Building a spiking neural network model of the Basal Ganglia on SpiNNaker

Basabdatta Sen Bhattacharya; Sebastian James; Oliver Rhodes; Indar Sugiarto; Andrew Rowley; Alan Stokes; Kevin Gurney; Stephen Furber

doi:10.1109/TCDS.2018.2797426

Building a spiking neural network model of the Basal Ganglia on SpiNNaker

Basabdatta Sen Bhattacharya, Sebastian James, Oliver Rhodes, Indar Sugiarto, Andrew Rowley, Alan Stokes, Kevin Gurney, Stephen Furber

Advanced Processor Technology

Research output: Contribution to journal › Article › peer-review

Abstract

We present a biologically-inspired and scalable
model of the Basal Ganglia (BG) simulated on the SpiNNaker
machine, a biologically-inspired low-power hardware platform
allowing parallel, asynchronous computing. Our BG model con-
sists of six cell populations, where the neuro-computational unit
is a conductance-based Izhikevich spiking neuron; the number
of neurons in each population is proportional to that reported in
anatomical literature. This model is treated as a single-channel
of action-selection in the BG, and is scaled-up to three channels
with lateral cross-channel connections. When tested with two
competing inputs, this three-channel model demonstrates action-
selection behaviour. The SpiNNaker-based model is mapped
exactly on to SpineML running on a conventional computer; both
model responses show functional and qualitative similarity, thus
validating the usability of SpiNNaker for simulating biologically-
plausible networks. Furthermore, the SpiNNaker-based model
simulates in real time for time-steps
1 ms; power dissipated
during model execution is ≈ 1.8 W.

Original language	English
Pages (from-to)	1-14
Number of pages	14
Journal	IEEE Transactions on Cognitive and Developmental Systems
DOIs	https://doi.org/10.1109/TCDS.2018.2797426
Publication status	Published - 8 Mar 2018

Keywords

asal Ganglia, SpiNNaker, biologically-inspired, SpineML, Spiking Neural Network, Izhikevich Neuron models

Access to Document

10.1109/TCDS.2018.2797426Licence: CC BY

Cite this

@article{53150f30754e4ec4bb0939605f223551,

title = "Building a spiking neural network model of the Basal Ganglia on SpiNNaker",

abstract = "We present a biologically-inspired and scalablemodel of the Basal Ganglia (BG) simulated on the SpiNNakermachine, a biologically-inspired low-power hardware platformallowing parallel, asynchronous computing. Our BG model con-sists of six cell populations, where the neuro-computational unitis a conductance-based Izhikevich spiking neuron; the numberof neurons in each population is proportional to that reported inanatomical literature. This model is treated as a single-channelof action-selection in the BG, and is scaled-up to three channelswith lateral cross-channel connections. When tested with twocompeting inputs, this three-channel model demonstrates action-selection behaviour. The SpiNNaker-based model is mappedexactly on to SpineML running on a conventional computer; bothmodel responses show functional and qualitative similarity, thusvalidating the usability of SpiNNaker for simulating biologically-plausible networks. Furthermore, the SpiNNaker-based modelsimulates in real time for time-steps1 ms; power dissipatedduring model execution is ≈ 1.8 W.",

keywords = "asal Ganglia, SpiNNaker, biologically-inspired, SpineML, Spiking Neural Network, Izhikevich Neuron models",

author = "{Sen Bhattacharya}, Basabdatta and Sebastian James and Oliver Rhodes and Indar Sugiarto and Andrew Rowley and Alan Stokes and Kevin Gurney and Stephen Furber",

year = "2018",

month = mar,

day = "8",

doi = "10.1109/TCDS.2018.2797426",

language = "English",

pages = "1--14",

journal = "IEEE Transactions on Cognitive and Developmental Systems",

issn = "2379-8939",

publisher = "IEEE",

}

TY - JOUR

T1 - Building a spiking neural network model of the Basal Ganglia on SpiNNaker

AU - Sen Bhattacharya, Basabdatta

AU - James, Sebastian

AU - Rhodes, Oliver

AU - Sugiarto, Indar

AU - Rowley, Andrew

AU - Stokes, Alan

AU - Gurney, Kevin

AU - Furber, Stephen

PY - 2018/3/8

Y1 - 2018/3/8

N2 - We present a biologically-inspired and scalablemodel of the Basal Ganglia (BG) simulated on the SpiNNakermachine, a biologically-inspired low-power hardware platformallowing parallel, asynchronous computing. Our BG model con-sists of six cell populations, where the neuro-computational unitis a conductance-based Izhikevich spiking neuron; the numberof neurons in each population is proportional to that reported inanatomical literature. This model is treated as a single-channelof action-selection in the BG, and is scaled-up to three channelswith lateral cross-channel connections. When tested with twocompeting inputs, this three-channel model demonstrates action-selection behaviour. The SpiNNaker-based model is mappedexactly on to SpineML running on a conventional computer; bothmodel responses show functional and qualitative similarity, thusvalidating the usability of SpiNNaker for simulating biologically-plausible networks. Furthermore, the SpiNNaker-based modelsimulates in real time for time-steps1 ms; power dissipatedduring model execution is ≈ 1.8 W.

AB - We present a biologically-inspired and scalablemodel of the Basal Ganglia (BG) simulated on the SpiNNakermachine, a biologically-inspired low-power hardware platformallowing parallel, asynchronous computing. Our BG model con-sists of six cell populations, where the neuro-computational unitis a conductance-based Izhikevich spiking neuron; the numberof neurons in each population is proportional to that reported inanatomical literature. This model is treated as a single-channelof action-selection in the BG, and is scaled-up to three channelswith lateral cross-channel connections. When tested with twocompeting inputs, this three-channel model demonstrates action-selection behaviour. The SpiNNaker-based model is mappedexactly on to SpineML running on a conventional computer; bothmodel responses show functional and qualitative similarity, thusvalidating the usability of SpiNNaker for simulating biologically-plausible networks. Furthermore, the SpiNNaker-based modelsimulates in real time for time-steps1 ms; power dissipatedduring model execution is ≈ 1.8 W.

KW - asal Ganglia, SpiNNaker, biologically-inspired, SpineML, Spiking Neural Network, Izhikevich Neuron models

U2 - 10.1109/TCDS.2018.2797426

DO - 10.1109/TCDS.2018.2797426

M3 - Article

SN - 2379-8939

SP - 1

EP - 14

JO - IEEE Transactions on Cognitive and Developmental Systems

JF - IEEE Transactions on Cognitive and Developmental Systems

ER -

Building a spiking neural network model of the Basal Ganglia on SpiNNaker

Abstract

Keywords

Access to Document

Fingerprint

Cite this