|
Author
|
Topic: ROI on flattened map
|
jimstef Junior Member
|
posted 14 October 2003 23:55
How do I create an ROI on a flattened map that is smaller than the surf patch area of 150 mm^2 (usually ~525 voxels)? This is way too large!Thanks- Jim Stefansic Vanderbilt University
|
Armand Member
|
posted 21 October 2003 22:09
dear Jim,If you open the "Mesh morphing and coloring" dialog and click on "Options", one of the options is named "Restrict selection area". The standard value here is 150. You can change this value and get accordingly smaller or larger minimal regions selected. Please let me know if you have any other questions. best regards Armin Thanks- Jim Stefansic Vanderbilt University[/B][/QUOTE]
|
Chun-Chia_Kung Member
|
posted 22 October 2003 19:12
Hi Armond:After your nice answer, can I follow on to ask a related question? When I click the activation in the flattened map, I can see the corresponding location in the axial-corronal-saggital slices. But I notice that the areas are the same, but the size are NOT! For example, the following pictures come from the same subject. When I click the flattened image, I get the corresponding roi in her axial slice, with the roi size 170 voxels. But there are 326 voxels originally in that roi (according what's shown on the option-> voi info window). Apparently some voxels are lost in the flattening process, but what cause those losses? Any comment or suggestion is welcome! Thanks, Chun-Chia [This message has been edited by Chun-Chia_Kung (edited 22 October 2003).] [This message has been edited by Chun-Chia_Kung (edited 22 October 2003).]
|
Axel Member
|
posted 23 October 2003 18:25
Hi Chun-Chia,I'm not Armin, but let me give you a quick answer: When you select a ROI in the surface, the readout is restricted to a certain depth. I assume this is why the ROI sizes differ in your case. You can try and play around with the readout: Go to "Mesh morphing and coloring", "Options" and select the readout depth in the "Functional data VTC access" part in the left upper corner. I hope this answers your question! Cheers, Axel
|
tschicke Member
|
posted 23 October 2003 19:26
hello, there may be two reasons for this: 1. if you did not use a cortex mask to calculate the GLM, then there will be activation outside cortex which is not projected onto the surface, therefore making your ROI smaller than in the VMR. From your picture, it looks like there was no cortex mask used. 2. if I am wrong, and you did use a cortex mask in the GLM, then the reason may be that the cortex mask is bigger (i.e. assumes more of the brain to belong to the cortex) than what you allow to be projected onto the cortex reconstruction mesh. The latter of these parameters can be chosen in the dialog Mesh=>Mesh Morphing and Coloring=>Options: in the upper left corner, you decide how many milimeters of brain activation are projected onto the cortex reconstruction. i do not know how to enter how many milimeters should be used for the cortex mask used in the GLM. Regards Tobias
|
astridr Junior Member
|
posted 03 November 2003 15:49
My problem is the complete opposite. I'm creating a really big ROI on an inflated map. Whatever I change in the dialogues, the reloaded ROI is not the same as that show when defined. Any idea how to correct this? Fast help is needed! Thanks Astrid Rohwedder
|
Chun-Chia_Kung Member
|
posted 12 November 2003 01:36
Hi all:Thanks for the previous suggestions, which are exactly where one can change the depth of the readVTC function (Meshes->Mesh morphing and coloring -> Options -> top left "Functional data VTC access"). But interesting, what really made me succesfully get the full ROI voxels is by checking the lower row of "Vertex > voxel > cluster access", instead of the upper "Surface patch > voxels access". Changing the starting or end "Read func depth" didn't help much. I did use the mask file in my data analysis. The only downside of being able to choose the whole ROI voxels in the flattened map is to lose the ability to highlight (by circumscribing the border of) the ROI. Thanks, Chun-Chia P.S. I never tried the approach specified in the last question, but sounds familiar to the procedure detailed in the "drawing regions in the meshes" section of uesr's guide. Have you tried that and figured it out already?
|