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 401–500 of 831
Criticality in the Brain: Evidence and Implications for Neuromorphic Computing
Trastoy, J., & Schuller, I. K. (2018) · ACS Chemical Neuroscience · doi:10.1021/acschemneuro.7b00507 · Cited by 11
Emergence of power laws in noncritical neuronal systems
Faqeeh, A., Osat, S., Radicchi, F., & Gleeson, J. P. (2019) · Physical review. E · doi:10.1103/physreve.100.010401 · Cited by 15
Exploring transfer function nonlinearity in echo state networks
Goudarzi, A., Shabani, A., & Stefanović, D. (2015) · IEEE Symposium on Computational Intelligence for Security and Defense Applications (CISDA) · doi:10.1109/cisda.2015.7208637 · Cited by 16
“ESN’s power in time series processing has been attributed to the reservoir’s memory [16], [17] and highdimensional projection of the input which acts like a temporal discriminant kernel [18] that is present in the critical dynamical regime, where input perturbations in the reservoir dynamics neither spread nor die out…”
BackgroundNonlinear Dynamics Analysis of a Self-Organizing Recurrent Neural Network: Chaos Waning
Eser, J., Zheng, P., & Triesch, J. (2014) · PLoS ONE · doi:10.1371/journal.pone.0086962 · Cited by 11
Stability of Gated Recurrent Unit Neural Networks: Convex Combination Formulation Approach
Stipanović, D. M., Kapetina, M., Rapaić, M. R., & Murmann, B. (2020) · Journal of Optimization Theory and Applications · doi:10.1007/s10957-020-01776-w · Cited by 11
Dopaminergic Modulation of Prefrontal Cortex Network Dynamics
Durstewitz, D. (2007) · Springer eBooks · doi:10.1007/978-0-387-72256-6_15 · Cited by 13
Connectivity and Dynamics in Local Cortical Networks
Beggs, J. M., Klukas, J., & Chen, W. (2007) · Understanding complex systems · doi:10.1007/978-3-540-71512-2_3 · Cited by 11
Gating creates slow modes and controls phase-space complexity in GRUs and LSTMsPreprint
Can, T., Krishnamurthy, K., & Schwab, D. J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2002.00025 · Cited by 22
“0 which is the critical value at which the zero fixed-point becomes unstable; benefits of training at the ”edge-of-chaos” has been noted in previous work Bertschinger and Natschl (2004); Glorot and Bengio (2010).”
BackgroundMethodInfluentialFull-Analog Reservoir Computing Circuit Based on Memristor With a Hybrid Wide-Deep Architecture
Chen, L., Wang, X., Yang, C., Chen, Z., Zhang, J., & Zeng, Z. (2023) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2023.3334267 · Cited by 14
“Besides, it shows complex dynamics that range from stable to chaos, and some researchers have found that the deep RC can perform the most computational capability at the edge of chaos [44].”
BackgroundA current-controlled magnonic reservoir for physical reservoir computing
Ustinov, A. B., Haponchyk, R. V., & Kostylev, M. (2024) · Applied Physics Letters · doi:10.1063/5.0189542 · Cited by 10
A Neurophilosophy of Libertarian Free Will
Tse, P. U. (2024) · doi:10.1093/oso/9780198876953.001.0001 · Cited by 10
Self-Organized Criticality in Asynchronously Tuned Elementary Cellular Automata
Gunji, Y. (2014) · Complex Systems · doi:10.25088/complexsystems.23.1.55 · Cited by 10
Guided Self-Organization of Input-Driven Recurrent Neural Networks
Obst, O., & Boedecker, J. (2013) · Emergence, complexity and computation · doi:10.1007/978-3-642-53734-9_11 · Cited by 10
“It has been argued that a tuning of network dynamics to criticality should bring the performance of static reservoir networks close to the optimal performance (Bertschinger and Natschläger, 2004).”
BackgroundMethodInfluentialAttaining and maintaining criticality in a neuronal network model
Peng, J., & Beggs, J. M. (2012) · Physica A Statistical Mechanics and its Applications · doi:10.1016/j.physa.2012.11.013 · Cited by 9
“Several studies, both computational [1,4–8] and experimental [9,10], have also shown that a broad range of information processing functions would be optimized at the critical point [11–13].”
Signal buffering in random networks of spiking neurons: Microscopic versus macroscopic phenomena
Mayor, J., & Wulfram Gerstner, B. (2005) · Physical Review E · doi:10.1103/physreve.72.051906 · Cited by 12
Long-Short Term Memory Neural Network Stability and Stabilization using Linear Matrix Inequalities
Deka, S. A., Stipanović, D. M., Murmann, B., & Tomlin, C. J. (2019) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas.2019.8702629 · Cited by 9
“Following the conjecture that trained RNNs operate efficiently at the edge of chaos [6], recent work showed that the edge of chaos for the nonlinear LSTMs had a clear correlation to the edge of (local asymptotic) stability of the linearized system [7].”
BackgroundKurikawa, T., Barak, O., & Kaneko, K. (2020) · Physical Review Research · doi:10.1103/physrevresearch.2.023307 · Cited by 12
Memory properties of electromigrated Au nanogaps to realize reservoir computing
Sakai, K., Yagi, M., Ito, M., & Shirakashi, J. (2021) · Applied Physics Letters · doi:10.1063/5.0055352 · Cited by 11
Self-organization and Emergence of Dynamical Structures in Neuromorphic Atomic Switch Networks
Stieg, A. Z., Avizienis, A. V., Sillin, H. O., Aguilera, R., Shieh, H.-H., Martin‐Olmos, C., Sandouk, E. J., Aono, M., & Gimzewski, J. K. (2019) · Springer eBooks · doi:10.1007/978-3-319-76375-0_14 · Cited by 10
Computational capacity of LRC , memristive, and hybrid reservoirs
Sheldon, F. C., Kolchinsky, A., & Caravelli, F. (2022) · Physical review. E · doi:10.1103/physreve.106.045310 · Cited by 20
“Reservoir computing has proven to be a powerful tool in a wide variety of signal processing tasks, including forecasting [16], pattern generation and classification [4], adaptive filtering and prediction of chaotic systems [18].”
BackgroundDistance to criticality undergoes critical transition before epileptic seizure attacks
Liu, S., Li, F., & Wan, F. (2023) · Brain Research Bulletin · doi:10.1016/j.brainresbull.2023.110684 · Cited by 10
“, 2010), computational power (Bertschinger and Natschläger, 2004; Gollo et al., 2013), dynamic range (Gollo et al.”
BackgroundOvchinnikov, I. V. (2011) · Physical Review E · doi:10.1103/physreve.83.051129 · Cited by 10
Yılmaz, Ö. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1503.00851 · Cited by 12
“The reservoir is expected to operate at the edge of chaos because the dynamical systems are shown to present high computational power at this mode [10, 11].”
BackgroundOptimizing Reservoir Computing Based on an Alternating Input-Driven Spin-Torque Oscillator
Wu, X., Tong, Z., & Shao, Q. (2023) · Physical Review Applied · doi:10.1103/physrevapplied.20.024069 · Cited by 9
Komatsu, H., Hosoda, N., & Ikuno, T. (2024) · ACS Applied Materials & Interfaces · doi:10.1021/acsami.4c11061 · Cited by 9
Information structure of heterogeneous criticality in a fish school
Niizato, T., Sakamoto, K., Mototake, Y.-I., Murakami, H., & Tomaru, T. (2024) · Scientific Reports · doi:10.1038/s41598-024-79232-2 · Cited by 11
“The advantages of applying IIT to a living system are twofold.”
Franović, I., Eydam, S., & Eroglu, D. (2024) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0247498 · Cited by 10
Biological computation through recurrence
Vidal-Saez, M. S., Vilarroya, Ó., & García‐Ojalvo, J. (2024) · Biochemical and Biophysical Research Communications · doi:10.1016/j.bbrc.2024.150301 · Cited by 11
“The measures employed include Lyapunov exponents 52,56–59 and Fisher information 60 of reservoir dynamics.”
MethodQiu, J., & Huang, H. (2024) · Communications in Theoretical Physics · doi:10.1088/1572-9494/ad8126 · Cited by 16
“The edge of chaos has thus great computational benefits [17, 38, 39].”
BackgroundNeuronal avalanches in complex networks
Hernandez-Urbina, V., & Herrmann, J. M. (2016) · Cogent Physics · doi:10.1080/23311940.2016.1150408 · Cited by 10
“It is in this boundary between order and randomness, known as the critical point, that brain networks reach maximum computational power (Bertschinger & Natschläger, 2004; Haldeman & Beggs, 2005; Kinouchi & Copelli, 2006).”
BackgroundEstevez-Rams, E., Estévez-Moya, D., Medina, K. G., & Lora-Serrano, R. (2019) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/1.5062364 · Cited by 11
Dynamic Computation in Visual Thalamocortical Networks
Moyal, R., & Edelman, S. (2019) · Entropy · doi:10.3390/e21050500 · Cited by 8
“Bertschinger and Natschläger [24] also observed increased network-mediated separation (a measure of input-related differentiation in the trajectories that takes into account the expected divergence due to chaotic sensitivity); when they supplied their randomly connected recurrent neural networks with real-time input,…”
BackgroundRobust cortical criticality and diverse dynamics resulting from functional specification
Gu, L., & Wu, R. (2021) · Physical review. E · doi:10.1103/physreve.103.042407 · Cited by 11
Stimulation and Control of Homeostasis
Szász, A. (2022) · Open Journal of Biophysics · doi:10.4236/ojbiphy.2022.122004 · Cited by 9
Carnell, A. (2008) · Neurocomputing · doi:10.1016/j.neucom.2008.07.012 · Cited by 11
“to perform some computable function F (ninput(t)) on the input, with the strong limiting factor being the size of the LSM, see [1].”
Randomized Recurrent Neural Networks
Gallicchio, C., Micheli, A., & Peter Tiňo, B. (2018) · CINECA IRIS Institutial research information system (University of Pisa) · Cited by 9
Local Homeostatic Regulation of the Spectral Radius of Echo-State Networks
Schubert, F., & Gros, C. (2021) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2021.587721 · Cited by 11
“It can happen that 51 the induced changes decrease performance by driving the network away from the critical 52 point, which is known to be a beneficial dynamical state for sequence learning 53 tasks [20–25].”
BackgroundDelirium, Caffeine, and Perioperative Cortical Dynamics
Kim, H., McKinney, A., Brooks, J., Mashour, G. A., Lee, U., & Vlisides, P. E. (2021) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2021.744054 · Cited by 9
“For example, neural network computation is optimized when residing near the point of criticality (Bertschinger and Natschläger, 2004), suggesting that critical dynamics support information processing.”
BackgroundLipreading Using Liquid State Machine with STDP-Tuning
Yu, X., Wan, Z., Shi, Z., & Wang, L. (2022) · Applied Sciences · doi:10.3390/app122010484 · Cited by 9
Seeking Synthesis: The Integrative Problem in Understanding Language and Its Evolution
Dale, R. A. C., Kello, C. T., & Schoenemann, P. T. (2016) · Topics in Cognitive Science · doi:10.1111/tops.12199 · Cited by 10
“Interestingly, studies have shown that dynamics giving rise to the kind of “balancing” we just described also maximize their memory and computational capacities (Bertschinger & Natschlager, 2004; Kello, 2013).”
Background(2011) · eScholarship (California Digital Library) · Cited by 8
Deka, S. A., Stipanović, D. M., Murmann, B., & Tomlin, C. J. (2018) · Journal of Optimization Theory and Applications · doi:10.1007/s10957-018-1447-6 · Cited by 8
Hydropathicity-based prediction of pain-causing NaV1.7 variants
Xenakis, M. N., Kapetis, D., Yang, Y., Gerrits, M. M., Heijman, J., Waxman, S. G., Lauria, G., Faber, C. G., Westra, R. L., Lindsey, P. J., et al. (2021) · BMC Bioinformatics · doi:10.1186/s12859-021-04119-2 · Cited by 6
Non-periodic input-driven magnetization dynamics in voltage-controlled parametric oscillator
Taniguchi, T. (2022) · Journal of Magnetism and Magnetic Materials · doi:10.1016/j.jmmm.2022.170009 · Cited by 8
Quantum Entanglement of the Brain, Dynamics of Information, and Intelligent Finance
Njegovanović, A. (2023) · Financial Markets Institutions and Risks · doi:10.61093/fmir.7(3).12-30.2023 · Cited by 9
Quantifying and Maximizing the Information Flux in Recurrent Neural Networks
Metzner, C., Yamakou, M. E., Voelkl, D., Schilling, A., & Krauß, P. (2024) · Neural Computation · doi:10.1162/neco_a_01651 · Cited by 13
“In particular, a dynamical regime called the ’edge of chaos’ at the transition from periodic to chaotic behavior [12] has been extensively studied and demonstrated to be important for computation [13, 14, 15, 16, 17, 18, 19, 20, 21, 22], and short-term memory [23, 24].”
BackgroundRegulation toward Self-organized Criticality in a Recurrent Spiking Neural Reservoir
Brodeur, S., & Rouat, J. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-33269-2_69 · Cited by 14
“It is known that edge-of-chaos regime in recurrent neural network leads to maximal computation capabilities [3].”
BackgroundLambda and the Edge of Chaos in Recurrent Neural Networks
Seifter, J., & Reggia, J. A. (2014) · Artificial Life · doi:10.1162/artl_a_00152 · Cited by 8
“This is consistent with evidence provided by others that networks using other types of node activation functions (than linear threshold neurons) yet still having edge-of-chaos dynamics are most effective as reservoirs during learning [1, 2, 12, 21, 22], although there are dissenting views [26].”
BackgroundResultZhao, Q., Yin, H., Chen, X., & Shi, W. (2015) · International Conference on Natural Computation (ICNC) · doi:10.1109/icnc.2015.7378039 · Cited by 8
Gunji, Y., & Uragami, D. (2020) · Entropy · doi:10.3390/e22091049 · Cited by 8
Chaotic Dynamics in Networks of Spiking Neurons in the Balanced State
Monteforte, M. (2011) · eDiss (Georg-August-Universität Göttingen) · doi:10.53846/goediss-3297 · Cited by 13
High-resolution dynamic consistency analysis of photonic time-delay reservoir computer
Oliverio, L., Rontani, D., & Sciamanna, M. (2023) · Optics Letters · doi:10.1364/ol.486383 · Cited by 7
Chaotic magnetization dynamics in magnetic Duffing oscillator
Tatsumi, R., Chiba, T., Komine, T., & Matsueda, H. (2025) · Physical review. E · doi:10.1103/physreve.111.064202 · Cited by 7
Network structure influences self-organized criticality in neural networks with dynamical synapses
Sugimoto, Y. A., Yadohisa, H., & Abe, M. (2025) · Frontiers in Systems Neuroscience · doi:10.3389/fnsys.2025.1590743 · Cited by 9
“…functions such as information processing capabilities and transmissions, suggesting that a brain benefits from being critical (Shew and Plenz, 2013; Bertschinger and Natschläger, 2004; Kinouchi and Copelli, 2006; Boedecker et al., 2012; Avramiea et al., 2022; Haldeman and Beggs, 2005; Safavi et…”
Device‐Level Implementation of Reservoir Computing With Memristors
Park, S., Kim, H., & Jang, H. W. (2026) · Advanced Intelligent Systems · doi:10.1002/aisy.202500843 · Cited by 9
“The separation should not be overly specific and lie near the so-called ‘ edge of chaos ’ [53, 54], in order to prevent the network from being adversely affected by noise.”
Progresses and outlook in neuromorphic devices
Wang, Y., Liu, C., Huang, R., & Yang, Y. (2019) · Chinese Science Bulletin (Chinese Version) · doi:10.1360/tb-2019-0739 · Cited by 6
“但是 , 构筑深度神经网络加速硬件需 要大规模的神经形态器件 , 若不考虑片上资源复用 , 仅 仅实现 LeNet [38] 就需要约 34 万人工突触和 1 万人工神经 [39] ; 而大脑在感知 和学习的过程中工作在混沌的边缘 [40] , 具有强大的全 局搜索能力和实时计算能力 [41] .”
Key role of neuronal diversity in structured reservoir computing
Thivierge, J., Giraud, É., Lynn, M. B., & Charbonneau, A. T. (2022) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0111131 · Cited by 8
Ikeda, N., Akita, D., & Takahashi, H. (2023) · Applied Physics Letters · doi:10.1063/5.0152633 · Cited by 10
LUTNet-RC: Look-Up Tables Networks for Reservoir Computing on an FPGA
Yoshioka, K., Tanaka, Y., & Tamukoh, H. (2023) · International Conference on Field-Programmable Technology (FPT) · doi:10.1109/icfpt59805.2023.00024 · Cited by 14
“S in [ t ] = 0 or 1 ( random ) , (14) The second task was a parity-check (PC) task used to measure memory capacity and nonlinearity simultaneously [29] and is represented by Eqs.”
MethodBrain Network Modularity and Resilience Signaled by Betweenness Centrality Percolation Spiking
Kotlarz, P. L., Febo, M., Nino, J. C., & Initiative, O. B. O. T. A. D. N. (2024) · Applied Sciences · doi:10.3390/app14104197 · Cited by 8
“In dynamic systems, such as neural networks, the boundary between order and chaos (i.e., edge of chaos) often contains states that allow maximized computational processing [16,17].”
Reservoir Computing Utilizing a Complementary Combination of n- and p-Channel FeFETs
Suzuki, R., Toprasertpong, K., Nakane, R., Nako, E., Takenaka, M., & Takagi, S. (2024) · IEEE Electron Device Letters · doi:10.1109/led.2024.3435422 · Cited by 7
“…performance, employed three tasks: the short-term memory (STM) task [24], evaluating the short-term memory function, the parity check (PC) task, evaluating both short-term memory and nonlinearity functions [25], and NARMA-2 task, evaluating both short-term memory and nonlinearity functions [26].”
MethodNonlinear transient computation as a potential “kernel trick” in cortical processing
Crook, N. T., & Goh, W. J. (2008) · Biosystems · doi:10.1016/j.biosystems.2008.05.010 · Cited by 7
Dynamics of oscillators coupled by a medium with adaptive impact
Daneshvar, R. (2013) · Complexity · doi:10.1002/cplx.21439 · Cited by 7
Evaluation of Memory Capacity and Time Series Prediction Using a Spin Hall Oscillator as Reservoir
Mathew, A. J., Mohan, J. R., Feng, R., Medwal, R., Gupta, S., Rawat, R. S., & Fukuma, Y. (2023) · IEEE Transactions on Magnetics · doi:10.1109/tmag.2023.3266517 · Cited by 12
“task, such as a three-bit parity check (PC) [8].”
BackgroundProduct reservoir computing: Time-series computation with multiplicative neurons
Goudarzi, A., Shabani, A., & Stefanović, D. (2015) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2015.7280453 · Cited by 10
“ESN’s power in time series processing has been attributed to the reservoir’s memory [8], [9] and highdimensional projection of the input, which acts like a temporal discriminant kernel [10] that is present in the critical dynamical regime, where input perturbations in the reservoir dynamics neither spread nor die out…”
BackgroundChaotic and non-chaotic phases in experimental responses of a single neuron
Marmari, H., Vardi, R., & Kanter, I. (2014) · Europhysics Letters (EPL) · doi:10.1209/0295-5075/106/46002 · Cited by 7
Plasticity, Learning, and Complexity in Spiking Networks
Kello, C. T., Rodny, J., Warlaumont, A. S., & Noelle, D. C. (2012) · Critical Reviews in Biomedical Engineering · doi:10.1615/critrevbiomedeng.2013006724 · Cited by 9
Self-organization to sub-criticality
Priesemann, V. (2015) · BMC Neuroscience · doi:10.1186/1471-2202-16-s1-o19 · Cited by 7
“A popular hypothesis is that neural networks assume a critical state [1,2], because in models criticality maximizes information processing capabilities [3,4].”
Comparative approximations of criticality in a neural and quantum regime
Bettinger, J. S. (2017) · Progress in Biophysics and Molecular Biology · doi:10.1016/j.pbiomolbio.2017.09.007 · Cited by 6
Tanaka, T., Nakajima, K., & Aoyagi, T. (2020) · Neuroscience Research · doi:10.1016/j.neures.2020.02.001 · Cited by 9
Beggs, J. M. (2007) · Nature Physics · doi:10.1038/nphys799 · Cited by 7
Reverberating activity in a neural network with distributed signal transmission delays
Omi, T., & Shinomoto, S. (2007) · Physical Review E · doi:10.1103/physreve.76.051908 · Cited by 7
Reservoir Computing in robotics: a reviewPreprint
Baldini, P. (2022) · OSF Preprints (OSF Preprints) · doi:10.31219/osf.io/wkz6v · Cited by 11
Optimizing Temporal Data Forecasting for Stiffness-Modulated MEMS Reservoir Computing
Guo, X., Yang, W., Bai, Y., Xiong, X., Wang, Z., & Zou, X. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3446672 · Cited by 6
“The task of the Parity benchmark is to predict whether the count of “ − 1” in a preceding sequence of binary signals is odd or even [23].”
BackgroundRecurrence resonance - noise-enhanced dynamics in recurrent neural networks
Metzner, C., Schilling, A., Maier, A., & Krauß, P. (2024) · Frontiers in Complex Systems · doi:10.3389/fcpxs.2024.1479417 · Cited by 13
“Specifically, a dynamic state known as the ’edge of chaos’, situated at the transition between periodic and chaotic behavior [13], has been extensively investigated and identified as crucial for computation [14, 15, 16, 17, 18, 19, 20, 21, 22, 23] and short-term memory [24, 25].”
BackgroundMagneto-Ionic Physical Reservoir Computing in Perpendicularly Magnetized Heterostructures
Rajib, M. M., Bhattacharya, D., Jensen, C., Chen, G., Chowdhury, F. F., Sarker, S., Liu, K., & Atulasimha, J. (2025) · Nano Letters · doi:10.1021/acs.nanolett.5c03889 · Cited by 8
Boosting reservoir computing with brain-inspired adaptive control of E-I balance
Srinivasan, K., Plenz, D., & Girvan, M. (2025) · Nature Communications · doi:10.1038/s41467-025-64978-8 · Cited by 18
“I balance, for example, through manipulation of input bias 18 or reservoir activity states 19 .”
What neuroscience can tell AI about learning in continuously changing environments
Durstewitz, D., Averbeck, B. B., & Koppe, G. (2025) · Nature Machine Intelligence · doi:10.1038/s42256-025-01146-z · Cited by 12
Rich spectrum of neural field dynamics in the presence of short-term synaptic depression
Wang, H., Lam, K., Fung, C. C. A., Wong, K. Y. M., & Wu, S. (2015) · Physical Review E · doi:10.1103/physreve.92.032908 · Cited by 7
Li, Y., Huang, B., & Zhang, H. (2016) · Neural Computing and Applications · doi:10.1007/s00521-016-2724-7 · Cited by 6
Kurashige, H., Yamashita, Y., Hanakawa, T., & Honda, M. (2018) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2018.00111 · Cited by 7
“The networks that make up the “edge of chaos” are known to possess the preferred characteristics for information processing (Bertschinger and Natschläger, 2004; Legenstein and Maass, 2007; Büsing et al., 2010; Boedecker et al., 2012).”
BackgroundControlling extended criticality via modular connectivity
Gutjahr, N., Hövel, P., & Viol, A. (2021) · Journal of Physics Complexity · doi:10.1088/2632-072x/ac202e · Cited by 7
Numerical simulation of performance of magnonic reservoir computer based on active-ring oscillator
Кондрашов, А. В., Nikitin, A. A., Nikitin, A. A., Kostylev, M., & Ustinov, A. B. (2022) · Journal of Magnetism and Magnetic Materials · doi:10.1016/j.jmmm.2022.169968 · Cited by 6
Emergent computations in trained artificial neural networks and real brains
Parga, N., Serrano-Fernández, L., & Falcó-Roget, J. (2023) · Journal of Instrumentation · doi:10.1088/1748-0221/18/02/c02060 · Cited by 8
“Then, they considered that a signal of sufficient amplitude and frequency induces a transition to from a chaotic to a nonchaotic state [52, 53, 54].”
Complexity and transition to chaos in coupled Adler-type oscillators
Estévez-Moya, D., Rams, E. E., & Kantz, H. (2023) · Physical review. E · doi:10.1103/physreve.107.044212 · Cited by 7
Statistical complexity is maximized in a small-world brain
Tan, T. L., & Cheong, S. A. (2017) · PLoS ONE · doi:10.1371/journal.pone.0183918 · Cited by 8
Controlling excitation avalanches in driven Rydberg gases
Klocke, K., & Buchhold, M. (2019) · Physical Review A · doi:10.1103/physreva.99.053616 · Cited by 7
Neuronal avalanches: Where temporal complexity and criticality meet
Dehghani-Habibabadi, M., Zare, M., Shahbazi, F., Mafahim, J. U., & Grigolini, P. (2017) · The European Physical Journal E · doi:10.1140/epje/i2017-11590-8 · Cited by 6
A random energy approach to deep learning
Xie, R., & Matteo Marsili, A. (2022) · Journal of Statistical Mechanics Theory and Experiment · doi:10.1088/1742-5468/ac7794 · Cited by 6
“There is extensive evidence that learning efficiency and criticality are related [12, 3, 21].”
Perturbations and phase transitions in swarm optimization algorithms
Vantuch, T., Zelinka, I., Adamatzky, A., & Marwan, N. (2019) · Natural Computing · doi:10.1007/s11047-019-09741-x · Cited by 6
Spike-phase coupling as an order parameter in a leaky integrate-and-fire model
Safari, N., Shahbazi, F., Dehghani-Habibabadi, M., Esghaei, M., & Zare, M. (2020) · Physical review. E · doi:10.1103/physreve.102.052202 · Cited by 5
Mayor, J., & Gerstner, W. (2004) · Journal of Physiology-Paris · doi:10.1016/j.jphysparis.2005.09.009 · Cited by 8
Hennequin, G., Ahmadian, Y., Rubin, D. B., Lengyel, M., & Miller, K. D. (2016) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/094334 · Cited by 7
“…2012; Ponce-Alvarez et al., 2013; Doiron and Litwin-Kumar, 2014; Mochol et al., 2015); and a stimulus may quench chaotic dynamics to produce non-chaotic dynamics (Molgedey et al., 1992; Bertschinger and Natschlger, 2004; Sussillo and Abbott, 2009; Rajan et al., 2010; Laje and Buonomano, 2013).”
BackgroundReservoir Computing Using Non-Uniform Binary Cellular AutomataPreprint
Nichele, S., & Gundersen, M. S. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1702.03812 · Cited by 6
“As discussed in [3], the reservoir should preferably exhibit edge of chaos behaviors [16], in order to allow for high computational power [10].”
BackgroundJu, H., Kim, J. Z., & Bassett, D. S. (2018) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/504761 · Cited by 6
“In a complementary line of theoretical work, power-law distributions in size and duration have been linked to self-organized criticality [4], and systems of this type are said to have optimal information storage [24], optimal information transmission [6], optimal dynamic range [36,57], and computational power [9].”
BackgroundReconstructing Computational Dynamics from Neural Measurements with Recurrent Neural NetworksPreprint
Durstewitz, D., Koppe, G., & Thurm, M. I. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.10.31.514408 · Cited by 6
Chen, Y., Chen, C.-C., Mei, Y.-Y., Zhou, N., Dongchuan, W., & Lee, T.-K. (2022) · Chinese Journal of Physics · doi:10.1016/j.cjph.2021.12.010 · Cited by 6
Singh, P., Hrishit, B. P., & Raman, B. (2025) · Chaos Solitons & Fractals · doi:10.1016/j.chaos.2025.117160 · Cited by 11
Edge-of-chaos-enhanced quantum-inspired algorithm for combinatorial optimization
Goto, H., Hidaka, R., & Tatsumura, K. (2026) · Physical Review Applied · doi:10.1103/2qd9-x6v8 · Cited by 15
Geadah, V., Horoi, S., Kerg, G., Wolf, G., & Lajoie, G. (2024) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1012567 · Cited by 7
“(3) Finally, system-level learning pressures drive biologically plausible adaptation strategies, namely activation function having biologically realistic configurations and more importantly, the implementation of biological SFA mechanisms such as gain scaling and input differentiation.”
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.