The signal intensity of the problem regions may be closer to that of
WM than typical cortical GM. Often, GM along the central sulcus is
difficult to separate from WM by eye. Because the segmentation in SPM
is pretty simple compared to the human visual system, it means that
these GM regions can be very difficult to segment accurately
(particularly with the not-so-great tissue probability priors used by
SPM).
Best regards,
-John
On 25 November 2011 10:02, Clara James <[log in to unmask]> wrote:
> Hallo,
>
> I perfomed VBM on 59 subjects in SPM8 using New Segment and DARTEL.
> Statistics (regression analyses according to 3 groups with different musical
> expertise: non musicians, amateur pianists and experts pianists) yielded
> plausible and nice results. However, I was surprised not to find differences
> in the motor areas. I then discovered that gray matter density values on top
> of the head are really low (mean around 0.2, for the rest of the head
>>0.7)..
>
> During the MPRAGE on a Siemens 3-T we corrected automatically for
> regularization bias: MPRAGE resolution > filter: prescan normalize
> activated. Furthermore, in New Segment I did a bias correction (very light
> regularisation (0.0001), 60mm cutoff), I saved the Bias Corrected, and
> compared the bias corrected images to the original, that improved slightly,
> and seemed to have relatively good overall intensities.
>
> I have 2 questions:
> - How to explain that the VBM values are so low on top of the brain
> - What is a reasonable relative mask value to use for the statistics
>
> Best, Clara James
>
> --
> Dr. Clara JAMES
> Faculty of Psychology and Educational Sciences
> Geneva Neuroscience Center
> University of Geneva
> Uni Mail, 40 Bd du Pont-d'Arve
> 1205 Geneva, Switzerland
> Tel. : 022-3799264
> Fax : 022-3799229
|