fmriqa_phantomqa.pl:
  If --timeselect specified, only use those timepoints that are selected, otherwise exclude the first 2 or 3 timepoints.
  If --nopercfluc specified, don't do percent fluctuation analysis.
  If --noroi specified, don't do Weiskoff/ROI-based analysis.
  If --summaryonly specified, don't write out any images to disk.
  
  input data types:
    RAW: just the (selected) input data, unaltered.
    MASKED: the (selected) input data, by default masked using a NxN ROI (default 15x15, can be changed with --roisize) in the middle slice only.  --maskfile specifies an alternate mask.
    AFNI: the (selected) input data, passed through the AFNI pipeline '3dvolreg', '3dDetrend -polort 2' (i.e. motion corrected and polynomially detrended), and masked with a mask generated using '3dvolreg', '3dTstat -mean', '3dAutomask -q'.
    DILATEDMASK: the (selected) input data, with a mask as used in AFNIMASKDETREND, but with the additional 3dAutomask option of '-dilate 4'.

  metrics [by input data type]:
    (nave [Static spatial noise], ave [Signal], sd [Temporal fluctuation noise], sfnr) images, (std, percent fluc, driftfit, rdc) summary values, (raw/fitted signal, spectrum, Weiskoff/roi-based analysis) plots [MASKED]: as described in Friedman, Glover (2006) "Report on a Multicenter fMRI Quality Assurance Protocol", with different mask (see MASKED above).  Note that some of the metrics are performed on detrended data or residuals, as described in the article.
    FWHM [AFNI]: plots FWHM as calculated by AFNI's 3dFWHMx
    Ghosting [DILATEDMASK]: plots "ghostiness" using algorithm in Greve et al. (2011) "A novel method for quantifying scanner instability in fMRI", Magnetic Resonance in Medicine, 65(4), 1053-1061, http://dx.doi.org/10.1002/mrm.22691

fmriqa_generate.pl:
  If --timeselect specified, only use those timepoints that are selected.
  
  input data types:
    RAW: just the (selected) input data, unaltered.
    MASKED: the (selected) input data, masked used a threshold derived from finding local minimums of a fitted histogram on voxel intensities.
    DETRENDED: the (selected) input data, filtered with a 60sec-period high-pass filter, but adding back the mean.
    MASKED, DETRENDED: the DETRENDED data, masked with the mask described above

  metrics [by input data type]:
    Clipped [RAW]: check for data that may have overflowed the voxel data type
    Volume means [RAW]: plots mean intensity across each full volume
    Masked, detrended volume means [MASKED, DETRENDED]: plots mean intensity of detrended data across each masked volume
    Mean volume difference [RAW]: plots difference of volume mean intensity of each volume from that of mean volume
    Center of mass [RAW]: plots centroid of each full volume
    Masked, detrended center of mass [MASKED, DETRENDED]: plots centroid of detrended data across each masked volume
    Running difference [MASKED, DETRENDED]: plots difference between one volume and the next of mean intensity of detrended data, across each masked volume
    Spectrum [RAW]: plots frequency spectrum
    Outlier count [MASKED]: plots outlier count as reported by AFNI's 3dToutcount
    FWHM [MASKED]: plots full-width half-max as reported by AFNI's 3dFWHMx, using the -demed [de-median] option
    Mean, stddev, SFNR [MASKED]: as computed by fmriqa_phantomqa.pl, using mask file