Pith. sign in

REVIEW 4 major objections 5 minor 62 references

Accessing the topological properties of human brain functional sub-circuits in Echo State Networks

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Echo-state networks whose reservoirs embed fMRI-derived brain functional sub-circuits outperform bipartite and degree-preserving null models, with the best readout network varying by task.

desk verdict The new idea is embedding Yeo functional networks as read-in/read-out modules on structural reservoirs, and it deserves a careful referee; but the node-mapping between parcellations is a load-bearing gap that needs explicit testing. read the letter →

arxiv 2412.14999 v2 pith:EK3D4RV6 submitted 2024-12-19 q-bio.NC

Bach Nguyen , Tianlong Chen , Shu Yang , Bojian Hou , Li Shen , Duy Duong-Tran This is my paper · ORCID
classification q-bio.NC
keywords echo-statenetworksreservoircomputingconnectome-basedfunctionalsub-circuitsstructuralconnectomeMRInetworktopologymemorycapacity
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

By embedding human structural connectomes as echo-state network reservoirs and using fMRI-derived functional sub-circuits (Yeo's seven networks plus subcortex) as the read-in and read-out node sets, this paper asks whether biological functional topology carries computational benefit. The authors show that these functional-sub-circuit-embedded reservoirs generally beat bipartite and degree-preserving null models across perceptual decision, delayed response, context-dependent decision, and memory-capacity tasks near criticality. They also find that the best-performing readout network depends on the task: default mode and somatomotor networks lead on three tasks, while dorsal attention and frontoparietal lead on perceptual decision-making. The paper concludes that performance is shaped by the graph-theoretic properties of the embedded functional sub-circuits, such as betweenness and communicability, and that reservoir performance is not strictly determined by the spectral radius (echo-state property).

What carries the argument

The pipeline is an echo-state network (ESN) whose reservoir adjacency matrix is a min-max-scaled, spectral-radius-normalized structural connectome, with input routed to subcortical nodes and the readout ridge-regression layer attached to nodes of one of the seven Yeo functional networks. The functional sub-circuits thus act as induced subgraphs defining read-in and read-out structure, and the comparison against null models (degree-preserving Maslov-Sneppen rewire, configuration model, uniform-weight randomization, and complete bipartite) ablates specific topological features. Communicability and group betweenness centrality quantify the dynamical and static properties of the readout subgraph that the paper links to performance.

What would settle it

Permute the assignment of the seven functional-network labels to the structural nodes and rerun the full ESN pipeline across many seeds: if the default-mode advantage on delayed-response, memory, and contextual tasks disappears under label permutation while the untreated connectome keeps it, then functional identity, not node size or placement, is what drives performance. The paper's own permutation null (Figure 8) suggests exactly this check is needed.

Watch

Extended reading notes

Core claim

The central claim is that the topology of the embedded fMRI-induced sub-circuits, not merely the reservoir's spectral radius, determines echo-state network performance, and that different functional networks are optimal for different computational demands. Concretely, when the structural connectome is the reservoir and a functional network is the readout node set, the default mode network (DMN) and somatomotor network outperform other networks on delayed-response, memory, and contextual tasks, whereas dorsal attention and frontoparietal networks win on the perceptual decision-making task; these rankings are tied to betweenness centrality and communicability of the readout subgraph. The paper further claims that reservoirs preserving the structural topology (control and induced-subgraph models) outperform degree-sequence-preserving nulls (Maslov-Sneppen rewire and configuration model) and the bipartite feed-forward-like null, contradicting the notion that the echo-state property, as measured by spectral radius, alone sets performance.

Load-bearing premise

That the nodes of the diffusion-MRI-derived structural connectome correspond one-to-one to the fMRI-derived Schaefer and Yeo parcels, despite coming from different cohorts and different parcellation schemes; if this mapping is wrong, every functional-network-specific performance ranking could be an artifact of mislabeled nodes.

Editorial extensions

If this is right

  • Task-specific readout choice becomes a design lever: choosing DMN or somatomotor readouts for memory-heavy tasks and dorsal attention or frontoparietal readouts for perceptual decisions can improve ESN performance near criticality.
  • Connectome-based reservoir performance cannot be predicted from spectral radius alone; topological features of the embedded subgraph must be considered.
  • Degree-preserving nulls (Maslov-Sneppen rewire and configuration model) underperform the true structural topology, implying that the specific wiring pattern of brain connectivity carries computational value beyond degree sequence.
  • The subgraph model performs almost identically to the full structural model on three of four tasks, suggesting information principally propagates through input-output pathways in this setup.
  • Uniform-weight nulls with preserved topology are competitive, indicating that edge weights matter less than topology for some tasks.

Reading between the lines

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

  • If the functional-identity result is real, a natural extension is to design neuromorphic reservoirs by optimizing readout subgraphs for communicability to the input nodes, rather than by tuning spectral radius alone; the paper does not propose such an optimization.
  • A same-subject multimodal dataset (diffusion MRI plus fMRI with identical parcellation) would directly test whether the reported rankings survive when structural and functional data come from the same brains; until then, cross-cohort node correspondence remains the main confound.
  • The paper's own permutation null appears to disrupt the DMN advantage pattern in the figures; systematically comparing that null against the control across seeds and tasks would clarify whether functional identity or node size and placement drive the ranking.
  • Because communicability is a global path-based measure, the authors' correlational analysis could be extended to predict the performance of arbitrary subgraphs in connectome reservoirs, allowing synthetic testbeds that separate size, placement, and biological identity.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes a pipeline for embedding human brain functional subcircuits, derived from fMRI, into echo-state networks (ESNs) whose reservoir connectivity is given by structural connectomes. The authors evaluate several reservoir topologies—structural connectome, vertex-induced subgraph, Maslov–Sneppen rewired, uniform-weight, complete bipartite, configuration model, and label-permutation null—on four synthetic tasks (PDM, PDMDR, CDM, MemCap) across a range of spectral radius values. They find that structural and subgraph reservoirs outperform degree-preserving nulls and bipartite feedforward-like nulls, and that the ranking of read-out functional networks differs across tasks, with DMN performing best in three of four tasks at near-criticality. The paper then correlates graph statistics (communicability, modularity, group betweenness) of the functional subcircuits with task performance to argue that neuro-physiological characteristics of the embedded subcircuits determine reservoir performance.

Significance. If the central finding holds, it would show that the specific mesoscale functional topology embedded in a reservoir—not just spectral radius or raw network size—shapes computational performance, offering a neuroscience-inspired design principle for reservoir computing. The experimental design has notable strengths: multiple null models ablate distinct structural properties, four tasks span different computation–memory balances, 1000 runs are used, and several spectral radius values are swept. However, the load-bearing identification of functional network labels on structural nodes is not established, and a critical control (the label-permutation null) is referenced but not described or numerically reported. Until these gaps are closed, the abstract's claim that performance optimums depend on the neuro-physiological characteristics of the subcircuits remains unverified. As a result, the paper's significance is conditional on a fixable but nontrivial methodological clarification.

major comments (4)
  1. [§4.1, §3.2] The mapping between the 463/1015 structural nodes from [2] and the Yeo-7 functional labels obtained from the Schaefer-500 parcellation ([49]) is never specified. The two node sets differ in size, parcellation, and cohort, and the text does not state how a structural node receives a Yeo label (nearest-parcel assignment, atlas overlap, or other). Without this mapping, the induced subgraphs of §3.3 and all functional-network-specific analyses in Figures 8–11 do not necessarily correspond to the named brain circuits. The manuscript's own §5.3 hypothesizes that Yeo networks derived from the Lausanne and Schaefer parcellations are 'not information-theoretically aligned,' which is an explicit admission that the central node-label correspondence is insecure.
  2. [§4.4, Figures 8–9] The label-permutation null is listed as a model and its results appear in Figures 8 and 9, but its construction is not described anywhere in the available text, and the pointer to the appendix is missing ('Appendix ??'). This null is the appropriate control for testing whether functional-network-specific rankings are artifacts of the node-label assignment. Without a description of how labels are permuted and what the outcomes are, the claim that performance differences reflect the neuro-physiological identity of the subcircuits is not supported by the reported evidence.
  3. [§5.1] The text cites the wrong figures when discussing the association between node count, betweenness, and best-performance counts: it refers to 'Figure 5' and 'Figure 4,' but those figures show performance versus spectral radius and reservoir configuration box plots, respectively. The relevant figures are Figures 10 and 8–9. The mis-citation obscures the basis for the claimed association between DMN's size, betweenness, and its top performance, and makes the analysis difficult to audit.
  4. [§5.2, Figure 11] The correlation between communicability and PDMDR performance is presented as a scatter plot with only seven data points (one per functional network), and the reported value ρ=2.245e-10 is stated to be the slope of the best linear fit, not a correlation coefficient. No p-value, confidence interval, or goodness-of-fit measure is given. The assertion that the betweenness panel shows 'four separate clusters' is made without a clustering criterion or any quantitative support. These analyses are too thin to sustain the strength of the conclusions drawn about which graph-theoretical properties determine ESN performance.
minor comments (5)
  1. [§4.5] The COVID-19 prediction experiment is referenced as 'Table ??' but no results table is provided in the manuscript; this section should either be completed with the actual numerical results or removed.
  2. [§4.2, §4.5] Several placeholders 'Appendix ??' appear in the text, indicating missing supplementary material. The experimental setup details and additional COVID-19 model descriptions promised in these references are not available for review.
  3. [Figure 7] The caption of Figure 7 states that F1-score performance is shown 'on three datasets,' but the displayed subplot is only labeled for ContextDecisionMaking; please clarify whether panels are omitted or add the missing subplots.
  4. [§5.3] The formula Q = ∑_{i,j}(a_{ij} − p_{ij})(δ_{σ_i,u}, δ_{σ_j,v}) is not self-contained: the Kronecker delta notation and the meaning of the partition vector σ are not defined in the text.
  5. [References] References [5] and [45] appear to cite the same Yeo et al. work with identical titles; please consolidate to avoid duplicate entries.

Circularity Check

0 steps flagged · score 2.0 of 10

No derivation in the paper reduces to its own inputs; the ESN comparisons are computed from fixed structural connectomes and standard ridge-regression readouts, while the graph-statistic correlations are post-hoc analyses rather than fitted predictions. Minor non-load-bearing self-citations occur, and the parcellation-alignment concern is a validity issue, not circularity.

full rationale

The central claim is that fMRI-induced sub-circuit-embedded ESNs outperform null models and that performance optimums depend on the graph properties of the functional sub-circuits. The derivation chain is not circular: the reservoir matrix WR is constructed from the structural connectome by fixed scaling (Eq. 5), the readout WO is trained by ridge regression against task targets (Eq. 4), and none of the graph statistics used in the explanatory analysis (node count, betweenness, communicability, modularity) are fitted parameters in the ESN. The null models—Maslov–Sneppen rewire, uniform weights, bipartite, and configuration model—are explicit ablations of topology or weights. The paper's use of the same Yeo functional networks to define read-out node sets and then to interpret performance through those networks' graph statistics is an empirical association, not a definitional equality: performance is measured from the reservoir dynamics, not constructed from the statistics. Self-citations such as refs. 27 and 29 provide terminology and prior context for 'functional sub-circuits' but are not the evidential basis for the ESN performance comparisons, which are independently computed and benchmarked against LSTM/RNN and real-world COVID-19 data. The serious concern about one-to-one alignment between 463/1015 structural nodes and Schaefer-500/Yeo labels, explicitly conceded in Section 5.3, is a data-mapping and external-validity problem, not circularity. The unreported label-permutation null is an incompleteness issue, also not a circular step. The score of 2 reflects the presence of non-load-bearing self-citations without any fitted-input-called-prediction or self-definitional reduction.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central empirical claim rests on hand-set hyperparameters (C, lambda, alpha), an assumed node correspondence between structural and functional parcellations, and the validity of the Yeo sub-circuit labels. These are not derived or ablated, but none is fitted to the benchmark outputs, so the circularity burden is low. No invented entities appear.

free parameters (3)
  • Input scaling constant C = not stated
    Eq. 3 fixes all nonzero input weights to C; its value and sensitivity are not reported.
  • Ridge regularization lambda = not stated
    Eq. 4 depends on lambda; no value or grid search is reported.
  • Spectral radius alpha = 0.8, 0.9, 0.95, 1.0, 1.1, 1.2 (headlines at 0.95)
    Alpha is swept rather than fitted, but the choice of 0.95 for the headline comparisons is manual and affects which model wins.
assumptions (5)
  • standard math Echo-state update x[t] = tanh(W_R x[t-1] + W_I u[t]) and ridge regression readout define the computational framework.
    Section 3.1; standard ESN definitions from Jaeger (ref 42), not re-derived.
  • domain assumption Yeo 7-network functional parcellation is a valid a priori partition and can be used to define read-in and read-out node sets.
    Section 3.2; the entire experiment uses these labels without testing alternative parcellations.
  • domain assumption Structural connectome nodes can be identified with functional parcellation nodes despite different source data and parcellation scales.
    Section 4.1; the combination of 463/1015 structural nodes with Schaefer 500 functional parcels is assumed node-compatible.
  • domain assumption Subcortical nodes are the correct input region, acting as relay stations for sensory signals.
    Section 3.2; no ablation or physiological justification beyond plausibility is provided.
  • domain assumption Synthetic NeuroGym tasks measure processing and memorization in a way that generalizes to real cognitive computation.
    Section 4.1; used as ground truth for all performance comparisons.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Accessing the topological properties of human brain functional sub-circuits in Echo State Networks." pith.science (2026). https://pith.science/paper/EK3D4RV6

@misc{pith2026241214999,
  author       = {Pith},
  title        = {Pith review of: Accessing the topological properties of human brain functional sub-circuits in Echo State Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EK3D4RV6}},
  note         = {Machine review of arXiv:2412.14999}
}
read the original abstract

Recent years have witnessed an emerging trend in neuromorphic computing that centers around the use of brain connectomics as a blueprint for artificial neural networks. Connectomics-based neuromorphic computing has primarily focused on embedding human brain large-scale structural connectomes (SCs), as estimated from diffusion Magnetic Resonance Imaging (dMRI) modality, to echo-state networks (ESNs). A critical step in ESN embedding requires pre-determined read-in and read-out layers constructed by the induced subgraphs of the embedded reservoir. As \textit{a priori} set of functional sub-circuits are derived from functional MRI (fMRI) modality, it is unknown, till this point, whether the embedding of fMRI-induced sub-circuits/networks onto SCs is well justified from the neuro-physiological perspective and ESN performance across a variety of tasks. This paper proposes a pipeline to implement and evaluate ESNs with various embedded topologies and processing/memorization tasks. To this end, we showed that different performance optimums highly depend on the neuro-physiological characteristics of these pre-determined fMRI-induced sub-circuits. In general, fMRI-induced sub-circuit-embedded ESN outperforms simple bipartite and various null models with feed-forward properties commonly seen in MLP for different tasks and reservoir criticality conditions. We provided a thorough analysis of the topological properties of pre-determined fMRI-induced sub-circuits and highlighted their graph-theoretical properties that play significant roles in determining ESN performance.

Figures

Figures reproduced from arXiv: 2412.14999 by the authors.

Figure 1
Figure 1. Overview of the pipeline. Memory, processing, or mixed datasets are fed into the model through a static input layer, which is then projected into higher dimensions with the reservoir layer, and readout using a ridge regression output layer. Structural, functional, and null model connectomes are embedded in the reservoir layer. training data {u[t], y[t]}t=1,...,T . The matrix WO is optimized using ridge regression th… view at source ↗
Figure 2
Figure 2. Human brain functional networks (sub-circuits) as parcellated by Yeo and colleagues5 . Figure is referenced from44 . each edge in the connectome exactly once using the Maslov￾Sneppen algorithm46. From the rewiring procedure, we ob￾tained a reference null connectome with the original constraint still moderately enforced. Uniform weights We preserved the underlying topology of the connectome while randomizing the conn… view at source ↗
Figure 3
Figure 3. General performance of various reservoir models tested on perceptual datasets. Performance is shown as a function of the reservoir’s spectral radius alpha. 0.8 0.9 1.0 1.1 1.2 alphas 0.4 0.6 0.8 Pearson correlation MemoryCapacity 0.8 0.9 1.0 1.1 1.2 alphas 0.60 0.65 0.70 0.75 F1-score ContextDecisionMaking Structural Subgraph MS rewire Uniform weights Bipartite null Configuration [PITH_FULL_IMAGE:figures/full_fig_p… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: General performance of various reservoir models tested on capacity and contextual datasets. Performance is shown as a function of the reservoir’s spectral radius alpha. 5 Functional sub-circuits’ topological per￾formance Analysis In this section, we analyze the neuro-p…
Figure 5
Figure 5. Figure 5: Performance of various reservoir models on perceptual datasets w.r.t. the reservoir configuration at spectral radius alpha near criticality (0.95), demonstrating the best performances of ESNs in most cases. (config: degree-sequence-preserving configuration random graph…
Figure 6
Figure 6. Figure 6: Performance of various reservoir models on capacity and contextual datasets w.r.t. the reservoir configuration at spectral radius alpha near criticality (0.95). (e.g., communicability) properties, quantifying their relation￾ship with model performance55–57 . 5.1 Read-o…
Figure 7
Figure 7. Figure 7: F1-score performance of various reservoir models on three datasets w.r.t. the reservoir configuration at spectral radius 0.95. ESN configurations are identical to the original paper’s description, LSTM and RNN models are followed by the size of the hidden layer. DA FP …
Figure 8
Figure 8. Figure 8: The number of best performance runs by model and nulls across 1000 runs, by each functional network on identical initialization of perceptual datasets. Note that the total of all functional networks from the same color (reservoir configuration) adds up to exactly 1000.…
Figure 9
Figure 9. Figure 9: The number of best performance runs by model and nulls across 1000 runs, by each functional network on identical initialization of capacity and contextual datasets. DMNSMVA VIS LIM DA FP functional network 0 20 40 60 80 100 120 #. of readout node Struct500, readout nod…
Figure 10
Figure 10. Figure 10: Struct500 (463 nodes) connectome, various graph statistics, 1000 different consensuses. 0 1 2 3 4 Communicability 1e8 0.55 0.60 0.65 0.70 0.75 0.80 F1-score = 2.245e-10 0.326 0.328 0.330 0.332 Modularity 0.64 0.66 0.68 0.70 0.72 0.74 0.76 0.78 F1-score = 0.188 0.025 0…
Figure 11
Figure 11. Figure 11: Structural 500 nodes, statistic-performance correlations, PDMDR, α = 0.95, with slope of best linear fit ρ. statistics and empirical performance on PDMDR. Communi￾cability spread shows that functional networks with high com￾municability generally perform well compared…
Figure 12
Figure 12. Figure 12: Modularity histogram: difference between structural and functional connectome. We performed cross-comparison between processed functional and structural connec￾tome on modularity measures, perform￾ing community de￾tection and compar￾ing adjusted mutual info score with…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 45 canonical work pages

  1. [2]

    E., Richards, B

    Suarez, L. E., Richards, B. A., Lajoie, G. & Misic, B. Learning function from structure in neuromorphic networks. Nat. Mach. Intell. 3, 771 – 786 (2020)

    work page 2020

  2. [49]

    Tipnis, U. et al. Functional connectome fingerprint gradients in young adults (2021). 2011.05212

    work page Pith review arXiv 2021

  3. [1]

    Damicelli, F., Hilgetag, C. C. & Goulas, A. Brain connectivity meets reservoir computing. PLOS Comput. Biol. 18, 1–17, DOI: 10.1371/journal.pcbi.1010639 (2022)

    doi:10.1371/journal.pcbi.1010639 2022

  4. [3]

    L., Villafañe–Delgado, M

    McDaniel, S. L., Villafañe–Delgado, M. & Johnson, E. C. Investigating echo state network performance with biologically-inspired hierarchical network structure. In 2022 International Joint Conference on Neural Networks (IJCNN) , 01–08, DOI: 10.1109/IJCNN55064. 2022.9892248 (2022)

    work page arXiv 2022

  5. [4]

    Suarez, L. E. et al. conn2res: A toolbox for connectome-based reservoir computing. bioRxiv DOI: 10.1101/2023.05.31.543092 (2023). https://www.biorxiv.org/content/early/2023/06/04/2023.05.31.543092.full.pdf

    work page doi:10.1101/2023.05.31.543092 2023

  6. [5]

    Yeo, B. T.et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J. neurophysiology (2011)

    work page 2011

  7. [6]

    & Stroobandt, D

    Verstraeten, D., Schrauwen, B., D’Haene, M. & Stroobandt, D. An experimental unification of reservoir computing methods. Neural Networks 20, 391–403, DOI: https://doi.org/10.1016/j.neunet.2007.04.003 (2007). Echo State Networks and Liquid State Machines

    doi:10.1016/j.neunet.2007.04.003 2007

  8. [7]

    & Subramanian, D

    Chattopadhyay, A., Hassanzadeh, P. & Subramanian, D. Data-driven predictions of a multiscale lorenz 96 chaotic system using machine-learning methods: reservoir computing, artificial neural network, and long short-term memory network. Nonlinear Process. Geophys. 27, 373–389, DOI: 10.5194/npg-27-373-2020 (2020)

    doi:10.5194/npg-27-373-2020 2020

Show all 62 references
  1. [8]

    Shahi, S., Fenton, F. H. & Cherry, E. M. Prediction of chaotic time series using recurrent neural networks and reservoir computing techniques: A comparative study. Mach. Learn. with Appl. 8, 100300, DOI: https://doi.org/10.1016/j.mlwa.2022.100300 (2022)

    2022

  2. [9]

    & Guéry-Odelin, D

    Bompas, S., Georgeot, B. & Guéry-Odelin, D. Accuracy of neural networks for the simulation of chaotic dynamics: Precision of training data vs precision of the algorithm. Chaos: An Interdiscip. J. Nonlinear Sci. 30, 113118, DOI: 10.1063/5.0021264 (2020). https://pubs.aip.org/ai...

    2020 doi

  3. [10]

    & Magri, L

    Huhn, F. & Magri, L. Gradient-free optimization of chaotic acoustics with reservoir computing. Phys. Rev. Fluids 7, 014402, DOI: 10.1103/PhysRevFluids.7.014402 (2022)

    2022 doi

  4. [11]

    A., Penny, S

    Platt, J. A., Penny, S. G., Smith, T. A., Chen, T.-C. & Abarbanel, H. D. A systematic exploration of reservoir computing for forecasting complex spatiotemporal dynamics. Neural Networks 153, 530–552, DOI: https://doi.org/10.1016/j.neunet.2022.06.025 (2022)

    2022 doi

  5. [12]

    & Kleemann, H

    Cucchi, M., Abreu, S., Ciccone, G., Brunner, D. & Kleemann, H. Hands-on reservoir computing: a tutorial for practical implementation. Neuromorphic Comput. Eng. 2, 032002, DOI: 10.1088/2634-4386/ac7db7 (2022)

    2022 doi

  6. [13]

    & Sohn, I

    Kaviani, S. & Sohn, I. Application of complex systems topologies in artificial neural networks optimization: An overview. Expert. Syst. with Appl. 180, 115073, DOI: https://doi.org/10.1016/j.eswa.2021.115073 (2021)

    2021

  7. [14]

    & Perc, M

    Erkaymaz, O., Ozer, M. & Perc, M. Performance of small-world feedforward neural networks for the diagnosis of diabetes. Appl. Math. Comput. 311, 22–28, DOI: https://doi.org/10.1016/j.amc.2017.05.010 (2017)

    2017 doi

  8. [15]

    & Ozer, M

    Erkaymaz, O. & Ozer, M. Impact of small-world network topology on the conventional artificial neural network for the diagnosis of diabetes. Chaos, Solitons & Fractals 83, 178–185, DOI: https://doi.org/10.1016/j.chaos.2015.11.029 (2016)

    2016 doi

  9. [16]

    Xie, S., Kirillov, A., Girshick, R. & He, K. Exploring randomly wired neural networks for image recognition. In 2019 IEEE/CVF International Conference on Computer Vision (ICCV) , 1284–1293, DOI: 10.1109/ICCV .2019.00137 (2019)

    2019

  10. [17]

    & Toschi, N

    Boccato, T., Ferrante, M., Duggento, A. & Toschi, N. Beyond multilayer perceptrons: Investigating complex topologies in neural networks. Neural Networks 171, 215–228, DOI: https://doi.org/10.1016/j.neunet.2023.12.012 (2024)

    2024 doi

  11. [18]

    & Shen, L

    Xu, F., Duong-Tran, D., Zhao, Y . & Shen, L. Caudal and thalamic segregation in white matter brain network communities in alzheimer’s disease population. In IEEE-EMBS International Conference on Biomedical and Health Informatics

  12. [19]

    Xu, F. et al. Consistency of graph theoretical measurements of alzheimer’s disease fiber density connectomes across multiple parcellation scales. In 2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) , 1323–1328 (IEEE, 2022)

    2022

  13. [20]

    Xu, F. H. et al. Topology-based clustering of functional brain networks in an alzheimer’s disease cohort. AMIA Summits on Transl. Sci. Proc. 2024, 449 (2024)

    2024

  14. [21]

    Garai, S. et al. Quantifying contributions from topological cycles in the brain network towards cognition. bioRxiv 2024–06 (2024)

    2024

  15. [22]

    Abbas, K. et al. Geff: Graph embedding for functional fingerprinting. NeuroImage 221, 117181 (2020)

    2020

  16. [23]

    Abbas, K. et al. Tangent functional connectomes uncover more unique phenotypic traits. Iscience 26 (2023)

    2023

  17. [24]

    Amico, E. et al. Towards a mathematical theory of communication for the human connectome. arXiv preprint arXiv:1911.02601 (2019)

    2019 arXiv

  18. [25]

    Amico, E. et al. Toward an information theoretical description of communication in brain networks. Netw. Neurosci. 5, 646–665 (2021)

    2021

  19. [26]

    Chiêm, B. et al. Improving functional connectome fingerprinting with degree-normalization. Brain Connect. 12, 180–192 (2022)

    2022

  20. [27]

    Duong-Tran, D. et al. Homological landscape of human brain functional sub-circuits. Mathematics 12, 455 (2024)

    2024

  21. [28]

    & Shen, L

    Duong-Tran, D., Magsino, M., Goñi, J. & Shen, L. Preserving human large-scale brain connectivity fingerprint identifiability with random projections. In 2024 IEEE International Symposium on Biomedical Imaging (ISBI) , 1–5 (IEEE, 2024)

    2024

  22. [29]

    Duong-Tran, D. et al. A principled framework to assess the information-theoretic fitness of brain functional sub-circuits. Mathematics 12, DOI: 10.3390/math12192967 (2024)

    2024 doi

  23. [30]

    Nguyen, N. et al. V olume-optimal persistence homological scaffolds of hemodynamic networks covary with meg theta-alpha aperiodic dynamics. In International Conference on Medical Image Computing and Computer-Assisted Intervention , 519–529 (Springer, 2024)

    2024

  24. [31]

    A., Zhao, Y

    Garai, S., Xu, F., Duong-Tran, D. A., Zhao, Y . & Shen, L. Mining correlation between fluid intelligence and whole-brain large scale structural connectivity. AMIA Summits on Transl. Sci. Proc. 2023, 225 (2023)

    2023

  25. [32]

    Garai, S. et al. Effect of brain network scale on persistence cycles: An adni comparative study. In Alzheimer’s Association International Conference (ALZ, 2024)

    2024

  26. [33]

    & Wang, X

    Song, D., Shen, L., Duong-Tran, D. & Wang, X. Causality-based subject and task fingerprints using fmri time-series data. InProceedings of the 15th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics , 1–10 (2024)

    2024

  27. [34]

    Power, J. D. et al. Functional network organization of the human brain. Neuron 72, 665–678 (2011)

    2011

  28. [35]

    Shine, J. M. et al. The dynamics of functional brain networks: integrated network states during cognitive task performance. Neuron 92, 544–554 (2016)

    2016

  29. [36]

    Shine, J. M. & Poldrack, R. A. Principles of dynamic network reconfiguration across diverse brain states. NeuroImage (2017)

    2017

  30. [37]

    Shine, J. M. et al. Human cognition involves the dynamic integration of neural activity and neuromodulatory systems. Nat. Neurosci. – (2019)

    2019

  31. [38]

    & Betzel, R

    Sporns, O. & Betzel, R. F. Modular brain networks. Annu. review psychology 67, 613–640 (2016)

    2016

  32. [39]

    Duong-Tran, D. et al. A morphospace framework to assess configural breadth based on brain functional networks. arXiv preprint arXiv:1901.10962 (2019). 11/12

    2019 arXiv

  33. [40]

    & Domenico, M

    d’Andrea, V ., Puppin, M. & Domenico, M. D. Complex topological features of reservoirs shape learning performances in bio-inspired recurrent neural networks (2022). 2211.00161

    2022 arXiv

  34. [41]

    J., Duncan, J

    Achterberg, J., Akarca, D., Strouse, D. J., Duncan, J. & Astle, D. E. Spatially embedded recurrent neural networks reveal widespread links between structural and functional neuroscience findings. Nat. Mach. Intell. 5, 1369–1381, DOI: 10.1038/s42256-023-00748-9 (2023)

    2023 doi

  35. [42]

    echo state

    Jaeger, H. The “echo state” approach to analysing and training recurrent neural networks-with an erratum note. Bonn, Ger. Ger. Natl. Res. Cent. for Inf. Technol. GMD Tech. Rep. 148, 13 (2001)

    2001

  36. [43]

    Wig, G. S. Segregated systems of human brain networks. Trends Cogn. Sci. 21, 981–996, DOI: https://doi.org/10.1016/j.tics.2017.09.006 (2017)

    2017 doi

  37. [44]

    Duong-Tran, D. et al. A morphospace of functional configuration to assess configural breadth based on brain functional networks. Netw. Neurosci. 1–36 (2021)

    2021

  38. [45]

    Thomas Yeo, B. T.et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J. Neurophysiol. 106, 1125–1165, DOI: 10.1152/jn.00338.2011 (2011). PMID: 21653723, https://doi.org/10.1152/jn.00338.2011

    2011

  39. [46]

    & Sneppen, K

    Maslov, S. & Sneppen, K. Specificity and stability in topology of protein networks.Science 296, 910–913, DOI: 10.1126/science.1065103 (2002). https://www.science.org/doi/pdf/10.1126/science.1065103

    2002 doi

  40. [47]

    Newman, M. E. J. The structure and function of complex networks. SIAM Rev. 45, 167–256, DOI: 10.1137/S003614450342480 (2003). https://doi.org/10.1137/S003614450342480

    2003 doi

  41. [48]

    F., Griffa, A., Hagmann, P

    Betzel, R. F., Griffa, A., Hagmann, P. & Miši ´c, B. Distance-dependent consensus thresholds for generating group-representative structural brain networks. Netw. Neurosci. 3, 475–496, DOI: 10.1162/netn_a_00075 (2019). https://direct.mit.edu/netn/article-pdf/3/2/ 475/1092648/ne...

    2019 doi

  42. [50]

    & Movshon, J

    Britten, K., Shadlen, M., Newsome, W. & Movshon, J. The analysis of visual motion: a comparison of neuronal and psychophysical performance. J. Neurosci. 12, 4745–4765, DOI: 10.1523/JNEUROSCI.12-12-04745.1992 (1992). https://www.jneurosci.org/content/12/ 12/4745.full.pdf

    1992

  43. [51]

    K., Fontolan, L., Romani, S

    Inagaki, H. K., Fontolan, L., Romani, S. & Svoboda, K. Discrete attractor dynamics underlies persistent activity in the frontal cortex. Nature 566, 212–217, DOI: 10.1038/s41586-019-0919-7 (2019)

    2019 doi

  44. [52]

    Mante, V ., Sussillo, D., Shenoy, K. V . & Newsome, W. T. Context-dependent computation by recurrent dynamics in prefrontal cortex. Nature 503, 78–84, DOI: 10.1038/nature12742 (2013)

    2013 doi

  45. [53]

    Short term memory in echo state networks, DOI: 10.24406/publica-fhg-291107 (2001)

    Jaeger, H. Short term memory in echo state networks, DOI: 10.24406/publica-fhg-291107 (2001)

    2001 doi

  46. [54]

    & Vazirgiannis, M

    Panagopoulos, G., Nikolentzos, G. & Vazirgiannis, M. Transfer graph neural networks for pandemic forecasting. Proc. AAAI Conf. on Artif. Intell. 35, 4838–4845, DOI: 10.1609/aaai.v35i6.16616 (2021)

    2021 doi

  47. [55]

    Freeman, L. C. A set of measures of centrality based on betweenness. Sociometry 40, 35–41 (1977)

    1977

  48. [56]

    & Hatano, N

    Estrada, E. & Hatano, N. Communicability in complex networks. Phys. Rev. E 77, DOI: 10.1103/physreve.77.036111 (2008)

    2008 doi

  49. [57]

    Clauset, A., Newman, M. E. & Moore, C. Finding community structure in very large networks. Phys. review E 70, 066111 (2004)

    2004

  50. [58]

    & Dolev, S

    Puzis, R., Elovici, Y . & Dolev, S. Fast algorithm for successive computation of group betweenness centrality.Phys. Rev. E 76, 056709, DOI: 10.1103/PhysRevE.76.056709 (2007)

    2007 doi

  51. [59]

    Clauset, A., Newman, M. E. J. & Moore, C. Finding community structure in very large networks. Phys. Rev. E 70, DOI: 10.1103/ physreve.70.066111 (2004)

    2004

  52. [60]

    Q., Clare Kelly, A., Biswal, B

    Uddin, L. Q., Clare Kelly, A., Biswal, B. B., Xavier Castellanos, F. & Milham, M. P. Functional connectivity of default mode network components: Correlation, anticorrelation, and causality. Hum. Brain Mapp. 30, 625–637, DOI: 10.1002/hbm.20531 (2009)

    2009 doi

  53. [61]

    Li, D. et al. DALK: Dynamic co-augmentation of LLMs and KG to answer Alzheimer‘s disease questions with scientific literature. In Al-Onaizan, Y ., Bansal, M. & Chen, Y .-N. (eds.)Findings of the Association for Computational Linguistics: EMNLP 2024 , 2187–2205, DOI: 10.18653/v...

    2024 doi

  54. [62]

    Lee, J. et al. Knowledge-driven feature selection and engineering for genotype data with large language models. arXiv preprint arXiv:2410.01795 (2024). 12/12

    2024 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.