 |
 |
Dear Anderson, Excuse me for delay. here it is the command I used to generate fwe corrected p-value: palm -i 'all_FA_skeletonised.nii' -d Design.mat -t Design.con -m 'mean_FA_skeleton_mask.nii' -n 500 -T -corrcon -demean -o PALM And I then looked up “PALM_tfce_tstat_fwep_c6.nii” As long as I needed just the pearson correlations I ran also: palm -i 'all_FA_skeletonised.nii' -d Design.mat -t Design.con -m 'mean_FA_skeleton_mask.nii' -n 500 -pearson -demean -o PALM And then looked up "PALM_vox_rstat_c6.nii”. Thanks so much for your helps. Bests Arman > Il giorno 15 mag 2018, alle ore 13:25, Anderson M. Winkler <[log in to unmask]> ha scritto: > > Hi Arman, > > Can you show the full command line of your two calls to PALM, and which output files for each run you are looking at? > > Thanks. > > All the best, > > Anderson > > > On 15 May 2018 at 03:56, Arman Bordbar <[log in to unmask] <mailto:[log in to unmask]>> wrote: > Dear Anderson, > > Thanks so much for your response. > > As a matter of fact, I also had tried with no -pearson option but I still obtained fwe corrected p value maps with no significant voxel. I have already plotted the relation between dependent and independent variables in some voxels and they are all in line with my hypothesis and I highly believe that there should be some my mistakes in the usage of palm. > > Bests > > Arman > > On Tue, May 15, 2018, 04:14 Anderson M. Winkler <[log in to unmask] <mailto:[log in to unmask]>> wrote: > Hi Arman, > > The -pearson option should probably be removed or, at least, not yield p-values. It isn't a pivotal statistic, and the results may not match the t-stats. It's ok to use the correlations saved, but for p-values, the default t-stat is ideal. > > All the best, > > Anderson > > On 9 May 2018 at 06:43, Arman Bordbar <[log in to unmask] <mailto:[log in to unmask]>> wrote: > Hi all, > > I ran a tbss analysis on my FA maps and obtained some significant results. Afterwards, I decided to calculate the Pearson coefficients using PALM. At the end of PALM analysis, all the tracts which had shown significance were disappeared and after fwe correction there is no voxel in the final static map with a p-value less than 0.05. My question is where I am making mistakes? > > Thanks in advance > > Bests > > Arman > > > On May 15, 2018 04:14, "Anderson M. Winkler" <[log in to unmask] <mailto:[log in to unmask]>> wrote: > Hi Arman, > > The -pearson option should probably be removed or, at least, not yield p-values. It isn't a pivotal statistic, and the results may not match the t-stats. It's ok to use the correlations saved, but for p-values, the default t-stat is ideal. > > All the best, > > Anderson > > On 9 May 2018 at 06:43, Arman Bordbar <[log in to unmask] <mailto:[log in to unmask]>> wrote: > Hi all, > > I ran a tbss analysis on my FA maps and obtained some significant results. Afterwards, I decided to calculate the Pearson coefficients using PALM. At the end of PALM analysis, all the tracts which had shown significance were disappeared and after fwe correction there is no voxel in the final static map with a p-value less than 0.05. My question is where I am making mistakes? > > Thanks in advance > > Bests > > Arman > > >