Kuramoto model in TVB

[Marta Arbizu Gomez]

Dear users,

I have recently started simulating with Kuramoto coupling and Kuramoto Oscillator local dynamic model. Nevertheless, I have tried varying parameters and the results I obtain are not the expected ones. In the picture you can see how the time series look like overall. Does anybody know why there's no fluctuation and how can I understand it?

Also, I have been searching for some papers where parameters with this model are explained, so that I have a starting point, but I haven't found anything useful. If anybody has information on this, and know how I can find it, it would be extremely useful.

Thank you in advanced.

Best regards, 

Marta

[Lia Domide]

Dear Marta,

Have you tried to zoom the signal?

Maybe use the other viewer https://www.thevirtualbrain.org/tvb/zwei/image-zoom/122446/17?stage=brainsimulator where it is easier to do zooming.

Maybe play with the models in the phase-plane page to understand when these are oscillating and when not https://www.thevirtualbrain.org/tvb/zwei/image-zoom/122362/35?stage=brainsimulator

Regarding papers. Have your tried the ones mentioned in the doc of the model https://github.com/the-virtual-brain/tvb-root/blob/master/tvb_library/tvb/simulator/models/oscillator.py#L398 ?

Best,

Lia.

[Marta Arbizu Gomez]

Dear Lia,

First of all, thank you for your response and suggestions.

I did try to zoom the signal with the scaling, but what I obtained didn't change (Picture 1). I also tried the other viewer, Time Series Visualizer SVG/d3, and at first what I obtained what the same, but when zooming I got a diagonal signal (Picture 2). This is why I don't see why I don't get oscillations in the Animated Time Series Visualizer, and also, don't know how to interpret the behavior of the model. I'm also going to add the parameters of the simulation cockpit (Pictures 3 and 4).

-Conduction speed: 10

-Coupling: kuramoto

-a: 5

-Local dynamic model: kuramoto oscillator

-w: 1

-Integration scheme: heun

-Integration step-size: 0,01220703125

-Sampling period: 0,9765625

-Simulation length: 10000,0

As for the phase plane, I did find it extremely useful when using other models. In this case, with the kuramoto oscillator model, there's just one model parameter and the phase plane is extremely simple, so it doesn't really help to adjust the value of the parameters (Picture 5).

Maybe I'm making a mistake and I can't see where, but I have tried different values in different simulations and don't know how to understand the results I am obtaining.

Regards, 

Marta

[DK]

Hi Marta,

The Kuramoto model is a phase model and does not "oscillate" in the sense that other neural mass models would. The output of the Kuramoto model is a steadily increasing phase (theta), which is what I believe you see in picture 2. However, the phases in that picture look a bit erratic. Try to run a simulation with the global coupling set to zero (all oscillators are isolated; or decrease coupling) and all regions should progress at their intrinsic frequency, you should see straight lines. I'm not sure if this is possible in the TVB GUI (I'm running TVB in Python) but usually people wrap this phase with the modulus 2*Pi, which allows you to see "fluctuations" (looks like sawtooth oscillations). You can also use sin(theta) to see oscillations with amplitude 1. Hope that helps a bit.

Best, Dominik

[Spase Petkoski]

Dear Marta

In the following collab you can find some more info about Kuramoto model

https://wiki.ebrains.eu/bin/view/Collabs/mouse-stroke-brain-network-model/Lab 

and in the drive there are some tutorials that could be helpful, such as:

https://lab.ch.ebrains.eu/hub/user-redirect/lab/tree/shared/TVB%20Mouse%20stroke%20Brain%20network%20model/KMstrokeAM-P1.ipynb

I hope this helps to begin with.

Best,

Spase

--
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/7da70354-861f-4cdd-898b-8ec5c9144a1dn%40googlegroups.com.

[Marta Arbizu Gomez]

Dear Spase,

Thank you for your answer. It was of great help. Has this mouse-stroke-brain-network-model resulted in any paper or any kind of source? I haven't found anything related and there are a few points where I have doubts. Sometimes having the code without further context can be messy, and questions pop up.

Best regards,

Marta

[Spase Petkoski]

Dear Marta,

Sorry if it wasn't clear enough, but it is stated that the collab should help to build the model used in 

https://doi.org/10.3389/fnsys.2020.00031 

Best,

Spase

To view this discussion on the web visit https://groups.google.com/d/msgid/tvb-users/e982bddf-bbe8-47df-8abe-0217bb937cf6n%40googlegroups.com.