Hello all, I have a question.

I am doing a fALFF analysis in conn, and at the end of the 1st-level analysis, when calculating the fALFF, I see a checkbox " ▢Normalization".
What does this mean to do? General Z transformation (=(x-mean(x). /std(x)))?
I would like to know what it means to add this process.
Thank you in advance,

Toshiharu

Hi Toshiharu,

Yes, it's a general Z transformation but non-linear, making not only the mean and standard deviation of the resulting values 0 and 1, respectively, but also enforcing normality of the resulting distribution. In particular all of the 'normalization' options in voxel-to-voxel analyses will take your original connectivity values (e.g. fALFF values) separately for each subject and for each condition and apply the following transformation, which makes the distribution of those values follow a N(0,1) distribution across voxels:

[a,~,k] = unique(x);
y = spm_invNcdf(k/(numel(a)+1));

where x are your original values, k is the rank order of each voxel (with ties assigned the same rank), and y are the output values (placed at the corresponding rank order from a N(0,1) distribution)

Best
Alfonso

Originally posted by Tosiharu Kamishikiryo:

Can you please help me find where the ALFF matric be located in the project folder? I mean by what name it is saved?