10.5061/DRYAD.J5R51
Mazzoni, Alberto
Sant'Anna School of Advanced Studies
Lindén, Henrik Anders
University of Copenhagen
Cuntz, Hermann
Ernst Strüngmann Institute for Neuroscience
Goethe University Frankfurt
Frankfurt Institute for Advanced Studies
Lansner, Anders
Royal Institute of Technology
Panzeri, Stefano
Italian Institute of Technology
Einevoll, Gaute Tomas
Norwegian University of Life Sciences
Lindén, Henrik
University of Copenhagen
Royal Institute of Technology
Data from: Computing the local field potential (LFP) from
integrate-and-fire network models
Dryad
dataset
2016
Local Field Potential
integrate and fire
simulations
2016-10-27T00:00:00Z
2016-10-27T00:00:00Z
en
https://doi.org/10.1371/journal.pcbi.1004584
392873882 bytes
1
CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
Leaky integrate-and-fire (LIF) network models are commonly used to study
how the spiking dynamics of neural networks changes with stimuli, tasks or
dynamic network states. However, neurophysiological studies in vivo often
rather measure the mass activity of neuronal microcircuits with the local
field potential (LFP). Given that LFPs are generated by spatially
separated currents across the neuronal membrane, they cannot be computed
directly from quantities defined in models of point-like LIF neurons.
Here, we explore the best approximation for predicting the LFP based on
standard output from point-neuron LIF networks. To search for this best
“LFP proxy”, we compared LFP predictions from candidate proxies based on
LIF network output (e.g, firing rates, membrane potentials, synaptic
currents) with “ground-truth” LFP obtained when the LIF network synaptic
input currents were injected into an analogous three-dimensional (3D)
network model of multi-compartmental neurons with realistic morphology,
spatial distributions of somata and synapses. We found that a specific
fixed linear combination of the LIF synaptic currents provided an accurate
LFP proxy, accounting for most of the variance of the LFP time course
observed in the 3D network for all recording locations. This proxy
performed well over a broad set of conditions, including substantial
variations of the neuronal morphologies. Our results provide a simple
formula for estimating the time course of the LFP from LIF network
simulations in cases where a single pyramidal population dominates the LFP
generation, and thereby facilitate quantitative comparison between
computational models and experimental LFP recordings in vivo.
Simulated laminar recordings for input intensity 0.5 sp/msSimulated
recordings (10101 ms) of the LFP generated by the 3D network described in
the paper, computed from LFPy. Each row is a different depth. Input
intensity is reported in the title.electrode_LFP5.outSimulated laminar
recordings for input intensity 1.0 sp/msSimulated recordings (10101 ms) of
the LFP generated by the 3D network described in the paper, computed from
LFPy. Each row is a different depth. Input intensity is reported in the
title.electrode_LFP10.outSimulated laminar recordings for input intensity
1.5 sp/msSimulated recordings (10101 ms) of the LFP generated by the 3D
network described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP15.outSimulated laminar recordings for input intensity
2.0 sp/msSimulated recordings (10101 ms) of the LFP generated by the 3D
network described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP20.outSimulated laminar recordings for input intensity
2.5 sp/msSimulated recordings (10101 ms) of the LFP generated by the 3D
network described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP25.outSimulated laminar recordings for input intensity
3.0 sp/msSimulated recordings (10101 ms) of the LFP generated by the 3D
network described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP30.outSimulated laminar recordings for input intensity
6.0 sp/msSimulated recordings (10101 ms) of the LFP generated by the 3D
network described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP60.outSimulated laminar recordings for homogeneous
synaptic distributionSimulated recordings (10101 ms) of the LFP generated
by the 3D network described in the paper when both inhibitory and
excitatory synapses are distributed homogeneously over the whole neuron
surface, computed from LFPy. Each row is a different depth. Input
intensity is 1.5 sp/ms.homelectrode_LFP15.outSimulated laminar recordings
for excitatory synapses only in the upper dendritic bushSimulated
recordings (10101 ms) of the LFP generated by the 3D network described in
the paper when excitatory synapses located only in the upper dendritic
bush, computed from LFPy. Each row is a different depth. Input intensity
is 1.5 sp/ms.excupelectrode_LFP15.outSimulated laminar recordings for
excitatory synapses only in the lower dendritic bushSimulated recordings
(10101 ms) of the LFP generated by the 3D network described in the paper
when excitatory synapses are located only in the lower dendritic bush,
computed from LFPy. Each row is a different depth. Input intensity is 1.5
sp/ms.excdownelectrode_LFP15.outSimulated laminar recordings for cortical
excitatory synapses only in the upper dendritic bushSimulated recordings
(10101 ms) of the LFP generated by the 3D network described in the paper
when cortical excitatory synapses are located only in the upper bush and
thalamic excitatory synapses are distributed homogeneneously over the
whole neuron surface, computed from LFPy. Each row is a different depth.
Input intensity is 1.5 sp/ms.cortupelectrode_LFP15.outLIF network output
variables for input 1.5 sp/msOutputs of Leaky Integrate and Fire Network
described in the paper for input intensity 1.5 sp/ms. Simulation lasts for
10100 ms sampled at 20kHz. Columns are 1) Membrane potential 2-5) local
AMPA, thalamic AMPA, long range cortical AMPA, local GABA over excitatory
neurons, 6-9) same as 2-5) but computed over inhibitory neurons, 10) input
firing rate, 11-12) firing rate of inhibitory neurons and excitatory
neuronslfp_15.1.outLIF network output variables for input 0.5 sp/msLIF
network output variables for input 0.5 sp/ms Outputs of Leaky Integrate
and Fire Network described in the paper for input intensity 1.5 sp/ms.
Simulation lasts for 10100 ms sampled at 20kHz. Columns are 1) Membrane
potential 2-5) local AMPA, thalamic AMPA, long range cortical AMPA, local
GABA over excitatory neurons, 6-9) same as 2-5) but computed over
inhibitory neurons, 10) input firing rate, 11-12) firing rate of
inhibitory neurons and excitatory neuronslfp_5.1.outLIF network output
variables for input 1.0 sp/msOutputs of Leaky Integrate and Fire Network
described in the paper for input intensity 1.0 sp/ms. Simulation lasts for
10100 ms sampled at 20kHz. Columns are 1) Membrane potential 2-5) local
AMPA, thalamic AMPA, long range cortical AMPA, local GABA over excitatory
neurons, 6-9) same as 2-5) but computed over inhibitory neurons, 10) input
firing rate, 11-12) firing rate of inhibitory neurons and excitatory
neuronslfp_10.1.outLIF network output variables for input 2.0 sp/msOutputs
of Leaky Integrate and Fire Network described in the paper for input
intensity 2.0 sp/ms. Simulation lasts for 10100 ms sampled at 20kHz.
Columns are 1) Membrane potential 2-5) local AMPA, thalamic AMPA, long
range cortical AMPA, local GABA over excitatory neurons, 6-9) same as 2-5)
but computed over inhibitory neurons, 10) input firing rate, 11-12) firing
rate of inhibitory neurons and excitatory neuronslfp_20.1.outLIF network
output variables for input 2.5 sp/msOutputs of Leaky Integrate and Fire
Network described in the paper for input intensity 2.5 sp/ms. Simulation
lasts for 10100 ms sampled at 20kHz. Columns are 1) Membrane potential
2-5) local AMPA, thalamic AMPA, long range cortical AMPA, local GABA over
excitatory neurons, 6-9) same as 2-5) but computed over inhibitory
neurons, 10) input firing rate, 11-12) firing rate of inhibitory neurons
and excitatory neuronslfp_25.1.outLIF network output variables for input
3.0 sp/msOutputs of Leaky Integrate and Fire Network described in the
paper for input intensity 3.0 sp/ms. Simulation lasts for 10100 ms sampled
at 20kHz. Columns are 1) Membrane potential 2-5) local AMPA, thalamic
AMPA, long range cortical AMPA, local GABA over excitatory neurons, 6-9)
same as 2-5) but computed over inhibitory neurons, 10) input firing rate,
11-12) firing rate of inhibitory neurons and excitatory
neuronslfp_30.1.outLIF network output variables for input 6.0 sp/msOutputs
of Leaky Integrate and Fire Network described in the paper for input
intensity 6.0 sp/ms. Simulation lasts for 10100 ms sampled at 20kHz.
Columns are 1) Membrane potential 2-5) local AMPA, thalamic AMPA, long
range cortical AMPA, local GABA over excitatory neurons, 6-9) same as 2-5)
but computed over inhibitory neurons, 10) input firing rate, 11-12) firing
rate of inhibitory neurons and excitatory neuronslfp_60.1.outSimulated
laminar recordings 125 microns from network centerSimulated recordings
(10101 ms) of the LFP generated by the 3D network described in the paper,
computed from LFPy. Virtual electrode is located 125 microns away from
network center. Each row is a different depth. Input intensity is 1.5
sp/mselectrode_LFP15.outSimulated laminar recordings 250 microns from
network centerSimulated recordings (10101 ms) of the LFP generated by the
3D network described in the paper, computed from LFPy. Virtual electrode
is located 250 microns away from network center. Each row is a different
depth. Input intensity is 1.5 sp/ms.electrode_LFP15.outSimulated laminar
recordings 375 microns from network centerSimulated recordings (10101 ms)
of the LFP generated by the 3D network described in the paper, computed
from LFPy. Virtual electrode is located 375 microns away from network
center. Each row is a different depth. Input intensity is 1.5
sp/ms.electrode_LFP15.outSimulated laminar recordings 500 microns from
network centerSimulated recordings (10101 ms) of the LFP generated by the
3D network described in the paper, computed from LFPy. Virtual electrode
is located 500 microns away from network center. Each row is a different
depth. Input intensity is 1.5 sp/ms.electrode_LFP15.outSimulated laminar
recordings 625 microns from network centerSimulated recordings (10101 ms)
of the LFP generated by the 3D network described in the paper, computed
from LFPy. Virtual electrode is located 625 microns away from network
center. Each row is a different depth. Input intensity is 1.5
sp/ms.electrode_LFP15.outSimulated laminar recordings 875 microns from
network centerSimulated recordings (10101 ms) of the LFP generated by the
3D network described in the paper, computed from LFPy. Virtual electrode
is located 875 microns away from network center. Each row is a different
depth. Input intensity is 1.5 sp/mselectrode_LFP15.outSimulated laminar
recordings 1000 microns from network centerSimulated recordings (10101 ms)
of the LFP generated by the 3D network described in the paper, computed
from LFPy. Virtual electrode is located 1000 microns away from network
center. Each row is a different depth. Input intensity is 1.5
sp/ms.electrode_LFP15.outSimulated laminar recordings for input intensity
0.5 sp/ms when synaptic model is conductance-basedSimulated recordings
(10101 ms) of the LFP generated by the 3D network with conductance-based
synapses described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP5.outSimulated laminar recordings for input intensity
1.0 sp/ms when synaptic model is conductance basedSimulated recordings
(10101 ms) of the LFP generated by the 3D network with conductance-based
synapses described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP10.outSimulated laminar recordings for input intensity
1.5 sp/ms when synaptic model is conductance basedSimulated recordings
(10101 ms) of the LFP generated by the 3D network with conductance-based
synapses described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP15.outSimulated laminar recordings for input intensity
2.0 sp/ms when synaptic model is conductance basedSimulated recordings
(10101 ms) of the LFP generated by the 3D network with conductance-based
synapses described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP20.outSimulated laminar recordings for input intensity
2.5 sp/ms when synaptic model is conductance basedSimulated recordings
(10101 ms) of the LFP generated by the 3D network with conductance-based
synapses described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
titleelectrode_LFP25.outSimulated laminar recordings for input intensity
3.0 sp/ms when synaptic model is conductance basedSimulated recordings
(10101 ms) of the LFP generated by the 3D network with conductance-based
synapses described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP30.outSimulated laminar recordings for input intensity
6.0 sp/ms when synaptic model is conductance basedSimulated recordings
(10101 ms) of the LFP generated by the 3D network with conductance-based
synapses described in the paper, computed from LFPy. Each row is a
different depth. Input intensity is reported in the
title.electrode_LFP60.outSimulated laminar recordings from reconstructed
morphologies for input intensity 0.5 sp/msSimulated recordings (10101 ms)
of the LFP generated by the 3D network with reconstructed (rather than
artificial) morphologies described in the paper, computed from LFPy. Each
row is a different depth. Input intensity is reported in the
title.electrode_LFP5.outSimulated laminar recordings from reconstructed
morphologies for input intensity 1.0 sp/msSimulated recordings (10101 ms)
of the LFP generated by the 3D network with reconstructed (rather than
artificial) morphologies described in the paper, computed from LFPy. Each
row is a different depth. Input intensity is reported in the
title.electrode_LFP10.outSimulated laminar recordings from reconstructed
morphologies for input intensity 1.5 sp/msSimulated recordings (10101 ms)
of the LFP generated by the 3D network with reconstructed (rather than
artificial) morphologies described in the paper, computed from LFPy. Each
row is a different depth. Input intensity is reported in the
title.electrode_LFP15.outSimulated laminar recordings from reconstructed
morphologies for input intensity 2.0 sp/msSimulated recordings (10101 ms)
of the LFP generated by the 3D network with reconstructed (rather than
artificial) morphologies described in the paper, computed from LFPy. Each
row is a different depth. Input intensity is reported in the
title.electrode_LFP20.outSimulated laminar recordings from reconstructed
morphologies for input intensity 2.5 sp/msSimulated recordings (10101 ms)
of the LFP generated by the 3D network with reconstructed (rather than
artificial) morphologies described in the paper, computed from LFPy. Each
row is a different depth. Input intensity is reported in the
title.electrode_LFP25.outSimulated laminar recordings from reconstructed
morphologies for input intensity 3.0 sp/msSimulated recordings (10101 ms)
of the LFP generated by the 3D network with reconstructed (rather than
artificial) morphologies described in the paper, computed from LFPy. Each
row is a different depth. Input intensity is reported in the
title.electrode_LFP30.outSimulated laminar recordings from reconstructed
morphologies for input intensity 6.0 sp/msSimulated recordings (10101 ms)
of the LFP generated by the 3D network with reconstructed (rather than
artificial) morphologies described in the paper, computed from LFPy. Each
row is a different depth. Input intensity is reported in the
title.electrode_LFP60.out