arxiv: 2604.18015 · v2 · submitted 2026-04-20 · ❄️ cond-mat.dis-nn · cs.ET

Recognition: unknown

Intrinsic Neuro-Synaptic Spiking Dynamics and Resonance in Memristive Networks

Georg A. Gottwald, Yinhao Xu, Zdenka Kuncic

Authors on Pith no claims yet

Pith reviewed 2026-05-10 03:38 UTC · model grok-4.3

classification ❄️ cond-mat.dis-nn cs.ET

keywords memristive networksspiking dynamicsnonlinear resonanceneuromorphic systemsself-organizing networksintrinsic dynamicscomputational optimality

0 comments

The pith

Memristive networks naturally produce brain-like spiking dynamics with resonance at their intrinsic frequency.

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

Self-organizing memristive networks adapt their connections in response to input signals through built-in neuro-synaptic properties. These networks generate spiking patterns at the population level that resemble those seen in biological neural systems for both constant and oscillating inputs. The nonlinear characteristics of these spikes reach their peak when the driving signal frequency aligns with the network's own natural response time, revealing a form of nonlinear resonance. The frequency best suited for computational tasks is identified as the highest one just before resonance begins.

Core claim

Self-organizing memristive networks with intrinsic neuro-synaptic dynamics and heterogeneous topology naturally generate neuronal population spiking dynamics similar to biological systems. For AC inputs, nonlinear spike-like features are maximized at the frequency matching the network's intrinsic dynamical timescale, where nonlinear resonance occurs. The optimal frequency for computation is the maximal frequency before the onset of this resonance.

What carries the argument

The combination of intrinsic neuro-synaptic dynamics in individual memristors with the heterogeneous network topology that enables self-organization and emergent spiking.

If this is right

Networks can emulate biological computation using only their physical properties.
Resonance provides a way to tune the network for enhanced spiking responses.
Computational use should avoid frequencies at or above the resonant point to maintain optimal performance.
Both DC and AC driving reveal the underlying intrinsic dynamics and resonance phenomena.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Physical implementations might exhibit similar resonance effects, enabling efficient analog computing hardware.
Further studies could explore how network size or connectivity heterogeneity affects the intrinsic timescale.
These dynamics may connect to broader principles in complex adaptive systems where resonance optimizes information processing.

Load-bearing premise

That the numerical models accurately represent real memristive device physics and that the generated spikes are meaningfully analogous to biological neuronal population activity.

What would settle it

Fabricating a physical memristive network and applying AC signals at frequencies around the predicted intrinsic timescale to observe whether spike nonlinearity maximizes exactly at that point.

Figures

Figures reproduced from arXiv: 2604.18015 by Georg A. Gottwald, Yinhao Xu, Zdenka Kuncic.

**Figure 2.** Figure 2: Ten arbitrary node readout voltages ri(t) from a simulated 100-node, 261-edge memristive network under a 0.2 V DC bias applied over 5 s. III. RESULTS & DISCUSSION [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Figure 3: Memristive network dynamics simulated on (a) a uniform lattice network and (b) a heterogeneous nanowire network (based on the connectivity of the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

**Figure 5.** Figure 5: Physical RC performance in a waveform regression task, where a [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗

**Figure 6.** Figure 6: The trajectories of xi and vi through phase space for an example reference edge (same as that of [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗

read the original abstract

Self-organizing memristive networks are physical circuits that dynamically reconfigure their circuitry in response to external input signals. Their adaptive behavior arises from intrinsic neuro-synaptic dynamics combined with a heterogeneous network topology. In this work, we demonstrate that such networks naturally generate neuronal population spiking dynamics similar to those observed in biological neuronal systems. This study investigates the intrinsic and emergent dynamics of memristive networks mathematically and numerically for both DC and AC input signals. Nonlinear spike-like features are maximized when the frequency of the input driving signal matches the network's intrinsic dynamical timescale, where nonlinear resonance is observed. Furthermore, the optimal frequency for computation is found to be the maximal frequency before the onset of resonance.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Memristive networks generate resonance-boosted spike-like dynamics, but the biological analogy and computational optimality rest on qualitative claims without quantitative metrics or device validation.

read the letter

The core observation is that self-organizing memristive networks produce neuronal-like population spiking, with nonlinear spike features maximized when the driving frequency matches the network's intrinsic timescale, and that the best frequency for computation sits just before resonance onset. The work backs this with both mathematical analysis and numerical simulations for DC and AC inputs, linking the behavior to the networks' adaptive reconfiguration and heterogeneous topology. That framing is useful because it treats the physical dynamics as the computing substrate rather than an add-on. The paper does a clean job of showing how the resonance emerges from the intrinsic neuro-synaptic rules without needing external programming. The soft spots are the missing quantitative checks. Similarity to biological spiking is stated but not measured with standard statistics such as inter-spike interval distributions, synchrony indices, or firing-rate correlations. Model fidelity to real memristors is also not demonstrated through direct comparison to experimental I-V curves, variability data, or switching statistics. The optimality claim for pre-resonance computation likewise lacks any task-based benchmark showing better performance at that frequency versus others. This is aimed at the neuromorphic and materials-computing community. Readers working on hardware that exploits physical dynamics will find the frequency rule worth testing, even if they have to supply their own metrics. The paper deserves a serious referee because the model is explicit and the resonance result is falsifiable; the gaps are fixable with added analysis rather than fundamental flaws. I would send it for review with the expectation that the authors add spike-train metrics and at least one simple computational benchmark.

Referee Report

3 major / 1 minor

Summary. The paper claims that self-organizing memristive networks naturally generate neuronal population spiking dynamics similar to those in biological systems. Through mathematical and numerical investigations of DC and AC inputs, nonlinear spike-like features are maximized when the driving frequency matches the network's intrinsic dynamical timescale (where nonlinear resonance occurs), and the optimal frequency for computation is identified as the maximal frequency before resonance onset.

Significance. If the numerical models faithfully reproduce physical memristor switching behavior and if the biological analogy is established with quantitative, falsifiable metrics, the work could be significant for neuromorphic and physical computing by suggesting that adaptive memristive hardware can exhibit emergent brain-like dynamics and self-tuned optimal operating frequencies without external parameter adjustment.

major comments (3)

[Abstract] Abstract: The manuscript asserts 'mathematical and numerical demonstration' of spiking similarity and resonance, yet the abstract (and by extension the core claims) contains no equations, no simulation protocols, no data figures, and no error analysis. This absence is load-bearing because the central claims about intrinsic dynamics and computational optimality cannot be evaluated or reproduced without the explicit model.
[Results] Throughout the manuscript (particularly any Model/Methods and Results sections): No quantitative metrics are provided to substantiate the claim that the generated spiking patterns are 'similar' to biological neuronal population dynamics. Standard measures such as inter-spike interval distributions, coefficient of variation, population synchrony indices, or firing-rate statistics are absent, leaving the analogy at a qualitative level that does not support the strong biological correspondence asserted.
[Model] Model description (likely §2 or equivalent): The specific equations governing the memristive elements (e.g., the state-variable update rule for resistance, the I-V nonlinearity, or the self-organization mechanism for network topology) are not presented. Without these, it is impossible to determine whether the reported spiking and resonance arise intrinsically from physical device physics or from particular modeling choices, directly undermining the 'intrinsic neuro-synaptic' claim.

minor comments (1)

[Abstract] The phrase 'maximal frequency before the onset of resonance' would benefit from an explicit operational definition (e.g., a threshold on the nonlinearity measure or a figure reference) to avoid ambiguity.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We are grateful to the referee for the detailed and constructive feedback. We have addressed each of the major comments below and will incorporate the suggested improvements in the revised manuscript to enhance clarity and support for our claims.

read point-by-point responses

Referee: [Abstract] Abstract: The manuscript asserts 'mathematical and numerical demonstration' of spiking similarity and resonance, yet the abstract (and by extension the core claims) contains no equations, no simulation protocols, no data figures, and no error analysis. This absence is load-bearing because the central claims about intrinsic dynamics and computational optimality cannot be evaluated or reproduced without the explicit model.

Authors: We note that abstracts are conventionally kept brief and free of technical details such as equations to ensure readability for a broad audience. The mathematical model, simulation protocols, figures, and any error analyses are fully detailed in the main body of the manuscript. To better convey the core claims in the abstract, we will revise it to include a concise reference to the key model components and the quantitative approaches used for demonstrating resonance and spiking dynamics. revision: yes
Referee: [Results] Throughout the manuscript (particularly any Model/Methods and Results sections): No quantitative metrics are provided to substantiate the claim that the generated spiking patterns are 'similar' to biological neuronal population dynamics. Standard measures such as inter-spike interval distributions, coefficient of variation, population synchrony indices, or firing-rate statistics are absent, leaving the analogy at a qualitative level that does not support the strong biological correspondence asserted.

Authors: We agree that incorporating quantitative metrics would provide stronger support for the biological analogy. The current manuscript includes qualitative comparisons and some basic statistics on spiking activity. In the revision, we will add explicit quantitative analyses, including inter-spike interval distributions, coefficient of variation for spike timing, population synchrony indices, and firing-rate statistics, with comparisons to established biological data where possible. revision: yes
Referee: [Model] Model description (likely §2 or equivalent): The specific equations governing the memristive elements (e.g., the state-variable update rule for resistance, the I-V nonlinearity, or the self-organization mechanism for network topology) are not presented. Without these, it is impossible to determine whether the reported spiking and resonance arise intrinsically from physical device physics or from particular modeling choices, directly undermining the 'intrinsic neuro-synaptic' claim.

Authors: The governing equations for the memristive elements and the network self-organization are presented in the Model section of the manuscript. We will revise this section to present the equations more clearly and prominently, perhaps with a summary table or additional explanatory text, to ensure readers can readily identify how the dynamics emerge from the physical model. revision: yes

Circularity Check

0 steps flagged

No circularity: claims rest on numerical demonstration rather than self-referential definitions or fits

full rationale

The abstract and context describe mathematical and numerical investigations showing that self-organizing memristive networks produce spike-like features maximized at resonance with the intrinsic timescale, with optimal computation frequency identified as the maximum before resonance onset. No equations, parameter-fitting procedures, or derivations are presented that would allow reduction of any 'prediction' to its own inputs by construction. No self-citations, uniqueness theorems, or ansatzes are invoked in the provided text to load-bear the central claims. The results are framed as emergent from the network dynamics under DC/AC inputs, independent of the target biological analogy or optimality statement. This is the normal case of a self-contained numerical study; external validation of model fidelity to physical memristors or quantitative spike metrics is a separate correctness issue, not circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Based solely on the abstract, no explicit free parameters, axioms, or invented entities are identified. The work appears to rely on standard memristor models and network topology assumptions without introducing new postulated particles or forces.

pith-pipeline@v0.9.0 · 5418 in / 1313 out tokens · 40613 ms · 2026-05-10T03:38:04.787033+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

27 extracted references

[1]

Brain-inspired computing needs a master plan,

A. Mehonic and A. J. Kenyon, “Brain-inspired computing needs a master plan,” vol. 604, no. 7905, pp. 255–260, 2022

2022
[2]

Opportunities for neuromorphic computing algorithms and applications,

C. D. Schuman, S. R. Kulkarni, M. Parsa, J. P. Mitchell, P. Date, and B. Kay, “Opportunities for neuromorphic computing algorithms and applications,”Nature Computational Science, vol. 2, no. 1, pp. 10–19, 2022

2022
[3]

Brain-inspired learning in artificial neural networks: A review,

S. Schmidgall, R. Ziaei, J. Achterberg, L. Kirsch, S. P. Hajiseyedrazi, and J. Eshraghian, “Brain-inspired learning in artificial neural networks: A review,”APL Machine Learning, vol. 2, no. 2, p. 021501, 2024

2024
[4]

Brain- inspired computing: A systematic survey and future trends,

G. Li, L. Deng, H. Tang, G. Pan, Y . Tian, K. Roy, and W. Maass, “Brain- inspired computing: A systematic survey and future trends,”Proceedings of the IEEE, vol. 112, no. 6, pp. 544–584, 2024

2024
[5]

Neuromorphic computing at scale,

D. Kudithipudi, C. Schuman, C. M. Vineyard, T. Pandit, C. Merkel, R. Kubendran, J. B. Aimone, G. Orchard, C. Mayr, R. Benosman,et al., “Neuromorphic computing at scale,”Nature, vol. 637, no. 8047, pp. 801– 812, 2025

2025
[6]

Networks of spiking neurons: the third generation of neural network models,

W. Maass, “Networks of spiking neurons: the third generation of neural network models,”Neural networks, vol. 10, no. 9, pp. 1659–1671, 1997

1997
[7]

Training spiking neural networks using lessons from deep learning,

J. K. Eshraghian, M. Ward, E. O. Neftci, X. Wang, G. Lenz, G. Dwivedi, M. Bennamoun, D. S. Jeong, and W. D. Lu, “Training spiking neural networks using lessons from deep learning,”Proceedings of the IEEE, vol. 111, no. 9, pp. 1016–1054, 2023

2023
[8]

A review of the integrate-and-fire neuron model: I. homogeneous synaptic input,

A. N. Burkitt, “A review of the integrate-and-fire neuron model: I. homogeneous synaptic input,”Biological cybernetics, vol. 95, no. 1, pp. 1–19, 2006

2006
[9]

Recent advances and future prospects for memristive materials, devices, and systems,

M.-K. Song, J.-H. Kang, X. Zhang, W. Ji, A. Ascoli, I. Messaris, A. S. Demirkol, B. Dong, S. Aggarwal, W. Wan,et al., “Recent advances and future prospects for memristive materials, devices, and systems,”ACS nano, vol. 17, no. 13, pp. 11994–12039, 2023

2023
[10]

Avalanches and criticality in self-organized nanoscale networks,

J. B. Mallinson, S. Shirai, S. Acharya, S. Bose, E. Galli, and S. Brown, “Avalanches and criticality in self-organized nanoscale networks,”Sci- ence advances, vol. 5, no. 11, p. eaaw8438, 2019

2019
[11]

Avalanches and edge-of-chaos learning in neuromorphic nanowire networks,

J. Hochstetter, R. Zhu, A. Loeffler, A. Diaz-Alvarez, T. Nakayama, and Z. Kuncic, “Avalanches and edge-of-chaos learning in neuromorphic nanowire networks,”Nature Communications, vol. 12, no. 1, p. 4008, 2021

2021
[12]

Toward a formal theory for computing machines made out of whatever physics offers,

H. Jaeger, B. Noheda, and W. G. van der Wiel, “Toward a formal theory for computing machines made out of whatever physics offers,”Nature Communications, vol. 14, no. 1, p. 4911, 2023

2023
[13]

Brain- inspired computing with self-assembled networks of nano-objects,

A. Vahl, G. Milano, Z. Kuncic, S. A. Brown, and P. Milani, “Brain- inspired computing with self-assembled networks of nano-objects,” Journal of Physics D: Applied Physics, vol. 57, no. 50, p. 503001, 2024

2024
[14]

Neuromorphic information processing with nanowire networks,

Z. Kuncic, O. Kavehei, R. Zhu, A. Loeffler, K. Fu, J. Hochstetter, M. Li, J. M. Shine, A. Diaz-Alvarez, A. Stieg, J. Gimzewski, and T. Nakayama, “Neuromorphic information processing with nanowire networks,” in 2020 IEEE International Symposium on Circuits and Systems (ISCAS), pp. 1–5, 2020

2020
[15]

Neuromorphic nanowire networks: Prin- ciples, progress and future prospects for neuro-inspired information processing,

Z. Kuncic and T. Nakayama, “Neuromorphic nanowire networks: Prin- ciples, progress and future prospects for neuro-inspired information processing,”Advances in Physics: X, vol. 6, no. 1, p. 1894234, 2021

2021
[16]

Online dynamical learning and sequence memory with neuromorphic nanowire networks,

R. Zhu, S. Lilak, A. Loeffler, J. Lizier, A. Stieg, J. Gimzewski, and Z. Kuncic, “Online dynamical learning and sequence memory with neuromorphic nanowire networks,”Nature Communications, vol. 14, no. 1, p. 6697, 2023

2023
[17]

Information dynamics in neuromorphic nanowire networks,

R. Zhu, J. Hochstetter, A. Loeffler, A. Diaz-Alvarez, T. Nakayama, J. T. Lizier, and Z. Kuncic, “Information dynamics in neuromorphic nanowire networks,”Scientific Reports, vol. 11, no. 1, p. 13047, 2021

2021
[18]

Learning chaotic dynamics with neuromorphic network dynamics,

Y . Xu, G. A. Gottwald, and Z. Kuncic, “Learning chaotic dynamics with neuromorphic network dynamics,”APL Machine Learning, vol. 3, p. 046104, 10 2025

2025
[19]

Dynamic reservoir computing with physical neuromorphic networks,

Y . Xu, G. A. Gottwald, and Z. Kuncic, “Dynamic reservoir computing with physical neuromorphic networks,” in2025 International Joint Conference on Neural Networks (IJCNN), pp. 1–8, 2025

2025
[20]

Topological properties of neuromorphic nanowire networks,

A. Loeffler, R. Zhu, J. Hochstetter, M. Li, K. Fu, A. Diaz-Alvarez, T. Nakayama, J. M. Shine, and Z. Kuncic, “Topological properties of neuromorphic nanowire networks,”Frontiers in Neuroscience, vol. 14, p. 184, 2020

2020
[21]

Ergodicity, lack thereof, and the performance of reservoir computing with memristive networks,

V . Baccetti, R. Zhu, Z. Kuncic, and F. Caravelli, “Ergodicity, lack thereof, and the performance of reservoir computing with memristive networks,”Nano Express, vol. 5, no. 1, p. 015021, 2024

2024
[22]

Complex dynamics of memristive circuits: Analytical results and universal slow relaxation,

F. Caravelli, F. L. Traversa, and M. Di Ventra, “Complex dynamics of memristive circuits: Analytical results and universal slow relaxation,” Physical Review E, vol. 95, no. 2, p. 022140, 2017

2017
[23]

Neuronal avalanches in neocortical circuits,

J. M. Beggs and D. Plenz, “Neuronal avalanches in neocortical circuits,” Journal of Neuroscience, vol. 23, no. 35, pp. 11167–11177, 2003

2003
[24]

A quantitative description of mem- brane current and its application to conduction and excitation in nerve,

A. L. Hodgkin and A. F. Huxley, “A quantitative description of mem- brane current and its application to conduction and excitation in nerve,” The Journal of physiology, vol. 117, no. 4, p. 500, 1952

1952
[25]

Hodgkin–huxley axon is made of memristors,

L. Chua, V . Sbitnev, and H. Kim, “Hodgkin–huxley axon is made of memristors,”International Journal of Bifurcation and Chaos, vol. 22, no. 03, p. 1230011, 2012

2012
[26]

Gerstner and W

W. Gerstner and W. M. Kistler,Spiking neuron models: Single neurons, populations, plasticity. Cambridge university press, 2002

2002
[27]

Emergent dynamics of neuromorphic nanowire networks,

A. Diaz-Alvarez, R. Higuchi, P. Sanz-Leon, I. Marcus, Y . Shingaya, A. Z. Stieg, J. K. Gimzewski, Z. Kuncic, and T. Nakayama, “Emergent dynamics of neuromorphic nanowire networks,”Scientific reports, vol. 9, no. 1, p. 14920, 2019

2019