Hello,
I am trying to simulate the network of a model with precise spike times as in Hanuschkin (2010), which is based on the balanced random network of Brunel (2000). As far as I understand the model of Hanuschkin exists already in NEST here: https://nest-simulator.readthedocs.io/en/v3.5/models/iaf_psc_exp_ps.html . However, I haven't found an example of the network simulation. I have tried using the Brunel network (https://nest-simulator.readthedocs.io/en/stable/auto_examples/brunel_delta_n...) with the iaf_psc_exp_ps model, but I got a firing rate (excitatory or inhibitory) of 0 Hz.
I would appreciate any advice on that!
Thanks in advance, Jenny
Hello Evgenia,
I assume you took the brunel_delta_nest model and just replaced the iaf_psc_delta neuron model with iaf_psc_exp_ps.
For the delta-model, each synaptic input causes an instantaneous jump of the membrane potential by J_ex or J_in, respectively. These are given by J in turn and with the default J=0.1 and g=5, this means excitatory jumps of +0.1 mV and inhibitory jumps of -0.5 mV.
In the iaf_psc_exp* models, on the other hand, the same input weight means that an incoming spike starts a postsynaptic current of 0.1 pA or -0.5 pA, respectively, which then decays exponentially. The resulting membrane potential change may be noticeably smaller. So I would suggest to increase J to see when you get spiking. Don't worry about increasing J by factors rather than small steps.
To look at details, you could also attach a voltmeter to a few neurons so you can see the membrane potential evolution.
Best regards, Hans Ekkehard
--
Prof. Dr. Hans Ekkehard Plesser
Department of Data Science Faculty of Science and Technology Norwegian University of Life Sciences PO Box 5003, 1432 Aas, Norway
Phone +47 6723 1560 Email hans.ekkehard.plesser@nmbu.nomailto:hans.ekkehard.plesser@nmbu.no Home http://arken.nmbu.no/~plesser
From: Evgenia Kartsaki evgenia.kartsaki@inria.fr Date: Friday, 25 August 2023 at 10:41 To: users@nest-simulator.org users@nest-simulator.org Subject: [NEST Users] Network simulation of Hanuschkin (2010) - precise spike times [Some people who received this message don't often get email from evgenia.kartsaki@inria.fr. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
Hello,
I am trying to simulate the network of a model with precise spike times as in Hanuschkin (2010), which is based on the balanced random network of Brunel (2000). As far as I understand the model of Hanuschkin exists already in NEST here: https://nest-simulator.readthedocs.io/en/v3.5/models/iaf_psc_exp_ps.html . However, I haven't found an example of the network simulation. I have tried using the Brunel network (https://nest-simulator.readthedocs.io/en/stable/auto_examples/brunel_delta_n...) with the iaf_psc_exp_ps model, but I got a firing rate (excitatory or inhibitory) of 0 Hz.
I would appreciate any advice on that!
Thanks in advance, Jenny _______________________________________________ NEST Users mailing list -- users@nest-simulator.org To unsubscribe send an email to users-leave@nest-simulator.org
Hello Prof. Ekkehard,
Thanks a lot for your reply.
I have tried two approaches: - Firstly, I have replaced the iaf_psc_delta neuron with the iaf_psc_exp_ps and I get an average firing rate of ~47Hz, which doesn't match the 10Hz that is mentioned in Hanuschkin (2010) (section 2.3.2). - Then, I noticed that the neuron_params in the Brunel network do not match the parameters of iaf_psc_exp_ps (Hanuschkin(2010) section 2.1.1). So, tried to simulate the network with the default parameters of iaf_psc_exp_ps and there I get 0 Hz. I have also tried increasing J but it doesn't change anything on the firing rate.
Would you have any further suggestions? Or is there any example of the Hanuschkin network available?
Best regards, Evgenia
Hello Evgenia,
From Hanuschkin et al 2.1.1, I would take the following parameter values
neuron_params = {"C_m": 250.0, "tau_m": 10.0, "t_ref": 2.0, "E_L": 0.0, "V_reset": 0.0, "V_m": 0.0, "V_th": 20, "tau_syn_ex": 1, "tau_syn_in": 1} # J = 0.1 # postsynaptic amplitude in mV J_ex = 32.29 # amplitude of excitatory postsynaptic potential J_in = -201.81 # amplitude of inhibitory postsynaptic potential
Could you try those values? I will reach out to colleagues to see if we can provide original scripts.
Best regards, Hans Ekkehard
--
Prof. Dr. Hans Ekkehard Plesser
Department of Data Science Faculty of Science and Technology Norwegian University of Life Sciences PO Box 5003, 1432 Aas, Norway
Phone +47 6723 1560 Email hans.ekkehard.plesser@nmbu.nomailto:hans.ekkehard.plesser@nmbu.no Home http://arken.nmbu.no/~plesser
From: Evgenia Kartsaki evgenia.kartsaki@inria.fr Date: Friday, 25 August 2023 at 14:35 To: users@nest-simulator.org users@nest-simulator.org Subject: [NEST Users] Re: Network simulation of Hanuschkin (2010) - precise spike times [Du mottar ikke ofte e-post fra evgenia.kartsaki@inria.fr. Finn ut hvorfor dette er viktig p? https://aka.ms/LearnAboutSenderIdentification ]
Hello Prof. Ekkehard,
Thanks a lot for your reply.
I have tried two approaches: - Firstly, I have replaced the iaf_psc_delta neuron with the iaf_psc_exp_ps and I get an average firing rate of ~47Hz, which doesn't match the 10Hz that is mentioned in Hanuschkin (2010) (section 2.3.2). - Then, I noticed that the neuron_params in the Brunel network do not match the parameters of iaf_psc_exp_ps (Hanuschkin(2010) section 2.1.1). So, tried to simulate the network with the default parameters of iaf_psc_exp_ps and there I get 0 Hz. I have also tried increasing J but it doesn't change anything on the firing rate.
Would you have any further suggestions? Or is there any example of the Hanuschkin network available?
Best regards, Evgenia _______________________________________________ NEST Users mailing list -- users@nest-simulator.org To unsubscribe send an email to users-leave@nest-simulator.org
Hello again,
In addition to these single neuron parameters and according to 2.3.2 in Hanuschkin (2010) I have modified the following parameters: - delay = 1.0 ms - order = 2520 so that NE=10080, NI=2520, CE=1008 and CI=252 - p_rate=2710 Hz of the additional excitatory Poissonian spike train ( vext=2.71kHz in the paper)
However, the average firing rate is still 0 Hz.
Thanks once again for your help.
Best regards, Evgenia
Hello Evgenia,
This seems strange. Have you tried to connect a voltmeter to a few neurons? If the membrane potential fluctuates and just always stays below threshold, that would point to an issue with connection strengths, if the membrane potential is just flat it would be a problem with synaptic input.
Best, Hans Ekkehard
--
Prof. Dr. Hans Ekkehard Plesser
Department of Data Science Faculty of Science and Technology Norwegian University of Life Sciences PO Box 5003, 1432 Aas, Norway
Phone +47 6723 1560 Email hans.ekkehard.plesser@nmbu.nomailto:hans.ekkehard.plesser@nmbu.no Home http://arken.nmbu.no/~plesser
From: Evgenia Kartsaki evgenia.kartsaki@inria.fr Date: Friday, 25 August 2023 at 15:41 To: users@nest-simulator.org users@nest-simulator.org Subject: [NEST Users] Re: Network simulation of Hanuschkin (2010) - precise spike times [Du mottar ikke ofte e-post fra evgenia.kartsaki@inria.fr. Finn ut hvorfor dette er viktig p? https://aka.ms/LearnAboutSenderIdentification ]
Hello again,
In addition to these single neuron parameters and according to 2.3.2 in Hanuschkin (2010) I have modified the following parameters: - delay = 1.0 ms - order = 2520 so that NE=10080, NI=2520, CE=1008 and CI=252 - p_rate=2710 Hz of the additional excitatory Poissonian spike train ( vext=2.71kHz in the paper)
However, the average firing rate is still 0 Hz.
Thanks once again for your help.
Best regards, Evgenia _______________________________________________ NEST Users mailing list -- users@nest-simulator.org To unsubscribe send an email to users-leave@nest-simulator.org
Hello Prof. Plesser,
I did use a voltmeter and the membrane potential fluctuated but always stayed below the threshold. However, it's not related to the connection strengths, but the external input. Specifically, when each neuron received an additional excitatory Poissonian spike train at 2.71 kHz then, it has to also receive a constant external direct current of 499 pA (section 2.3.1 of the paper). This requires the parameter I_e to be set to 499 pA in the neuron_params. This resolves my issue and gives an average firing rate of 10Hz.
Thanks a lot for your help!
Best regards, Evgenia
Excellent that you found the solution!
Best, Hans Ekkehard
--
Prof. Dr. Hans Ekkehard Plesser
Department of Data Science Faculty of Science and Technology Norwegian University of Life Sciences PO Box 5003, 1432 Aas, Norway
Phone +47 6723 1560 Email hans.ekkehard.plesser@nmbu.nomailto:hans.ekkehard.plesser@nmbu.no Home http://arken.nmbu.no/~plesser
From: Evgenia Kartsaki evgenia.kartsaki@inria.fr Date: Monday, 28 August 2023 at 11:58 To: users@nest-simulator.org users@nest-simulator.org Subject: [NEST Users] Re: Network simulation of Hanuschkin (2010) - precise spike times [Du mottar ikke ofte e-post fra evgenia.kartsaki@inria.fr. Finn ut hvorfor dette er viktig p? https://aka.ms/LearnAboutSenderIdentification ]
Hello Prof. Plesser,
I did use a voltmeter and the membrane potential fluctuated but always stayed below the threshold. However, it's not related to the connection strengths, but the external input. Specifically, when each neuron received an additional excitatory Poissonian spike train at 2.71 kHz then, it has to also receive a constant external direct current of 499 pA (section 2.3.1 of the paper). This requires the parameter I_e to be set to 499 pA in the neuron_params. This resolves my issue and gives an average firing rate of 10Hz.
Thanks a lot for your help!
Best regards, Evgenia _______________________________________________ NEST Users mailing list -- users@nest-simulator.org To unsubscribe send an email to users-leave@nest-simulator.org
-
Evgenia Kartsaki -
Hans Ekkehard Plesser