Cognitive Neurodynamics · 2026 · Journal article

Redefining spiking neural networks through the lens of dynamical superspace

Guozhang Chen, Junrong Qi, Xingyu Liu, Qixuan Fan, Zhuoheng Gao, Songcheng Xu, Xuliang Wang, Yujie Wu, Dongping Yang, Douglas Zhou, Tiejun Huang

This perspective introduces dynamical superspace, a theoretical framework organizing spiking neural networks by temporal density and non-equilibrium state-space complexity.

The scientific question

How can spiking network theory unite deep learning with biological non-equilibrium dynamics?

Synchronous rate-based spiking networks overlook the computational capacity of asynchronous biological timing, heterogeneity, and transient non-equilibrium trajectories.

Dynamical superspace of neural computing.
Figure 1 from the published paper ↗

The core idea

Hierarchy of Temporal and State Complexity

The review organizes neural computation along a continuous superspace spanning clock-driven rate equilibria to asynchronous spike timing and complex non-equilibrium transient dynamics.

What we found

Roadmap for Native Neuromorphic Intelligence

The perspective outlines how heterogeneity, criticality, and evolutionary priors can bridge global optimization with local execution for low-latency event perception and long-context memory.

Continue the question

Related reading

Explore our research

Cite this work

Guozhang Chen, Junrong Qi, Xingyu Liu, Qixuan Fan, Zhuoheng Gao, Songcheng Xu, Xuliang Wang, Yujie Wu, Dongping Yang, Douglas Zhou, Tiejun Huang. Redefining spiking neural networks through the lens of dynamical superspace. Cognitive Neurodynamics, 2026. doi:10.1007/s11571-026-10470-y.