Re: [caret-users] Displaying different sets of activations without RGB
John, It sounds like your best bet would be to use border contours to outline the 2nd and 3rd activations. You can do this using Surface: Region of Interest Operations: Operate on Selected Nodes: Operation: Create Borders Around Clusters. Start by selecting the nodes of interest for the 2nd activation, then apply the Create Borders around Clusters operation. Choose a suitable border color file. Then do the same for the 3rd cluster, but choosing a different color if you like. Once the border(s) around the selected nodes are generated, project the borders, then save them. You can then adjust whether they are points or lines and set the line/point size/thickness. Does this meet your needs? David VE Does this meet your needs? On Apr 25, 2007, at 4:42 PM, John Arsenault wrote: Hi John, Sorry for not clarifying our problem more specifically. We have used the metric to RGB function and it is quite helpful in most cases but in this specific instance it would be be nice to have one metric with a normal scale that ranges between 2 colors (orange-yellow; so as to vividly show the maxima and outlying areas which is important to our argument) while 2 other metrics are simply denoted by a circle outlining the areas above a threshold. We have tried to display this data using the three columns of RGB but in the end it becomes a bit too muddled and the scale of the first activation which is very important is reduced to half of the range available with the two colors. Consequently a solution where we are able to see one activation across a color scale of 2 colors and 2 other secondary activations were it would be most important to see there position relative to the first activation. We are now able to extract the points from a metric above a threshold. Therefore is it possible that we could somehow connect these outermost points creating an outline of different foci for the 2nd and 3rd activations sets outline. Is this possible in Caret and if so how? Or is there any alternative where a main activation can be shown and 2 secondary sets can be shown without taking away from the information of the first. Thank you for all of your help, John John, On page 12 of the Analysis Guide (http://brainmap.wustl.edu/caret/ pdf/ Caret_5.5_UserGuide_Analysis.pdf) are the steps that allow assignment of up to three metric columns to the red, green, and blue channels of an RGB paint file. So, assign one column of metric data to red and another metric column to blue. Where the two activations overlay, you will see a shade of magenta. Where the two activations do not overlap, you will see a shade or red or blue. -- John Harwell [EMAIL PROTECTED] 314-362-3467 Department of Anatomy and Neurobiology Washington University School of Medicine 660 S. Euclid Ave.Box 8108 St. Louis, MO 63110 USA On Apr 25, 2007, at 2:50 PM, John Arsenault wrote: Hello, I am trying to show the similarities/differences between two seperate activations that share many common foci. The RGB works quite well especially when presented on in the same space in powerpoint or similiar programs where you can toggle back and forth and see the regions of overlap and how the activations are similiar/different. The problem I am experiences is that when displayed in a static figure it is difficult to see exactly were the overlaps are and where they are not. Consequently it would be ideal to display one set of activations through the normal metric color scale and overlay some sort of outline denoting the second set of activations(perhaps concentric blobs denoting T-score). Is it possible to generate these outlines in caret and if so how? Thank you for your help, John Arsenault ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users
[caret-users] spm99 metrics projection
Hello, We have 18 subjects that have been analysed with spm99. Their structural images have been normalized into a template from Marseille (http://irmfmrs.free.fr/formation/traitement_des_donnees/GuideSPM.htm) in MNI space. Since, we didn't use spm99 standard template, I am not sure we can use Human.PALS_B12.LR.MULTI-FIDUCIAL_SPM99_fMRI-MAPPER.B1-12.LEFT.73730.spec for projecting the metrics. What would be the correct way of projecting the metric files of those subjects into a Caret/PALS surface? Is there a Human.PALS_B12.LR.AVERAGE-FIDUCIAL_SPM96.RIGHT.73730.spec for spm99? Thank you for your help. Mateus
[caret-users] Surface ROI report
Hi folks, I'd like to extract stats for individual Brodmann regions (warped from the Colin atlas to an individual) from a functional experiment. I can query each individual region, but that's tedious and I was hoping that Caret's Statistical Report on Paint Subregions would do the trick (see screenshot). Unfortunately, the resulting report only contains data for one entry of my Paint file, BROD.UNASSIGNED. I'm running v.5.503, fresh from the web. Maybe I'm misinterpreting what this option is supposed to do? All hints appreciated! Thanks a lot, Johannes inline: caret_stats_report.png
Re: [caret-users] spm99 metrics projection
Hi Mateus, Wasn't SPM96 the single subject colin target? I'm asking, because I really don't know the answer, but I think David may have told me this while working on his foci project. It turns out my caret/data_files/fmri_mapping_files directory *does* have AVERAGE-FIDUCIAL_SPM95 and AVERAGE-FIDUCIAL_SPM96 spec files that almost certainly have to do with David's foci project. Your version may be older and not yet have these files. But a worthwhile task might be to do this: cd /usr/local/caret/data_files/fmri_mapping_files caret5 Then, open whichever of these specs you think has the best chance of aligning: Human.PALS_B12.LR.AVERAGE-FIDUCIAL_SPM95.LEFT.73730.spec Human.PALS_B12.LR.AVERAGE-FIDUCIAL_SPM96.LEFT.73730.spec Human.colin.SPM2.LEFT.71785.spec Human.colin.SPM99.LEFT.71785.spec Human.PALS_B12.LR.MULTI-FIDUCIAL_FLIRT_fMRI-MAPPER.B1-12.LEFT.73730.spec Human.PALS_B12.LR.MULTI-FIDUCIAL_MRITOTAL_fMRI-MAPPER.B1-12.LEFT.73730.spec Human.PALS_B12.LR.MULTI-FIDUCIAL_SPM2_fMRI-MAPPER.B1-12.LEFT.73730.spec Human.PALS_B12.LR.MULTI-FIDUCIAL_SPM99_fMRI-MAPPER.B1-12.LEFT.73730.spec Then File: Open Data File: Volume anatomy file: and open your Marseille template Switch to volume view ALL D/C: overlay/underlay volume: settings: Toggle on Show surface contour checkbox Scroll around and evaluate how well the fiducial contour intersects the surface Perhaps even get a capture, if the alignment is excellent, and use this as your reason for using this target. If none align well, then I don't have a good plan B for you. Donna On 04/26/2007 05:15 AM, Mateus Joffily wrote: Hello, We have 18 subjects that have been analysed with spm99. Their structural images have been normalized into a template from Marseille (http://irmfmrs.free.fr/formation/traitement_des_donnees/GuideSPM.htm) in MNI space. Since, we didn't use spm99 standard template, I am not sure we can use Human.PALS_B12.LR.MULTI-FIDUCIAL_SPM99_fMRI-MAPPER.B1-12.LEFT.73730.spec for projecting the metrics. What would be the correct way of projecting the metric files of those subjects into a Caret/PALS surface? Is there a Human.PALS_B12.LR.AVERAGE-FIDUCIAL_SPM96.RIGHT.73730.spec for spm99? Thank you for your help. Mateus ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users
Re: [caret-users] Surface ROI report
Hi Johannes, Try using statistical report on selected nodes instead. Below is an example for another user that might help: When you use Statistical Report on Paint Subregions, whatever column you have selected on the ROI menu is affecting the areas reported for each column. If you select the CS of subject A's paint column, then that subject/column will have no ??? area listed, e.g.: Paint Column 0 deformed CS 1 Paint NameAreaPercent Area COG - X COG - Y COG - ZLatitude Longitude CS2107.198 100.000 31.448 -22.408 54.292 0.000 0.000 But each of the other subjects' non-selected columns has its paint split into overlap and non-overlap areas: Paint Column 29 deformed CS 30 Paint NameAreaPercent Area COG - X COG - Y COG - ZLatitude Longitude ???287.918 13.664 30.901 -22.159 61.147 0.000 0.000 CS1819.280 86.336 31.548 -22.453 53.045 0.000 0.000 This could work to your advantage if you have a centroid paint selected. Then it will report the percent overlap for each subject. But if you want the real CS area for each subject, select all nodes and use statistical report on selected nodes. Donna On 04/26/2007 06:50 AM, Johannes Klein wrote: Hi folks, I'd like to extract stats for individual Brodmann regions (warped from the Colin atlas to an individual) from a functional experiment. I can query each individual region, but that's tedious and I was hoping that Caret's Statistical Report on Paint Subregions would do the trick (see screenshot). Unfortunately, the resulting report only contains data for one entry of my Paint file, BROD.UNASSIGNED. I'm running v.5.503, fresh from the web. Maybe I'm misinterpreting what this option is supposed to do? All hints appreciated! Thanks a lot, Johannes ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users
Re: [caret-users] spm99 metrics projection
Ok, I will try it. Thank you very much! Mateus Donna Dierker wrote: On 04/26/2007 08:29 AM, Mateus Joffily wrote: Hi Donna, Wasn't SPM96 the single subject colin target? I'm asking, because I really don't know the answer, but I think David may have told me this while working on his foci project. I don't know, but I will check it. It turns out my caret/data_files/fmri_mapping_files directory *does* have AVERAGE-FIDUCIAL_SPM95 and AVERAGE-FIDUCIAL_SPM96 spec files that almost certainly have to do with David's foci project. Your version may be older and not yet have these files. Yes. I also have an AVERAGE-FIDUCIAL_SPM95. But I was wondering if it exists an AVERAGE-FIDUCIAL_SPM99. But a worthwhile task might be to do this: cd /usr/local/caret/data_files/fmri_mapping_files caret5 Then, open whichever of these specs you think has the best chance of aligning: Human.PALS_B12.LR.AVERAGE-FIDUCIAL_SPM95.LEFT.73730.spec Human.PALS_B12.LR.AVERAGE-FIDUCIAL_SPM96.LEFT.73730.spec Human.colin.SPM2.LEFT.71785.spec Human.colin.SPM99.LEFT.71785.spec Human.PALS_B12.LR.MULTI-FIDUCIAL_FLIRT_fMRI-MAPPER.B1-12.LEFT.73730.spec Human.PALS_B12.LR.MULTI-FIDUCIAL_MRITOTAL_fMRI-MAPPER.B1-12.LEFT.73730.spec Human.PALS_B12.LR.MULTI-FIDUCIAL_SPM2_fMRI-MAPPER.B1-12.LEFT.73730.spec Human.PALS_B12.LR.MULTI-FIDUCIAL_SPM99_fMRI-MAPPER.B1-12.LEFT.73730.spec Then File: Open Data File: Volume anatomy file: and open your Marseille template Switch to volume view ALL D/C: overlay/underlay volume: settings: Toggle on Show surface contour checkbox Scroll around and evaluate how well the fiducial contour intersects the surface Ok. I will try it. Finaly, if I choose to use a MULTI-FIDUCIAL spec, I will need to project each individual metric to each Buck_Case01:12 surface? After, how can I get an average fiducial map of the activations? No: If you decide MULTI-FIDUCIAL_SPM99 aligns nicely, then select SPM99 from the list of spaces in the Map to Atlas menu. Then you'll see the entries that correspond to this spec file, select them, and select either map to average fiducial only; map to each of the subjects; and/or generate the average of the individual mappings. In other words, the process is exactly the same; the only issue is first determining whether any of the existing surfaces intersects your template well enough, and if so, which intersects best. Thank you very much. Mateus Perhaps even get a capture, if the alignment is excellent, and use this as your reason for using this target. If none align well, then I don't have a good plan B for you. Donna On 04/26/2007 05:15 AM, Mateus Joffily wrote: Hello, We have 18 subjects that have been analysed with spm99. Their structural images have been normalized into a template from Marseille (http://irmfmrs.free.fr/formation/traitement_des_donnees/GuideSPM.htm) in MNI space. Since, we didn't use spm99 standard template, I am not sure we can use Human.PALS_B12.LR.MULTI-FIDUCIAL_SPM99_fMRI-MAPPER.B1-12.LEFT.73730.spec for projecting the metrics. What would be the correct way of projecting the metric files of those subjects into a Caret/PALS surface? Is there a Human.PALS_B12.LR.AVERAGE-FIDUCIAL_SPM96.RIGHT.73730.spec for spm99? Thank you for your help. Mateus ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users
Re: [caret-users] Displaying different sets of activations without RGB
Hi David, This solution works very well if you only want to show the location of secondary activations upon a main activation. As well by manipulating the threshold of the ROI you can create a contour map defining the significance distribution within this region. Thank you so much for the help, John John, It sounds like your best bet would be to use border contours to outline the 2nd and 3rd activations. You can do this using Surface: Region of Interest Operations: Operate on Selected Nodes: Operation: Create Borders Around Clusters. Start by selecting the nodes of interest for the 2nd activation, then apply the Create Borders around Clusters operation. Choose a suitable border color file. Then do the same for the 3rd cluster, but choosing a different color if you like. Once the border(s) around the selected nodes are generated, project the borders, then save them. You can then adjust whether they are points or lines and set the line/point size/thickness. Does this meet your needs? David VE Does this meet your needs? On Apr 25, 2007, at 4:42 PM, John Arsenault wrote: Hi John, Sorry for not clarifying our problem more specifically. We have used the metric to RGB function and it is quite helpful in most cases but in this specific instance it would be be nice to have one metric with a normal scale that ranges between 2 colors (orange-yellow; so as to vividly show the maxima and outlying areas which is important to our argument) while 2 other metrics are simply denoted by a circle outlining the areas above a threshold. We have tried to display this data using the three columns of RGB but in the end it becomes a bit too muddled and the scale of the first activation which is very important is reduced to half of the range available with the two colors. Consequently a solution where we are able to see one activation across a color scale of 2 colors and 2 other secondary activations were it would be most important to see there position relative to the first activation. We are now able to extract the points from a metric above a threshold. Therefore is it possible that we could somehow connect these outermost points creating an outline of different foci for the 2nd and 3rd activations sets outline. Is this possible in Caret and if so how? Or is there any alternative where a main activation can be shown and 2 secondary sets can be shown without taking away from the information of the first. Thank you for all of your help, John John, On page 12 of the Analysis Guide (http://brainmap.wustl.edu/caret/ pdf/ Caret_5.5_UserGuide_Analysis.pdf) are the steps that allow assignment of up to three metric columns to the red, green, and blue channels of an RGB paint file. So, assign one column of metric data to red and another metric column to blue. Where the two activations overlay, you will see a shade of magenta. Where the two activations do not overlap, you will see a shade or red or blue. -- John Harwell [EMAIL PROTECTED] 314-362-3467 Department of Anatomy and Neurobiology Washington University School of Medicine 660 S. Euclid Ave.Box 8108 St. Louis, MO 63110 USA On Apr 25, 2007, at 2:50 PM, John Arsenault wrote: Hello, I am trying to show the similarities/differences between two seperate activations that share many common foci. The RGB works quite well especially when presented on in the same space in powerpoint or similiar programs where you can toggle back and forth and see the regions of overlap and how the activations are similiar/different. The problem I am experiences is that when displayed in a static figure it is difficult to see exactly were the overlaps are and where they are not. Consequently it would be ideal to display one set of activations through the normal metric color scale and overlay some sort of outline denoting the second set of activations(perhaps concentric blobs denoting T-score). Is it possible to generate these outlines in caret and if so how? Thank you for your help, John Arsenault ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users ___ caret-users mailing list caret-users@brainvis.wustl.edu http://pulvinar.wustl.edu/mailman/listinfo/caret-users
