A shell script was used to invoke the following commands on each case:
flirt -in avg152T1.hdr -ref Human_Buckner_Case1.hdr -omat mni152_to_buck1.mat -bins 256 -cost corratio -searchrx -90 90 -searchry -90 90 -searchrz -90 90 -dof 12 -interp trilinear convert_xfm -matonly -omat buck1_to_mni152.mat -inverse mni152_to_buck1.mat flirt -in Human_Buckner_Case1.hdr -applyxfm -init buck1_to_mni152.mat -out flirt_Human_Buckner_Case1 -paddingsize 0.0 -interp trilinear -ref avg152T1.hdr
Note that a left-on-left version of avg152T1.hdr was used (i.e., SPM99's template), rather than the right-on-left version in FSL's etc subdirectory. Likewise, the input volume and surface were both in SureFit/Caret/AFNI LPI orientation.
Because the VE lab used to map vtk files in "native" space, Bob Cox wrote Vecwarp to output coord files that way. So we have to shift the coord files from AC-origin to "full brain grid" (i.e., origin is left-most dataset corner, posterior-most dataset corner, inferior-most dataset corner); apply the FLIRT affine transform; and then shift the coords back to AC-origin. (Actually, since FLIRT's origin is the 0,0,0 voxel coordinate, which is also the 0,0,0 mm coordinate, and is the left-most, bottom-most coordinate, we needed to do this anyway. See this relevant FSL list message.)
1 0 0 88 0 1 0 123 0 0 1 74
1 0 0 -90 0 1 0 -126 0 0 1 -72
CLOSEDtopo_file Human.sphere_6.LEFT_HEM.73730.topo FIDUCIALcoord_file Human_Buck_Case1.L.F.RegToPALS_B12.LR.FIDUCIAL_FLIRT.73730.coord volume_anatomy_file flirt_Human_Buckner_Case1.hdr