ADNI3 ASL

[ExploreASL]

On 6 Jul 2020, at 20:30, Amaryllis Tsiknia <tsi...@usc.edu> wrote:

Hi Henk, 

Thank you for getting back to me. Your input was very helpful. I am currently working on an exploratory analysis of ADNI3 ASL data. This is my first time working with ASL data and ExploreASL was undoubtedly the most well-documented software tool I could find. 

I do have another question regarding the analysis of the M0 image. I am working with a Siemens 3D PASL sequence which I believe used the "GRAY-WHITE" background suppression mode. I inspected the ASL volumes, and they do not seem to contain a calibration volume where no background suppression was used, and there's an even number of 3D volumes arranged in control-label pairs. All Siemens 3D PASL scans from ADNI3 are accompanied by a Perfusion Weighted Image acquired from those same ASL scan parameters. Would it be appropriate to use that separate perfusion weighted image as an M0 instead of the mean_control ASL image as an M0 in my analysis? 

Any input would be greatly appreciated. 

Best wishes, 

Amaryllis 

[Henk-Jan Mutsaerts]

Dear Amaryllis,

Good to hear, no, the M0 and perfusion-weighted image are two different things! Hopefully my lecture will help getting you started a bit! You can find it as first link here (first link).

We’re working on finding a more exact solution, for the time being I would insert the following in your DataPar.json file:

“M0”: “UseControlAsM0”

“M0_GMScaleFactor”: 0.069

Then don’t specify M0PositionInASL4D

This will use the control image as M0, and correct for the expected signal loss due to background suppression.

You can check that you get appropriate GM CBF values!

BTW: A student of ours might also process ADNI ASL data, would you be interested in collaborating?

BW, Henk