Zonghua Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Zonghua Liu has a total of 196 co-authors. Here are the top co-authors by number of shared publications:
Ming Tang
9 shared publicationsMuhua Zheng
State Key Laboratory of Precision Spectroscopy and Department of Physics, East China Normal University, Shanghai 200062, China
9 shared publicationsShuguang Guan
9 shared publications- Xiyun Zhang
Department of Physics, Jinan University, Guangzhou, China
9 shared publications - Yong Zou
East China Normal University
9 shared publications Jie Zhou
Department of Physics, East China Normal University, Shanghai 200062, People's Republic of China
7 shared publications
Latest Publications
- Open Access
- 20 October 2026
- Published by Springer Nature in Communications Physics
Multi-scale higher-order contagion dynamics in complex networks
- 20 October 2026
- Published by Springer Nature in Communications Physics
Open AccessMulti-scale higher-order contagion dynamics in complex networks
- Open Access
- 14 September 2026
- Published by American Physical Society (APS) in Physical Review Research
- Vol. 8 (3)
Detecting remote synchronization from empirical data of brain networks
The emergence of functional connectivity between distant brain regions that lack a direct structural link remains a long-standing question in neuroscience. Remote synchronization (RS) offers a promising framework for uncovering the underlying mechanisms. So far, RS has been well studied in artificial network topologies such as star or starlike graphs but little attention has been paid to realistic brain networks with complicated topologies, due to their completely different synchronization pathways. In this study, we address this gap by introducing an efficient method, termed , to detect RS using empirical structural and functional brain network data. Under this approach, two nodes are considered to exhibit RS if they are functionally synchronized despite being structurally disconnected. That is, they share neither a direct structural link nor an indirect path composed of synchronized nodes. We demonstrate that the proportion of RS nodes in a brain network strongly depends on the thresholds used to define both structural and functional connectivity, with optimal threshold values existing for each. Furthermore, we show that the method can be generalized: It applies not only to empirical time series but also to oscillator-based network models, and it admits a theoretical interpretation. This indicates that the algorithm offers a general framework to detect RS in complex networks.
- 14 September 2026
- Published by American Physical Society (APS) in Physical Review Research
- Vol. 8 (3)
Open AccessDetecting remote synchronization from empirical data of brain networks
The emergence of functional connectivity between distant brain regions that lack a direct structural link remains a long-standing question in neuroscience. Remote synchronization (RS) offers a promising framework for uncovering the underlying mechanisms. So far, RS has been well studied in artificial network topologies such as star or starlike graphs but little attention has been paid to realistic brain networks with complicated topologies, due to their completely different synchronization pathways. In this study, we address this gap by introducing an efficient method, termed , to detect RS using empirical structural and functional brain network data. Under this approach, two nodes are considered to exhibit RS if they are functionally synchronized despite being structurally disconnected. That is, they share neither a direct structural link nor an indirect path composed of synchronized nodes. We demonstrate that the proportion of RS nodes in a brain network strongly depends on the thresholds used to define both structural and functional connectivity, with optimal threshold values existing for each. Furthermore, we show that the method can be generalized: It applies not only to empirical time series but also to oscillator-based network models, and it admits a theoretical interpretation. This indicates that the algorithm offers a general framework to detect RS in complex networks.
- 1 August 2026
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 36 (8)
Remote propagation of external signals
Recent studies in nonlinear dynamics and complex networks have significantly advanced our understanding of brain functions, particularly in areas such as chimera states, remote synchronization, and remote firing propagation. However, how external signals propagate in networks, a process relevant to cognitive mechanisms, remains less explored. In this work, we propose a model to investigate the propagation of an external signal in a network with heterogeneously distributed natural frequencies. We begin with a star network in which the hub node has a higher natural frequency than the leaf nodes. An external signal is applied to one leaf node (the source node). Interestingly, we observe a phenomenon of remote propagation (RP), where the external signal becomes phase-locked not to the source or hub node but to an indirectly connected leaf node. We further demonstrate that RP persists in more complex network topologies. Additionally, RP can be enhanced by using multiple source nodes and can even lead to double phase-locking, i.e., a finding that may offer insights into signal transmission in brain networks. Finally, we provide a theoretical analysis to explain the RP mechanism and evaluate signal transmission efficacy.
- 1 August 2026
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 36 (8)
Remote propagation of external signals
Recent studies in nonlinear dynamics and complex networks have significantly advanced our understanding of brain functions, particularly in areas such as chimera states, remote synchronization, and remote firing propagation. However, how external signals propagate in networks, a process relevant to cognitive mechanisms, remains less explored. In this work, we propose a model to investigate the propagation of an external signal in a network with heterogeneously distributed natural frequencies. We begin with a star network in which the hub node has a higher natural frequency than the leaf nodes. An external signal is applied to one leaf node (the source node). Interestingly, we observe a phenomenon of remote propagation (RP), where the external signal becomes phase-locked not to the source or hub node but to an indirectly connected leaf node. We further demonstrate that RP persists in more complex network topologies. Additionally, RP can be enhanced by using multiple source nodes and can even lead to double phase-locking, i.e., a finding that may offer insights into signal transmission in brain networks. Finally, we provide a theoretical analysis to explain the RP mechanism and evaluate signal transmission efficacy.
- Open Access
- 2 April 2026
- Published by American Physical Society (APS) in Physical Review Research
- Vol. 8 (2)
Higher-order interactions induce chimera states in globally coupled oscillators
We show that globally coupled phase oscillators inherently possess a phase frustration when four-body interactions are considered. This intrinsic frustration arises when oscillators have asymmetric distributions of initial conditions spanning across multiple fixed points. Moreover, when the frequency distribution is bimodal, the two groups of oscillators may evolve toward distinct macroscopic dynamical behaviors, with one group being synchronized and the other remaining asynchronous. This constitutes a mechanism for the emergence of chimera states, and one that is specifically characteristic of higher-order interactions.
- 2 April 2026
- Published by American Physical Society (APS) in Physical Review Research
- Vol. 8 (2)
Open AccessHigher-order interactions induce chimera states in globally coupled oscillators
We show that globally coupled phase oscillators inherently possess a phase frustration when four-body interactions are considered. This intrinsic frustration arises when oscillators have asymmetric distributions of initial conditions spanning across multiple fixed points. Moreover, when the frequency distribution is bimodal, the two groups of oscillators may evolve toward distinct macroscopic dynamical behaviors, with one group being synchronized and the other remaining asynchronous. This constitutes a mechanism for the emergence of chimera states, and one that is specifically characteristic of higher-order interactions.
- 11 February 2026
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 207
Noise-induced remote synchronization in star networks
- 11 February 2026
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 207
Noise-induced remote synchronization in star networks
- Open Access
Entrainment of the suprachiasmatic nucleus network by a light-dark cycle
The synchronization of biological activity with the alternation of day and night (circadian rhythm) is performed in the brain by a group of neurons, constituting the suprachiasmatic nucleus (SCN). The SCN is divided into two subgroups of oscillating cells: the ventro-lateral (VL) neurons, which are exposed to light (photic signal) and the dorso-medial (DM) neurons which are coupled to the VL cells. When the coupling between these neurons is strong enough, the system synchronizes with the photic period. Upon increasing the cell coupling, the entrainment of the DM cells has been recently shown to occur via a very sharp (jumping) transition when the period of the photic input is larger than the intrinsic period of the cells. Here, we characterize this transition with a simple realistic model. We show that two bifurcations possibly lead to the disappearance of the endogenous mode. Using a mean field model, we show that the jumping transition results from a supercritical Hopf-like bifurcation. This finding implies that both the period and strength of the stimulating photic signal, and the relative fraction of cells in the VL and DM compartments are crucial in determining the synchronization of the system.
Open AccessEntrainment of the suprachiasmatic nucleus network by a light-dark cycle
The synchronization of biological activity with the alternation of day and night (circadian rhythm) is performed in the brain by a group of neurons, constituting the suprachiasmatic nucleus (SCN). The SCN is divided into two subgroups of oscillating cells: the ventro-lateral (VL) neurons, which are exposed to light (photic signal) and the dorso-medial (DM) neurons which are coupled to the VL cells. When the coupling between these neurons is strong enough, the system synchronizes with the photic period. Upon increasing the cell coupling, the entrainment of the DM cells has been recently shown to occur via a very sharp (jumping) transition when the period of the photic input is larger than the intrinsic period of the cells. Here, we characterize this transition with a simple realistic model. We show that two bifurcations possibly lead to the disappearance of the endogenous mode. Using a mean field model, we show that the jumping transition results from a supercritical Hopf-like bifurcation. This finding implies that both the period and strength of the stimulating photic signal, and the relative fraction of cells in the VL and DM compartments are crucial in determining the synchronization of the system.
- 5 July 2025
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 199
Adaptive coupling induced recurrence of phase coalescence
- 5 July 2025
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 199
Adaptive coupling induced recurrence of phase coalescence
- 3 April 2025
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 196
Remote synchronization induced by external signals in star networks
- 3 April 2025
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 196
Remote synchronization induced by external signals in star networks
- 1 April 2025
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 35 (4)
Chaotic time series classification by means of reservoir-based convolutional neural networks
We propose a novel Reservoir Computing (RC) based classification method that distinguishes between different chaotic time series. Our method is composed of two steps: (i) we use the reservoir as a feature extracting machine that captures the salient features of time series data; (ii) the readout layer of the reservoir is subsequently fed into a Convolutional Neural Network (CNN) to facilitate classification and recognition. One of the notable advantages is that the readout layer, as obtained by randomly generated empirical hyper-parameters within the RC module, provides sufficient information for the CNN to accomplish the classification tasks effectively. The quality of extracted features by RC is independently evaluated by the root mean square error, which measures how well the training signal may be reconstructed from the input time series. Furthermore, we propose two ways to implement the RC module, namely, a single shallow RC and parallel RC configurations, to further improve the classification accuracy. The important roles of RC in feature extraction are demonstrated by comparing the results when the CNN is provided with either ordinal pattern probability features or unprocessed raw time series directly, both of which perform worse than RC-based method. In addition to CNN, we show that the readout of RC is good for other classification tools as well. The successful classification of electroencephalogram recordings of different brain states suggests that our RC-based classification tools can be used for experimental studies.
- 1 April 2025
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 35 (4)
Chaotic time series classification by means of reservoir-based convolutional neural networks
We propose a novel Reservoir Computing (RC) based classification method that distinguishes between different chaotic time series. Our method is composed of two steps: (i) we use the reservoir as a feature extracting machine that captures the salient features of time series data; (ii) the readout layer of the reservoir is subsequently fed into a Convolutional Neural Network (CNN) to facilitate classification and recognition. One of the notable advantages is that the readout layer, as obtained by randomly generated empirical hyper-parameters within the RC module, provides sufficient information for the CNN to accomplish the classification tasks effectively. The quality of extracted features by RC is independently evaluated by the root mean square error, which measures how well the training signal may be reconstructed from the input time series. Furthermore, we propose two ways to implement the RC module, namely, a single shallow RC and parallel RC configurations, to further improve the classification accuracy. The important roles of RC in feature extraction are demonstrated by comparing the results when the CNN is provided with either ordinal pattern probability features or unprocessed raw time series directly, both of which perform worse than RC-based method. In addition to CNN, we show that the readout of RC is good for other classification tools as well. The successful classification of electroencephalogram recordings of different brain states suggests that our RC-based classification tools can be used for experimental studies.
- 21 February 2025
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 194
A remote synchronization model of community networks with homogeneous frequencies
- 21 February 2025
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 194
A remote synchronization model of community networks with homogeneous frequencies