Real-Time Computation at the Edge of Chaos in Recurrent Neural Networks
The DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index.
Abstract
Depending on the connectivity, recurrent networks of simple computational units can show very different types of dynamics, ranging from totally ordered to chaotic. We analyze how the type of dynamics (ordered or chaotic) exhibited by randomly connected networks of threshold gates driven by a time-varying input signal depends on the parameters describing the distribution of the connectivity matrix. In particular, we calculate the critical boundary in parameter space where the transition from ordered to chaotic dynamics takes place. Employing a recently developed framework for analyzing real-time computations, we show that only near the critical boundary can such networks perform complex computations on time series. Hence, this result strongly supports conjectures that dynamical systems that are capable of doing complex computational tasks should operate near the edge of chaos, that is, the transition from ordered to chaotic dynamics.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Intrinsic fluctuations that are poised ''at the edge of chaos'' are thought to maximize the computational efficiency of such systems and the flexibility to switch between adaptive regimes (Bertschinger & Natschla ¨ger, 2004).”
From A potential mechanism for Gibsonian resonance: behavioral entrainment emerges from local homeostasis in an unsupervised reservoir network (2024) · cites this paper as Bertschinger & Natschläger, 2004
In the citing paper
“Finally, we show that a randomly connected circuit model, akin to a “reservoir” network13-15, can reproduce the observed distribution of timescales, provided that the network operates at the critical point (or “edge of chaos”)16-18.”
From A reservoir of time constants for memory traces in cortical neurons (2011) · cites this paper as [17]
Citations
Cites
Self-organized criticality: An explanation of the 1/f noise
Bak, P., Tang, C., & Wiesenfeld, K. (1987) · Physical Review Letters · doi:10.1103/physrevlett.59.381 · Cited by 7,591
The origins of order; self organization and selection in evolution
(1994) · International Journal of Biochemistry · doi:10.1016/0020-711x(94)90119-8 · Cited by 6,649
Introduction to the theory of neural computation
(1994) · Neural Networks · doi:10.1016/0893-6080(94)90014-0 · Cited by 6,401
Bak, P., Tang, C., & Wiesenfeld, K. (1988) · Physical Review A · doi:10.1103/physreva.38.364 · Cited by 4,713
Maass, W., Natschläger, T., & Markram, H. (2002) · Neural Computation · doi:10.1162/089976602760407955 · Cited by 4,025 · cited 10×
“(Maass et al., 2002) and indepenently by Jaeger (Jaeger, 2002) who called it “echo state networks”.”
BackgroundInfluentialSelf-citationChaos in Neuronal Networks with Balanced Excitatory and Inhibitory Activity
van Vreeswijk, C., & Sompolinsky, H. (1996) · Science · doi:10.1126/science.274.5293.1724 · Cited by 1,811
Computation at the edge of chaos: Phase transitions and emergent computation
Langton, C. G. (1990) · Physica D: Nonlinear Phenomena · doi:10.1016/0167-2789(90)90064-v · Cited by 1,527 · cited 5×
“These results provide further evidence for the idea of “computation at the edge of chaos” (Langton, 1990).”
BackgroundResultInfluentialFading memory and the problem of approximating nonlinear operators with Volterra series
Boyd, S., & Chua, L. (1985) · IEEE Transactions on Circuits and Systems · doi:10.1109/tcs.1985.1085649 · Cited by 991
Random Networks of Automata: A Simple Annealed Approximation
Derrida, B., & Pomeau, Y. (1986) · Europhysics Letters (EPL) · doi:10.1209/0295-5075/1/2/001 · Cited by 706
Dynamic Patterns in Complex Systems
(1988) · doi:10.1142/0534 · Cited by 548
Research works printed but not linked: 1
- Mitchell M., 1993, Complex Systems, 7, 89
Cited by
Items 301–400 of 831
Dynamics and Information Import in Recurrent Neural Networks
Metzner, C., & Krauß, P. (2022) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2022.876315 · Cited by 27
“Since medium-level correlations are thought to be optimally suited for information processing (Bertschinger and Natschläger, 2004; Natschläger et al., 2005; Legenstein and Maass, 2007; Schrauwen et al., 2009; Büsing et al., 2010; Toyoizumi and Abbott, 2011; Dambre et al., 2012), it is remarkable…”
BackgroundInfluentialSpike-time reliability of layered neural oscillator networks
Lin, K., Shea‐Brown, E., & Young, L.-S. (2009) · Journal of Computational Neuroscience · doi:10.1007/s10827-008-0133-3 · Cited by 21
“For discrete-state networks, [53, 54, 10, 36] also describe how recurrent connectivity can create instabilities related to those observed here in large, randomly connected networks.”
BackgroundA Nonvolatile Fractional Order Memristor Model and Its Complex Dynamics
Wu, J., Wang, G., Iu, H. H., Shen, Y., & Zhou, W. (2019) · Entropy · doi:10.3390/e21100955 · Cited by 19
“Researches show that artificial neural networks working on chaotic or chaotic edges can optimize neural computations [7].”
BackgroundReservoir Computing using Cellular AutomataPreprint
Yılmaz, Ö. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1410.0162 · Cited by 47
“It is claimed that many of these type of dynamical systems possess high computational power [9, 10].”
BackgroundYamaguchi, T., Tsunegi, S., Nakajima, K., & Taniguchi, T. (2023) · Physical review. B./Physical review. B · doi:10.1103/physrevb.107.054406 · Cited by 20
“Such investigations in spintronics [29, 30] have recently focused on computing near the chaotic state because the edge of chaos sometimes provides a boundary of high computational capability [31, 32].”
BackgroundExcitation-Inhibition Balance Controls Information Encoding in Neural Populations
Barzon, G., Busiello, D. M., & Nicoletti, G. (2025) · Physical Review Letters · doi:10.1103/physrevlett.134.068403 · Cited by 28
Evolutionary pre-training for CRJ-type reservoir of echo state networks
Yuenyong, S., & Nishihara, A. (2014) · Neurocomputing · doi:10.1016/j.neucom.2014.08.065 · Cited by 19
“It is often claimed that at such edge of chaos, dynamical systems posses high computational power [8].”
Information propagation within the Genetic Network of Saccharomyces cerevisiae
Chowdhury, S., Lloyd‐Price, J., Smolander, O., Baici, W., Hughes, T. R., Yli-Harja, O. P., Chua, G., & Ribeiro, A. S. (2010) · BMC Systems Biology · doi:10.1186/1752-0509-4-143 · Cited by 18
“A study of information propagation in Boolean networks where all nodes are driven by a common input signal found that critical RBNs best distinguished differences between inputs states and were able to perform the most complex computations on time series [11].”
BackgroundElectrical Coupling Controls Dimensionality and Chaotic Firing of Inferior Olive Neurons
Hoang, H., Lang, E. J., Hirata, Y., Tokuda, I. T., Aihara, K., Toyama, K., Kawato, M., & Schweighofer, N. (2019) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1008075 · Cited by 20
“In contrast to the scenario of strong and weak couplings, moderate interactions of the neurons via intermediate coupling strengths induce asynchronous and irregular spiking activity and thus maximize the entropy [85].”
BackgroundCriticality in reservoir computer of coupled phase oscillators
Wang, L., Fan, H., Xiao, J., Lan, Y., & Wang, X. (2022) · Physical review. E · doi:10.1103/physreve.105.l052201 · Cited by 24
“For instance, it has been demonstrated that the computational capacity of RC is maximized at “the edge of chaos (stability)” [23, 25, 26], namely the critical state at the transition from the ordered to chaotic states, and, to achieve the optimal performance, the machine should be tuned to the vicinity of this critica…”
BackgroundBachmann, C., Tetzlaff, T., Duarte, R., & Morrison, A. (2020) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1007790 · Cited by 22
Familiarity Detection and Memory Consolidation in Cortical Assemblies
Zhang, X., Yeh, F.-C., Ju, H., Jiang, Y., Foo, G., & VanDongen, A. M. J. (2020) · eNeuro · doi:10.1523/eneuro.0006-19.2020 · Cited by 19
“4G) seems to fall at the “edge of chaos,” which has been predicted to have optimal computational power (Bertschinger and Natschläger, 2004; Legenstein and Maass, 2007).”
BackgroundProgressive development of spin wave chaos in active-ring oscillators
Ustinov, A. B., Кондрашов, А. В., Tatsenko, I. Y., Nikitin, A. A., & Kostylev, M. (2021) · Physical review. B./Physical review. B · doi:10.1103/physrevb.104.l140410 · Cited by 19
Ag2S island network reservoir that works with direct optical signal inputs
Shimizu, Y., Minegishi, K., Tanaka, H., & Hasegawa, T. (2022) · Japanese Journal of Applied Physics · doi:10.35848/1347-4065/acab0a · Cited by 19
Edge of Stability Echo State Network
Ceni, A., & Gallicchio, C. (2024) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2024.3400045 · Cited by 29
“An ideal situation would be for the RNN to stay in a regime of balance between stable contractive dynamics and unstable chaotic dynamics, i.e., along the edge of chaos [16].”
Repertoires of Spike Avalanches Are Modulated by Behavior and Novelty
Ribeiro, T. L., Ribeiro, S. T. G., & Copelli, M. (2016) · Frontiers in Neural Circuits · doi:10.3389/fncir.2016.00016 · Cited by 18
“It has been shown that a critical brain would have desirable advantages such as optimal computational capabilities (Bertschinger and Natschläger, 2004), information transmission (Beggs and Plenz, 2003; Rämö et al., 2007) and diversity (Nykter et al., 2008), size of memory repertoires (Beggs and…”
BackgroundSelf-organized Criticality via Retro-Synaptic Signals
Hernandez-Urbina, V., & Herrmann, J. M. (2017) · Frontiers in Physics · doi:10.3389/fphy.2016.00054 · Cited by 18
“It has been reported that neural networks whose dynamics are poised at the border of a phase transition, and whose activity is identified by power-law distribution of events, acquire several benefits for neural computation which include: optimal information transmission and maximum dynamic range [18], maximum informat…”
BackgroundMR. Estimator, a toolbox to determine intrinsic timescales from subsampled spiking activity
Spitzner, F. P., Dehning, J., Wilting, J., Hagemann, A., Neto, J. P., Zierenberg, J., & Priesemann, V. (2021) · PLoS ONE · doi:10.1371/journal.pone.0249447 · Cited by 24
“For artificial recurrent networks, such a tuning capabilitywas already shown to be attainable by operating around the critical point (of a dynamic second order phase transition) [24, 15, 45, 39].”
BackgroundHarnessing adaptive dynamics in neuro-memristive nanowire networks for transfer learning
Zhu, R., Hochstetter, J., Loeffler, A., Álvarez, A. D., Stieg, A. Z., Gimzewski, J., Nakayama, T., & Kuncic, Z. (2020) · IEEE International Conference on Rebooting Computing (ICRC) · doi:10.1109/icrc2020.2020.00007 · Cited by 18
“I − V trajectories begin to diverge), which may be optimal for information processing [46].”
BackgroundSymmetry-aware reservoir computing
Barbosa, W. A. S., Griffith, A., Rowlands, G. E., Govia, L. C. G., Ribeill, G. J., Nguyen, M.-H., Ohki, T. A., & Gauthier, D. J. (2021) · Physical review. E · doi:10.1103/physreve.104.045307 · Cited by 18
Baroni, F., Torres, J. J., & Varona, P. (2010) · PLoS ONE · doi:10.1371/journal.pone.0015023 · Cited by 21
“This behavior is due to their lack of attractors and ‘‘edge of chaos’’ dynamics [102], and to the reverberation of activity through positive feedback.”
Sulfasalazine-associated encephalopathy in a patient with Crohn's disease.
Schoonjans, R. S., Mast, A., den Abeele, G. V., Dewilde, D., Achten, E., Maele, V. V., & Pauwels, W. (1993) · PubMed · Cited by 19
Falandays, J. B., Yoshimi, J., Warren, W. H., & Spivey, M. J. (2024) · Cognitive Neurodynamics · doi:10.1007/s11571-023-09988-2 · Cited by 19
“Intrinsic fluctuations that are poised ‘‘at the edge of chaos’’ are thought to maximize the computational efficiency of such systems and the flexibility to switch between adaptive regimes (Bertschinger & Natschla¨ger, 2004).”
“How do humans make sense?” multiscale dynamics and emergent meaning
Dale, R., & Kello, C. T. (2018) · New Ideas in Psychology · doi:10.1016/j.newideapsych.2017.09.002 · Cited by 15
Hydrodynamic Stabilization of Self-Organized Criticality in a Driven Rydberg Gas
Klocke, K., Wintermantel, T. M., Lochead, G., Whitlock, S., & Buchhold, M. (2021) · Physical Review Letters · doi:10.1103/physrevlett.126.123401 · Cited by 23
Lévy Walk in Swarm Models Based on Bayesian and Inverse Bayesian Inference
Gunji, Y.-P., Kawai, T., Murakami, H., Tomaru, T., Minoura, M., & Shinohara, S. (2020) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2020.11.045 · Cited by 20
Reservoir computing with higher-order interactive coupled pendulums
Li, X., SMALL, M. P., & Lei, Y. (2023) · Physical review. E · doi:10.1103/physreve.108.064304 · Cited by 19
Quantitative Analysis of the Effective Functional Structure in Yeast Glycolysis
la Fuente, I. M. D., & Cortés, J. M. (2012) · PLoS ONE · doi:10.1371/journal.pone.0030162 · Cited by 19
“which show that when a dynamical system operates in the frontier between order (periodic behavior) and chaos its complexity is maximal [40,41].”
BackgroundIntermittent and metastable chaos in a memristive artificial neuron with inertia
Wojtusiak, A., Balanov, A. G., & Savel'ev, S. E. (2020) · Chaos Solitons & Fractals · doi:10.1016/j.chaos.2020.110383 · Cited by 17
Criticality-Driven Evolution of Adaptable Morphologies of Voxel-Based Soft-Robots
Talamini, J., Medvet, E., & Nichele, S. (2021) · Frontiers in Robotics and AI · doi:10.3389/frobt.2021.673156 · Cited by 21
“, 1988), which naturally evolve toward a critical point of phase transition between chaotic and non-chaotic regimes (Bertschinger and Natschläger, 2004).”
BackgroundWhy Layering in Recurrent Neural Networks? A DeepESN Survey
Gallicchio, C., & Micheli, A. (2018) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2018.8489368 · Cited by 20
“Indeed, as reported in several literature works [45], [46], [47], RNNs whose dynamics are brought near to the stable-unstable transition naturally tend to develop rich representations of their driving input signals and show better predictive performances in several classes of problems, including those requiring long t…”
BackgroundMetabolic Dissipative Structures
de la Fuente, I. M. (2013) · Springer series in biophysics · doi:10.1007/978-3-642-38505-6_8 · Cited by 20
Theory of the Computational Function of Microcircuit Dynamics
Maass, W., & Markram, H. (2006) · The MIT Press eBooks · doi:10.7551/mitpress/4596.003.0020 · Cited by 18
“been demonstrated by Bertschinger and Natschläger (2004) for synchronized”
BackgroundDynamical Systems as Temporal Feature SpacesPreprint
Tiňo, P. (2019) · University of Birmingham Research Portal (University of Birmingham) · doi:10.48550/arxiv.1907.06382 · Cited by 22
“This confirms previous findings in the ESN literature on the importance of tuning the dynamical system at the edge of stability [5].”
BackgroundFunctional optimization in complex excitable networksPreprint
Johnson, S., Marro, J., & Torres, J. J. (2008) · Cited by 16
Burrows, D. R. W., Diana, G., Pimpel, B., Moeller, F., Richardson, M. P., Bassett, D. S., Meyer, M. P., & Rosch, R. E. (2023) · Journal of Neuroscience · doi:10.1523/jneurosci.0171-22.2023 · Cited by 23
“…vitro and in silico studies have demonstrated that (1) the net-work-mediated separation (NMS), a network ’ s ability to discriminate between inputs (Bertschinger and Natschläger, 2004); and (2) the dynamic range ( d ), a network ’ s ability to represent a wide range of inputs (Kinouchi and…”
InfluentialSeeing double with a multifunctional reservoir computer
Flynn, A., Tsachouridis, V. A., & Amann, A. (2023) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0157648 · Cited by 16
Rucco, R., Bernardo, P., Lardone, A., Baselice, F., Pesoli, M., Polverino, A., Bravaccio, C., Granata, C., Mandolesi, L., Sorrentino, G., et al. (2020) · Frontiers in Psychology · doi:10.3389/fpsyg.2020.550749 · Cited by 12
A devil’s advocate view on ‘self-organized’ brain criticality
Gros, C. (2021) · Journal of Physics Complexity · doi:10.1088/2632-072x/abfa0f · Cited by 19
Khajeh, R., Fumarola, F., & Abbott, L. F. (2022) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1008836 · Cited by 21
Neuronal avalanche dynamics and functional connectivity elucidate information propagation in vitro
Heiney, K., Ramstad, O. H., Fiskum, V., Sandvig, A., Sandvig, I., & Nichele, S. (2022) · Frontiers in Neural Circuits · doi:10.3389/fncir.2022.980631 · Cited by 18
“Among the computational benefits of this state are a widened dynamic range (Kinouchi and Copelli, 2006) and increased sensitivity to inputs (Bertschinger and Natschläger, 2004).”
BackgroundAdaptive functional systems: Learning with chaos
Komarov, M., Osipov, G. V., & Burtsev, M. (2010) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/1.3521250 · Cited by 16
The Ongoing Search for the Neuronal Correlate of Consciousness
Singer, W. (2015) · MPG.PuRe (Max Planck Society) · doi:10.15502/9783958570344 · Cited by 17
Differential Effects of Simulated Cortical Network Lesions on Synchrony and EEG Complexity
Ibáñez-Molina, A. J., Iglesias‐Parro, S., & Escudero, J. (2018) · International Journal of Neural Systems · doi:10.1142/s0129065718500247 · Cited by 16
Extended Anderson Criticality in Heavy-Tailed Neural Networks
Wardak, A., & Gong, P. (2022) · Physical Review Letters · doi:10.1103/physrevlett.129.048103 · Cited by 35
Input-driven chaotic dynamics in vortex spin-torque oscillator
Imai, Y., Nakajima, K., Tsunegi, S., & Taniguchi, T. (2022) · Scientific Reports · doi:10.1038/s41598-022-26018-z · Cited by 19
“For example, the short-term memory capacity is maximized near the edge of chaos in some physical reservoirs 30,31 , although this is not the present case; see also Supplementary Information, where the Lyapunov exponent and the short-term memory capacity for various parameters are summarized.”
Reservoir computing based on solitary-like waves dynamics of liquid film flows: A proof of concept
Maksymov, I. S., & Pototsky, A. (2023) · Europhysics Letters (EPL) · doi:10.1209/0295-5075/acd471 · Cited by 15
Robustness and Information Propagation in Attractors of Random Boolean Networks
Lloyd‐Price, J., Gupta, A., & Ribeiro, A. S. (2012) · PLoS ONE · doi:10.1371/journal.pone.0042018 · Cited by 15
“Information propagation in RBNs has been also assessed by other methods, which also found a maximization at the critical regime [4,6,7,9,13].”
MethodTowards a Calculus of Echo State Networks
Goudarzi, A., & Stefanović, D. (2014) · Procedia Computer Science · doi:10.1016/j.procs.2014.11.101 · Cited by 15
“Several studies attributed this property to the dynamical regime of the reservoir and showed it to be optimal when the system operates in the critical dynamical regime—a regime in which perturbations to the system’s trajectory in its phase space neither spread nor die out [1–3, 11 , 16 ].”
BackgroundOn the Characteristics and Structures of Dynamical Systems Suitable for Reservoir Computing
Inubushi, M., Yoshimura, K., Ikeda, Y., & Nagasawa, Y. (2021) · Natural computing series · doi:10.1007/978-981-13-1687-6_5 · Cited by 15
Effect of excited state lasing on the chaotic dynamics of spin QD-VCSELs
Georgiou, P., Tselios, C., Mourkioti, G., Skokos, C., & Alexandropoulos, D. (2021) · Nonlinear Dynamics · doi:10.1007/s11071-021-06955-6 · Cited by 15
Axonal geometry as a tool for modulating firing patterns
Ofer, N., & Shefi, O. (2015) · Applied Mathematical Modelling · doi:10.1016/j.apm.2015.10.017 · Cited by 18
“Moreover, it has been shown that in other biological systems, including neural networks, activity at the edge of chaos close to critical dynamics is optimal for information processing purposes and robustness [74,75].”
BackgroundNearly Extensive Sequential Memory Lifetime Achieved by Coupled Nonlinear Neurons
Toyoizumi, T. (2012) · Neural Computation · doi:10.1162/neco_a_00324 · Cited by 23
“In addition to memory, it is also important for the reservoir to map input to a high dimensional “feature space” so that a linear readout has access to useful features (Bertschinger and Natschläger, 2004; Büsing et al., 2010; Jaeger and Haas, 2004; Maass et al., 2002; Sussillo and Abbott, 2009).”
BackgroundInferring Stimulus Selectivity from the Spatial Structure of Neural Network Dynamics
Rajan, K., Abbott, L. F., & Sompolinsky, H. (2010) · neural information processing systems · Cited by 17
Spatiotemporal signal classification via principal components of reservoir states
Prater, A. (2017) · Neural Networks · doi:10.1016/j.neunet.2017.04.008 · Cited by 16
Frustration induced transient chaos, fractal and riddled basins in coupled limit cycle oscillators
Kanagaraj, S., Karthiga, S., Chandrasekar, V. K., Senthilkumar, D. V., & Lakshmanan, M. (2019) · Communications in Nonlinear Science and Numerical Simulation · doi:10.1016/j.cnsns.2019.01.024 · Cited by 15
“Particularly, information processing based on edge of chaos and nonlinear transient computing are promising contemporary research topics both from computational neuroscience and dynamical systems points of view [19, 27, 28, 29, 30, 31], where the edge of chaos that lies between ordered and chaotic regimes is proved to…”
BackgroundFunctional optimization in complex excitable networks
Johnson, S., Marro, J., & Torres, J. J. (2008) · Europhysics Letters (EPL) · doi:10.1209/0295-5075/83/46006 · Cited by 20
Nearly extensive sequential memory lifetime achieved by coupled nonlinear neuronsPreprint
Toyoizumi, T. (2014) · Cited by 15
Papadopoulos, L., Jo, S., Zumwalt, K., Wehr, M., Jaramillo, S., McCormick, D. A., & Mazzucato, L. (2025) · Neuron · doi:10.1016/j.neuron.2025.11.011 · Cited by 21
Some Local Stability Properties of an Autonomous Long Short-Term Memory Neural Network Model
Stipanović, D. M., Murmann, B., Causo, M., Lekić, A., Rubies-Royo, V., Tomlin, C. J., Beigné, É., Thuries, S., Zarudniev, M., & Lesecq, S. (2018) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas.2018.8350958 · Cited by 15
“This would also provide an approximation of the boundary between stable and chaotic behaviors from the side of the domain of initial conditions for which the system is stable and provide relevant information for the networks’ training [7].”
BackgroundInfluentialFading memory as inductive bias in residual recurrent networks
Dubinin, I., & Effenberger, F. (2024) · Neural Networks · doi:10.1016/j.neunet.2024.106179 · Cited by 19
Chaotic recurrent neural networks for brain modelling: A review
Mattera, A., Alfieri, V., Granato, G., & Baldassarre, G. (2024) · Neural Networks · doi:10.1016/j.neunet.2024.107079 · Cited by 25
Direct model of memory properties and the linear reservoir topologies in echo state networks
Ma, Q., Chen, W., Wei, J., & Yu, Z. (2014) · Applied Soft Computing · doi:10.1016/j.asoc.2014.04.038 · Cited by 14
Heritability lost; intelligence found
Richardson, K. (2012) · EMBO Reports · doi:10.1038/embor.2012.83 · Cited by 16
Should all disaster risks be reduced? A perspective from the systems concept of the edge of chaos
Coetzee, C., & van Niekerk, D. (2018) · Environmental Hazards · doi:10.1080/17477891.2018.1463912 · Cited by 17
“Another noteworthy aspect of the information sensitivity and exchange characteristics of edge of chaos is that it controls the actual rate of adaption within the system (Bertschinger & Natschläger, 2004, pp. 20–21).”
BackgroundOn the Criticality of Adaptive Boolean Network Robots
Braccini, M., Roli, A., Barbieri, E., & Kauffman, S. A. (2022) · Entropy · doi:10.3390/e24101368 · Cited by 17
“Dynamical criticality has been identified as a feature of complex natural as well as arti-ficial systems in plenty of settings [1–3], from cells [4–7] to artificial neural networks [8–10].”
BackgroundSelf-Organized Behavior Generation for Musculoskeletal Robots
Der, R., & Martius, G. (2017) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2017.00008 · Cited by 17
“This working regime may be associated with the somewhat vague “edge of chaos” concept (Langton, 1990; Mitchell et al., 1993; Kauffman, 1995; Bertschinger and Natschläger, 2004; Natschläger et al., 2005).”
BackgroundLi, K., & Príncipe, J. C. (2018) · Frontiers in Neuroscience · doi:10.3389/fnins.2018.00194 · Cited by 14
“SP is quantified by a kernel-quality measure proposed in Maass et al. (2005) that is based on the rank of a matrix formed by the system states corresponding to different input signals (Bertschinger and Natschläger, 2004).”
BackgroundPinto, Í. L. D., & Copelli, M. (2019) · Physical review. E · doi:10.1103/physreve.100.062416 · Cited by 16
Critical line in random-threshold networks with inhomogeneous thresholds
Rohlf, T. (2008) · Physical Review E · doi:10.1103/physreve.78.066118 · Cited by 16
Nagase, S., Nezu, S., & Sekiguchi, K. (2024) · Physical Review Applied · doi:10.1103/physrevapplied.22.024072 · Cited by 18
Suppression of chaos in a partially driven recurrent neural network
Takasu, S., & Aoyagi, T. (2024) · Physical Review Research · doi:10.1103/physrevresearch.6.013172 · Cited by 17
“Such an RNN is commonly referred to as being at the ”edge of chaos”, and it is known empirically that reservoirs in this regime have the highest computational capacity [7].”
BackgroundStatistical mechanics for networks of real neurons
Meshulam, L., & Bialek, W. (2025) · Reviews of Modern Physics · doi:10.1103/jcrn-3nrc · Cited by 32
Hudson, N. J., Porto-Neto, L. R., Kijas, J., McWilliam, S. M., Taft, R. J., & Reverter, A. (2014) · BMC Bioinformatics · doi:10.1186/1471-2105-15-66 · Cited by 12
“The overall position of all the human populations supports a controversial concept from complex systems science [55,57-61] that genomes are poised at or close to ‘The Edge of Chaos.”
BackgroundChrol-Cannon, J., & Jin, Y. (2014) · PLoS ONE · doi:10.1371/journal.pone.0101792 · Cited by 17
“We use the calculation method defined in [27] which was formulated based on theory described in [12].”
BackgroundMethodDynamical complexity and computation in recurrent neural networks beyond their fixed point
Márquez, B. A., Larger, L., Jacquot, M., Chembo, Y. K., & Brunner, D. (2018) · Scientific Reports · doi:10.1038/s41598-018-21624-2 · Cited by 16
Sato, D., Shima, H., Matsuo, T., Yonezawa, M., Kinoshita, K., Kobayashi, M., Naitoh, Y., Akinaga, H., Miyamoto, S., Nokami, T., et al. (2023) · ACS Applied Materials & Interfaces · doi:10.1021/acsami.3c08638 · Cited by 16
“To evaluate the short-term memory (STM) and nonlinear data transformation capabilities of IL-PRD, we discussed the role of Faradaic currents in the physical reservoir computation performance of STM and parity check tasks in the IL-PRD, 15,39 respectively.”
BackgroundComputational capacity and energy consumption of complex resistive switch networks
rger, J. B., Goudarzi, A., Stefanović, D., & Teuscher, C. (2015) · AIMS Materials Science · doi:10.3934/matersci.2015.4.530 · Cited by 15
Critical branching neural computation
Kello, C. T., & Mayberry, M. R. (2010) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2010.5596813 · Cited by 13
“Memory capacity is plotted from τ ∈ [1] [15], and m = 6 for this simulation.”
BackgroundMethodResultInfluentialEchlin, M., Aguilar, B. H., Notarangelo, M., Gibbs, D. L., & Shmulevich, I. (2018) · Entropy · doi:10.3390/e20120954 · Cited by 12
“Critical dynamics, which lie at the border between order and chaos, have been shown to be important in computation and information processing, biological and otherwise [51–55].”
BackgroundDu, K.-L., & Swamy, M. N. S. (2019) · Springer eBooks · doi:10.1007/978-1-4471-7452-3_12 · Cited by 17
Zwanzig-Mori projection operators and EEG dynamics: deriving a simple equation of motion
Hsu, D., & Hsu, M. (2009) · PMC Biophysics · doi:10.1186/1757-5036-2-6 · Cited by 15
Intrinsic noise and deviations from criticality in Boolean gene-regulatory networks
Villegas, P., Ruiz‐Franco, J., Hidalgo, J., & Muñoz, M. A. (2016) · Scientific Reports · doi:10.1038/srep34743 · Cited by 13
“This conjecture (which was developed in the machine-learning and neural-network community 15–17 ), proposes that –by operating nearby criticality– networks exhibit an optimal tradeoff between stability to perturbations and sensitivity/responsiveness to signals.”
BackgroundEditorial: On the relation of dynamics and structure in brain networks
Papo, D., Goñi, J., & Buldú, J. M. (2017) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/1.4981391 · Cited by 12
Psychiatric Illnesses as Disorders of Network Dynamics
Durstewitz, D., Huys, Q. J. M., & Koppe, G. (2020) · Biological Psychiatry Cognitive Neuroscience and Neuroimaging · doi:10.1016/j.bpsc.2020.01.001 · Cited by 100
Maximal Memory Capacity Near the Edge of Chaos in Balanced Cortical E-I Networks
Kanamaru, T., Hensch, T. K., & Aihara, K. (2023) · Neural Computation · doi:10.1162/neco_a_01596 · Cited by 18
Fleddermann, L., Herzog, S., & Parlitz, U. (2025) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0258250 · Cited by 14
Neutral Stability, Rate Propagation, and Critical Branching in Feedforward Networks
Cayco-Gajic, N. A., & Shea‐Brown, E. (2013) · Neural Computation · doi:10.1162/neco_a_00461 · Cited by 12
Bai, K., Li, J., Hamedani, K., & Yi, Y. (2018) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1145/3195970.3196044 · Cited by 13
“The computational performance in ordered, edge of chaos, and chaotic regimes are evaluated in [8, 9].”
BackgroundP-CRITICAL: a reservoir autoregulation plasticity rule for neuromorphic hardware
Balafrej, I., Alibart, F., & Rouat, J. (2022) · Neuromorphic Computing and Engineering · doi:10.1088/2634-4386/ac6533 · Cited by 16
A Cationic Oligomer as an Organic Template for Direct Synthesis of Aluminosilicate ITH Zeolite
López, A. R., Carocci, S., Martin, T., Molinari, A., Alexander, J., Martínez, P., Bianchi, G., Feller, L., Bernardi, P., Conte, E., et al. (2014) · DIGITAL.CSIC (Spanish National Research Council (CSIC)) · doi:10.13039/501100000781 · Cited by 15
Chaos in spin-torque oscillator with feedback circuit
Kamimaki, A., Kubota, T., Tsunegi, S., Nakajima, K., Taniguchi, T., Grollier, J., Cros, V., Yakushiji, K., Fukushima, A., Yuasa, S., et al. (2021) · Physical Review Research · doi:10.1103/physrevresearch.3.043216 · Cited by 18
Criticality in FitzHugh-Nagumo oscillator ensembles: Design, robustness, and spatial invariance
Beattie, B. A., Feketa, P., Ochs, K., & Kohlstedt, H. (2024) · Communications Physics · doi:10.1038/s42005-024-01530-y · Cited by 18
Theory of Input Driven Dynamical Systems
Gandhi, M., Tiňo, P., & Jaeger, H. (2012) · The European Symposium on Artificial Neural Networks · Cited by 16
Neuroevolution on the edge of chaos
Matzner, F. (2017) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3071178.3071292 · Cited by 15
“Before the main experiment, we will replicate the original results of Bertschinger and Natschläger [3] and Boedecker et al.”
BackgroundMethodResultInfluentialFunctional differentiations in evolutionary reservoir computing networks
Yamaguti, Y., & Tsuda, I. (2021) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0019116 · Cited by 18
Network structure of cascading neural systems predicts stimulus propagation and recovery
Ju, H., Kim, J. Z., Beggs, J. M., & Bassett, D. S. (2020) · Journal of Neural Engineering · doi:10.1088/1741-2552/abbff1 · Cited by 20
Multi-Scale Expressions of One Optimal State Regulated by Dopamine in the Prefrontal Cortex
Hu, G., Huang, X.-H., Jiang, T., & Yu, S. (2019) · Frontiers in Physiology · doi:10.3389/fphys.2019.00113 · Cited by 9
Criticality in Cortex: Neuronal Avalanches and Coherence Potentials
Plenz, D. (2014) · Wiley eBooks · doi:10.1002/9783527651009.ch2 · Cited by 12
Dynamical Properties of Artificially Evolved Boolean Network Robots
Roli, A., Villani, M., Serra, R., Benedettini, S., Pinciroli, C., & Birattari, M. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-24309-2_4 · Cited by 14
“This conjecture has been introduced with the expression “computation at the edge of chaos” [4,13,18] and it is supported by results on different computational models such as -machines [30], cellular automata [9], and neural networks of different kinds [2,12,14].”
Result
Metadata
- Authors
- Bertschinger, N., & Natschläger, T.
- Affiliations
- Graz University of Technology (AT)
- Software Competence Center Hagenberg (Austria) (AT)
- ORCID
- Bertschinger, Nils
- Venue
- Neural Computation
- Publisher
- MIT Press - Journals
- Published
- 2004-07-01
- Volume
- 16
- Issue
- 7
- Pages
- 1413-1436
- ISSN
- 0899-7667 (ISSN-L)
- 1530-888X
- OpenAlex
- W2057755457
- Cited by
- 833
- References
- 26
- Citations by year
- 2026: 45 · 2025: 73 · 2024: 61 · 2023: 48 · 2022: 66 · 2021: 73 · 2020: 52 · 2019: 53 · 2018: 47 · 2017: 48 · 2016: 26 · 2015: 33 · 2014: 29 · 2013: 33 · 2012: 27
- FWCI
- 13.53
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Neural Computation (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
- Physical Sciences › Computer Science › Computational Theory and Mathematics › Cellular Automata and Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Reservoir Computing
- Keywords
- edge of chaos, recurrent neural networks, critical connectivity, threshold gates, random connectivity, real-time processing, dynamical systems
- MeSH terms
- Animals, Artificial Intelligence, Computer Simulation, Feedback, Memory, Models, Neurological, Time Factors, Neural Networks, Computer, Nonlinear Dynamics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.