We are also using Philips 3T Achieva and my understanding is just like what
you described. I am using no-overplus-high and all gradient tables turn out
the same. I have confirmed this with Philips' clinical science people, too.
Just like the imaging gradients, the DTI gradients are also rotated to
accommodate the oblique acquisition. So, the same gradient table is applied
for all subjects in the imaging plane. That makes sense because you
generally choose a set of anatomical landmarks to align your DTI acquisition
so that you standardize acquisition orientation and minimize inter-subject
variance. Applying the same gradient tables in the imaging plane will also
help minimize the inter-subject variance.
The matlab software from KKI also produces the same tables for all subjects.
But I recently started using yes-overplus-high for a different study and the
tables turned out to be different for each subject. I wonder why they did
not use the same approach.
The gradient tables in the par file are in magnet coordinates, so they have
to be rotated based on the angulation and offsets described in the par file.
Tugan
On Tue, 5 May 2009 09:21:44 -0700, Wayne Su <[log in to unmask]> wrote:
>I am also confusing on this issue. According to Jonathan Farrell's website (
>http://godzilla.kennedykrieger.org/~jfarrell/software_web.htm), the DTI
>gradient tables are same for all subjects regardless of the slice
>angulations if using "No Overplus" option. He also said " The V4.1 par files
>list the diffusion weighting directions. HOWEVER, these directions are not
>in pixel space. Specifically, in the case of gradient overplus = YES
>tables, the directions need to be rotated and corrected for slice
>angulation." My scans are using "No Overplus" option, If I am going to use
>the gradient table listed in the PAR file, should I correct it for the FOV
>or not?
>
>So for the Philips 3T scanner, if using "No overplus" option, the gradient
>tables are same for all subjects because they have been corrected already on
>the scanner for the FOV. For my case, the bvec files used in FDT are same.
>Am I right?
>
>I haven't reached Jonathan Farrell. I would appreciate if anyone can confirm
>on this issue.
>
>Regards,
>
>Wayne
>
>
>On 05/05/09 7:03 AM, "Matt Glasser" <[log in to unmask]> wrote:
>
>> Well, regardless of what option you select in the scanner, the gradients
>> outputted by DICOM to NIFTI conversion programs like MRIConvert or MRICron's
>> converter may vary from subject to subject if you have rotated the FOV. The
>> gradients applied by the scanner are the same; they are just corrected for
>> these rotations.
>>
>> Peace,
>>
>> Matt.
>>
>> -----Original Message-----
>> From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
>> Of Yuzheng HU
>> Sent: Monday, May 04, 2009 11:36 PM
>> To: [log in to unmask]
>> Subject: Re: [FSL] gradient table for Philips 3T scanner
>>
>> Hi,
>> Thanks for answer.
>>
>> I was told that '... if all subjects use same protocal, the gradients sould
>> be
>> same. however,for Philips, if you select "Overplus" option during scanning
>> the
>> gradients will be changed dynamically accordin to the tilt-angles...'
>> The gradients I have might be in such cases.
>>
>> Best!
>>
>>
>> Yuzheng
|