I want to use 3dNetCorr with the -push_thru_many_zeros option (due to several voxels outside of the brain). I just want to verify what this option does with the null time series.
Does it include voxels of zero activation in the average, or does it exclude them?
If the former, then is there a parameter to set that will calculate the average without null voxels?
3dNetCorr does not include any time series that are all zero (= “null time series”) in an ROI average.
To help warn people that their ROIs might be poorly aligned, or masking might be mismatched, or some other badness, the program will not run if an ROI has more than 10% of voxels that are all zero. The “-push_thru_many_zeros” option results in the program not stopping if more than 10% of voxels are all zero. It still does not include null time series in the ROI average.
That’s good news for me - thanks!