Forwarding Stefano's answers...
-------- Forwarded Message -------- Subject: Re: [Freesurfer] SAMSEG longitudinal processing questions Date: Thu, 28 Jan 2021 16:45:39 +0000 From: Stefano Cerri stce@dtu.dk To: Douglas N. Greve dgreve@mgh.harvard.edu CC: Koen Van Leemput leempko@gmail.com
External Email - Use Caution
1) SAMSEG long should compute more consistent longitudinal structural volumes than SAMSEG while detecting better volumetric group differences (cf. Table 1 and Fig 3 of https://arxiv.org/pdf/2008.05117.pdf). Yes, using SBTIV for correcting volume changes from different scanners would be beneficial and most probably necessary.
2) The lesion version of SAMSEG long is available in the development version. Enabling the lesion extension can be done by adding the same flag(s) as for the cross-sectional lesion version (--lesion, see the Wiki for details: https://surfer.nmr.mgh.harvard.edu/fswiki/Samseg)
3) We never conducted extensive experiments for the number of burn-in and samples for the MCMC scheme, but we found that 50 burn-in and 50 samples lead to robust and replicable lesion segmentation results. Increasing these two numbers might increase the lesion segmentation performance, although differences might be minimal and the overall computational cost high. We suggest to not decrease these two numbers as lesion performance could potentially decrease.
Best, Stefano
------------------------------------------------------------------------ *From:* Douglas N. Greve dgreve@mgh.harvard.edu *Sent:* 28 January 2021 16:30 *To:* Koen Van Leemput; Stefano Cerri *Subject:* Fwd: [Freesurfer] SAMSEG longitudinal processing questions
-------- Forwarded Message -------- Subject: [Freesurfer] SAMSEG longitudinal processing questions Date: Thu, 28 Jan 2021 09:37:32 -0500 From: XY XY kamsonmd@gmail.com Reply-To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu To: freesurfer@nmr.mgh.harvard.edu
External Email - Use Caution Hi FreeSurfers,
Thank you for sharing the new SAMSEG tool, which I consider a major breakthrough in image processing! I have a few questions about using SAMSEG vs SAMSEG_Long:
1. What volumetric differences should I expect between running SAMSEG on individual images, vs. using SAMSEG_Long for the longitudinal analysis? Could I use SBTIV to adjust differences between scans? (all images are FLAIRs preprocessed using mrI_robust_template)
2. Is there currently a way to integrate lesion segmentation into the SAMSEG_long’s longitudinal analysis?
3. What is the expected impact of increasing or decreasing -Burn-In and -Sample numbers for lesion segmentation?
Thank you for all the help!
Best,
Dave Kamson
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Mass General Brigham Compliance HelpLine at http://www.massgeneralbrigham.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.