T1 - ASL Coregistration problem

100 views
Skip to first unread message

Francesco Neri

unread,
Dec 16, 2020, 5:56:50 AM12/16/20
to ASLtbx discussion board
Hi,
I am Francesco Neri, psychologist and PhD student.
I am very new in analysing CBF signal and I have tried to optimize your pipeline to analyze ASL images from elderly patients with primary progressive aphasia.
I have some problems with my images after the preprocessing.

The scans were acquired with a Siemens 1.5T Avanto.

1: I have extracted my M0 from my PASL sequence and I got the raw PASL image with 90 slices + 1 image of the M0.

2: I put the raw T1 under the STRUC folder and PASL+M0 images under the FUNC folder.
I put all the parameters of my images and the preprocessing using batch_run gave me no errors

3: at the end I see that the ASL image (cbf_0_sASLflt_rPASL_4D) has not been coregistered with my T1 and I obtained a processed ASL image with the  same coordinate of the raw ASL image.
I tried to coregister my ASL image with the T1, but I got an image without the 4D, only with 1 "volume". And if I coregister my T1 into the ASL image space I got a very low resolution and shrinked image

Here I share the screenshot of the images of the processed CBF, my RAW_T1 and folder of my structural files and functional files I got at the end of the process.  

Have I done something wrong before the preprocessing? Or have I to do something after the process that I missed?

Thank for your precious help
Francesco
FUNC_images.PNG
cbf_0_sASLflt_rPASL_4D.PNG
STRUC_images.PNG
RAW_T1.PNG

Francesco Neri

unread,
Dec 17, 2020, 4:59:12 AM12/17/20
to ASLtbx discussion board
Hello again, 
sorry, I have to make a correction to my previous message. I took the "wmeanCBF_0_sASLflt_rPASL" image and I think this is the right image to look at, am I right?
Here I upload my output. 
One last question: is this image flipped L/R?
Thanks for your help.
Francesco

wmeanCBF_0_sASLflt_rPASL.PNG

Francesco Neri

unread,
Dec 18, 2020, 11:39:19 AM12/18/20
to ASLtbx discussion board
Hello again,
sorry for bothering you,
I share the image of my wm CBF maps overlaid with a grey mask. Is it ok that lot of the voxels have negative values? 
Thanks again
Francesco

cbf_sbj_1.PNG
cbf_sbj_1_with_negative_values.PNG

Francesco Neri

unread,
Dec 23, 2020, 11:49:04 AM12/23/20
to ASLtbx discussion board
Hello again and sorry for my questions.
I share a short presentation showing the output files that I got. From your experience, do you think that there is something wrong with my processing? Can you help me?
Thanks for your precious help.
Francesco

Siemens_Avanto_15.T.pptx

redhatw

unread,
Dec 25, 2020, 10:02:33 PM12/25/20
to asltbx-disc...@googlegroups.com, neri...@gmail.com
wCBF means spatially normalized CBF map. wcmeanCBF means the spatially normalized outlier cleaned mean CBF map. It is not LR flipped. PASL data acquired in 1.5T will be very challenging to get reliable results since ASL signal depends on arterial blood T1 which is much shorter in 1.5T than 3T. Also PASL has lower SNR than pCASL. Not sure how you can get 90 slices? Did you mean 90 images? Negative value in PASL is not surprising because the low SNR. It can exist in 3T pCASL data too when large motion occurs.

--
You received this message because you are subscribed to the Google Groups "ASLtbx discussion board" group.
To unsubscribe from this group and stop receiving emails from it, send an email to asltbx-discussion...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/asltbx-discussion-board/01bae5e6-6f40-4999-86f0-3dc927e9149bn%40googlegroups.com.


--
Ze Wang, PhD

redhatw

unread,
Dec 25, 2020, 10:05:32 PM12/25/20
to asltbx-disc...@googlegroups.com, neri...@gmail.com
Interestingly, your CBF map looks fine to me.  Remember to change the magnetic field option to be 1.5T. Also, you should run outlier cleaning. For old subjects, you may also need partial volume correction.
--
Ze Wang, PhD

Francesco Neri

unread,
Dec 30, 2020, 10:24:52 AM12/30/20
to ASLtbx discussion board
Thanks for your response, 
- yes, I mean 90 images (24 slices), sorry for my mistake.
- I run the outlier cleaning and the images look better.
- I will try to perform PVC
 thanks for all your advice.
Francesco

Francesco Neri

unread,
Dec 30, 2020, 11:54:03 AM12/30/20
to ASLtbx discussion board
I have some problems with the parameters set in par.m and asl_perf_subtract.m.
I don't know exactly how to set TI1,  LabelTime and DelayTime.
I can't find this in my .json file this information. 
Do the "1800" of my sequence name refer to TI2 and it is equivalent to the DelayTime? Thanks again for your precious help!
Francesco

redhatw

unread,
Jan 3, 2021, 8:33:41 PM1/3/21
to asltbx-disc...@googlegroups.com
TI2 = post labeling delay time here. 



--
Ze Wang, PhD

Reply all
Reply to author
Forward
0 new messages