External Email - Use Caution
Hello Freesurfer Team:
Because of ensuring that results are exactly the same from run to run, norandomness flag is used with recon-all.
To better understand the randomness, we have a few questions listed below.
1. Which algorithms inside the recon-all use the seeded random number? And what randomness is about? 2. Is there a way to use a seed value to ensure the same data can be reproduced? 3. Whether we are compromising anything by going with the non-random option?
Thank you very much for your time and looking forward to hearing from you.
Best regards Xin
Hi Xin,
In Freesurfer 7.3, recon-all is running with ‘-norandomness’ by default unless ‘-randomness’ is specified in the command line.
‘-randomness’ uses current time and date as the seed for random number generation. This results in slightly different surfaces each run.
‘-norandomness’ ensures consistency in surface creation. By default, all seed critical binaries will run with identical seeds ‘1234’, which can be changed with ‘-rng-seed <seed>’. Here is the list of binaries that recon-all passes the seed: mri_normalize mris_smooth mris_sphere mris_fix_topology mris_topo_fixer mris_curvature mris_ca_label mri_segstats
Best,
Yujing
From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu On Behalf Of Xin Qi Sent: Tuesday, November 15, 2022 8:50 PM To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] about using -norandomness flag with recon-all
External Email - Use Caution Hello Freesurfer Team:
Because of ensuring that results are exactly the same from run to run, norandomness flag is used with recon-all.
To better understand the randomness, we have a few questions listed below.
1. Which algorithms inside the recon-all use the seeded random number? And what randomness is about? 2. Is there a way to use a seed value to ensure the same data can be reproduced? 3. Whether we are compromising anything by going with the non-random option?
Thank you very much for your time and looking forward to hearing from you.
Best regards Xin
freesurfer@nmr.mgh.harvard.edu