Alle,
Please refer to the original pickatlas paper in NeuroImage 19 (2003),
1233-1239. On page 1236, in the section titled "Atlas modifications":
The last uniquely labeled slice returned from the Talairach Daemon is at
Talairach Z coordinate 65. The information returned from slice 66 is
identical to that at slice 65, and nearly identical at slice 67. There
is approximately 1 cm of additional brain tissue in the Z direction that
is unlabeled. This region encompasses the superior-most extent of the
sensorimotor cortex and includes portions of Brodmann areas 1, 2, 3, 4,
5, and 6. In MNI space, the Talairach Daemon atlas labels terminate in
a partial slice at MNI Z coordinate 72. In order to complete the atlas
to the vertex in MNI space, we replicated the last completely labeled
slice (MNI Z coordinate 71) with a 0.9 demagnification factor for
successive slices in the Z direction (MNI Z coordinate 72-78). Since
Brodmann areas 1, 2, 3, 4, 5, and 6 continue to the vertex, this
approach affords a good approximation for the intracranial contents
(Fig. 1).
The "artifact" you note is actually a "feature" in that the Talairach
Daemon (from which the pickatlas was derived) terminated before the top
of the vertex. If you look at some of the other atlases in the
pickatlas (that are not from the Talairach Daemon - e.g., AAL), I don't
think you will see this "feature."
Kathy Pearson
WFUHS Radiology
-----Original Message-----
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]]
On Behalf Of Alle Meije Wink
Sent: Friday, July 27, 2007 8:15 AM
To: [log in to unmask]
Subject: [SPM] WFU files -segmentation problem?
Dear all,
Playing around with the WFU pickatlas label files, I saw this strange
shape at the top of the brain. Has this been reported before?
The attached image shows the EPI template, overlayed in green+brown is
the tissue type map (grey/white matter/csf), and in blue the hemisphere
map (left/right cortex/cerebellum etc) from the WFU pickatlas.
At the top of the brain, these maps have weird artifacts, which I think
are due to the segmentation. Is there a fix for this? Could similar
errors also be in the label files?
Thanks
Alle Meije
|