All standard tensor analysis use a glm, they just don't often call it that. The design matrix are columns formed from the cross products of the gradients (of which there are 6) plus a column of 1s, and the log() of the input data is taken before the glm. After that, it's just a glm.

Siddharth Srivastava wrote:

Hi Douglas, Thanks for the reply. can i get a link to a publication or report that describes the GLM analysis applied to tensors? thanks, sid.

On Mon, Feb 9, 2009 at 10:22 AM, Douglas N Greve <greve@nmr.mgh.harvard.edu mailto:greve@nmr.mgh.harvard.edu> wrote:

It will probably be the case that it just gives the entire volume
(the "masking" routine just looks for 0-valued voxels). I don't
think it actually uses the mask for anything, though.

glmfit is run because that's what the tensor analysis is.

Siddharth Srivastava wrote:

    Hi everyone,
                      I have recently started using the dt_recon
    processing pipeline
    in freesurfer. I just ran a dataset through it, and had a few
    questions about
    the methodology
    1) How does the script calculate the mask from the DW data
    set. Is there any way
    to influence the mask calculation routine? I am familiar with
    bet/bet2 and i have
    an set of parameters tuned to my dataset. within dt_recon, i
    get the entire volume
    as a mask, while fsl generated a nice inclusive mask. Is there
    any way i can use
    the bet generated mask with dt_recon?
    2) why is a glm fit done at the end of the processing stage?

    best regards,
    sid.
    ------------------------------------------------------------------------

    _______________________________________________
    Freesurfer mailing list
    Freesurfer@nmr.mgh.harvard.edu
    <mailto:Freesurfer@nmr.mgh.harvard.edu>
    https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

---

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer