| |
Discussion and Future Work
186
·
The registration algorithm, which uses elastic matching based on linear elasticity the-
ory can easily be plugged in to mrFindBorders. Then, the user could choose this algo-
rithm from the list boxes in the options menu.
8.5 Other Fields of Application
·
A general approach for fitting two-dimensional parameterized models (atlases) to
measured images, which in this case were based on fMRI signals, has been developed.
The methods explained in this document are not limited to visual areas V1, V2, and
V3, but may also be applied to other brain regions. For example, it might be possible
to identify other visual areas, such as V3a, V7, and MT by modifying the atlases.
·
It might be possible to calculate the cortical area that corresponds to certain parts of a
flat map by using iso-angle lines and iso-eccentricity lines. This would involve going
back to gray nodes because of distortions introduced during the flattening procedure.
·
mrFindBorders allows saving the deformed atlases (flat maps) as phase images. These
atlases might be transformed back to the volume coordinate system (3-space), where
they could be used to measure cortical magnification functions. Currently, cortical
magnification functions are fit to measured data. According to experienced research-
ers, this idea seems good enough to pursue. The transformation from 2-space (flat
map) to 3-space (gray nodes) is necessary because of distortions introduced during the
flattening procedure. Using data from a deformed atlas instead of the original meas-
ured image is advantageous since measured data is very noisy.
8.6 Quality
·
Another quality measure could be based on the boundaries only, and not on the ROIs
that mark the visual areas in the flat maps.
|  |
|
| |
|
|
