Hi Randy,

The default option is 0 for each input scan as no assumption on the MRI contrast and order of the input scans is made.  We strongly recommend users to use the value of 1 for both T2w and FLAIR scans and 0 for T1w scans. 

The mask is reducing false-positive lesions by applying intensity constraints, but it also indirectly leads to better lesion segmentation overall as the model parameters are better estimated. So in general it's hard to disentangle its effect.

Hope it helps,
Stefano


From:
Subject: Re: [Freesurfer] SAMSEG (Lesion Mask Pattern) ... ah, [0 0]
To: "freesurfer@nmr.mgh.harvard.edu" <freesurfer@nmr.mgh.harvard.edu>
Message-ID: <1991518757.1033.1648594478958@mail.yahoo.com>
Content-Type: text/plain; charset="utf-8"

        External Email - Use Caution       

Dear experts,?
I think have answered my own question, but it does raise one more.
>From the early lines in the reporting thread, I see that without "--lesion-mask-pattern" flag, the default appears to be [0 0].??
New question:? From an accuracy/precision standpoint, what do you think the impact would be on lesion detection?
Thanks again (and in advance)
-Randy


-----Original Message-----
From: daedalus39@aol.com
To: freesurfer@nmr.mgh.harvard.edu <freesurfer@nmr.mgh.harvard.edu>
Sent: Tue, Mar 29, 2022 4:24 pm
Subject: SAMSEG (Lesion Mask Pattern)

Dear experts,
I inadvertently omitted the lesion mask pattern [0 1] when running SAMSEG, and while it ran to completion, there were fewer voxels present in the lesion mask (both at thresh 0.3) when I re-ran SAMSEG using the LMP.? ?The SAMSEG run included a T1 and a co-registered T2-FLAIR image. I am curious as to what the software defaults to in the absence of the --lesion-mask-pattern information.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/freesurfer/attachments/20220329/7b75d64c/attachment-0001.html