- -a1 analysisname1 : name of functional analysis -a2 analysisname2 : name of functional analysis -amerged analysisname : name of merged analysis
- -fsdmerged fsd -perrun analysis -jackknife analysis -space spacename : tal or sph; if sph, need -hemi -hemi hemisphere : lh or rh; with -space sph only -spacedir spacedir : if other than tal or sph
Session Arguments (Required)
- -sf sessidfile ... -df srchdirfile ... -s sessid ... -d srchdir ...
Other Arguments (Optional)
- -help : as in H-E-L-P, as in look here FIRST for answers
This will merge two analyses together. While the analyses must be in the same session, they do not need to be in the same functional subdirectory (FSD). The parameter estimates are concatenated. The merged noise variance estimate and covairance matrix is computed as the weighted average of the individual analyses. The mean offsets are simply averaged together. The number of non-null conditions in each analysis may vary from analysis to analysis. The number of estimates per condition must be the same across analyses. The number of non-null conditions in the merged analysis will equal the sum of the number of non-null conditions in the input analyses. The order of the conditions in the merged analysis is the same as if the conditions from the individual analyses were concatenated together.
If the user specifies spherical space (-space sph), then a hemisphere must also be specified using -hemi.
With few exceptions, the merged analysis can be treated as with any other analysis (ie, an analysis directory is created in the study folder and the session FSD, it can be operated on by mkcontrast-sess, stxgrinder-sess, sliceview-sess, func2tal-sess, etc).
If a new FSD is not specified (with -fsdmerged), then the FSD of the first analysis is used (or the single FSD found in the analysis named by -perrun or -jackknife).
The input analyses can be specified in one of three ways. First, by specifying them explicitly with -a1 and -a2 (in which case only two analyses can be merged). Second and third, with -perrun and -jackknife, in which case the argument of either of those flags is the analysis name, and the input analyses will be analysis-RRR (perrun) or analysis-jkRRR (jacknife).
If the analysis named by -amerge already exists, then the parameters within the analysis are checked to make sure that they agree with those specified on the command-line. If they do not agree, then the user is informed and the program exits. The user can then either delete or rename the analysis or change the argument of -amerge.