Potential pitfalls and
1: Construction of mouse 3D brain atlases and navigation software
using fiducials located by the correlation filter may fail. Should
implicit fiducials defined using a matched filter prove insufficient for
reconstruction of any given atlas, point fiducials will be obtained using
bisected cells or blood vessels (the iScope will be used to acquire images
of these). While this approach is straightforward and is likely to yield
highly accurate reconstruction, it requires a high degree of operator
interaction and thus is not as attractive as using matched filters. If
this approach is required, it will add another 6 months to the atlas
construction phase of the work. However, it will introduce only a moderate
delay in the alignment phase of Specific Aim 2, which can commence when
one atlas becomes available.
to achieve acceptable cross-atlas alignment as proposed will necessitate
use of a larger number of fiducials. Since each brain will eventually be
delineated, all VOIs can be used for this purpose.