This could well be due to poorer SNR in your data, with the inference tending to revert to the prior mean value in WM. For that scenario I could see an argument for using a less informative prior to get a sense of whether there is information in the data to say something about WM ATT, but chances are that you will find you would have porr confidence (i.e. large variances) associated with the estimates.
Michael
On 13 Feb 2019, at 16:15, Sang-Young Kim <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Hello, Michael:
Thanks for your suggestion and explanation.
I have one more question regarding oxford_asl output.
In general, ATT values in WM is longer than those in GM. But the estimated mean of ATT value in GM is longer than WM in our data (all of our subjects, n=11). This might be due to low SNR of WM? Do you have a clear explanation for that?
Any insights would be greatly appreciated.
Best,
Sang-Young
On Feb 13, 2019, at 6:27 AM, Michael Chappell <[log in to unmask]<mailto:[log in to unmask]>> wrote:
I wouldn’t generally recommend changing the priors unless you have a good reason to do so. You might have justification for changing the priors in specific circumstances based on prior knowledge, but it is normally a good idea to play with them and make decisions based on the output
You effectively alter the mean of some of the priors via the options in BASIL (i.e. the mean of the ATT is set by the —bat option), but we haven’t specifically expose the settings of the variances.
If you are editing the course code, which is a direct way of altering them, you will then need to recompile fabber and make sure that oxford_asl is using that new version.
Michael
On 12 Feb 2019, at 17:56, Sang-Young Kim <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Dear FSL experts:
I’m working on analyzing multi-PLD ASL data using oxford_asl.
But I would like to change prior knowledge for variational Bayesian inference to see how prior knowledge affect the estimated parameters.
I have tried to change parameters (e.g, precision and mean values) defined in fabber source code (e.g., fwdmodel_asl_pvc.cc<http://fwdmodel_asl_pvc.cc/>),
but it seems like the results are the same even if I changed the values.
Could you please help me how to change prior knowledge?
Thanks in advance.
Best wishes,
Sang-Young Kim
________________________________
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
---
Michael Chappell MEng DPhil
W: http://www.ibme.ox.ac.uk/qubic
T: +44 1865 617657
Associate Professor, Institute of Biomedical Engineering, University of Oxford.
http://www.ibme.ox.ac.uk<http://www.ibme.ox.ac.uk/>
Director of Training, EPSRC-MRC CDT in Biomedical Imaging
http://www.onbicdt.ox.ac.uk<http://www.onbicdt.ox.ac.uk/>
Governing Body Fellow, Wolfson College, Oxford.
http://www.wolfson.ox.ac.uk<http://www.wolfson.ox.ac.uk/>
PhysiologyforEngineers.org<http://physiologyforengineers.org/>
NeuroimagingPrimers.org<http://neuroimagingprimers.org/>
________________________________
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
________________________________
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
---
Michael Chappell MEng DPhil
W: http://www.ibme.ox.ac.uk/qubic
T: +44 1865 617657
Associate Professor, Institute of Biomedical Engineering, University of Oxford.
http://www.ibme.ox.ac.uk
Director of Training, EPSRC-MRC CDT in Biomedical Imaging
http://www.onbicdt.ox.ac.uk
Governing Body Fellow, Wolfson College, Oxford.
http://www.wolfson.ox.ac.uk
PhysiologyforEngineers.org<http://PhysiologyforEngineers.org>
NeuroimagingPrimers.org<http://NeuroimagingPrimers.org>
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
