----- Forwarded Message From: Bruce Fischl fischl@nmr.mgh.harvard.edu Date: Wed, 19 Jul 2006 08:42:08 -0400 (EDT) To: Mike Chee mchee@pacific.net.sg
Hi Mike
A. the wm parcellation isn't released yet, but I think we'll do so soon (Doug: what do you think). The current tools should allow you to correlate demographic/neuropsych with parcellation units (e.g. superior frontal gyrus).
B. The improved bias correction should help both the cortical and subcortical streams.
C. The control points range of effect depends on the other control points (auto and manual) near it). If a control point is anterior to all others it can have an effect on the entire frontal region.
you should post these to the list - other people can answer some of them (like A) better than I can.
cheers, Bruce
On Wed, 19 Jul 2006, Mike Chee wrote:
Hi Bruce,
I have a couple of questions: A. We are interested in correlating parcellated brain volumes of
brain lobes
(not subcortical structures for openers) with neuropsych test
scores. For
example, middle frontal gyrus volume with executive function.
Presently, it
seems that it is possible to evaluate cortical thickness globally
and then
pick our areas that regress to a significant extent with variables of interest. At some point in time, it would be good to see if the
combination
of GM + the underlying WM would give additional / complimentary
information.
In this regard, I am not entirely clear how we are to realize this
goal.
According to David’s HBM 2006 poster, this would take place by
adding the
volume of the cortical rim and its underlying WM. B. If so, this would necessitate good quality segmentation along the volumetric pipeline (that generates the .aseg images). This in
turn would
require some adjustments in the segmentation of white matter using
the
volume processing stream. By dropping control points, we active
improved
segmentation along the surface processing stream (correct?) If
so, how are
benefits to volume measurement realized? As I understand our data,
the
surface processing pipeline is more or less satisfactory as is. So
I wonder
what the marginal benefit of dropping control points is? C. A final question relates to how laying control points in WM
works. Jenni
told us that dropping more points increases intensity in an additive fashion. If so what is the intensity shift and how extensive is
the induced
change? If you think these questions serve a general audience, feel free
to post
this question in the help website.
Mike
------ End of Forwarded Message