Flat

From VISTA LAB WIKI

Jump to: navigation, search

This page describes using the mrVista Flat view to visualize data projected onto a flattened cortical surface. We use mrFlatMesh to create a flat representation of the cortical surface.

From the Flat view we fit visual field atlases to retinotopic data. Atlas fitting uses the atlasFitGUI code; that process uses deformation code contributed by the SAFIR group (Jens Heyder).

Contents

[edit] Opening a Flat view

Once an unfold has been installed into a session, you can open a flat window by selecting the menu Window | Open Flat Window. A window will open with the unfolded flat patches displayed. You can flip between the left and right hemispheres using the radio buttons at the bottom of the window.

The intensities of the underlay image for the unfold show the estimated curvature of each gray matter node (relative to its neighbors). Higher curvature values are more convex with respect to the outside of the brain, and associated with gyri. Low curvature values are more concave, and are associated with sulci.

The full range of curvatures is displayed by default, which can make the flat maps look different from conventional gray matter representations. Select Color Map | Threshold Anat Curvature to make the color map threshold between more convex and concave regions, making the gyral and sulcal patterns clearer. Color Map | Reset Defaults will restore the old settings.

[edit] Orienting yourself on the flat patches

The first challenge in using these flat patches is often identifying which part of cortex is represented. Because of the disk representations, there aren't many immediate landmarks to identify which direction is which. Complicating things, the flat patch can be displayed at an arbitrary orientation with respect to anterior/posterior, dorsal/ventral or other directions. Occasionally, the flattening code will even have a patch appear to be flipped left/right with respect to actual cortex (as if you were looking at it from inside the brain instead of outside!) These ambiguities can be corrected as described below.

First, it's good to have a reference as to the 3D pattern of gyri and sulci on an inflated brain. Open or build a 3D Mesh from the Volume window, and compare it to the flat patch. Note the pattern of gyri and sulci. Inflate if you need to (selected one of the Mesh | Smooth options in the 3D Window) to show the patterns more clearly. Now compensate for orientation. The 'rotation' slider on the bottom of the flat window allows you to rotate the patch for the selected hemisphere. If the patch appears to be left/right flipped, you can flip the patch by using the toggle menus in View | Flip Hemisphere Left/Right. The flip and rotations are set separately for each hemisphere, so do this for both if you're analyzing both. (Projecting ROIs and data maps from the volume also helps with orienting.)

When you have the flat patches oriented nicely, be sure to save the settings (File | Save Prefs) so that they will be remembered for future times you use the window.

[edit] Projecting data from the volume

All of the projections into the flat map come via the volume or gray views, and are accessible via the Xform menu.

[edit] Viewing ROIs

ROIs can be defined on the flat data or projected from the volume/gray views. You may notice that, because of the gridding process, the ROIs appear patchy on the flat surface. This does not necessarily reflect the actual ROIs. One way to show the ROIs more clearly is to render them with a thicker line perimeter. To do this, select the menu ROIs | Hide/Show ROIs | Set ROI Options (or use Ctrl-3). You will be presented with a dialog for setting ROI display options. Select the "filled perimeter" option, and the ROI outlines will be thicker.


[edit] Creating high-resolution images of flat patches suitable for publication

The gridded display of the flat patch in the window may not always be clear for presenting results. To produce an upsampled version of the flat patch suitable for publication, select the menu Plots | Current Scan | Publish Figure. The option to publish and set params will bring up a dialog allowing you to set the parameters for how to publish the figure (e.g. setting the line width of ROIs, the degree of spatial blurring on the overlay map, whether to show a legend with the color bar or ROI names, etc.).

Comparison of the Flat window with the results of the publishFigure command.

Hint: If you have an ROI named 'mask' in the flat view, when publishing, only the overlay data contained within the ROI will be displayed. (This is also true of projecting data onto surface meshes).

[edit] Atlas

The Atlas method fits a model of visual field representation in the Flat view to retinotopic data. Fitting an atlas provides a principled way to define visual areas and measures how well a visual field model (e.g., a hemifield or quarterfield) fits the data/

The method creates an atlas standardized space that can be used to combine data across subjects. This is the method Kaoru Amano used to identify two visual field maps (TO-1, TO-2) in motion-selective cortex.

[edit] Screenshots

File:FlatViewExample.png
Example Flat view window, with default curvature settings
Example Flat view window, thresholded curvature
Comparison of the Flat window with the results of the 'publishFigure command.
Comparing a flat patch to a 3D Mesh
Personal tools