Dear all,

 when I choose

1)    the Multi-session(group)ICA in 'Perform'
2)    input 122 4D nifti files

however, here I can not choose the mask, only custom mask can be chose.

I want to choose the MNI Structural Atlas Cerebellum, how should I do?

thanks a lot

yours

lin

Dear Lin,

I would suggest to check two things.

• The mICA Toolbox looks for the atlases in the directory ${FSLDIR}/data/atlases. Could you check if your atlas directory is there?
• The toolbox also checks, whether the matrix size and the orientation of the nifti files correspond to the MNI-space. The matrix size should be either 91x109x91 or 182x218x182 or 364x436x364; and the qform_name should be MNI_152 (qform_code 4). If this is not the case, the mICA Toolbox will only allow to choose custom masks.

Kind regards,

Jorge

Dear Lin,
in general you have to make sure that your input data are already normalized to the template that your mask(s) come from. That means if you want to use an MNI space based atlas, your data need to be normalized to MNI space. Please not that there are differences in MNI templates between FSL and SPM. Our toolbox is based on FSL. As Jorge has pointed out, mICA will check the bounding box of your input data and only allow MNI-based atlases if it matches that of the template. In every other case, it will only allow custom masks.

Hope that helps,

Florian

Dear Jorge,

Thank you so very much. Now, my data set [1] 122 participants(83patients & 39healthy) [2] preprocessing have been done using FSL MELODIC (I have been completed the ICA and dual regression) .

My question: 

1) If I want to know the relationship between the brain stem and other whole brain regions using mICA, how should I do?
2) In the paper 'MICA-A toolbox for masked independent component analysis of fMRI data'  Fig. 4 (a)(b), is this necessary for me to decide the number of component? 
does this figure mean that if I donot sure the number of component I can use the 'mICA reproducibility analysis'?

thanks a lot 
Kind regards,

lin

Dear Florian,

Thank you so very much. Now, my data set [1] 122 participants(83patients & 39healthy) [2] preprocessing have been done using FSL MELODIC (I have been completed the ICA and dual regression) .
My question:
1) If I want to know the relationship between the brain stem and other whole brain regions using mICA, how should I do?
2) In your research 'MICA-A toolbox for masked independent component analysis of fMRI data' Fig. 4 (a)(b), is this necessary for me to decide the number of component?
does this figure mean that if I donot sure the number of component I can use the 'mICA reproducibility analysis'?

thanks a lot
Kind regards,


lin