provide better options for multiple test correction

[ebrahimebrahim]

Bonferroni is overly conservative because voxels are not independent

We should look into FDR-based / random field methods (?)

[tao558]

Great summary on random field theory Two roadblocks:

• Most of the resources I can find online give the equations for z-scores. We’ll be using t and f statistics. The analogous equations exist, just need to find them. Second paragraph of 3.2 of the above gives some next steps
• The above mentions that we’ll need to somehow figure out the smoothness of our images. This is typically done using the residuals of the statistical analysis (last paragraph of 3.4) Above resource mentions SPM, might be good to look at that.