[Mrtrix-discussion] Slice collapse problems

[Donald Tournier]

Hi Dorian,

Romain's pretty much covered all the bases already, but I just thought I'd
throw in my 2 cents. In my experience corruption isn't really specific to
any manufacturer, and is usually caused by subject motion. Either gross
motion, which is often associated with slice dropout due to spin history
effects, or pulsatile motion due to the cardiac cycle, particularly near
the 4th ventricle in front of the cerebellum. The bed vibration problem is
typically very reproducible, and causes problems in specific areas (mid-sag
parietal in our case), so doesn't really fit with what you're describing.
I've never personally had issues with gradient heating, so can't comment.
I've also very occasionally seen slices just drop out completely for no
apparent reason, which I can only attribute to some kind of hardware
problem (RF issues maybe?).

As to the Romain's mention of a fix for this in the next release of MRtrix,
I have to set the record straight: this is very much at the planning stage
right now! I'll be working on it, but there's nothing at this point.
Sorry...

Cheers,
Donald

--
Dr J-Donald Tournier (PhD)

Senior Lecturer, Biomedical Engineering
Division of Imaging Sciences & Biomedical Engineering
King's College London

A: Department of Perinatal Imaging & Health, 1st Floor South Wing, St
Thomas' Hospital, London. SE1 7EH
T: +44 (0)20 7188 7118 ext 53613
W:
http://www.kcl.ac.uk/medicine/research/divisions/imaging/departments/biomedengineering

On 6 Nov 2013 14:54, "Dorian P." <alb.net at gmail.com> wrote:

> Hi Romaine,
>
> Our scanner is Philips Achieva 3T. Similar to your explanation, I was told
> vibrations can be a cause, and I was pointed to a paper by Gallichan et al.
> on HBM.
> Subject motion should also play a role, otherwise I don't explain why some
> subjects don't have bad slices.
>
> The fourier transform is a new reason I never knew about. Is there a way
> to change fourier parameters in the scanner?
>
> It would be great to have mrtrix fix or remove bad slices. But... I think
> this needs concurrent motion correction. In fact, if motion correction is
> done in other software before bringing the data to mrtrix, bad slices will
> contaminate nearby slices. At that point instead of having one bad slice we
> would have two or more. I had this experience with regular DTI, where bad
> slices I flagged before realignment were spread in adjacent slices making
> them more blurry after ralignment.
>
> Despite this, do you know when is the expected release of mrtrix with
> slice correction?
>
> Thank you.
>
> Dorian
> TJU
>
>
> 2013/11/6 romain valabregue <romain.valabregue at upmc.fr>
>
>>  Hello
>>
>> Yes signal dropouts is a common feature of Siemens dti acquisition. but
>> it is not clear what is the exact cause. I think there may be different
>> cause
>> _ subject motion (when there is a lot of subject motion then I see
>> several dropout on the same volume
>> _ bed vibration a know problem on TRIO 3T (I heard they had a solution
>> but we could not convince them to install it ...)
>> _ gradient heating
>> _ partial fourier : Jesper on fsl list : " My personal experience is for
>> example that 5/8 is to severe and that 3/4 is better for reducing drop-out."
>>
>> For the gradient heating you can try to increase the TR (take 1 or 2
>> second abouve  the minimal TR ). The bvec order may also play a role : make
>> sure you the directions where the absolute amplitude on a given axis (it is
>> often the case)
>>
>> in the next mrtrix release there will be a correction for those 'bad'
>> slice
>>
>> Cheers
>>
>> Romain
>>
>> Le 05/11/2013 15:13, Dorian P. a écrit :
>>
>> Dear MRtrix list,
>>
>>  My HARDI sequence is frequently contaminated by slices with collapsed
>> signal. This happens in 2-3 slices of 61 volumes. Slices and volumes
>> affected are a bit random, not always the same.
>>
>>  In concomitance I see shifts in anterior-posterior (phase) direction
>> from one volume to the other. I have been told this may happen because
>> gradients heat to much.
>>
>>  My question at this point is:anybody had the same experience with slice
>> dropouts?
>>
>>  Do you think slice dropouts are related to the b-value. Would
>> decreasing b from 3000 to 2500 make a difference? Could slices collapse
>> because of motion causing bigger artifacts with high b-values?
>>
>>  Finally, do you think the slice collapse is related to the
>> anterior-posterior shift (gradient heating) problem?
>>
>>  Any comment may be useful.
>> Thank you.
>>
>>  Dorian
>> TJU
>>
>>
>> _______________________________________________
>> Mrtrix-discussion mailing listMrtrix-discussion at www.nitrc.orghttp://www.nitrc.org/mailman/listinfo/mrtrix-discussion
>>
>>
>>
>> _______________________________________________
>> Mrtrix-discussion mailing list
>> Mrtrix-discussion at www.nitrc.org
>> http://www.nitrc.org/mailman/listinfo/mrtrix-discussion
>>
>>
>
> _______________________________________________
> Mrtrix-discussion mailing list
> Mrtrix-discussion at www.nitrc.org
> http://www.nitrc.org/mailman/listinfo/mrtrix-discussion
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.nitrc.org/pipermail/mrtrix-discussion/attachments/20131106/fcec48db/attachment-0001.html>