Re: [TVB] Oscillation in ReducedWongWangExcInh model
Julie Courtiol
Le 30 mars 2021 à 09:36, Celien Iliaens <celien....@gmail.com> a écrit :
Hi,
I am trying to work with the ReducedWongWangEI model on the default data in the GUI of TVB (version 2.1a1 on Windows) but I can’t seem to get it to oscillate. The only way I get an oscillation is by adding noise (integration scheme= Stochastic Heun) but when I do that and run a Fourier Transform on my EEG output, it doesn’t look like it should (see attachment). Is it possible that my output is now showing purely noise? And if so, does anyone know how to get the model to oscillate without noise? (I tried to find an oscillation without noise by changing the parameters in the phase plane but without success)
Thanks in advance!
Celien
--
You received this message because you are subscribed to the Google Groups "TVB Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tvb-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tvb-users/fff7b81c-4ce9-429c-bbaf-e57034c278can%40googlegroups.com.
<FFT_EEG_with noise.PNG><EEG_without noise.PNG><EEG_with noise.PNG>
Daniele Marinazzo
To view this discussion on the web visit https://groups.google.com/d/msgid/tvb-users/8F7FE20B-8353-4345-BC95-94BAC6031824%40gmail.com.
WOODMAN Michael
hi,
What was meant is that oscillation or limit cycle attractors require two or more dimensions in ordinary differential equation systems. This is not the case in e.g. a delay differential equation, but TVB assumes the node local mass model is an ODE, with delays
entering through network connectivity.
In this case, from the figures shown, it's a good guess that the fluctuations seen in the stochastic simulation are not sustained oscillations. The model is likely set to a fixed point attractor. This can be seen in the deterministic simulation, namely a transient to flat line. This transient has a specific time scale which is then found in the stochastic simulation, as a series of noise driven transients. In any case, the horizontal axis of the FFT is unfortunate, since it almost certainly hides this characteristic time scale which would appear as a bump amid 1/f noise, probably on a slow time scales of 10s of Hz.
Lastly, this model consists of two dimensionals both with sigmoidal nonlinearities, so it should be possible to configure it to produce sustained oscillations, even without noise. The relevant phase plane portrait would have 4 quadrants with a rotating vector
field (E,I high: I decreasing; E high, I low: E,I decreasing; E,I low: I increasing; I high, E low: E increasing). The phase plane tool can help with this, changing parameters to move the nullclines around.
Sent: Monday, April 5, 2021 15:10
To: tvb-...@googlegroups.com
Subject: Re: [TVB] Oscillation in ReducedWongWangExcInh model
Celien Iliaens
Dear all,
Thank you for the replies!
I don’t really have a technical background so I’m not sure if I completely understand your explanation about the relevant phase plane. I did try changing all kinds of combinations of the parameters in the phase plane but nothing lead to any oscillations. Do you know if creating a phase plane with 4 quadrants and a rotating vector field is something that is achievable using the phase plane in the GUI? Or is this something more advanced for which I would need to use the scripting interface?
Kind regards,
Celien
WOODMAN Michael
Hi
I just checked in the GUI, and for this model, the parameter ranges are limited so you can't quite get to the oscillations through the phase plane tool. If you want to get into the scripting then it should be straightforward to do it.
cheers,
Marmaduke
Sent: Tuesday, April 6, 2021 4:33:49 PM
To: TVB Users