Autocorrelations from emergent bistability in homeostatic spiking neural 
networks on neuromorphic hardware

Cramer, Benjamin; Kreft, Markus; Billaudelle, Sebastian; Karasenko, Vitali; Leibfried, Aron; Müller, Eric; Spilger, Philipp; Weis, Johannes; Schemmel, Johannes; Muñoz, Miguel A.; Priesemann, Viola; Zierenberg, Johannes

doi:10.1103/PhysRevResearch.5.033035

Local TagsRelease HistoryDetailsSummary

Autocorrelations from emergent bistability in homeostatic spiking neural networks on neuromorphic hardware

Cramer, B., Kreft, M., Billaudelle, S., Karasenko, V., Leibfried, A., Müller, E., et al. (2023). Autocorrelations from emergent bistability in homeostatic spiking neural networks on neuromorphic hardware. Physical Review Research, 5(3): 033035. doi:10.1103/PhysRevResearch.5.033035.

Item is Released

show all hide all

Basic

show hide

Item Permalink: https://hdl.handle.net/21.11116/0000-000D-A76F-B Version Permalink: https://hdl.handle.net/21.11116/0000-000D-A770-8

Genre: Journal Article

Files

show Files

hide Files

:

PhysRevResearch.5.033035.pdf (Publisher version), 3MB

View Save

File Permalink:
https://hdl.handle.net/21.11116/0000-000D-A771-7

Name:
PhysRevResearch.5.033035.pdf

Description:
-

OA-Status:
Gold

Visibility:
Public

MIME-Type / Checksum:
application/pdf / [MD5]

Technical Metadata:

View

Copyright Date:
-

Copyright Info:
-

License:
https://creativecommons.org/licenses/by/4.0/

Locators

show

Creators

show

hide

Creators:
Cramer, Benjamin, Author
Kreft, Markus, Author
Billaudelle, Sebastian, Author
Karasenko, Vitali, Author
Leibfried, Aron, Author
Müller, Eric, Author
Spilger, Philipp, Author
Weis, Johannes, Author
Schemmel, Johannes, Author
Muñoz, Miguel A., Author
Priesemann, Viola¹, Author
Zierenberg, Johannes¹, Author

Affiliations:
1Max Planck Research Group Complex Systems Theory, Max Planck Institute for Dynamics and Self-Organization, Max Planck Society, ou_2616694

Content

show

hide

Free keywords: -

Abstract: A fruitful approach towards neuromorphic computing is to mimic mechanisms of the brain in physical devices, which has led to successful replication of neuronlike dynamics and learning in the past. However, there remains a large set of neural self-organization mechanisms whose role for neuromorphic computing has yet to be explored. One such mechanism is homeostatic plasticity, which has recently been proposed to play a key role in shaping network dynamics and correlations. Here, we study—from a statistical-physics point of view—the emergent collective dynamics in a homeostatically regulated neuromorphic device that emulates a network of excitatory and inhibitory leaky integrate-and-fire neurons. Importantly, homeostatic plasticity is only active during the training stage and results in a heterogeneous weight distribution that we fix during the analysis stage. We verify the theoretical prediction that reducing the external input in a homeostatically regulated neural network increases temporal correlations, measuring autocorrelation times exceeding 500 ms, despite single-neuron timescales of only 20 ms, both in experiments on neuromorphic hardware and in computer simulations. However, unlike theoretically predicted near-critical fluctuations, we find that temporal correlations can originate from an emergent bistability. We identify this bistability as a fluctuation-induced stochastic switching between metastable active and quiescent states in the vicinity of a nonequilibrium phase transition. Our results thereby constitute a complementary mechanism for emergent autocorrelations in networks of spiking neurons with implications for future developments in neuromorphic computing.

Details

show

hide

Language(s): eng - English

Dates: Published Online: 2023-07-19Date issued: 2023

Publication Status: Issued

Pages: -

Publishing info: -

Table of Contents: -

Rev. Type: Peer

Identifiers: DOI: 10.1103/PhysRevResearch.5.033035

Degree: -

Event

show

Legal Case

show

Project information

show hide

Project name : This work has received funding from the European Union Sixth Framework Programme (FP6/2002-2006) un- der Grant Agreement No. 15879 (FACETS), the European Union Seventh Framework Programme (FP7/2007-2013) un- der Grant Agreements No. 604102 (HBP), No. 269921 (BrainScaleS), and No. 243914 (Brain-i-Nets), the Horizon 2020 Framework Programme (H2020/2014-2020) under Grant Agreements No. 720270, No. 785907, and No. 945539 (HBP), the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strat- egy EXC 2181/1-390900948 (the Heidelberg STRUCTURES Excellence Cluster), the Helmholtz Association Initiative and Networking Fund [Advanced Computing Architectures (ACA)] under Project No. SO-092, and the Manfred Stärk Foundation. M.A.M. acknowledges support from the Span- ish Ministry and Agencia Estatal de Investigación (AEI) through Project I+D+i (Reference No. PID2020-113681GB- I00), financed by MICIN/AEI/10.13039/501100011033 and FEDER “A way to make Europe,” as well as the Consejería de Conocimiento, Investigación Universidad, Junta de An- dalucía, and European Regional Development Fund, Project No. P20-00173, for financial support. V.P. and J.Z. were sup- ported by the Max Planck Society. J.Z. received financial support from the Joachim Herz Stiftung and the Plan Pro- pio de Investigación y Transferencia de la Universidad de Granada. The authors acknowledge support from the state of Baden-Württemberg through bwHPC and the DFG through Grant No. INST 39/963-1 FUGG (bwForCluster NEMO).

Grant ID : -

Funding program : -

Funding organization : -

Project name : FACETS

Grant ID : 15879

Funding program : Sixth Framework Programme (FP6/2002-2006)