I've done simulations for an experiment to compare 2 different things, let's call them A and B. The rows of A and B correspond to different experiments. Their sizes are 100*200, so there are 100 experiments and 200 data points. These are plotted against 'trials', lets call it T, which is just 1:200. I was interested to see where along T are A and B significantly different. So, I wrote some code to determine at each instance T, are A and B significantly different.
Edit: I think what I'm looking for is "contour heat map" essentially.
That is:
for kk=1:200
[h,p]=ttest2(A(:,kk),B(:,kk),0.05);
pval(kk)=p;
hval(kk)=h; end
So now I have a vector p that gives me the probability that A and B are significantly different at a given instance T. What I would like to do is generate some kind of map (I guess a heat map would be approprite here? I'm open to ideas though) that shows how the p values change along the T dimension and the A/B dimension. I'd like to draw some kind of shape, or just have some type of indication, that specifically shows the region of p=X (for example, p=0.05). I want to make it obvious that anything touching or inside a given region is a p=X region.
I hope I articulated my question well. I've tried to generate this plot a few times now and it's surprisingly giving me a hard time. I guess I still need to get used to matlab more.
Any help would be appreciated.
