markallenthornton/matlabtfce

Standalone MATLAB implementation of permutation TFCE correction

matlabtfce's Issues

false positive rate detail

This is a technical detail with (probably) no repercursions but I am wondering if:

sum(fp)/nsim % false positive rate
in the test scripts, should actually be:

sum(fp>0)/nsim % false positive rate
Sorry for the nitpicking - just trying to clear up some confusion I was having.

Thanks

output images from two-tail test of one-way t-tests (matlab_TFCE)

Hello,

I have a question regarding the image output when using the function "matlab_TFCE" with analysis "oneway". As I understand from the documentation, there should be two images as output (one for the negative tail and one for the positive tail):
% If tails == 2, two such output images will be returned, one for the
% 'positive' tail and one for the 'negative' tail of the test,
% respectively.
However, when I run the function, only one output image is produced. Can you help me with this? Thank you very much.

Georgette

independent t-test variance normalisation for group 2

Hi Mark,

thanks for your code. Just a quick thought, in matlab_tfce_ttest_independent.m line 42:

truestat = (mean(imgs1,4)-mean(imgs2,4))./sqrt(var(imgs1,0,4)/nsub1+var(imgs2,0,4)/nsub1);

shouldn't this be

truestat = (mean(imgs1,4)-mean(imgs2,4))./sqrt(var(imgs1,0,4)/nsub1+var(imgs2,0,4)/nsub2);

this same line is correct when you implement it in the permutation test (line 77):

rstats = (mean(rimgs1,2)-mean(rimgs2,2))./sqrt(var(rimgs1,0,2)/nsub1+var(rimgs2,0,2)/nsub2);

ian

Pearson/Spearman correlation

Hi - thank you for your contribution. I have been using the correlation option in this, and it seems to work well and gives me results consistent with other methods.

In my data I am not really expecting a linear relationship between voxel activity and the parameters I am correlating.

I am thinking that swapping out the Pearson correlations for Spearman correlations should be fine - I was mainly concerned that the following Fisher transformation would somehow be inappropriate (although a quick google search suggests that it is:
https://onlinecourses.science.psu.edu/stat509/book/export/html/157).

Essentially, I was just looking for a second opinion on this in case there is something I am missing - seems like a simple change for an increase in robustness.

Thank you for your time.

EDIT: OK, I see that using Pearson is maybe 25x faster than Spearman (on my data).