For quantitative volume measurements, you'll want to use the tool
referenced by Lars to fix the distortions produced by the non-
linearity in the scanner's gradients. FYI, the gradient non-linearity
distortions are significant on our Siemens Allegra with head-only
gradients. They are much less noticeable on a whole body scanner.
As pointed out by Dave, the B0 induced distortions small, but in my
experience not as small as one might hope, especially for high res
acquisitions with long readouts. These are possibly correctible with
a field map, although more likely needing an acquisition with reversed
gradients. All highly non-standard. That said, if you're at 1x1x1 mm
resolution and you don't use excessively low bandwiths/long readout
times, then you are probably okay to ignore this for most applications.
Cheers,
Souheil
---------------------------------
Souheil Inati, PhD
Research Associate Professor
Center for Neural Science and Department of Psychology
Chief Physicist, NYU Center for Brain Imaging
New York University
4 Washington Place, Room 809
New York, N.Y., 10003-6621
Office: (212) 998-3741
Fax: (212) 995-4011
Email: [log in to unmask]
On Apr 7, 2009, at 5:34 AM, Lars Tjelta Westlye wrote:
> Hi Matt,
>
> I'm not sure what kind of distortions you are thinking about, but
> you may
> want to check out this tool for correcting geometrical distortions
> due to
> gradient non-linearity, which at least works great with our avanto
> mprages
> (though the distortions are small). See ref below.
>
> http://www.nbirn.net/tools/gradient_non_linearity/index.shtm
>
> Cheers
> Lars
>
>
> Jorge Jovicich, Silvester Czanner; Douglas Greve; Elizabeth Haley;
> Andre
> van der Kouwe; Randy Gollub; David Kennedy; Franz Schmitt; Gregory
> Brown;
> James MacFall; Bruce Fischl; Anders Dale. Reliability in Multi-Site
> Structural MRI Studies: Effects of Gradient Non-linearity Correction
> on
> Phantom and Human Data, Neuroimage. 2006 Apr 1;30(2):436-43.
>
>
On Apr 7, 2009, at 5:03 AM, David Lythgoe wrote:
> Hi,
>
> Yes, it’s true that structural images have some distortion.
> Unfortunately, fieldmaps won’t help in this case, as this is caused
> by the difficulty in building MR gradients that are linear over
> large fields of view. Typically, gradients are linear near the
> centre of the magnet coordinate system, but the field due to the
> gradients falls off more rapidly towards the edges of the FOV.
>
> I believe Siemens & and I know that GE perform a correction for this
> gradient non-linearity. I don’t know the details for Siemens
> scanners, but on GE, the correction is only applied in-plane, so
> there is no correction in the slice direction. Philips, I believe
> claim their gradients are linear, so they don’t need to apply a
> correction.
>
> The effect can be quite large, with shifts on the order of a
> centimetre at the edge of the FOV in head images, and can vary with
> head position and angle. If you wish to look at longitudinal changes
> in structure, it’s probably worthwhile taking care to position
> subjects reproducibly in the scanner at each visit.
>
>
>
> Dave
>
>>> On Mon, Apr 6, 2009 at 8:26 PM, Matt Glasser <[log in to unmask]>
>>> wrote:
>>>> Hi,
>>>>
>>>> I was recently told that typical T1 images (e.g. MPRAGE) have some
>>>> distortion. Having previously thought that they are essentially
>>>> undistorted,
>>>> I am curious of the magnitude of any distortion in typical
>>>> structural
>>>> images,
>>>> and the mechanism behind it if it does indeed exist.
>>>>
>>>> Thanks,
>>>>
>>>> Matt.
|