Sample dataset for the NIS package


This is a small dataset for a simple finger-tapping fMRI experiment.
There are 48 images in the set, alternating between "tap" and "no-tap"
conditions every four images.  The sample design file shows what
block, condition, and scan correspond to each image, which can be used
as input to nisanova.  The "files" file is simply the last column of
the design file, and is used by several programs as input.  The
PsyScope datafile for this experiment is called tap.data.  The images
are of dimension 64x64x16 (XYZ) with voxel sizes 3.75x3.75x3.8 (in mm)
and are in ANALYZE format.  The following list of command calls shows
how the commands can be used to analyze this data.  You can also step
through these commands (plus the generation of the design file) with
the Java graphical user interfaces provided with this package.

For information about the options that a command takes, give the -h
flag to the command (for example, type 'nisanova -h').  See the NIS
web page for more information about these programs.  NOTE: the images
must be obtained from the web page before running these commands.


First, nisair is run so that all the images will be lined up spatially
with the first image.  This call gives the nisair program a list of
images as input, a threshold for doing the alignment, smoothing
parameters, and specifies a reslice directory.

nisair -f airfilelist -t 800 -b 2 -B 2 -k -r air_images.3D

Next, nisanova computes an analysis of variance per the factors
specified in the design file.  This call gives nisanova a stem for the
output filename, a threshold for the anova, and factor names.

nisanova -v -s sample -t 800 -f sample.df BL CD SC >& sample.outlog

Now niscrit is used to produce critical F values and image thresholds
for various probabilities.  The program takes the probabilities as
input parameters and reads factors from the design file, which is
given as standard input.

niscrit -v -f sample.df -o critFs .01 .001 .0001

Next, the resulting threshold from niscrit is used in nisroi to find
regions of interest (ROIs) in the images produced by nisanova.  The -3
flag tells nisroi to compute the ROI analysis in three dimensions, the
-v flag specifies that it should be verbose in its output, and the
next three parameters are the significance threshold, minimum ROI size
(in pixels), and the filename stem of the file to process.

nisroi -3v 977 6 sample.CD_x_SC

Finally, nistime is used to generate a list of activations per ROI
across time.  The program takes as input a list of images as well as
the .plist (particle list) file generated by nisroi.

nistime -f files sample.CD_x_SC.roi.3D.plist


The program nismaps is used to produce a variance map for the set of
images.  It takes as input a list of images to process, the type of
map to create (variance in this case) and a stem for the output
filename, and the name of the directory in which to place the
resulting map.

nismaps -v -f files -r variance -d maps > variance.log

The program nisnorm can be used to normalize the images so as to make
sure that the average intensities in all the images is the same.  The
program takes an output stem, a threshold, and an output directory as
input and reads the names of the files from standard input.

nisnorm -s norm_ -t 1000 -f files -d normed