Virtually all AFNI programs that read AFNI datasets will read .1D files. They interpret them as row index = spatial index, column index = sub-brick index.

If you need to swap rows and columns in a .1D file, then the utility program 1dtranspose will do that for you.

The types of input datasets allow are not documented with each program separately. All AFNI programs use the same set of functions to read datasets in, so they all take the same types of files as allowed inputs. At the current time coordinate (Stardate -1205.3), these types of files are:

.mnc = MINC files
.hdr = ANALYZE filepairs
.mri = CTF-formatted MRI files (CTF is a MEG company)
.svl = CTF-formatted SAM volumes
.1D = AFNI .1D files (columns of ASCII numbers)
.3D = columns of ASCII numbers with an XML header (cf. 3dAFNIto3D)
.nii = NIfTI-formatted files (a new format, to be announced soonish)
.mpg = MPEG-1/2 files (as of 04 Dec 2003)
.HEAD or .BRIK = Classical AFNI files -- cf. [afni.nimh.nih.gov]

If an input filename does not end in one of these suffixes, then AFNI will try to append .HEAD to it and read again. Thus, you can specify an AFNI dataset on the command line as prefix+tlrc (say), without the .HEAD. But if the file is some other type (e.g., .mnc), then you'd have to put the full filename on the command line. Also, AFNI can read gzip-compressed .BRIK.gz files, but won't read any other compressed formats (unless you count MPEG).

3dcalc is an exception -- it treats 1D files in a special way. If you want to do expression analysis on 1D files, the 1deval program is the tool to use. 3dcalc should read the rest of these file formats as AFNI datasets.

Now, if you are using 3dANOVA (for example), then all the input datasets need to have the same number of voxels in space -- if the inputs are .1D files, this means they need to have the same number of rows.

You can specify column (sub-brick) selectors on .1D input files the same as for AFNI datasets, using the 'zork.1D[index]' notation. You can also use row (voxel index) selectors on .1D files, using a notation like 'zork{a..b}' to extract rows #a to #b. You can combine these two notations to analyze only certain rows and columns, as in 'zork.1D[3]{100..1000}'.

AFNI programs output AFNI formatted datasets. You can convert these to some of the other formats using the various 3dAFNIto* programs. In the future, AFNI will support direct output of NIfTI formatted datasets. This is a new format that SPM and FSL will support as well (a variant on ANALYZE). Don't ask when.

Hope this helps -- bob cox