Very briefly, you can download here a Matlab code to simulate the neural activity in a single, 3-channel, cortico-striatal loop. The parameters used in this example have been hand tuned to replicate the dynamics described in Fiore et al. 2016,Changing pattern in the basal ganglia: motor switching under reduced dopaminergic drive. http://www.nature.com/articles/srep23327
As a start, you can easily change the main parameter regulating the simulated DA release in the main file (RUN_MOTOR_TEST.m)
|
1 2 |
%input and dopamine variability da_baseline = 0.4; % .2 -- .42 |
or you can manipulate, in the parameter file (motor_loop_par.m), the strength of the connections between the two GP parts, currently set to 0.29. See also the mentioned article for an in depth explanation of these phenomena.
|
1 2 3 4 5 6 7 8 |
%%optimised Sci Rep: genType1=[0.3 0.3 ...str2gpi, str2gpe (2 paths) 0.15 0.05 0.1 0.29 ...stn2gpi, stn2gpe gpe2stn gpe2gpi 0.1 0.1 0.12 0.00 ...gpi2thal thal2crx crxl1_2crxl2 crxl2_2crxl1 0.1 0.1 0.35 ...crx2thal crx2str crx2stn 0.2 0.2 0.35 0 ...baselines: strd1 strd2 stn(x2) gpi 0 0.15 0 0 ...baselines: gpe thal(x2) crx1 crx2 0.2 0.5 0.6 0.25]; ...epsilon and lambda, strd1, strd2 |
The selection dynamics are *very* sensitive to these parameters, if they are altered alone: an increase in the value of the gpe2gpi connection by 0.02 should be sufficient to recreate the oscillations described in Scientific Reports.
As a reminder, since the original libraries used for the paper have been written in C++, this is not an exact copy of the program used for the simulations published in Scientific Reports, though it is sufficiently close to replicate the results. Please also consider the input is different and the system is tuned for 3 channels rather than 4. Feel free to play with the parameters to see how the selection pattern/behaviour of the agent changes!
DOWLOAD >>> mean_field_BG_complete_loop
Here an example of the results you should see plotted using the basic parameters.
1 comment
Very nice! Thanks a bunch…