GitHub - ModelDBRepository/150245: Sensorimotor cortex reinforcement learning of 2-joint virtual arm reaching (Neymotin et al. 2013)

ModelDBRepository / 150245 Public

Notifications You must be signed in to change notification settings
Fork 3
Star 0

Sensorimotor cortex reinforcement learning of 2-joint virtual arm reaching (Neymotin et al. 2013)

modeldb.science/150245

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
arm.hoc		arm.hoc
basestdp.hoc		basestdp.hoc
col.hoc		col.hoc
colors.hoc		colors.hoc
declist.hoc		declist.hoc
decmat.hoc		decmat.hoc
decnqs.hoc		decnqs.hoc
decvec.hoc		decvec.hoc
default.hoc		default.hoc
drline.hoc		drline.hoc
drspk.mod		drspk.mod
filtutils.hoc		filtutils.hoc
geom.hoc		geom.hoc
grvec.hoc		grvec.hoc
hinton.hoc		hinton.hoc
infot.hoc		infot.hoc
infot.mod		infot.mod
init.hoc		init.hoc
intf6_.mod		intf6_.mod
labels.hoc		labels.hoc
misc.h		misc.h
misc.mod		misc.mod
mosinit.hoc		mosinit.hoc
network.hoc		network.hoc
nload.hoc		nload.hoc
nqs.hoc		nqs.hoc
nqsnet.hoc		nqsnet.hoc
nrnoc.hoc		nrnoc.hoc
nstim.mod		nstim.mod
params.hoc		params.hoc
python.hoc		python.hoc
pywrap.hoc		pywrap.hoc
readme.html		readme.html
run.hoc		run.hoc
samutils.hoc		samutils.hoc
screenshot.png		screenshot.png
sense.hoc		sense.hoc
setup.hoc		setup.hoc
simctrl.hoc		simctrl.hoc
stats.hoc		stats.hoc
stats.mod		stats.mod
stim.hoc		stim.hoc
syncode.hoc		syncode.hoc
trainedplast.nqs		trainedplast.nqs
units.hoc		units.hoc
vecst.mod		vecst.mod
xgetargs.hoc		xgetargs.hoc

Repository files navigation

<html><pre>
This simulation was used in an article in press at Neural Computation
Citation: Neymotin SA, Chadderdon GL, Kerr CC, Francis JT, Lytton WW
(2013).  Reinforcement learning of 2-joint virtual arm reaching in a
computer model of sensorimotor cortex.

This simulation was tested/developed on LINUX systems, test on Mac OS,
but doesn't run yet on Microsoft Windows.

To run, you will need the NEURON simulator (available at
http://www.neuron.yale.edu)

Unzip the contents of zip file to a new directory.

Compile the mod files from the command line with (Linux):
 nrnivmodl

This will produce an architecture-dependent folder with a script
called special.  On 64 bit systems the folder is x86_64. To run the
simulation from the command line:
 nrniv
then NEURON will start and load the mechanisms (cell types, etc.)
then from the NEURON prompt:

 load_file("mosinit.hoc")

That will load the simulation and all required files. Network and
inputs will be setup. A graphical user interface will be displayed
with the option to run either the trained network or a naive network.
The middle panel will display the motion of the virtual arm, as
controlled by the network. After the simulation is complete, the
trajectory will be drawn in the middle panel (similar to Figure 6A)
and the arm will be positioned over the target (black 'x'). The right
panel will display a raster plot (similar to Figure 3A or B).

The trained network run should look like this:

<img src="./screenshot.png" alt="screenshot">

For questions/comments on the simulation contact:
 Sam Neymotin (email: samn at neurosim dot downstate dot edu)

20130628 A values.h include statement in misc.h was commented out so
that the simulation would run on Mac OS X.

20160921 Includes updates from the Lytton lab that allow the model to
run on mac OS X.

20220517 Updated MOD files to contain valid C++ and be compatible
with the upcoming versions 8.2 and 9.0 of NEURON. Updated to use
post ~2011 signature of mcell_ran4_init function and fix hashseed2
argument.

20230228 Fix to ensure printed strings are still valid, avoiding junk
being printed in https://github.com/neuronsimulator/nrn-modeldb-ci

20230420 Fixes for compatibility with new data structures in the
upcoming version 9.0 of NEURON.
</pre></html>