Hi Mimi,
The outlier file is not a graph of motion, although it can be useful in identifying timepoints with excessive motion. It is a measure of the number of voxels containing an outlier value at each successive timepoint. A timepoint can become an outlier because of motion, a scanner malfunction, or anything else that affects a large number of voxels at a small number of timepoints.
Go to Latest AFNI News [afni.nimh.nih.gov]
and scroll down to August 16, 2001 for a more thorough description. Also helpful is the documentation to 3dToutcount [afni.nimh.nih.gov]
The other graph that you mention is the output of 3dvolreg. This output gives six parameters that specify how each successive image is rotated and translated in order to make it line up with the designated base image. Sources of information about this are:
[afni.nimh.nih.gov] [afni.nimh.nih.gov]
I'm not in a position to say what constitutes excessive motion - I hope you'll hear from others about that. In part, it has to do with how badly you need that particular set of data, and how successfully you're able to separate the motion artifact from the true signal in the analysis.
Hope this helps.
Sally