How to convert fwe_pvalue from fixel space into voxel space?

Hello @jdtournier ,
I am using [ fixel2voxel fwe_pvalue_fixel_space mean pvalue_voxel_space ] command to convert fwe_pvalue from fixel space into voxel space. However, the p-values are stored in 1-pvalue format in fixel space, and we lower threshold to 0.95.

My question is: how can I threshold pvalue in voxel space? I am using 0.05 as the threshold for statistical significant right now.

Many Thanks,
Suren

Hi @suren,

I don’t think there’s a single answer here, it depends on what you’re trying to do, and why.

  • If you’re after the p-value of the least significant fixel, you can use:

    fixel2voxel fwe_1mpvalue.mif min pvalue_voxel_space.mif
    
  • If you’re after the p-value of the most significant fixel, you can use the same command with the max operation

  • if you’re after the number of significant fixels in that voxel, you could use (assuming a p<0.05 threshold for significance):

    mrthreshold fwe_1mpvalue.mif -abs 0.95 - | fixel2voxel - sum num_sig_fixels_in_voxel.mif
    
  • if you just want to know whether any fixels where significant in each voxel, you could use the same as above, but additionally threshold the output:

    mrthreshold fwe_1mpvalue.mif -abs 0.95 - | fixel2voxel - sum - | mrthreshold - -abs 1 any_sig_fixels_in_voxel.mif
    

There are probably other combinations that may also make sense depending on exactly what you’re after.

But this all depends on why you would need to do this?

Cheers,
Donald.