Page 1 of 1

Unexpected 3D behaviour

Posted: Thu Jan 31, 2019 3:24 pm
by RBJ
Sorry. me again. I am confused by the output I get with 3D RXD. Figure I below shows how I get a nice as diffusion if I just add some factor to central nodes with diffusion . It seems to indicate to me I am on the right track.
However, now look what happens (2D scatter graphs) when I add a rate reaction.
The first graph here is with solve set to 1D.
There is a dear buildup of substance where it is being produced by the rate reaction. (Red is time zero, blue after 1 second).
If I simply change solve type to 3D, I expect to see the same build up of substance in the R.H.S.? see below, I do not.

There seem to be no parameters I can use, where species build up when produced by a rate or reaction. I am so sorry, any idea what I an doing wrong? code below:

Code: Select all

# -*- coding: utf-8 -*-
Created on Mon Jan 21 19:54:56 2019

@author: Richard

from neuron import h, rxd
from matplotlib import pyplot

# needed for standard run system

left = h.Section(name='left')

left.nseg= 11
left.L = 10
left.diam= 5

cytosol = rxd.Region(h.allsec(),dx=1)#change dx here.. default 0.25
ctfact = rxd.Species(cytosol, d=.01, initial=0.1)

def right_only(node):
    return 1 if node.x3d > left.L/2 else 0

production_region = rxd.Parameter(cytosol, initial=right_only)
reaction = rxd.Rate(ctfact, 1* production_region)
print('about to finalize')

def plot_it(color='b'):
    y = ctfact.nodes.concentration
    x = [node.x3d for node in ctfact.nodes]
    pyplot.scatter(x, y, marker='o', color=color, s=5)

# plot the initial situation

# we will plot every 25ms up to 100ms
pyplot.title('time {}s'.format(1))

Re: Unexpected 3D behaviour

Posted: Mon Feb 04, 2019 3:05 pm
by adamjhn
This is due to a bug in reactions in 3D rxd. Thank you for identifying it.
We have a fix for it, you can clone and build from source.

Alternatively replace your <neuron python library dir>/neuron/rxd/ with ...

E.g. on Windows, by default, this file would be
And on mac;

Re: Unexpected 3D behaviour

Posted: Mon Feb 04, 2019 4:19 pm
by RBJ
Oh that's perfect, thanks so much! I did try and debug but of course I got nowhere. Just copying in the new file worked fine-I didn't need to compile.
kind Regards R.