[caret-users] Caret visualization - cluster size thresholds

R. Matthew Hutchison rhutchis at uwo.ca
Fri Aug 26 12:22:38 CDT 2011


I have used the steps you described successfully before and just repeated it
now with no problems.

Have you tried using extremely large values- just as a proof of principal?
All of the clusters may be larger than your current threshold value.
Have you adjusted your values for the "cluster value threshold"?

-Matt

On Fri, Aug 26, 2011 at 12:19 PM, Muriel LOBIER <muriel.lobier at gmail.com>wrote:

> I'm not sure I've made myself very clear.
>
> What I'm trying to do is the following:
> I have mapped my SPM data to Caret, but it shows all activated clusters.
> I want to be able to display only those clusters that are larger than 50
> vxls.
> >From what I had understood, that would be done in the metric modification
> window (attributes/metric/clustering and smoothing ), where there is a box
> where you can select "Cluster Minimum Size" either by min nb of nodes or min
> surface area.
> It is those values that I've tried modifying without any luck: I create a
> new metric column with my cluster min size, change the column name in my
> visualization, but all my clusters remain.
>
>
>
>
>
>
> On Fri, Aug 26, 2011 at 18:06, Donna Dierker <donna at brainvis.wustl.edu>wrote:
>
>> Hmmm.  I don't know that Caret can do this.  There are several
>> cluster-based tests that find the minimum area required to correct for
>> multiple comparisons, given an input cluster-forming threshold (say,
>> t=+/-3.0):
>>
>> caret_command.txt:      caret_command -metric-statistics-anova-one-way
>> caret_command.txt:      caret_command -metric-statistics-anova-two-way
>> caret_command.txt:      caret_command
>> -metric-statistics-coordinate-difference
>> caret_command.txt:      caret_command
>> -metric-statistics-interhemispheric-clusters
>> caret_command.txt:      caret_command -metric-statistics-one-sample-t-test
>>
>> caret_command.txt:      caret_command -metric-statistics-paired-t-test
>> caret_command.txt:      caret_command -metric-statistics-two-sample-t-test
>>
>>
>> These are permutation tests that find the smallest alpha*iterations
>> cluster at the specified threshold.  And they output a paint/region file
>> showing the clusters that survived significance cut-off.
>>
>> But interactively in the GUI, I'm not aware of such a feature.  If you've
>> checked the metric/smoothing dialog, then Surface: Region of Interest is the
>> other place I'd look.  And I don't recall it.
>>
>> Since we adopted TFCE, we don't use the cluster-based tests anymore.
>>
>>
>> On Aug 26, 2011, at 8:40 AM, Muriel LOBIER wrote:
>>
>> Hi all,
>>
>> I'm just beginning to use CARET to map activations from SPM5 to 3D brains.
>> There is still one thing I can't quite succeed in doing, and that's
>> limiting visible clusters to clusters above a certain size threshold. I've
>> used the
>> appropriate attributes/metric/clustering and smoothing command as I've
>> seen in tutorials and previous posts to the mailing lists, but it doesn't
>> ever seem to apply changes I make to the visualization (I've made sure I use
>> the modified column).
>>
>> Is there anything obvious I could have missed ?
>>
>> Thanks,
>>
>> Muriel
>> --
>> -- - - -- - - -- - - -- - - -- - - -- - - -- - - -- - - -- - - -- - - --
>> Muriel LOBIER
>> Doctorante en  Sciences cognitives psychologie et neurocognition
>> PhD student in Cognitive science, psychology, and neurocognition
>>
>> muriel.lobier at gmail.com
>> muriel.lobier at upmf-grenoble.fr
>> tel : +33 6 88 67 15 55
>>
>> Laboratoire de Psychologie et Neurocognition (CNRS UMR 5105)
>> Bureau 220
>> Tel : +33 4 76 82 58 93
>>
>> Bâtiment Sciences de l'Homme et des Mathématiques
>> BP 47, 38040 Grenoble Cedex 9
>>
>>
>> Web : http://webu2.upmf-grenoble.fr/LPNC/membre_muriel_lobier
>> _______________________________________________
>> caret-users mailing list
>> caret-users at brainvis.wustl.edu
>> http://brainvis.wustl.edu/mailman/listinfo/caret-users
>>
>>
>>
>> _______________________________________________
>> caret-users mailing list
>> caret-users at brainvis.wustl.edu
>> http://brainvis.wustl.edu/mailman/listinfo/caret-users
>>
>>
>
>
> --
> -- - - -- - - -- - - -- - - -- - - -- - - -- - - -- - - -- - - -- - - --
> Muriel LOBIER
> Doctorante en  Sciences cognitives psychologie et neurocognition
> PhD student in Cognitive science, psychology, and neurocognition
>
> muriel.lobier at gmail.com
> muriel.lobier at upmf-grenoble.fr
> tel : +33 6 88 67 15 55
>
> Laboratoire de Psychologie et Neurocognition (CNRS UMR 5105)
> Bureau 220
> Tel : +33 4 76 82 58 93
>
> Bâtiment Sciences de l'Homme et des Mathématiques
> BP 47, 38040 Grenoble Cedex 9
>
>
> Web : http://webu2.upmf-grenoble.fr/LPNC/membre_muriel_lobier
>
> _______________________________________________
> caret-users mailing list
> caret-users at brainvis.wustl.edu
> http://brainvis.wustl.edu/mailman/listinfo/caret-users
>
>


-- 
R. Matthew Hutchison, PhD. Candidate
Centre for Functional and Metabolic Mapping
Robarts Research Institute
Cuddy Wing - 9.4T Suite
100 Perth Drive
London, Ontario, Canada  N6A 5K8
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://brainvis.wustl.edu/pipermail/caret-users/attachments/20110826/e62a0680/attachment-0002.html>


More information about the caret-users mailing list