Convergence of regular spiking and intrinsically bursting Izhikevich neuron models as a function of discretization time with Euler method

Harish Gunasekaran, Giacomo Spigler, Alberto Mazzoni, Enrico Cataldo, Calogero Maria Oddo

Research output: Contribution to journalArticleScientificpeer-review

Abstract

This study investigates the trade-off between computational efficiency and accuracy of Izhikevich neuron models by numerically quantifying their convergence to provide design guidelines in choosing the limit time steps during a discretization procedure. This is important for bionic engineering and neuro-robotic applications where the use of embedded computational resources requires the introduction of optimality criteria. Specifically, the regular spiking (RS) and intrinsically bursting (IB) Izhikevich neuron models are evaluated with step inputs of various amplitudes. We analyze the convergence of spike sequences generated under different discretization time steps (10 µs to 10 ms), with respect to an ideal reference spike sequence approximated with a discretization time step of 1 µs. The differences between the ideal reference and the computed spike sequences were quantified by Victor–Purpura (VPd) and van Rossum (VRd) distances. For each distance, we found two limit discretization times (lower dt 1 and upper dt 2 ), as a function of the applied input and thus firing rate, beyond which the convergence is lost for each neuron model. The estimated limit time steps were found to be consistent regardless of metric used (VPd and VRd) and neuron type (RS and IB), but also to depend on the average inter-spike interval (ISI) produced by the neurons. However, in most cases, a good trade-off between the quality of the convergence of the models dynamics and the computational load required to simulate them numerically was found for values of the discretization time step between dt 1 ∈ (0.1 ms, 1 ms) and dt 2 ∈ (2 ms, 3 ms).
Original languageEnglish
Pages (from-to)237-247
Number of pages11
JournalNeurocomputing
Volume350
DOIs
Publication statusPublished - 20 Jul 2019

Fingerprint

Neurons
Bionics
Computational efficiency
Dynamic models
Robotics

Keywords

  • CUSUM
  • Convergence
  • Discretization
  • Euler method
  • Izhikevich Neuron model
  • Victor–Purpura distance
  • van Rossum distance

Cite this

Gunasekaran, Harish ; Spigler, Giacomo ; Mazzoni, Alberto ; Cataldo, Enrico ; Oddo, Calogero Maria. / Convergence of regular spiking and intrinsically bursting Izhikevich neuron models as a function of discretization time with Euler method. In: Neurocomputing. 2019 ; Vol. 350. pp. 237-247.
@article{a3e7bf9d843044ec8c64af7699de605e,
title = "Convergence of regular spiking and intrinsically bursting Izhikevich neuron models as a function of discretization time with Euler method",
abstract = "This study investigates the trade-off between computational efficiency and accuracy of Izhikevich neuron models by numerically quantifying their convergence to provide design guidelines in choosing the limit time steps during a discretization procedure. This is important for bionic engineering and neuro-robotic applications where the use of embedded computational resources requires the introduction of optimality criteria. Specifically, the regular spiking (RS) and intrinsically bursting (IB) Izhikevich neuron models are evaluated with step inputs of various amplitudes. We analyze the convergence of spike sequences generated under different discretization time steps (10 µs to 10 ms), with respect to an ideal reference spike sequence approximated with a discretization time step of 1 µs. The differences between the ideal reference and the computed spike sequences were quantified by Victor–Purpura (VPd) and van Rossum (VRd) distances. For each distance, we found two limit discretization times (lower dt 1 and upper dt 2 ), as a function of the applied input and thus firing rate, beyond which the convergence is lost for each neuron model. The estimated limit time steps were found to be consistent regardless of metric used (VPd and VRd) and neuron type (RS and IB), but also to depend on the average inter-spike interval (ISI) produced by the neurons. However, in most cases, a good trade-off between the quality of the convergence of the models dynamics and the computational load required to simulate them numerically was found for values of the discretization time step between dt 1 ∈ (0.1 ms, 1 ms) and dt 2 ∈ (2 ms, 3 ms).",
keywords = "CUSUM, Convergence, Discretization, Euler method, Izhikevich Neuron model, Victor–Purpura distance, van Rossum distance",
author = "Harish Gunasekaran and Giacomo Spigler and Alberto Mazzoni and Enrico Cataldo and Oddo, {Calogero Maria}",
year = "2019",
month = "7",
day = "20",
doi = "10.1016/j.neucom.2019.03.021",
language = "English",
volume = "350",
pages = "237--247",
journal = "Neurocomputing",
issn = "0925-2312",
publisher = "Elsevier Science BV",

}

Convergence of regular spiking and intrinsically bursting Izhikevich neuron models as a function of discretization time with Euler method. / Gunasekaran, Harish; Spigler, Giacomo; Mazzoni, Alberto; Cataldo, Enrico; Oddo, Calogero Maria.

In: Neurocomputing, Vol. 350, 20.07.2019, p. 237-247.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - Convergence of regular spiking and intrinsically bursting Izhikevich neuron models as a function of discretization time with Euler method

AU - Gunasekaran, Harish

AU - Spigler, Giacomo

AU - Mazzoni, Alberto

AU - Cataldo, Enrico

AU - Oddo, Calogero Maria

PY - 2019/7/20

Y1 - 2019/7/20

N2 - This study investigates the trade-off between computational efficiency and accuracy of Izhikevich neuron models by numerically quantifying their convergence to provide design guidelines in choosing the limit time steps during a discretization procedure. This is important for bionic engineering and neuro-robotic applications where the use of embedded computational resources requires the introduction of optimality criteria. Specifically, the regular spiking (RS) and intrinsically bursting (IB) Izhikevich neuron models are evaluated with step inputs of various amplitudes. We analyze the convergence of spike sequences generated under different discretization time steps (10 µs to 10 ms), with respect to an ideal reference spike sequence approximated with a discretization time step of 1 µs. The differences between the ideal reference and the computed spike sequences were quantified by Victor–Purpura (VPd) and van Rossum (VRd) distances. For each distance, we found two limit discretization times (lower dt 1 and upper dt 2 ), as a function of the applied input and thus firing rate, beyond which the convergence is lost for each neuron model. The estimated limit time steps were found to be consistent regardless of metric used (VPd and VRd) and neuron type (RS and IB), but also to depend on the average inter-spike interval (ISI) produced by the neurons. However, in most cases, a good trade-off between the quality of the convergence of the models dynamics and the computational load required to simulate them numerically was found for values of the discretization time step between dt 1 ∈ (0.1 ms, 1 ms) and dt 2 ∈ (2 ms, 3 ms).

AB - This study investigates the trade-off between computational efficiency and accuracy of Izhikevich neuron models by numerically quantifying their convergence to provide design guidelines in choosing the limit time steps during a discretization procedure. This is important for bionic engineering and neuro-robotic applications where the use of embedded computational resources requires the introduction of optimality criteria. Specifically, the regular spiking (RS) and intrinsically bursting (IB) Izhikevich neuron models are evaluated with step inputs of various amplitudes. We analyze the convergence of spike sequences generated under different discretization time steps (10 µs to 10 ms), with respect to an ideal reference spike sequence approximated with a discretization time step of 1 µs. The differences between the ideal reference and the computed spike sequences were quantified by Victor–Purpura (VPd) and van Rossum (VRd) distances. For each distance, we found two limit discretization times (lower dt 1 and upper dt 2 ), as a function of the applied input and thus firing rate, beyond which the convergence is lost for each neuron model. The estimated limit time steps were found to be consistent regardless of metric used (VPd and VRd) and neuron type (RS and IB), but also to depend on the average inter-spike interval (ISI) produced by the neurons. However, in most cases, a good trade-off between the quality of the convergence of the models dynamics and the computational load required to simulate them numerically was found for values of the discretization time step between dt 1 ∈ (0.1 ms, 1 ms) and dt 2 ∈ (2 ms, 3 ms).

KW - CUSUM

KW - Convergence

KW - Discretization

KW - Euler method

KW - Izhikevich Neuron model

KW - Victor–Purpura distance

KW - van Rossum distance

UR - http://www.mendeley.com/research/convergence-regular-spiking-intrinsically-bursting-izhikevich-neuron-models-function-discretization

U2 - 10.1016/j.neucom.2019.03.021

DO - 10.1016/j.neucom.2019.03.021

M3 - Article

VL - 350

SP - 237

EP - 247

JO - Neurocomputing

JF - Neurocomputing

SN - 0925-2312

ER -