The conditions are as follows:
1 -- fixation
2 -- PP preferred
3 -- PP unpreferred
4 -- RRC preferred
5 -- RRC unpreferred
6 -- blocked RR pref (ignore)
7 -- blocked RR pref (ignore)
There are 4 fixation tokens and 10 tokens each for conditions 2 - 5. The 4 fixation tokens will each be 20 images. The number of images for each token in conditions 2 - 5 will vary between 16 and 18.There are 840 separate images ranging from 0-839.
For each subject we also have also have a <subjectnmuber>tsp1
file that contains a separate line for each of the 840 images.The first
column is the image number, the second column is a zero and is meaningless,
the third corresponds to the condition of the and the fourth is the image
number within a token.This file cannot be used as a split file in FIASCO.
However, a split file can be created by using the third column for the
condition number and adding a column of 1's as the image per token number.
.. 02882.LB.fs.list(ascii
list of relevant voxels for left Broca)
.. 02882.LT.fs.list(ascii
list of relevant voxels for left Temporal)
the data in the ".list"
files is indexed based on the first voxel being x=1,y=1,z=1.
The voxinfo directoy
also contains a subdirectory named mroi contains information about the
roi
LB, LT, RB, RT (subdirectories
for four ROIs)
.. .. .. 02882_??_Z0
to 02882_??_Z6(theseascii
files contain x and y
coordinates
of the ROI for each
slice
numbered 0 to 6)
The data in the "mroi"
files is indexed based on the first voxel being x=0,y=0,z=0.
The first two files
(02882removed.dat and 02930removed.dat) list the
images that are to
be ignored (usually we set condition = 0).Listed
is
the image number.It
is based on starting at 0 and going to 839.
Data Manipulation
Utilities: More help is available
if pff2pgm is run without any arguments.
The following command
generates data files for the lb area: The ".loc" file generated
by the program utilizes a co-ordinate system By default the cropping
file is assumed to utilize a co-ordinate system
To view the data
as images, we need to use pff2ppm. It converts data from a PFF format
file to a series of PGM images. pff2ppm takes as input the detrend.dat
file and a range of values for z and t and outputs a series of images for
the given z and t values. All z slices for a given time t can also be merged
in a single image. A sample run would be:
The following command
generates 10 ppm files with the base name "test" of t = 0-9, z=3:
pff2ppm /afs/cs/project/theo-35/fmri/02882_34/detrend.dat
test 0 9 3 3
To use the fMRI data
in any learning algorithm, we need to convert the pff file to an ascii
file. Pff2txt (also called voxelcropper)utility does exactly that.
Also it can take a list of voxels (x,y,z coordinates) in which weare interested
and generate an ascii file only containing information about those voxels.
The list of voxels is stored in the .fs.list files and the program generates
a .loc file which which contains the x,y,z locations of the cropped voxels
and the .hdr file which contains various human readable information about
the data
pff2txt -cf /afs/cs/project/theo35/fmri/02882_34/02882_voxinfo/02882.LB.fs.list
-tf /afs/cs/project/theo-35/fmri/02882_34/2882tsp1 /afs/cs/project/theo-35/fmri/02882_34/detrend.dat
/afs/cs/project/theo35/fmri/02882_34/lb_dat/02882_lb
where the first voxel
is x=0,y=0,z=0.
where the first voxel
is x=1,y=1,z=1.This can be changed
with the "-i0" option.
The convention for
naming the ascii files generated by pff2txt is the following. In each of
these directories there is one file for each of the trials, where each
line of the file is a set of 15 floating point numbers in ascii, separated
by spaces.An example filename is
as follows:
02882_lb_i525_c05_n17.031
02882
= subject number
lb=
brain areas
i525=
original image index (time) of the first image
in
this subsequence
c05=
experimental condition
n17=
number of images in this subsequence
.031=
index of the trial for this subject (begins with trial 000)
To use the program
pffvar:
Calculate
the mean and variance of the data in a pff file.
For
help, just type: pffvar:
The
following command generates "pff" files containing the mean, variance,
and
standard deviation of the entire sequence of data in the
"02882_34"
directory:
pffvar
-mf 02882_mean.pff -vf 02882_var.pff -sf 02882_stdev.pff /afs/cs/project/theo-35/fmri/02882_34/detrend.dat
------------------------------------------------------------------------------
To use the program
pffhist:
Calculate
a histogram of the data in a pff file.
For
help, just type: pffhist
The
following command generates a histogram for the file "02882_var.pff"
cotaining
variance data:
pffhist
-ts 0 -te 0 02882_var.pff > 02882_var.hist
------------------------------------------------------------------------------
Pretty voxel plotting
of LB MSE data - full brain version:
pff2ppm -sc 2 -pa
-1 -ds 10000 -cm 5 -pf lb-predictability.vox -zm 1 /afs/cs/project/theo-35/fmri/02882_34/stats/02882_mean.pff
lb-pred 0 0 0 6
rm -f lb-pred_t000.gif
ppmquant 256 lb-pred_t000.ppm
| ppmtogif > lb-pred_t000.gif
------------------------------------------------------------------------------
Pretty voxel plotting
of LB MSE data - cropped version:
pff2ppm -sc 4 -xs
35 -xe 55 -ys 10 -ye 30 -pa -1 -ds 10000 -cm 5 -pf lb-predictability.vox
-zm 1 /afs/cs/project/theo-35/fmri/02882_34/stats/02882_mean.pff lb-pred-crop
0 0 0 6
rm -f lb-pred-crop_t000.gif
ppmquant 256 lb-pred-crop_t000.ppm
| ppmtogif > lb-pred-crop_t000.gif
------------------------------------------------------------------------------
Pretty voxel plotting
of LB MSE data - full brain version - straight to gif:
rm -f lb-pred_t000.gif
pff2ppm -sc 2 -pa
-1 -ds 10000 -cm 5 -pf lb-predictability.vox -zm 1 /afs/cs/project/theo-35/fmri/02882_34/stats/02882_mean.pff
stdout 0 0 0 6 | ppmquant 256 | ppmtogif > lb-pred_t000.gif
------------------------------------------------------------------------------
Pretty voxel plotting
of LB MSE data - cropped version - staright to gif:
rm -f lb-pred-crop_t000.gif
pff2ppm -sc 4 -xs
35 -xe 55 -ys 10 -ye 30 -pa -1 -ds 10000 -cm 5 -pf lb-predictability.vox
-zm 1 /afs/cs/project/theo-35/fmri/02882_34/stats/02882_mean.pff stdout
0 0 0 6 | ppmquant 256 | ppmtogif > lb-pred-crop_t000.gif
------------------------------------------------------------------------------