Hi everyone,
I have some doubts about which statistical test to apply in order to test the probability of eventA to happen at the same time of non-eventB.
Basically I have my fMRI group with subjects exposed to independent stimuli: A and B. I computed the Z map for A and the one for B. If I compare the activation maps overlapping them I see that some brain regions are activated by both the stimuli and some other brain regions are activated only by A. But this is only a visual comparison of the two maps. What I would like to obtain is a statistical map showing the onlyA activation (and not A>B or A
Brain activity often lights up regardless of stimuli, in other words, we often bias our perception of regional brain activity to be associated to the stimulus we apply.
I would suggest a difference transform, take the difference of each group A and B and form a single set of values, then perform a t-test to determine any significance does in fact exist within the difference. Your outliers will certainly widen the variance, hence, you may find it very difficult to find significance, but there may be no significance whatsoever, after-all.
Dear Gabriele, MATLAB has a number of fitting surfaces functions, well implemented and with high accuracy. Probably this is desirable to the classical statistical tests, as you are looking for performance, not only significance.
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping
- Maps