Deletions are marked like this. | Additions are marked like this. |
Line 46: | Line 46: |
'''isxavg-re''' is a program for intersubject averaging using a random-effects model (see ["isxavg-fe"] for fixed effects model). It actually performs statistical analysis as well. It works by computing one number for each subject (at each voxel) based on the contrast matrix. This yields a set of numbers (at each voxel), one for each subject. The average and standard deviation of the set are computed and used to compute a t-statistic the siginficance of H_0: t=0$ is computed using N-1 degrees of freedom where N is the number of subjects. When jackknifing is turned on, a new set of N numbers is computed by successively averaging N-1 subjects while excluding a different subject each time. This reduces the effects of small sample sizes on the correctness of the siginficance levels. Jackknifing does increase the time to process data. It is highly recommended and the default in '"isxavg-re"'. Note that '''isxavg-re''' can be run on any data set, not just the output of ["selxavg"]. |
Index TableOfContents
Name
isxavg-re-sess - computes random effects analysis for group data
Synopsis
isxavg-re-sess -analysis (analysisname) -contrast (contrastname) -group (gropuname) -space (spacename) -sf (subjectfilename)
Arguments
Positional Arguments
none |
Required Flagged Arguments
-analysis analysisname |
session-level functional analysis name |
-group groupname |
name of group |
-space spacename |
space in which to average (native, tal, sph, roidef) |
-spacedir spacedirname |
space in which to average (eg, spm-smooth) |
-contrast contrastname |
contrast name |
-sf sessidfile ... |
|
-df srchdirfile ... |
|
-s sessid ... |
|
-d srchdir ... |
Optional Flagged Arguments
-hemi hemisphere |
with sph space <lh rh> |
-pctsigch |
use percent signal change |
-nojackknife |
do not use jackknifing |
-trunc sign |
truncation (pos or neg; neg = set neg vals to 0) |
-scriptonly |
don't run, just generate a script |
-version |
print version and exit |
Outputs
Description
isxavg-re is a program for intersubject averaging using a random-effects model (see ["isxavg-fe"] for fixed effects model). It actually performs statistical analysis as well. It works by computing one number for each subject (at each voxel) based on the contrast matrix. This yields a set of numbers (at each voxel), one for each subject. The average and standard deviation of the set are computed and used to compute a t-statistic the siginficance of H_0: t=0$ is computed using N-1 degrees of freedom where N is the number of subjects. When jackknifing is turned on, a new set of N numbers is computed by successively averaging N-1 subjects while excluding a different subject each time. This reduces the effects of small sample sizes on the correctness of the siginficance levels. Jackknifing does increase the time to process data. It is highly recommended and the default in '"isxavg-re"'. Note that isxavg-re can be run on any data set, not just the output of ["selxavg"].
Examples
Example 1
Example 2
Bugs
None
See Also
["paint-sess"], ["surf-sess"]
Links
Methods Description
References
["References/Lastname###"]
Reporting Bugs
Report bugs to <analysis-bugs@nmr.mgh.harvard.edu>
Author/s
Doug Greve, PhD