recordings.

Figures

Figures reproduced from arXiv: 2509.02799 by the authors.

**Figure 2.** FIG. 2. Reconstruction of MPR dynamics using QIF simulations. (a) Phase diagram comparing the ground truth (thick [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** FIG. 3. Phase diagram. (a) Co-dimension two bifurcation [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (7 more)

**Figure 4.** Figure 4: FIG. 4. Comparison of the dynamics of whole-brain models using the ground truth model (MPR) or the reconstructed model [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: FIG. 5. Posterior [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

**Figure 6.** Figure 6: FIG. 6. Posterior distributions estimated from the fMRI [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

**Figure 8.** Figure 8: FIG. 8. Values of the partial derivatives on the phase [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

**Figure 9.** Figure 9: FIG. 9. Firing rate times series of QIF simulations un [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]

**Figure 10.** Figure 10: FIG. 10. Loss curves for different datasets. (a) Training and [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]

**Figure 11.** Figure 11: FIG. 11. Quality of the reconstruction of the cusp bifurcation [PITH_FULL_IMAGE:figures/full_fig_p010_11.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Mechanistic bridges from receptors to whole-brain dynamics: mean-field reductions, validity domains, and computational trade-offs
q-bio.NC 2026-07 conditional novelty 5.0 of 10

This review reconstructs the mathematical lineage from spiking circuits to receptor-aware whole-brain mean-field models and introduces operation- and memory-cost benchmarks showing global covariance propagation is imp...

Reference graph

Works this paper leans on

81 extracted references · 70 canonical work pages · cited by 1 Pith paper

[1]

Hashemi, D

M. Hashemi, D. Depannemaecker, M. Saggio, P. Triebkorn, G. Rabuffo, J. Fousek, A. Ziaeemehr, V. Sip, A. Athanasiadis, M. Breyton, M. Woodman, H. Wang, S. Petkoski, P. Sorrentino, and V. Jirsa, IEEE Reviews in Biomedical Engineering , 1–29 (2025)

work page 2025
[2]

Breakspear, Nature Neuroscience 20, 340–352 (2017)

M. Breakspear, Nature Neuroscience 20, 340–352 (2017)

work page 2017
[3]

H. E. Wang, P. Triebkorn, M. Breyton, B. Dollomaja, J.-D. Lemarechal, S. Petkoski, P. Sorrentino, D. Depan- nemaecker, M. Hashemi, and V. K. Jirsa, National Sci- ence Review 11 (2024), 10.1093/nsr/nwae079. 11

work page doi:10.1093/nsr/nwae079 2024
[4]

Jirsa, H

V. Jirsa, H. Wang, P. Triebkorn, M. Hashemi, J. Jha, J. Gonzalez-Martinez, M. Guye, J. Makhalova, and F. Bartolomei, The Lancet Neurology 22, 443–454 (2023)

work page 2023
[5]

Sanz-Leon, S

P. Sanz-Leon, S. A. Knock, A. Spiegler, and V. K. Jirsa, NeuroImage 111, 385–430 (2015)

work page 2015
[6]

G. Deco, V. K. Jirsa, P. A. Robinson, M. Breakspear, and K. Friston, PLoS Computational Biology4, e1000092 (2008)

work page 2008
[8]

Breyton, J

M. Breyton, J. Fousek, G. Rabuffo, P. Sorrentino, L. Kusch, M. Massimini, S. Petkoski, and V. Jirsa, (2024), 10.7554/elife.98920.1

work page doi:10.7554/elife.98920.1 2024
[9]

V. K. Jirsa, W. C. Stacey, P. P. Quilichini, A. I. Ivanov, and C. Bernard, Brain 137, 2210–2230 (2014)

work page 2014

Show all 81 references

[10]

M. L. Saggio, D. Crisp, J. M. Scott, P. Karoly, L. Kuhlmann, M. Nakatani, T. Murai, M. D¨ umpelmann, A. Schulze-Bonhage, A. Ikeda, M. Cook, S. V. Gliske, J. Lin, C. Bernard, V. Jirsa, and W. C. Stacey, eLife 9 (2020), 10.7554/elife.55632

2020 doi
[11]

Ponce-Alvarez and G

A. Ponce-Alvarez and G. Deco, Scientific Reports 14 (2024), 10.1038/s41598-024-53105-0

2024 doi
[12]

Xavier, P

M. Xavier, P. Figueiredo, G. Deco, A. I. Luppi, and J. Cabral, Entropy 26, 1048 (2024)

2024
[13]

B. H. Jansen and V. G. Rit, Biological Cybernetics 73, 357–366 (1995)

1995
[14]

H. R. Wilson and J. D. Cowan, Biophysical Journal 12, 1 (1972)

1972
[15]

Stefanovski, P

L. Stefanovski, P. Triebkorn, A. Spiegler, M.-A. Diaz- Cortes, A. Solodkin, V. Jirsa, A. R. McIntosh, and P. Ritter, Frontiers in Computational Neuroscience 13 (2019), 10.3389/fncom.2019.00054

2019
[16]

V. Sip, M. Hashemi, T. Dickscheid, K. Amunts, S. Petkoski, and V. Jirsa, Science Advances 9 (2023), 10.1126/sciadv.abq7547

2023 doi
[17]

R. E. Baker, J.-M. Pena, J. Jayamohan, and A. J´ erusalem, Biology letters14, 20170660 (2018)

2018
[18]

Personalized virtual brains of alzheimer’s disease link dynamical biomarkers of fMRI with increased local excitability,

B. H. Yal¸ cınkaya, A. Ziaeemehr, J. Fousek, M. Hashemi, M. Lavanga, A. Solodkin, A. R. McIntosh, V. K. Jirsa, and S. Petkoski, “Personalized virtual brains of alzheimer’s disease link dynamical biomarkers of fMRI with increased local excitability,” (2023), pages: 2023.01.11.23284438

2023
[19]

A com- putational approach to evaluate how molecular mecha- nisms impact large-scale brain activity,

A. Destexhe, M. Sacha, F. Tesler, and R. Cofre, “A com- putational approach to evaluate how molecular mecha- nisms impact large-scale brain activity,” (2024)

2024
[20]

El Boustani and A

S. El Boustani and A. Destexhe, Neural Computation 21, 46–100 (2009)

2009
[21]

Coombes, Frontiers in Applied Mathematics and Statistics 9 (2023), 10.3389/fams.2023.1128224

S. Coombes, Frontiers in Applied Mathematics and Statistics 9 (2023), 10.3389/fams.2023.1128224

2023
[22]

Schwalger, M

T. Schwalger, M. Deger, and W. Gerstner, PLOS Com- putational Biology 13, e1005507 (2017)

2017
[23]

Montbri´ o, D

E. Montbri´ o, D. Paz´ o, and A. Roxin, Physical Review X 5 (2015), 10.1103/physrevx.5.021028

2015 doi
[24]

Devalle, E

F. Devalle, E. Montbri´ o, and D. Paz´ o, Physical Review E 98 (2018), 10.1103/physreve.98.042214

2018 doi
[25]

Pietras, P

B. Pietras, P. Clusella, and E. Montbri´ o, Physical Re- view E 111 (2025), 10.1103/physreve.111.014422

2025 doi
[26]

R. Gast, T. R. Kn¨ osche, and H. Schmidt, Physical Re- view E 104 (2021), 10.1103/physreve.104.044310

2021 doi
[27]

R. Gast, S. A. Solla, and A. Kennedy, Physical Review E 107 (2023), 10.1103/physreve.107.024306

2023 doi
[28]

Pyragas and K

V. Pyragas and K. Pyragas, Physics Letters A 480, 128972 (2023)

2023
[29]

How connec- tion probability shapes fluctuations of neural population dynamics,

N. E. Greven, J. Ranft, and T. Schwalger, “How connec- tion probability shapes fluctuations of neural population dynamics,” (2024), arXiv:2412.16111

2024
[30]

E. M. Izhikevich, Dynamical Systems in Neuroscience: The Geometry of Excitability and Bursting (The MIT Press, 2006)

2006
[31]

Stimberg, R

M. Stimberg, R. Brette, and D. F. Goodman, eLife 8, e47314 (2019)

2019
[32]

Durstewitz, G

D. Durstewitz, G. Koppe, and M. I. Thurm, Nature Reviews Neuroscience 24, 693 (2023)

2023
[33]

Yu and R

R. Yu and R. Wang, Proceedings of the National Academy of Sciences 121, e2311808121 (2025)

2025
[34]

M., Pattern Recognition and Machine Learning , 1st ed., Information Science and Statistics (Springer, New York, NY, 2006)

C. M., Pattern Recognition and Machine Learning , 1st ed., Information Science and Statistics (Springer, New York, NY, 2006)

2006
[35]

P. E. Kloeden and E. Platen, Numerical Solution of Stochastic Differential Equations (Springer Berlin Hei- delberg, 1992)

1992
[36]

BifurcationKit.jl,

R. Veltz, “BifurcationKit.jl,” (2020), hal-02902346

2020
[37]

V. K. Jirsa and H. Haken, Physical Review Letters 77, 960–963 (1996)

1996
[38]

V. K. Jirsa and J. A. S. Kelso, Physical Review E 62, 8462–8465 (2000)

2000
[39]

Sporns, G

O. Sporns, G. Tononi, and R. K¨ otter, PLoS Computa- tional Biology 1, e42 (2005)

2005
[40]

Rabuffo, H.-A

G. Rabuffo, H.-A. Lokossou, Z. Li, A. Ziaee-Mehr, M. Hashemi, P. P. Quilichini, A. Ghestem, O. Arab, M. Esclapez, P. Verma, et al. , Proceedings of the Na- tional Academy of Sciences 122, e2405706122 (2025)

2025
[41]

Friston, A

K. Friston, A. Mechelli, R. Turner, and C. Price, Neu- roImage 12, 466–477 (2000)

2000
[42]

E. C. A. Hansen, D. Battaglia, A. Spiegler, G. Deco, and V. K. Jirsa, Neuroimage 105, 525 (2015)

2015
[43]

G. Deco, M. L. Kringelbach, A. Arnatkeviciute, S. Old- ham, K. Sabaroedin, N. C. Rogasch, K. M. Aquino, and A. Fornito, Science Advances 7, eabf4752 (2021)

2021
[44]

X. Kong, R. Kong, C. Orban, P. Wang, S. Zhang, K. An- derson, A. Holmes, J. D. Murray, G. Deco, M. van den Heuvel, and B. T. T. Yeo, Nature Communications 12 (2021), 10.1038/s41467-021-26704-y

2021 doi
[45]

Hashemi, A

M. Hashemi, A. N. Vattikonda, V. Sip, M. Guye, F. Bar- tolomei, M. M. Woodman, and V. K. Jirsa, NeuroImage 217, 116839 (2020)

2020
[46]

J. Jha, M. Hashemi, A. N. Vattikonda, H. Wang, and V. Jirsa, Machine Learning: Science and Technology 3, 035016 (2022)

2022
[47]

van de Schoot, S

R. van de Schoot, S. Depaoli, R. King, B. Kramer, K. M¨ artens, M. G. Tadesse, M. Vannucci, A. Gelman, D. Veen, J. Willemsen, and C. Yau, Nature Reviews Methods Primers 1 (2021), 10.1038/s43586-020-00001-2

2021 doi
[48]

Cranmer, J

K. Cranmer, J. Brehmer, and G. Louppe, Proceedings of the National Academy of Sciences 117, 30055 (2020)

2020
[49]

To do this, first, a set of random simu- lations is run for parameter values θ sampled from prior distributions

to estimate an invertible mapping between the pa- rameter space of the model and low-dimensional feature space of choice. To do this, first, a set of random simu- lations is run for parameter values θ sampled from prior distributions. Second, a set of features x is extracted f...
[50]

Masked Autoregressive Flow for Density Estimation,

G. Papamakarios, T. Pavlakou, and I. Murray, “Masked Autoregressive Flow for Density Estimation,” (2018), arXiv:1705.07057 [cs, stat]

2018 arXiv
[51]

P. J. Gon¸ calves, J.-M. Lueckmann, M. Deistler, M. Non- nenmacher, K. ¨Ocal, G. Bassetto, C. Chintaluri, W. F. Podlaski, S. A. Haddad, T. P. Vogels, D. S. Greenberg, and J. H. Macke, eLife 9 (2020), 10.7554/elife.56261

2020 doi
[52]

Q. Liu, J. Xu, R. Jiang, and W. H. Wong, Proceedings of the National Academy of Sciences 118, e2101344118 12 (2021)

2021
[53]

Hashemi, A

M. Hashemi, A. N. Vattikonda, J. Jha, V. Sip, M. M. Woodman, F. Bartolomei, and V. K. Jirsa, Neural Net- works 163, 178–194 (2023)

2023
[54]

Calibrating model-based inferences and decisions,

M. Betancourt, “Calibrating model-based inferences and decisions,” (2018), arXiv:1803.08393 [stat.ME]

2018 arXiv
[55]

Lavanga, J

M. Lavanga, J. Stumme, B. H. Yalcinkaya, J. Fousek, C. Jockwitz, H. Sheheitli, N. Bittner, M. Hashemi, S. Petkoski, S. Caspers, and V. Jirsa, NeuroImage 283, 120403 (2023)

2023
[56]

Fousek, G

J. Fousek, G. Rabuffo, K. Gudibanda, H. Sheheitli, S. Petkoski, and V. Jirsa, Scientific Reports 14 (2024), 10.1038/s41598-024-83542-w

2024 doi
[57]

Rabuffo, J

G. Rabuffo, J. Fousek, C. Bernard, and V. Jirsa, eneuro 8, ENEURO.0283 (2021)

2021
[58]

H. Ko, S. B. Hofer, B. Pichler, K. A. Buchanan, P. J. Sj¨ ostr¨ om, and T. D. Mrsic-Flogel, Nature473, 87 (2011)

2011
[59]

D’Angelo and V

E. D’Angelo and V. Jirsa, Trends in Neurosciences 45, 777–790 (2022)

2022
[60]

Chen and S

L. Chen and S. A. Campbell, Journal of Computational Neuroscience 50, 445–469 (2022)

2022
[61]

Markram, E

H. Markram, E. Muller, S. Ramaswamy, M. W. Reimann, M. Abdellah, C. A. Sanchez, A. Ailamaki, L. Alonso- Nanclares, N. Antille, S. Arsever, G. A. A. Kahou, T. K. Berger, A. Bilgili, N. Buncic, A. Chalimourda, G. Chindemi, J.-D. Courcol, F. Delalondre, V. De- lattre, S. Druckman...

2015
[62]

H. K. Yao, A. Guet-McCreight, F. Mazza, H. Moradi Chameh, T. D. Prevot, J. D. Griffiths, S. J. Tripathy, T. A. Valiante, E. Sibille, and E. Hay, Cell Reports 38, 110232 (2022)

2022
[63]

Romani, A

A. Romani, A. Antonietti, D. Bella, J. Budd, E. Gi- acalone, K. Kurban, S. S´ aray, M. Abdellah, A. Ar- naudon, E. Boci, C. Colangelo, J.-D. Courcol, T. Dele- montex, A. Ecker, J. Falck, C. Favreau, M. Gevaert, J. B. Hernando, J. Herttuainen, G. Ivaska, L. Kanari, A.-K. Kaufma...

2024
[64]

J. J. J. Hjorth, A. Kozlov, I. Carannante, J. Frost Nyl´ en, R. Lindroos, Y. Johansson, A. Tokarska, M. C. Dorst, S. M. Suryanarayana, G. Silberberg, J. Hellgren Ko- taleski, and S. Grillner, Proceedings of the National Academy of Sciences 117, 9554–9565 (2020)

2020
[65]

Guet-McCreight, H

A. Guet-McCreight, H. M. Chameh, F. Mazza, T. D. Prevot, T. A. Valiante, E. Sibille, and E. Hay, Commu- nications Biology 7 (2024), 10.1038/s42003-024-05907-1

2024 doi
[66]

Carannante, M

I. Carannante, M. Scolamiero, J. J. J. Hjorth, A. Ko- zlov, B. Bekkouche, L. Guo, A. Kumar, W. Chach´ olski, and J. H. Kotaleski, Network Neuroscience 8, 1149–1172 (2024)

2024
[67]

JAX: com- posable transformations of Python+NumPy programs,

J. Bradbury, R. Frostig, P. Hawkins, M. J. Johnson, C. Leary, D. Maclaurin, G. Necula, A. Paszke, J. Vander- Plas, S. Wanderman-Milne, and Q. Zhang, “JAX: com- posable transformations of Python+NumPy programs,” (2018)

2018
[68]

Tejero-Cantero, J

A. Tejero-Cantero, J. Boelts, M. Deistler, J.-M. Lueck- mann, C. Durkan, P. Gon¸ calves, D. Greenberg, and J. Macke, Journal of Open Source Software 5, 2505 (2020)

2020
[69]

Durstewitz, G

D. Durstewitz, G. Koppe, and M. I. Thurm, Nature Reviews Neuroscience 24, 693–710 (2023)

2023
[70]

Stimberg, R

M. Stimberg, R. Brette, and D. F. Goodman, eLife 8 (2019), 10.7554/elife.47314

2019 doi
[71]

Pyragas and K

V. Pyragas and K. Pyragas, Physical Review E 110 (2024), 10.1103/physreve.110.064211

2024 doi
[72]

Dumont and B

G. Dumont and B. Gutkin, PLOS Computational Biol- ogy 15, e1007019 (2019)

2019
[73]

Byrne, R

A. Byrne, R. D. O’Dea, M. Forrester, J. Ross, and S. Coombes, Journal of Neurophysiology123, 726 (2020)

2020
[74]

Yamazaki, V.-K

K. Yamazaki, V.-K. Vo-Ho, D. Bulsara, and N. Le, Brain Sciences 12, 863 (2022)

2022
[75]

R. Gast, S. A. Solla, and A. Kennedy, Proceed- ings of the National Academy of Sciences 121 (2024), 10.1073/pnas.2311885121

2024 doi
[76]

Pyragas and K

V. Pyragas and K. Pyragas, Physical Review E 105 (2022), 10.1103/physreve.105.044402

2022 doi
[77]

Gandolfi, J

D. Gandolfi, J. Mapelli, S. M. G. Solinas, P. Triebkorn, E. D’Angelo, V. Jirsa, and M. Migliore, Nature Compu- tational Science 3, 264–276 (2023)

2023
[78]

Ott and T

E. Ott and T. M. Antonsen, Chaos: An Inter- disciplinary Journal of Nonlinear Science 18 (2008), 10.1063/1.2930766

2008 doi
[79]

Baldy, M

N. Baldy, M. Breyton, M. M. Woodman, V. K. Jirsa, and M. Hashemi, Neural Computation 36, 2030 (2024)
[80]

Di Volo, A

M. Di Volo, A. Romagnoni, C. Capone, and A. Destexhe, Neural computation 31, 653 (2019)

2019
[81]

Hashemi, A

M. Hashemi, A. Ziaeemehr, M. M. Woodman, J. Fousek, S. Petkoski, and V. K. Jirsa, Machine Learning: Science and Technology 5, 035019 (2024)

2024
[82]

Rabuffo, A

G. Rabuffo, A. Bandyopadhyay, C. Calabrese, K. Gudibanda, D. Depannemaecker, L. M. Takarabe, M. L. Saggio, M. Desroches, A. Ivanov, M.-L. Linne, et al. , eLife 14 (2025)

2025

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

Hashemi, D

M. Hashemi, D. Depannemaecker, M. Saggio, P. Triebkorn, G. Rabuffo, J. Fousek, A. Ziaeemehr, V. Sip, A. Athanasiadis, M. Breyton, M. Woodman, H. Wang, S. Petkoski, P. Sorrentino, and V. Jirsa, IEEE Reviews in Biomedical Engineering , 1–29 (2025)

work page 2025

[2] [2]

Breakspear, Nature Neuroscience 20, 340–352 (2017)

M. Breakspear, Nature Neuroscience 20, 340–352 (2017)

work page 2017

[3] [3]

H. E. Wang, P. Triebkorn, M. Breyton, B. Dollomaja, J.-D. Lemarechal, S. Petkoski, P. Sorrentino, D. Depan- nemaecker, M. Hashemi, and V. K. Jirsa, National Sci- ence Review 11 (2024), 10.1093/nsr/nwae079. 11

work page doi:10.1093/nsr/nwae079 2024

[4] [4]

Jirsa, H

V. Jirsa, H. Wang, P. Triebkorn, M. Hashemi, J. Jha, J. Gonzalez-Martinez, M. Guye, J. Makhalova, and F. Bartolomei, The Lancet Neurology 22, 443–454 (2023)

work page 2023

[5] [5]

Sanz-Leon, S

P. Sanz-Leon, S. A. Knock, A. Spiegler, and V. K. Jirsa, NeuroImage 111, 385–430 (2015)

work page 2015

[6] [6]

G. Deco, V. K. Jirsa, P. A. Robinson, M. Breakspear, and K. Friston, PLoS Computational Biology4, e1000092 (2008)

work page 2008

[7] [8]

Breyton, J

M. Breyton, J. Fousek, G. Rabuffo, P. Sorrentino, L. Kusch, M. Massimini, S. Petkoski, and V. Jirsa, (2024), 10.7554/elife.98920.1

work page doi:10.7554/elife.98920.1 2024

[8] [9]

V. K. Jirsa, W. C. Stacey, P. P. Quilichini, A. I. Ivanov, and C. Bernard, Brain 137, 2210–2230 (2014)

work page 2014