Single-Entity Spiking Neuron Models: Survey

2026-07-09 · Source: cs.NE updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Advanced, long

Summary

This survey reviews mathematical models for single-entity spiking neurons, classifying them by features and use cases to simulate biologically plausible neural systems. It details Integrate and Fire (IF) models, including basic, leaky, adaptive, and quadratic variations, noting their simplicity for neuromorphic processors despite limited biological accuracy. The paper then covers Hodgkin-Huxley (HH) based models, such as the FitzHugh-Nagumo and computationally efficient Izhikevich (IZ) neuron, highlighting their biophysical plausibility but computational demands. It also examines Dendritic Neuron Models (DNM), which incorporate dendrite shape and synaptic interactions, and discusses Artificial Neural Networks (ANNs) like CNN-LSTM, which can simulate complex neuron behavior and benefit from parallelization, though they face training costs and out-of-domain data challenges. The work aims to guide model selection.

Key takeaway

For AI Scientists or Machine Learning Engineers designing biologically plausible neural networks, you should carefully match model complexity to your simulation goals. If computational efficiency is paramount for large-scale brain simulations, prioritize Izhikevich neuron models. Conversely, if detailed ion-channel dynamics are critical, be aware that Integrate and Fire models are insufficient, necessitating more complex Hodgkin-Huxley based approaches despite their higher computational cost.

Key insights

The survey categorizes single-entity neuron models by biological plausibility, computational efficiency, and specific use cases, guiding selection.

Principles

• Integrate and Fire models prioritize simplicity over biological detail.
• Hodgkin-Huxley models offer high accuracy but demand computational resources.
• Dendritic Neuron Models capture complex synaptic interactions.

Method

The paper categorizes neural system models into single-entity and composite, focusing on the former. It characterizes models based on common features, special use cases, and correlated problems they solve, including mathematical formulations.

In practice

• Use IF models for neuromorphic hardware implementation.
• Employ Izhikevich neurons for efficient large-scale brain simulations.
• Consider DNMs for analyzing synapse and dendrite mechanisms.

Topics

• Spiking Neuron Models
• Integrate and Fire Models
• Hodgkin-Huxley Model
• Izhikevich Neuron
• Dendritic Neuron Models
• Computational Neuroscience

Best for: AI Scientist, Machine Learning Engineer, Research Scientist

Related on AIssential

• Memory-Augmented Spiking Networks: Synergistic Integration of Complementary Mechanisms for Neuromorphic Vision — Architectural balance of memory mechanisms in SNNs yields synergistic performance gains over individual optimizations.
• Neurons receive precisely tailored teaching signals as we learn — The brain uses neuron-specific error signals, similar to machine learning's backpropagation, for precise learning.
• Robust Parameter and State Estimation in Multiscale Neuronal Systems Using Physics-Informed Neural Networks — PINNs offer robust parameter and state estimation in multiscale neuronal systems, overcoming limitations of traditional methods.
• Bridging the gap between neuromorphic ionic computing and more efficient AI — The brain's efficient neuromorphic computing inspires AI for better efficiency.
• Selective Neuron Amplification for Training-Free Task Enhancement — Weak neuron activation, not missing knowledge, often causes LLM failures, addressable via inference-time amplification.
• Linear Spiking Focus Network — The LSFNN achieves O(N) computational cost for SNNs by transforming an O(N²) focus field into a linear identity.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.NE updates on arXiv.org.