Hi Mingrui,

Thanks for this very useful tool! I am also trying to visualize subcortex activations (ventral striatum) using the ROI drawing function, with the hope to still remaining the activation heat map. Because Brainnet viewer detects different integers as different ROIs, I multiplied my original activation map by a number (N) and take the integer as the new activation value, in this case same ROIs detected by Brainnet are the voxels with the same activation level and N defines the number of activation levels. By inputting a colormap I am able to see the gradient of activation but the number N can not be over 9 in my case because then the size of ROI (voxels with the same activation level) will become too small with larger N and got stuck in the smooth process (the below code loop), and even I commented these codes it will not run because it can not get the right coordinates for the small ROIs. 

while size(fv{i,1}.vertices,1) == 0 %% Added by Mingrui Xia, expand ROI while it is too small
if EC.vol.roi.smooth_kernal == 1
vol = smooth3(vol);
else
vol = smooth3(vol,'gaussian');
end% Modified by Mingrui Xia 20160921, using Gaussian kernel instead of the 'box' kernel
vol(vol>0) = 1;
fv{i,1} = isosurface(vol);
end

Putting the image with N = 9 here: https://drive.google.com/file/d/10ZmG8nvj-41r9RQFYKNYc6pWjq8vh7Ml/view?usp=sharing, as you can see it's too pixellated.

I wonder if you think this subcortex activation visualization is possible and how you would think I can edit the script to make it work.

Any suggestion would be really helpful.

Thank you!

Best,
Xue