Dear Matthieu,
On Wed, 9 Jan 2019 14:46:17 +0000, Matthieu Vanhoutte <[log in to unmask]> wrote:
>Dear CAT12's experts,
>
>I think I found how to choose the surfaces used during projection of volume
>to surface thanks to the parameters "StartPoint", "StepSize" and "EndPoint".
>
>1) For example, if I want to project native volume with data ranging from
>35 to 65% of the cortical thickness with a step *t* = 5%, should I specify
>"StartPoint=0.35", "StepSize=0.05" and "EndPoint=0.65" combined with
>"Relative position within a tissue class" parameter ?
I don't know what kind of fMRI data you have, but for the common voxel size of 2-3mm you probably overestimate the resulting spatial resolution of that approach. The default values in the GUI work fine even for higher resoluted data. Only for layer-specific fMRI data (<= 1mm voxel size) there is probably need to fine.tune the parameters.
>
>2) Is it possible to choose the weights associated to each surface used in
>the projection ?
No, and I don't think it's necessary here. Even the difference between the standard and the Gaussian weighting is rather small, but the Gaussian weighting results in less noisy maps. For fMRI data I would also think about rather using the absolute max function for contrast maps in native space. Please also keep in mind that this approach is still experimental and I cannot give any general recommendations.
>
>3) The "Map Volume (Native space) to Individual" surface fonction specifies
>that functional volume has to be coregistered in the same space as the T1
>image. Does it suppose that native volume has to be "coregistered and
>resliced" or just "coregistered without reslicing" in the same space as the
>T1 image?
Resampling is not necessary.
Best,
Christian
>
>Best,
>Matt
>
>Le mar. 8 janv. 2019 à 11:38, Matthieu Vanhoutte <
>[log in to unmask]> a écrit :
>
>> Dear CAT12's experts,
>>
>> I noticed that in the last version of CAT12 it was possible to project
>> data from native volume to surface according mean signal or gaussian
>> weighted signal (recommanded for fMRI).
>>
>> Could you inform me how many weighted are used for mean or gaussian
>> weighting ?
>>
>> Would it be possible to personalize the weights and the associated
>> surfaces used ?
>>
>> Best,
>> Matt
>>
>
|