Contribution to a conference proceedings/Contribution to a book

FZJ-2026-00334

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Biologically Inspired Multi-Timescale Sequence Processing in Hybrid CMOS-Memristive Hardware

Yang, M.-J.FZJ* ; Nieters, P. ; Hellwig, J.FZJ* ; Spithouris, D.FZJ* ; Dittmann, R.FZJ* ; Pipa, G. ; Strachan, J. P.FZJ*

2025
IEEE

This record in other databases:  

Please use a persistent id in citations: doi:10.1109/ICECS66544.2025.11270719

Abstract: emporal sequences of spiking activity are a primary means of neural information encoding, making the extraction of their structure a fundamental computational challenge in neuroscience and artificial intelligence. Recent advances in neuroscience have introduced the concept of dendritic plateau computation (DPC), a mechanism that emerges from the interaction of localized plateau potentials across functionally compartmentalized dendritic branches. Inspired by this biological paradigm, we propose a memristive dendritic architecture incorporating a multi-compartmental neuron model. Critical sequence features are encoded through tunable resistances at memristive crosspoints, while the adjustable retention time of volatile memristors is utilized to modulate plateau duration, thereby aligning temporal dynamics with application-specific requirements. With the memristive hardware dendrite, we constructed a two-layer network of neuron populations for sequence classification. This work demonstrates the potential advantage of incorporating the active dendritic processes as fundamental operation for implementing efficient brain-inspired computing hardware.

---

Contributing Institute(s):

1. Neuromorphic Compute Nodes (PGI-14)

Research Program(s):

1. 5234 - Emerging NC Architectures (POF4-523) (POF4-523)
2. DFG project G:(GEPRIS)441959088 - Memristiver Zeitabstandskodierer (MemTDE) (441959088) (441959088)

Appears in the scientific report 2025

---