Dear FSL Experts,
I am going through feat2_part1.pdf. Could you please clarify the following? Your help is greatly appreciated. Thank you.
1. p.5 Why some of the graphs are Trustworthy while others are dodgy? When I compared the two graphs in the first column, I thought the top figure is trustworthy because the data have more variance. However, when I looked at the two graphs in the 2nd column, the situation is the opposite. For the graphs in the 3rd column, I assume that the bottom one is dodgy because there are only a few data points (possible repeated ones stacking on top of each other)?
2. p. 22 "An example experiment": What is the star behind t8 = 2.64 in Item 3?
3. p. 30 "Family-wise error". Why do we want to threshold the data so that only once (do you mean one?) in 20 studies do we find a voxel above this threshold? Is there anything to do with 1/20 = 0.05?
4. p. 32 "And if we reject anything, we will definitely reject the most "extreme" value (max(z)) in the brain." Could you please explain why?
5. p. 38 "This is the distribution we want to use for our FWE control."
a. Why?
b. Which distribution? The one with a smooth curve?
c. Actually what are the two graphs on the left?
d. Why there is a red cross on top of a graph?
6. p. 42 Could you please explain the two curves on the left?
b. Why are they shifted to produce the other?
c. How?
d. What is the "Now" stands for here?
7. p. 47 At the bottom there is a link to an url. I went there but did not see any tutorial. Which one are you referring to?
8. p. 49 What is EC(R)?
9. p. 51 "GRF for voxel-wise tests"
a. Could you please explain what is mean by "Hinges on strict assumptions about the distribution of the data (~N(0,^))"?
b. What is ~N(0,^)?
c. Could you please explain why "Is too lenient when volume is small (SVC)?" What is SVC?
10. p. 67 "Distribution of Max Cluster Size"
a. Could you please let me know how you got the number 76?
b. Why this is the level we want to threshold at?
Thank you.
|