The following steps must be performed:
- Selection of regions of interest (ROI):
As a first step, locations for the regions of interest must be defined. ROIs are defined as spheres whose centers and radii must be supplied by the user. In general, the same ROIs should be used for all subjects and conditions. The centers of the ROIs may for instance be obtained by using local maxima of a zmap. - Creation of a ROI-mask:
A ROI-mask is created using 'vmaskedit'or 'vledit'. 'vledit' is an interactive drawing program. For more information go to its help page.In 'vmaskedit', the ROIs are defined as spheres that are given in an ASCII file containing a list of points (sphere centers). The format of this file is the same as that for the program 'vshowpts'.
Each line specifies one point and has the following format:
x y z id radius
Example:
10 20 35 1 2
10 23 24 1 2
18 33 17 2 3
This line produces two spheres of radius 2 and id '1', and one of radius 3 with id '2'. Spheres with identical id's are treated as one entity in subsequent analyses steps.
The dimension of the output image (its matrix size) can be specified using the parameters '-nslices', '-nrows', '-ncolumns'. If set to '0', a default value is used.
Example:
vmaskedit -in pts.txt -system talairach -resolution 3 -out roi.v
The output 'roi.v' is a file containing all ROIs specified in 'pts.txt'. The parameter '-resolution' defines the voxel-size in mm. In this example, voxels in the output-file are of size 3x3x3mm.
Hint: The resulting file 'roi.v' can be visualized as follows:
vconvert -in roi.v -repn float -map linear -out roi_1.v
vlview -in normbrain.v -zmap roi_1.v
For details of vconvert see via-help.
- Averaging:
Within each subject and blob, average contrasts are calculated with 'vmaskave'. This has to be done separately for the experimental conditions:vmaskave -in *_norm_a.v -mask roi.v -id 1 -report a1.stat
vmaskave -in *_norm_b.v -mask roi.v -id 0 -report b.statThe input of 'vmaskave' are normalized individual contrast images. Wildcards are allowed and recommended. '-id' determines which blob of the blobmask has to be analyzed. If '-id 0' is specified, all blobs are analyzed. In this example, the output files are 'a1.stat' and 'b.stat'. For each blob these ASCII-files contain the subject-wise mean contrasts.
- ANOVA:
Finally, the mean contrasts are analyzed utilizing SAS or SPSS.