help
help > RE: Problem with looking fdr corrected maps
Mar 14, 2016 06:03 PM | Alfonso Nieto-Castanon - Boston University
RE: Problem with looking fdr corrected maps
Hi Shounak,
A reasonable choice would be to display the correlation coefficient values but limited only to suprathreshold areas. You can do that, for example, using spm_imcalc and selecting both the corr*.img and the pFDR*.img files for a given subject/condition, and then entering an evaluation formula like "i1.*(i2<.05)" (i.e. this displays the values in i1 -correlation coefficients-, but it sets to 0 those values that do not pass the p-FDR<.05 threshold).
Hope this helps
Alfonso
Originally posted by shounak dasgupta:
A reasonable choice would be to display the correlation coefficient values but limited only to suprathreshold areas. You can do that, for example, using spm_imcalc and selecting both the corr*.img and the pFDR*.img files for a given subject/condition, and then entering an evaluation formula like "i1.*(i2<.05)" (i.e. this displays the values in i1 -correlation coefficients-, but it sets to 0 those values that do not pass the p-FDR<.05 threshold).
Hope this helps
Alfonso
Originally posted by shounak dasgupta:
Hi Alfonso,
One more thing- output of the spm_icalc uses the logic of 0 or 1. So the voxels of the pFDR corrected map goes through a threshold and those above it will have value 1 and others 0. The problem is the output image doesn't contain any information about how strongly the voxels are correlated ( no colour weighting if I visualize it in MRIcron). It shows a uniform colour all through out. Is there a way out to keep FDR corrected p values less than 0.05 as it is so that there is gradation of colour when I visualize it?
-Shounak
One more thing- output of the spm_icalc uses the logic of 0 or 1. So the voxels of the pFDR corrected map goes through a threshold and those above it will have value 1 and others 0. The problem is the output image doesn't contain any information about how strongly the voxels are correlated ( no colour weighting if I visualize it in MRIcron). It shows a uniform colour all through out. Is there a way out to keep FDR corrected p values less than 0.05 as it is so that there is gradation of colour when I visualize it?
-Shounak
Threaded View
| Title | Author | Date |
|---|---|---|
| shounak dasgupta | Mar 8, 2016 | |
| shounak dasgupta | Mar 11, 2016 | |
| Alfonso Nieto-Castanon | Mar 14, 2016 | |
| Alfonso Nieto-Castanon | Mar 10, 2016 | |
| shounak dasgupta | Mar 10, 2016 | |