Neuroimaging Data Processing/Processing/Steps/Discarding First Volumes
Discarding first volumes edit
The signal in first volume of an EPI sequence can be off-scale due to longitudinal magnetization not having reached steady-state yet. Therefore, they are often discarded and not used for analysis. A visual inspection of voxel time-series can give a good idea of which volumes are affected.
Keep in mind that discarding the first volumes will have an effect on the number of volumes in your scan as well as on the timing information in stimulus onset files and physiological parameter files.
3dTcat is AFNIs function to discard the first volumes. It basically copies the original data, leaving out specified volumes. A simple command could look similar to this:
3dTcat [options] INPUTFILE[2..$]
which in this case writes all but the first two volumes (0 and 1) of the original data (inputfile) into the outputfile. This step is typically used to copy the (interesting part) of the raw data into the results directory for further preprocessing, so that the original data remains untouched. Note that volume number and timing information might have to be adjusted accordingly.
3dTcat can also be used for linear detrending using the -rlt option. Check manual page  for more details and options.
In afni_proc.py this is a default step, however with the number of TRs to remove set to 0. The respective option to remove the first n TRs is