Niscorrect

Niscorrect takes a pgh txyz file (and an analyze image if baseline correction is selected) as input and performs one or more of several timeseries corrections to the pgh file.

Invocation

java Niscorrect

Niscorrect window

Information to be entered

• Input file/directory
  Enter the name of an ANALYZE filelist, the name of a directory with ANALYZE images, or the name of a pgh1.0 file in txyz order. For a pgh file, either .mri or .dat extension can be specified, but neither is necessary.
• Output file/directory
  For pgh format, enter the basename for all output files. Different suffixes will be added to this basename to create the output and auxiliary files. The main output files contain your results, auxiliary files contain intermediate or supplementary files that may be helpful in interpreting results. For ANALYZE input format, enter the name of an output directory; this name will also be used as the basename for auxiliary files, which will be created in the same directory where the output directory resides.
• main output
  • basename.{mri,dat}: Pgh1.0 txyz dataset. This is the corrected output.
• polynomial detrend / outlier correction auxiliary output
  • basename_reg_a.{mri,dat}: pgh1.0 txyz format, regressors used for detrend.
  • basename_dout_a.{mri,dat}: pgh1.0 txyz format, outlier corrected data used internally for fitting the polynomial.
  • basename_diff_m.{img.hdr}: ANALYZE image. This mask contains a measure of the magnitude of the adjustments made by the internal outlier correction for each voxel.
  • basename_dout_m.{img.hdr}: ANALYZE image. This mask contains a count of the number of outliers per timecourse for each voxel found in the internal outlier correction.
  • basename_out_m.{img.hdr}: ANALYZE image. This mask contains a count of the number of outliers per timecourse for each voxel.
  • basename_reg_m.{img.hdr}: ANALYZE image. This mask contains a measure of the magnitude of the y range of the regressor (eg for 1st order regressors the magnitude of the slope of the line) at each voxel.
• Blocking size(s)
  If a blocking size, b, is specified, each b timepoints in the timeseries will be adjusted separately. Typically, you would enter the number of images per pfile as b, so that each pfile is adjusted independently. You may also enter several blocking sizes (separated by spaces) corresponding to the number of images per pfile if your pfiles have a varying number of images per pfile. If there's a repeated pattern in the block sizes (such as 36 24 36 24 36 24), only the pattern needs to be entered (36 24).
• Threshold
  Vectors in which ANY element is below the given threshold will not be acted upon. Default is to operate on all vectors.
• Save auxiliary files
  Select this box to save the auxiliary output files created by the detrend and outlier correction steps.
• Perform baseline correction
  For each voxel (x,y,z) in the pgh1.0 input file, this option will take all t timepoints for that voxel, calculate their mean, and then adjust that mean by adding/subtracting to each voxel the difference between that mean and the corresponding (x,y,z) value in the analyze input file, so that the mean value of each voxel's timeseries matches the intensity value for that voxel given in the analyze image. Any voxels that are negative after the baseline correction is performed are set to zero, and the algorithm accounts for this when performing the baseline correction.
• Mean image
  ANALYZE image with the same xyz dimensions as the timeseries data.
• Perform polynomial detrend
  This option applies a polynomial detrend to each vector in the pgh1.0 txyz format file. Polynomial detrend code was taken from Doug Noll and Scott Peltier, with help from Tom Nichols.
• Polynomial order
  Select what order polynomial to fit to the timecourse. 1 = linear, 2 = 2nd order, etc.
• Internal outlier correction
  Outlier correction is applied to an internal copy of the timecourse, in order to get a better fit of the polynomial. The final output results will not have outlier correction. In this box, select how many standard deviations mark the boundary for outliers. Data points beyond this many standard deviations from the mean will be set to this many standard deviations from the mean.
• Perform outlier correction
  
• Outlier correction
  Outlier correction is applied to the timecourse. In this box, select how many standard deviations mark the boundary for outliers. Data points beyond this many standard deviations from the mean will be set to this many standard deviations from the mean.

• Menubar
• Toolbar
• Output window