Related Documentation:
Convert a GENESIS 2 Simulation to GENESIS 3
Creating GENESIS 3 Simulations with Python
Index of GENESIS 3 User Tutorials
This describes some GENESIS 2 single cell and channel models that have been converted to G-3 NDF format for inclusion in the '/usr/local/neurospaces/models/library' directory. The list of model categories may be seen with the g-shell command 'library_show'. The cells and channels are listed with the commands 'library_show ndf cells' and 'library_show ndf channels'.
Except as noted, NDF format files in the cells/ directory were produced with the gshell 'sli_load" and 'ndf_save' commands from GENESIS 2 SLI scripts that create a cell '/cell' from prototpe compartments in '/library'. A typical set of commands to generate the file 'simplecell-nolib.g would be:
sli_load simplecell-nolib.g delete /proto delete /output delete /library ndf_save /cell simplecell-nolib.ndf quit
Presently, the command ndf_save /cell or ndf_save /cell/** saves not only the '/cell' element tree, but also the '/library' and the elements created by default in GENESIS 2 '/proto' and '/output'. These trees were deleted before the models were saved.
Also, NDF files generated in this manner have FORWARDPARAMETERS blocks that are incompatible with their use with 'ndf_namespace_load'. As noted below, most of these files were then edited manually to replace FORWARDPARAMETERS blocks with equivalent PARAMETERS blocks within the elements that are named in the block. In addition, the GROUP block was renamed CELL. These changes allow them to be used in network simulations, and gshell test scripts have been provided to test them in a simple two cell network. The edited NDF files also have short versions of the model description in comments at the beginning.
With the exception of squidcell.ndf, all of these cell models have synchans and spikegens to allow synaptic connections. Unless noted, they may be used in networks with the 'ndf_namespace_load' command.
squidcell.ndf - is based on the model of the squid giant axon segment that was used in the original GENESIS 'Squid' tutorial by Mark Nelson in 1989. Recent versions of the tutorial scripts accompany the GENESIS 2.3 distribution. The 'cell' is a cylindrical segment with length = diameter = 500 um, with the parameters that were used in the original Hodgkin-Huxley model and the Squid tutorial. The Hodgkin-Huxley Na and K channels are implemented as tabchannels. Unlike the GENESIS 2 version, the G-3 NDF representation uses SI units instead of physiological units, and sets the resting potential to -0.070 volts, instead of zero. This model will be used in future G-3 versions of the Squid tutorial.
simplecell.ndf - The basic two compartment 'simplecell' model that is used in many of the G-3 ns-sli and gshell test scripts, and in the GENESIS 2 Modeling Tutorial section Building a cell the easy way. It also forms the basis of recent GENESIS 3 Python scripting tutorials, e.g. Creating GENESIS 3 Simulations with Python.
The soma compartment contains "squid-like" Na and K channels formed from the channel prototype script 'hh_tchan.g' in the GENESIS 2 neurokit/prototypes library. The gate tables of these channels are filled using the 'setupalpha' command with no added options. The dendrite compartment contains an excitatory synchan 'Ex_channel' and an inhibitory synchan 'Inh_channel'. Because of its simplified morphology, this is useful model for testing variations of GENESIS 2 scripting commands that are used in more complex models, and for testing their conversion to G-3.
simplecell-nolib.ndf is an older version, not suitable for networks.
RScell.ndf - RScell is a very simple one-compartment model of a neocortical regular spiking pyramidal cell that is used in the RSnet simulation. This example simulation is described in the GENESIS 2 Modeling Tutorial section Creating large networks with GENESIS. and the G-3 tutorial Modeling Synaptic Connections and Large Networks with G-3. The simplicity of the cell model makes it suitable for use in large network models that need to run quickly.
In addition to a fast sodium current and delayed rectifier potassium current, the model uses a Muscarinic potassium current (KM) in order to achieve spike frequency adaption. The GENESIS implementation by David Beeman is based on the NEURON demonstration 'FLUCT' by Alain Destexhe, as described by Destexhe, et al. (2001). This uses channel models by Destexhe and Par (1999), which use modified Traub and Miles (1991) sodium and potassium conductances, and a muscarinic potassium 'M current' from Mainen and Sejnowski (1996). These are implemented in GENESIS 2 and 3 as tabchannels with a 'setupalpha' parameterized form to represent the channel rate parameters.
RScell-nolib2.ndf is an older equivalent version, without comments.
RScell-nolib.ndf is an older version, not suitable for networks.
VA_HHcell.ndf - The GENESIS implementation by David Beeman of the model cell that was used as a benchmark for neural simulators in the review by Brette et al. (2007). This is a dual exponential conductance version of the the Vogels and Abbott (2005) model with single compartment neurons having Hodgkin-Huxley dynamics. The fast sodium conductance Na_traub_mod and non-inactivating potassium conductance K_traub_mod are the same as those used in RScell. The GENESIS 2 version of the benchmark can be found on ModelDB and at http://www.genesis-sim.org/GENESIS/UGTD/Tutorials/networks/Vogels-Abbott_net.
BDK5cell2.ndf - The 'BDK5cell2' model is a branched layer 5 cortical pyramidal cell with 9 compartments and 9 voltage or calcium activated channels in the soma. This large variety of conductances makes the model a good candidate for improved fits to experiment using automated parameter searches.
The morphology and passive parameters are based on the reduced neocortical pyramidal cell models of Bush and Sejnowski (1993). These are simplified models of much larger (about 400 compartment) models of pyramidal cells from area 17 of cat visual cortex (Koch, Douglas and Wehmeier 1990; Bernander, et al. 1991). A collapsing method was used that conserves the axial resistance and makes some adjustments in the passive membrane parameters in order to faithfully reproduce the electrical responses of the larger models.
A detailed description of the channel models and the fitting of parameters that were used in the GENESIS implementation by David Beeman is given by Bernander, Douglas and Koch (1992) in the unpublished Caltech CNS Memo 16 (large PDF).
The implementation of the passive morphology used in BDK5cell2 has asymmetric compartments instead of the symmetric compartments of the Bush and Sejnowski model. This is done by setting the axial resistance to be the average of the original value and that of the 'child' compartment to which it is connected. The original values of the 'specific' passive parameters RM, RA, CM, which do not depend on the compartment dimensions, and the channel conductance densities are preserved by a transformation that alters the length and diameter of each compartment to keep the surface area the same and to yield the desired axial resistance.
This transformation of the cell morphology produces a slight shortening of the dendrites, but no detectable changes in the passive properties. The resulting cell has an input resistance of 47 Mohm and a membrane time constant of 10 msec, as does the original version with symmetric compartments. The total apical dendrite length is reduced from 150 um to 975 um, the length of the oblique dendrite from 150 um to 125 um, and the total length of basal dendrites from 200 um to 180 um in the asymmetric model. The gate tables are filled using 'setupalpha'. This version produces results consistent with the GENESIS 2 version.
The original version of this model can be found in the genesis-sim.org Models archive under 'corticalcells'.
deep_pyrcell.ndf - This represents a simplified deep pyramidal cell from the thalamorecipient layer of primary auditory cortex (AI). The model, implemented by David Beeman (GENESIS 2 version April 2010; G-3 version Sept 2012), uses a morphology based on the Bush and Sejnowski (1993) reduced 9 compartment layer 5 cat pyramidal cell from visual cortex that was used in BDK5cell2.
The voltage activated channels used here are a small set of modified Traub et al. (1991) hippocampal CA3 region channels with activation and inactivation time constants scaled to give dynamics typical of neocortical cells. Parameter searches were performed manually and with the GENESIS 2 parameter search library simulated annealing method (Vanier and Bower, 1999) to approximately fit current clamp results by Nowak et al. (2003) for regular spiking cells in cat visual cortex, and by Hefti and Smith (2000) for rat layer V primary auditory cortex.
The channels used are:
Na_pyr Fast sodium Kdr_pyr Potassium delayed rectifier Ca_hip_traub91 High threshold calcium channel Kahp_pyr Potassium AHP channel dependent on Ca concentration Ca_conc concentration element to convert Ca current to [Ca]
The firng patterns under current clamp conditiions show firing rates and spike frequency adaptation typical of neocortical deep pyramidal cells. However, as with most simple models involving only a subset of the channels that have been found in cortical neurons, the latency to the first spike after current injection is shorter than that seen experimentally.
baskcell.ndf - 'baskcell' is a very simple 'ad hoc' model of a fast spiking inhibitory neocortical interneuron, such as a basket cell, implemented by David Beeman (August 2008). The 30 um soma contains fast sodium and delayed rectifier potassium conductances, and an inhibitory GABA synaptic conductance. The single 200 x 2 um dendritic cylinder contains an excitatory AMPA synaptic conductance. The input resistance of the cell is 113 Mohm, with a membrane time constant of 10 msec. The sodium and potassium conductances were derived from the the Destexhe and Par (1999) model that was used in RScell, using parameters obtained with parameter searches using the GENESIS 2 'param' library. These produced rough agreement with spike frequency vs. current injection measurements of fast spiking cells in cat visual cortex by Nowak, et al. (2003).
The test scripts for the models described above are in the directory ~/neurospaces_project/gshell/source/snapshots/0/tests/scripts/
They are executed within that directory using the command:
genesis-g3 <script-name>
All of them produce output to a file '/tmp/output'.
test-squidcell.g3 - Performs a 'ndf_load' of squidcell.ndf and plots the Na activation and inactication variables (state_m and state_h) and the K activation (state_n) during a steady current injection of 100 nA.
The following tests are based on the 'two-cells.g3' script that is described in the tutorial Modeling Synaptic Connections and Large Networks with G-3. In each case, cell 1 has a steady injection current, and soma action potentials generate spikegen events that are passed to cell 2 via a synaptic connection to the synchan Ex_channel. The connection between the cells uses a very large (30 msec) propagation delay in order to easily see the effect. This simple circuit is created by loading the cell model into a namespace with 'ndf_namespace_load', creating copies for the two cells, and creating a connection within a 'projection' element, as described in the tutorial. If the simulation is running correctly, the second cell will begin firing slightly after 30 msec of run time. The tutorial version uses RScell.ndf.
two-baskcells.g3 - test of baskcell.ndf
two-BDK5cells.g3 - test of BDK5cell2.ndf
two-pyrcells.g3 - test of deep_pyrcell.ndf
two-simplecells.g3 - test of simplecell.ndf
two-VAcells.g3 - test of VA_HHcell.ndf
two-cells1.g3 - test of RScell-nolib2.ndf
traub91-nolib.ndf - The traub91 model is a burst-firing CA3 region hippocampal pyramidal cell, using a linear arrangement of 19 compartments containing active conductances in all compartments. The original GENESIS 2 model is distributed with GENESIS 2.3. It is based upon the paper by Traub et al. (1991).
traub94-nolib.ndf - A burst-firing hippocampal pyramidal cell using 64 asymmetric compartments in a branched geometry, containing active conductances in all compartments. The channels are very similar to the ones used in the traub91 model. The model is based on the paper by Traub et al. (1994) and was implemented by Pulin Sampat (Brandeis University) and Patricio Huerta (MIT) with help from Dr. Roger Traub. A fuller description is given in the genesis-sim.org Models archive.
traub95-nolib.ndf - A fast spiking hippocampal interneuron, using 51 branched asymmetric compartments containing active conductances in all compartments. The model is based in the paper by Traub and Miles (1995) and was implemented by Eliot Menschik. This, and the traub94 model pyramidal cell were used in the papers by Menschik and Finkel (1998, 1999). A fuller description is given in the genesis-sim.org Models archive.
The 'channels' directory contains the files 'Na_hh_tchan.ndf' and 'K_hh_tchan.ndf'. These represent the Hodgkin-Huxley squid fast Na and delayed rectifier K channels used in the simplecell model and many tutorial scripts. The files were generated with sli_load/ndf_save in the same manner as the cell NDF files, with a SLI script that created the two channels. As with the cells, all other elements except the channel to be saved were deleted before saving the channel to NDF.
Bernander O, Douglas RJ, Martiin AC, Koch C (1991) Synaptic background activity influences spatiotemporal integration in single pyramidal cells. Proc. Natl. Acad. Sci. 88:11569-11573.
Bernander O, Douglas RJ, Koch C (1992) A model of regular-firing cortical pyramidal neurons, CNS Memo 16, Caltech
Brette R, Rudolph M, Carnevale T, Hines M, Beeman D, Bower JM, Diesmann M, Morrison A, Goodman PH, Harris Jr FC, Zirpe M, Natschlager T, Pecevski D, Ermentrout B, Djurfeldt M, Lansner A, Rochel O, Vieville T, Muller E, Davison AP, El Boustani S, and Destexhe A (2007) Simulation of networks of spiking neurons: a review of tools and strategies. J. Comput. Neurosci. 23: 349-398.
Bush PC and Sejnowski TJ (1993) Reduced compartmental models of neocortical pyramidal cells, J. Neurosci. Methods 46:159-166.
Bush PC and Sejnowski TJ (1996) Inhibition synchronizes sparsely connected cortical neurons within and between columns in realistic network models, J. Comput. Neurosci. 3:91-110.
Destexhe A and Par D (1999) Impact of network activity on the integrative properties of neocortical pyramidal neurons in vivo. J. Neurophysiol. 81: 1531-1547.
Destexhe A, Rudolph M, Fellous JM and Sejnowski TJ (2001) Fluctuating synaptic conductances recreate in-vivo-like activity in neocortical neurons. Neuroscience 107: 13-24.
Hefti BH and Smith PH (2000) Anatomy, Physiology, and Synaptic Responses of Rat Layer V Auditory Cortical Cells and Effects of Intracellular GABAA Blockade. J. Neurophysiol. 83:2626-2638.
Koch C, Douglas RJ and Wehmeier U (1990) Visibility of synaptically induced conductance changes: theory and simulations of anatomically characterized cortical pyramidal cells. J. Neurosci. 10:1728-1744.
Mainen ZF and Sejnowski TJ (1996) Influence of dendritic structure on firing pattern in model neocortical neurons. Nature 382:363-366.
Menschik ED and Finkel L H (1998) Neuromodulatory control of hippocampal function: Towards a model of Alzheimer's disease. Artificial Intelligence in Medicine 13:99-121.
Menschik ED and Finkel LH (1999) Cholinergic neuromodulation and Alzheimer's disease: from single cells to network simulations. Progress in Brain Research, 121:19-45.
Nowak LG, Azouz R, Sanchez-Vives MV, Gray CM and McCormick DA (2003) Electrophysiological classes of cat primary visual cortical neurons in vivo as revealed by quantitative analyses. J. Neurophysiol. 89:1541-1566.
Traub RD and Miles R (1991) Neuronal Networks of the hippocampus. Cambridge University Press.
Traub RD, Wong RKS, Miles R, Michelson H. (1991) A Model of a CA3 Hippocampal Neuron Incorporating Voltage-Clamp Data on Intrinsic Conductances. Neurophysiol. 66:635-650.
Traub RD, Jefferys JG, Miles R, Whittington MA and Toth K (1994) A branching dendritic model of a rodent CA3 pyramidal neurone. J Physiol. 481:79-95.
Traub RD and Miles R (1995) Pyramidal cell-to-inhibitory cell spike transduction explicable by active dendritic conductances in inhibitory cell. J Comput Neurosci. 2:291-298.
Vanier MC and Bower JM (1999) A Comparative Survey of Automated Parameter-Search Methods for Compartmental Neural Models. J. Computat. Neurosci. 7:149-171.
Vogels TP and Abbott LF (2005) Signal propagation and logic gating in networks of integrate-and-fire neurons. J. Neurosci. 25: 10786-10795.