FATCAT: fat_mvm_prep.py

AFNI Message Board
1

Hey there,

So I was trying to use the command: fat_mvm_prep.py from the FATCAT packages.

and I noticed that the input options are:
fat_mvm_prep.py -p PREFIX -c CSV_FILE \
{ -m MATR_FILES | -l LIST}

but my question is, if I input a a list (as one of the arguments), will the fat_mvm_prep command automaticaLly match each of the subjects in my list to the matching subject name in the csv file? What if the subjects are not listed in the same order as the csv file? Is there a way where I can match the subject name of the list to the subject name of the csv file?

Also how does this command handle missing information about a subject in the csv file?

Thanks for your help.

2

Hi, Sondos-

There are two ways of providing the CSV list file and matrix files. You can either:

  1. just list those after “-c” and “-m”, respectively, and the program will try to match the subject IDs in the CSV file with the matrix files (-> by looking for subj IDs uniquely in the files paths of the *grid files).
  2. provide the CSV file with “-c” and then provide a text file after “-l” that has two columns: first the matrix files (with paths) and then the matched subject ID; these must be matched row-by-row.

From the help file, for the latter:


-l, --list_match=LIST      :another way of inputting the matrix
                                  (*.grid or *.netcc) files-- by explicit
                                  path, matched per file with a CSV
                                  subject ID.
                                  The LIST text file contains two columns:
                                  col 1: path to subject matrix file.
                                  col 2: CSV IDs,
                                  (first line can be a '#'-commented one.

Re. missing data:
Missing data is not handled in the MVM modeling. Columns with missing data (“NA” values) should be removed (e.g., making a copy of the CSV file without those columns) or not selected for the analysis. If you include a subject who has missing data in one of the variables selected for the model, that subject is removed from the analysis.

–pt

3

Thanks for the clarification, Paul.