An electrophysiological model of GABAergic double bouquet cells (Chrysanthidis et al. 2019)

 Download zip file 
Help downloading and running models
Accession:257610
We present an electrophysiological model of double bouquet cells (DBCs) and integrate them into an established cortical columnar microcircuit model that implements a BCPNN (Bayesian Confidence Propagation Neural Network) learning rule. The proposed architecture effectively solves the problem of duplexed learning of inhibition and excitation by replacing recurrent inhibition between pyramidal cells in functional columns of different stimulus selectivity with a plastic disynaptic pathway. The introduction of DBCs improves the biological plausibility of our model, without affecting the model's spiking activity, basic operation, and learning abilities.
Reference:
1 . Chrysanthidis N, Fiebig F, Lansner A (2019) Introducing double bouquet cells into a modular cortical associative memory model Journal of Computational Neuroscience
Citations  Citation Browser
Model Information (Click on a link to find other models with that property)
Model Type: Realistic Network;
Brain Region(s)/Organism:
Cell Type(s): Neocortex U1 interneuron basket PV GABA cell; Neocortex U1 L2/6 pyramidal intratelencephalic GLU cell; Abstract integrate-and-fire adaptive exponential (AdEx) neuron; Neocortex layer 2-3 interneuron; Neocortex bitufted interneuron;
Channel(s):
Gap Junctions:
Receptor(s):
Gene(s):
Transmitter(s):
Simulation Environment: NEST;
Model Concept(s): Learning;
Implementer(s): Chrysanthidis, Nikolaos [nchr at kth.se]; Fiebig, Florian [fiebig at kth.se]; Lansner, Anders [ala at kth.se];
Search NeuronDB for information about:  Neocortex U1 L2/6 pyramidal intratelencephalic GLU cell; Neocortex U1 interneuron basket PV GABA cell;
/
ChrysanthidisEtAl2019
BCPNN_NEST_Module
module-100725
autom4te.cache
libltdl
sli
aclocal.m4 *
aeif_cond_exp_multisynapse.cpp *
aeif_cond_exp_multisynapse.h *
bcpnn_connection.cpp *
bcpnn_connection.h *
bcpnn_connection_backup.cpp *
bcpnn_connection_backup.h *
bootstrap.sh *
compile *
config.guess *
config.sub *
configure *
configure.ac *
depcomp *
iaf_cond_alpha_bias.cpp *
iaf_cond_alpha_bias.h *
iaf_cond_exp_bias.cpp *
iaf_cond_exp_bias.h *
install-sh *
ltmain.sh *
Makefile.am *
Makefile.in *
missing *
pt_module.cpp *
pt_module.h *
pt_module_config.h.in *
pt_module_names.cpp *
pt_module_names.h *
                            
/*
 *  pt_module.cpp
 *  This file is part of NEST.
 *
 *  Copyright (C) 2008 by
 *  The NEST Initiative
 *
 *  See the file AUTHORS for details.
 *
 *  Permission is granted to compile and modify
 *  this file for non-commercial use.
 *  See the file LICENSE for details.
 *
 */

// include necessary NEST headers
#include "config.h"
#include "network.h"
#include "model.h"
#include "dynamicloader.h"
#include "genericmodel.h"
#include "generic_connector.h"
#include "booldatum.h"
#include "integerdatum.h"
#include "tokenarray.h"
#include "exceptions.h"
#include "sliexceptions.h"
#include "nestmodule.h"

// include headers with your own stuff
#include "pt_module.h"
#include "bcpnn_connection.h"
#include "iaf_cond_alpha_bias.h"
#include "iaf_cond_exp_bias.h"
#include "aeif_cond_exp_multisynapse.h"


// -- Interface to dynamic module loader ---------------------------------------

/*
 * The dynamic module loader must be able to find your module. 
 * You make the module known to the loader by defining an instance of your 
 * module class in global scope. This instance must have the name
 *
 * <modulename>_LTX_mod
 *
 * The dynamicloader can then load modulename and search for symbol "mod" in it.
 */
 
mynest::Pt_Module pt_module_LTX_mod;

// -- DynModule functions ------------------------------------------------------

mynest::Pt_Module::Pt_Module()
  { 
#ifdef LINKED_MODULE
     // register this module at the dynamic loader
     // this is needed to allow for linking in this module at compile time
     // all registered modules will be initialized by the main app's dynamic loader
     nest::DynamicLoaderModule::registerLinkedModule(this);
#endif     
   }

mynest::Pt_Module::~Pt_Module()
   {
   }

   const std::string mynest::Pt_Module::name(void) const
   {
     return std::string("Pt NEST Module"); // Return name of the module
   }

   const std::string mynest::Pt_Module::commandstring(void) const
   {
     /* 1. Tell interpreter that we provide the C++ part of Ml_Module with the
           current revision number. 
        2. Instruct the interpreter to check that ml_module-init.sli exists,
           provides at least version 1.0 of the SLI interface to Ml_Module, and
           to load it.
      */
     return std::string(
//             "(pt_module) run"
			// FOR NEST 2.4.2 use only the upper one
			// FOR NEST 2.2.2 use the lower two lines
		   "/pt_module /C++ ($Revision: 8512 $) provide-component "
		   "/pt_module /SLI (7165) require-component"
       );
   }

   /* BeginDocumentation
      Name: StepPatternConnect - Connect sources and targets with a stepping pattern
      
      Synopsis:
      [sources] source_step [targets] target_step synmod StepPatternConnect -> n_connections
      
      Parameters:
      [sources]     - Array containing GIDs of potential source neurons
      source_step   - Make connection from every source_step'th neuron
      [targets]     - Array containing GIDs of potential target neurons
      target_step   - Make connection to every target_step'th neuron
      synmod        - The synapse model to use (literal, must be key in synapsedict)
      n_connections - Number of connections made
      
      Description:
      This function subsamples the source and target arrays given with steps
      source_step and target_step, beginning with the first element in each array,
      and connects the selected nodes.
      
      Example:
      /first_src 0 /network_size get def
      /last_src /iaf_neuron 20 Create def  % nodes  1 .. 20
      /src [first_src last_src] Range def
      /last_tgt /iaf_neuron 10 Create def  % nodes 21 .. 30
      /tgt [last_src 1 add last_tgt] Range def
      
      src 6 tgt 4 /drop_odd_spike StepPatternConnect 
  
      This connects nodes [1, 7, 13, 19] as sources to nodes [21, 25,
      29] as targets using synapses of type drop_odd_spike, and
      returning 12 as the number of connections.  The following
      command will print the connections (you must paste the SLI
      command as one long line):

      src { /s Set << /source s /synapse_type /static_synapse >> FindConnections { GetStatus /target get } Map dup length 0 gt { cout s <- ( -> ) <- exch <-- endl } if ; } forall
      1 -> [21 25 29]
      7 -> [21 25 29]
      13 -> [21 25 29]
      19 -> [21 25 29]
      
      Remark:
      This function is only provided as an example for how to write your own 
      interface function. 
      
      Author:
      Hans Ekkehard Plesser
      
      SeeAlso:
      Connect, ConvergentConnect, DivergentConnect
   */
   void mynest::Pt_Module::StepPatternConnect_Vi_i_Vi_i_lFunction::execute(SLIInterpreter *i) const
   {
     // Check if we have (at least) five arguments on the stack.
     i->assert_stack_load(5);

     // Retrieve source, source step, target, target step from the stack
     const TokenArray sources = getValue<TokenArray> (i->OStack.pick(4)); // bottom
     const long src_step      = getValue<long>       (i->OStack.pick(3));
     const TokenArray targets = getValue<TokenArray> (i->OStack.pick(2));
     const long tgt_step      = getValue<long>       (i->OStack.pick(1));  
     const Name synmodel_name = getValue<std::string>(i->OStack.pick(0)); // top
     
     // Obtain synapse model index
     const Token synmodel 
       = nest::NestModule::get_network().get_synapsedict().lookup(synmodel_name);
     if ( synmodel.empty() )
       throw nest::UnknownSynapseType(synmodel_name.toString());
     const nest::index synmodel_id = static_cast<nest::index>(synmodel);

     // Build a list of targets with the given step
     TokenArray selected_targets;
     for ( size_t t = 0 ; t < targets.size() ; t += tgt_step )
       selected_targets.push_back(targets[t]);
     
     // Now connect all appropriate sources to this list of targets
     size_t Nconn = 0;  // counts connections
     for ( size_t s = 0 ; s < sources.size() ; s += src_step )
     {
       // We must first obtain the GID of the source as integer
       const nest::long_t sgid = getValue<nest::long_t>(sources[s]);

       // nest::network::divergent_connect() requires weight and delay arrays. We want to use
       // default values from the synapse model, so we pass empty arrays.
       nest::NestModule::get_network().divergent_connect(sgid, selected_targets, 
							 TokenArray(), TokenArray(),
							 synmodel_id);
       Nconn += selected_targets.size();
     }

     // We get here only if none of the operations above throws and exception.
     // Now we can safely remove the arguments from the stack and push Nconn
     // as our result. 
     i->OStack.pop(5);
     i->OStack.push(Nconn);
     
     // Finally, we pop the call to this functions from the execution stack.
     i->EStack.pop();
   }

  //-------------------------------------------------------------------------------------

  void mynest::Pt_Module::init(SLIInterpreter *i, nest::Network*)
  {
    /* Register a neuron or device model.
       Give node type as template argument and the name as second argument.
       The first argument is always a reference to the network.
       Return value is a handle for later unregistration.
    */

    /* nest::register_model<izhik_cond_exp>(nest::NestModule::get_network(),
                                                "izhik_cond_exp"); */

    nest::register_model<iaf_cond_exp_bias>(nest::NestModule::get_network(),
                                                "iaf_cond_exp_bias");

    nest::register_model<iaf_cond_alpha_bias>(nest::NestModule::get_network(),
                                                "iaf_cond_alpha_bias");

    nest::register_model<aeif_cond_exp_multisynapse>(nest::NestModule::get_network(),
                                                "aeif_cond_exp_multisynapse");

    /* Register a synapse type.
       Give synapse type as template argument and the name as second argument.
       The first argument is always a reference to the network.
    */
    nest::register_prototype_connection<BCPNNConnection>(nest::NestModule::get_network(), 
                                                       "bcpnn_synapse");

    /* Register a SLI function.
       The first argument is the function name for SLI, the second a pointer to
       the function object. If you do not want to overload the function in SLI,
       you do not need to give the mangled name. If you give a mangled name, you
       should define a type trie in the ml_module-init.sli file.
    */
    i->createcommand("StepPatternConnect_Vi_i_Vi_i_l", 
                     &stepPatternConnect_Vi_i_Vi_i_lFunction);

  }  // Ml_Module::init()