Dear Eric and All,

I have several questions regarding pestica estimators which are needed to know which independent components are related the noise (by comparing the correlation) I wnat to know whether these maps are acquired in 3 tesla, 4 tesla, 1.5 tesla ??

I assume that it should be 3 tesla, because the original article of PESTICA (Adaptive cyclic physiologic noise modeling and correction in functional MRI) is done in 3 tesla. but actually in that artcile you didnt mention directly that those estimation maps are acquired in 3 tesla. (since the article is mostly about the IRF part)

Now my question regarding these maps are, if they are acquired from 3 tesla. Is'nt there anything wrong to you those maps for y experiment (4 tesla), or, for making it more general, for other typs like 7 tesla?

My second question is: what kind of pre-processing you did before getting those estimation maps (in particuler did you do motion correction?) Actually, based on the original article you did motion correction after Physiological noise correction. But in the tutorial you said that its better using Pestica after motion correction?! 

I expect these maps would be different if you get them after motion correction or before that. and Now my question is in what stage you got those estimators (estimated maps) ?

Hi Pouya,

These maps were created using 3 Tesla data, but should work for 1.5 and 4T field strengths.  I don't know yet if they'll work with 7T data, but should know in a few months (we're getting a 7T this summer, and should have data by August)!  The spatial pattern is of most importance, and this should be largely preserved across field strength, with the caveat that there is some dependence of the effect size of physiologic noise on field strength and a dependence of warping on field strength.  7T maps with GRAPPA R=2 should result in close enough warping to 3T non-GRAPPA EPI data but I will have to check later this summer.

In the original articles I ran PESTICA without any pre-processing.  However, if there is more than negligible motion, the performance of PESTICA will start to degrade.  I don't have a clear threshold, but to-date I've found that its safest to just run motion correction on all subjects before running PESTICA.  I don't run any other pre-processing besides motion correction.
Erik