Scaled Hippocampus-inspired Neural Networks on Neuromorphic Memristive Hardware

Computer Science > Neural and Evolutionary Computing
[Submitted on 14 Sep 2026]
Title:Scaled Hippocampus-inspired Neural Networks on Neuromorphic Memristive Hardware
View PDFAbstract:The hippocampus, a key brain region for learning and memory, exhibits rich structural diversity, sparse communication, and robust dynamics with incredible energy efficiency. It offers promising insights for novel computing capabilities, particularly when co-designed with emerging hardware technologies. In this work, we draw inspiration from the rodent CA3 hippocampal subregion to develop the first spiking neural network with neuronal diversity and biologically-realistic resting state dynamics demonstrated on memristor hardware. We propose a network downscaling methodology utilizing a 4-prong objective function and demonstrate a small-scale CA3-inspired network with 179 Izhikevich-modeled neurons, 3 neuronal types and 17,996 synapses with similar resting-state dynamics as the orders-of-magnitude larger full-scale network. The small-scale network is mapped to an FPGA/memristor platform using a greedy algorithm and 18,316 memristors. Benefiting from memristor noise, the hardware implementation shows continuous periodic behavior, outperforming simulated hardware. This work showcases the potential of biologically-realistic algorithms on emerging hardware for neuromorphic computing.
References & Citations
Loading...
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

* The work draws inspiration from the rodent CA3 hippocampal subregion.
* A spiking neural network with neuronal diversity and biologically-realistic resting state dynamics was developed on memristor hardware.
* A downscaling methodology utilizing a 4-prong objective function was proposed.
* The small-scale network consisted of 179 Izhikevich-modeled neurons, 3 neuronal types, and 17,996 synapses.
* The small network's resting-state dynamics matched those of a larger network orders of magnitude larger.
* The small-scale network was mapped to an FPGA/memristor platform using a greedy algorithm.
* The implementation used 18,316 memristors.
* The hardware implementation showed continuous periodic behavior, outperforming simulated hardware.

The research introduces a method for scaling hippocampal-inspired spiking neural networks onto memristor hardware by leveraging biological inspiration from the rodent CA3 subregion. The work proposes a downscaling methodology using a four-prong objective function to manage complexity. It demonstrates a small-scale network featuring 179 Izhikevich-modeled neurons, three neuronal types, and 17,996 synapses that exhibit resting-state dynamics similar to a full-scale network. This small model was mapped onto an FPGA/memristor platform using a greedy algorithm, utilizing 18,316 memristors. The hardware implementation exhibited continuous periodic behavior, which surpassed the results from simulated hardware.

The research bridges the gap between biologically plausible network dynamics and emerging neuromorphic hardware by focusing on efficient downscaling methods for complex memory structures. The core implication lies in demonstrating that mimicking the sparse communication and dynamic properties of the hippocampus—such as resting-state behavior—can be successfully realized in physical, energy-efficient systems like memristors. The success in achieving periodic behavior that surpasses simulation suggests a functional alignment between the abstract biological model and the physical implementation when constraints are correctly applied. A critical area for deeper investigation involves understanding how the specific noise characteristics inherent in memristor devices influence these emergent dynamics, as the text notes this behavior is superior to pure simulation. Further inquiry should explore whether replicating subtle mechanisms of synaptic plasticity or long-term potentiation within this constrained hardware environment yields functional advantages beyond mere behavioral similarity. How does the observed periodicity scale with network size and device stochasticity? What are the specific noise characteristics that enable this outperformance, and can these properties be systematically exploited for more complex computations than simple dynamic simulation?

Scaled Hippocampus-inspired Neural Networks on Neuromorphic Memristive Hardware

Facts Only

Executive Summary

Full Take

Scaled Hippocampus-inspired Neural Networks on Neuromorphic Memristive Hardware

Facts Only

Executive Summary

Full Take