Hi,
I've been trying to extract de individual (subject-value) Maximum mfALFF Peak Value from a Two Sample T-Test mfALFF map extracted by REST, but I haven't found de way to do it. I think that using SPM Pet it is posible, but my T-Test maps are been obtained usinf REST.
Particularly the problem is that I can't extract the values from every individual mfALFF Map because I have introduced 3 covariates in the comparison between groups (and therefore the two Sample T Test map in wich I want to extract the values is been modulated by these covariates).
Does anyone known if it's posible to obtain these values without using directly SPM? (this would force me to repeat the analysis...)
Thanks,
Andriu
I've been trying to extract de individual (subject-value) Maximum mfALFF Peak Value from a Two Sample T-Test mfALFF map extracted by REST, but I haven't found de way to do it. I think that using SPM Pet it is posible, but my T-Test maps are been obtained usinf REST.
Particularly the problem is that I can't extract the values from every individual mfALFF Map because I have introduced 3 covariates in the comparison between groups (and therefore the two Sample T Test map in wich I want to extract the values is been modulated by these covariates).
Does anyone known if it's posible to obtain these values without using directly SPM? (this would force me to repeat the analysis...)
Thanks,
Andriu
Forums
Re: Individual PEAK value
1. You need to firstly find the peak voxel. Using either SPM/xjview on your T map, you can get the information for the peak voxel. Or you can use REST Slice Viewer, and click "Cl. Report", it will call xjview to generate the cluster report, then you can find the coordinates of the peak voxel.
2. Use REST->Utilities->Extract ROI signal to extract the mfALFF value. Define ROI as the coordinates you obtained in Step 1, and set the radius as small as possible to get an ROI with only one voxel (e.g., radius = 1.5 for voxel size of 3*3*3).
3. You have got the individual mfALFF values through Step 2. If you want to get the fitted value (regress out the 3 covariates), I will suggest you to perform a regression analysis through MATLAB's regstats (or regress, the differene is if the constant need to be modelled) on the signal you obtained. And then Remove the fitted response of the 3 covariates from the signal, e.g., s - b3*c3 - b4*c4 - b5*c5.
Best,
Chao-Gan