Vierling-Claassen et al 2010 model in mswin, network view

Discussions of particular models.

Moderator: tom_morse

Post Reply
tom_morse
Posts: 40
Joined: Wed May 18, 2005 10:23 pm
Location: Yale University School of Medicine
Contact:

Vierling-Claassen et al 2010 model in mswin, network view

Post by tom_morse » Mon Jan 28, 2013 11:19 am

________________________________________
From: liu kefei [lkefei@ion.ac.cn]
Sent: Thursday, January 24, 2013 4:05 AM
To: Morse, Thomas
Subject: ModelDB:

Dear Mr. or Miss

I am wrting to ask a question about auto-run. I saw this in your website:
Some models can be run directly from the web page. For these models there will be a button labeled "Auto-launch".
but for this model:
http://senselab.med.yale.edu/ModelDB/Sh ... del=141273
when I run it directly from the web, it said that mosinit.hoc was missed. Error was also produced when I downloaded it
and run it in my computer:

nrniv: FDSExp2Syn is not a template
in lightgamma_network_DB.hoc near line 27
s = new FDSExp2Syn(0.5)
^
xopen("lightgamma_network_DB.hoc" )
execute1("{xopen("lightgamma_network_DB.hoc")}" )
load_file("lightgamma_network_DB.hoc" )
xopen("lightgamma_init_DB.hoc" )
and others
0
nrniv: syntax error
in lightgamma_network_DB.hoc near line 28
s.tau1=0.5
^
xopen("lightgamma_init_DB.hoc" )
execute1("{xopen("lightgamma_init_DB.hoc")}" )
load_file("lightgamma_init_DB.hoc" )
xopen("light_batch_FSdrive.hoc" )
and others
0
nrniv: syntax error

Is the problem of the code or the file was indeed missed?
Thank you for your time. Best wishes!!


________________________________
Ke-fei Liu (刘克飞)
Laboratory of Visual Neurophysiology
Institute of Neuroscience
Shanghai Institute for Biological Science
Chinese Academy of Science

发件人: Morse, Thomas<mailto:tom.morse@yale.edu>
发送时间: 2013-01-24 23:32
收件人: lkefei<mailto:lkefei@ion.ac.cn>
主题: RE: ModelDB:
Dear Liu Kefei,

The mosinit.hoc is not included because the model is too time intensive for a demonstration run. The error that you received was due to not having the mod files compiled before you started the simulation.

to compile the mod files use

(mswin platform) start mknrndll and change directory to the Vierling-Claassen_lightgamma_DB_9-2011 folder. Choose to make the nrnmech.dll

(mac os x platform) drag and drop the Vierling-Claassen_lightgamma_DB_9-2011 folder onto the mknrndll icon

(unix/linux platform) Type the following two commands at the shell prompt (the first one assumes you are in the folder in which you expanded the zip file)
cd Vierling-Claassen_lightgamma_DB_9-2011
nrnivmodl

Once the mod files are compiled start the simulation and it should run OK.

Please send any more questions,
Tom Morse
ModelDB Adminsitrator

________________________________________
From: liu kefei [lkefei@ion.ac.cn]
Sent: Thursday, January 24, 2013 10:18 PM
To: Morse, Thomas
Subject: 回复: RE: ModelDB:

Dear Morse

Thank you for your reply.
I have compiled the code. I apologize that I haven't provide the detailed information.
After I downloaded the zip file. I unziped it, and run 'mknrndll' to generate a dll file called ' nrnmech.dll '. It said
'nrnmech.dll was built successfully. Press Return key to exit'.
So this may not be caused by the dll.
Then I open 'nrngui' and load 'light_batch_FSdrive.hoc', then the error info appear:
nrniv: FDSExp2Syn is not a template
in lightgamma_network_DB.hoc near line 27
s = new FDSExp2Syn(0.5)
^
xopen("lightgamma_network_DB.hoc" )
execute1("{xopen("lightgamma_network_DB.hoc")}" )
load_file("lightgamma_network_DB.hoc" )
xopen("lightgamma_init_DB.hoc" )
and others
0
nrniv: syntax error
in lightgamma_network_DB.hoc near line 28
s.tau1=0.5
^
xopen("lightgamma_init_DB.hoc" )
execute1("{xopen("lightgamma_init_DB.hoc")}" )
load_file("lightgamma_init_DB.hoc" )
xopen("light_batch_FSdrive.hoc" )
and others
0
nrniv: syntax error
same error appear when I load the dll file first, or double click the hoc file.
My operating system was windows 7 64bit.




________________________________
Ke-fei Liu (刘克飞)
Laboratory of Visual Neurophysiology
Institute of Neuroscience
Shanghai Institute for Biological Science
Chinese Academy of Science
________________________________
Ke-fei Liu (刘克飞)
Laboratory of Visual Neurophysiology
Institute of Neuroscience
Shanghai Institute for Biological Science
Chinese Academy of Science

发件人: Morse, Thomas<mailto:tom.morse@yale.edu>
发送时间: 2013-01-25 11:59
收件人: lkefei<mailto:lkefei@ion.ac.cn>
主题: RE: RE: ModelDB:
I don't have mswin at home right now however I think the following will work for you:

Although created the nrnmech.dll is not being loaded. After starting the nrngui, use the NEURON main menu
File -> Change working dir
and then select the folder Vierling-Claassen_lightgamma_DB_9-2011
(that will load the nrnmech.dll and you should see a list of mechanism files appear in the oc> prompt window).

finally load the light_batch_FSdrive.hoc

See if this works
________________________________________
From: liu kefei [lkefei@ion.ac.cn]
Sent: Thursday, January 24, 2013 11:09 PM
To: Morse, Thomas
Subject: 回复: RE: ModelDB:

The dll file was load successfully, i think. After loading it respond:
loading membrane mechanisms from nrnmech.dll
Additional mechanisms from files
Gfluct.mod ar.mod ca.mod cad.mod cat.mod fdsexp2syn.mod gnetstim.mod kca.mod km.mod kv.mod na_2.mod
There's no error here. Then I load the hoc file, then the error appear and the neuron program closed (at some time it didn't close).

________________________________________
From: Morse, Thomas
Sent: Thursday, January 24, 2013 11:24 PM
To: lkefei
Subject: RE: RE: ModelDB:

I wil try it tomorrow on a windows system at work and let you know what happens. I have run it previously on a linux system so you could try it on that if you have one available.
发件人: Morse, Thomas<mailto:tom.morse@yale.edu>
发送时间: 2013-01-26 03:29
收件人: lkefei<mailto:lkefei@ion.ac.cn>
抄送: Hines, Michael<mailto:michael.hines@yale.edu>
主题: RE: RE: ModelDB:
Ke-fei Liu,

There might be a mknrndll bug in the compilation of this models mod files under mswin NEURON 7.2.
It is possible that when you tried it earlier and the mod files compiled successfully it was because the folder above the proper folder was selected. When the proper folder is selected the error message was (attached):

...
rebase -b 0x64000000 -v nrnmech.dll
make: *** [nrnmech.dll] Segmentation fault (core dumped)
make: *** Deleting file `nrnmech.dll'

There was an error in the process of creating nrnmech.dll
Press Return key to exit
...

One reason why it may have been hard to compile the mod files in the proper folder (with mknrndll) is that the top-level folder name was so long that it was not visible in the mknrndll file browser (perhaps another bug?).
I am going to shorten the folder name from Vierling-Claassen_lightgamma_DB_9-2011 to Vierling-ClaassenEtAl2010
I ask Michael to debug the mswin mknrndll compilation of this model when he gets a chance:

http://senselab.med.yale.edu/ModelDB/Sh ... del=141273

I am glad you got it working under linux in the meantime.
-Tom
Tom Morse
ModelDB Admistrator

Michael Hines quickly provided this message mentioning that an alpha version of NEURON does not have the above problem:

From: Hines, Michael
Date: 2013-01-26 03:41
To: Morse, Thomas; lkefei
Subject: RE: RE: ModelDB:
When I use the alpha version
http://www.neuron.yale.edu/ftp/neuron/v ... -setup.exe
on a Windows 7 OS
I do not see the problem and the latter part of the mknrndll output reads:

rebase -b 0x64000000 -v nrnmech.dll
/cygdrive/c/Users/Guest/Desktop/Vierling-Claassen_lightgamma_DB_9-2011/nrnmech.dll: new base = 64000000, new size = 30000

nrnmech.dll was built successfully.
Press Return key to exit

________________________________________
From: liu kefei [lkefei@ion.ac.cn]
Sent: Saturday, January 26, 2013 3:42 AM
To: Morse, Thomas
Subject: 回复: RE: ModelDB:

Morse,
Sorry to bother you again.
Well when I use tools-model view to look at the structure of the network, I just saw isolated cells.
Could you tell me how to view the connection of the network? I am new to this and haven't found a tutorial mention that.

________________________________
Ke-fei Liu (刘克飞)
Laboratory of Visual Neurophysiology
Institute of Neuroscience
Shanghai Institute for Biological Science
Chinese Academy of Science



Ke-fei Liu,

The modelview tool does not reveal the network connectivity. Their are diverse ways to create a network and a program that analyzes a model after it is instantiated may present data that is harder to understand than the underlying (algorithmic) structure of network models. Currently the best way to understand it is to get it from the paper (and then also looking at the code). I notice there is a Network Architecture section of the paper and that the files lightgama_wiring_DB.hoc contains functions which connect the cells, and lightgamma_localconn_DB.hoc sets up the network by calling these functions.
Also the rconnect proc in lightgamma_network_DB.hoc is used.

Depending at your NEURON level you may find these tutorials helpful:
http://www.neuron.yale.edu/phpbb/viewto ... =15&t=2306
Best wishes,
Tom

Post Reply