Synchronization and semantization in deep spiking networks

2025-08-18

AI Summary

Establishing a computational link between spiking neural network dynamics and in vivo cortical observations—specifically, elucidating how synchronous spiking and semantic representation co-emerge in visual cortex. Method: We trained a multilayer spiking neural network on natural image classification using spike-time coding and error-correcting plasticity (ECP), a biologically plausible learning rule that enables supervised learning without predefined architecture. Contribution/Results: The network spontaneously self-organizes hierarchical synchronous activity and semantically selective spike pathways—early layers exhibit divergent spatiotemporal patterns, while deeper layers develop selective γ-band–like synchrony and convergent excitatory pathways. This work provides the first empirical demonstration in deep spiking networks of the coupled evolution of synchronization and semantic separation. It offers an interpretable, learnable computational framework for cortical phenomena such as γ-synchrony, thereby bridging spiking dynamics and cortical computation.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Computer Vision: Representation Learning for Vision
• Machine Learning: Multimodal Learning

Application Category

• Economics, Online Markets and Human Computation: Social networks and social learning
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Recent studies have shown how spiking networks can learn complex functionality through error-correcting plasticity, but the resulting structures and dynamics remain poorly studied. To elucidate how these models may link to observed dynamics in vivo and thus how they may ultimately explain cortical computation, we need a better understanding of their emerging patterns. We train a multi-layer spiking network, as a conceptual analog of the bottom-up visual hierarchy, for visual input classification using spike-time encoding. After learning, we observe the development of distinct spatio-temporal activity patterns. While input patterns are synchronous by construction, activity in early layers first spreads out over time, followed by re-convergence into sharp pulses as classes are gradually extracted. The emergence of synchronicity is accompanied by the formation of increasingly distinct pathways, reflecting the gradual semantization of input activity. We thus observe hierarchical networks learning spike latency codes to naturally acquire activity patterns characterized by synchronicity and separability, with pronounced excitatory pathways ascending through the layers. This provides a rigorous computational hypothesis for the experimentally observed synchronicity in the visual system as a natural consequence of deep learning in cortex.

Problem

• Understanding spiking networks' emergent patterns in deep learning
• Linking model dynamics to observed cortical computation in vivo
• Exploring synchronicity and semantization in hierarchical spiking networks

Innovation

• Spike-time encoding for visual classification
• Multi-layer spiking network training
• Emergent synchronicity and separability patterns