[color=#000000]Hi Kaija,[/color]
Not exactly. Since group-ICA is basically a group-level analysis, running ICA twice, each on different groups of subjects, is not the same running it once on the entire dataset. I would recommend to first resample your data to the same resolution using one of the methods below, and then re-run ICA on the entire dataset.
[color=#000000]method 1) if your data is already in MNI-space, then select in [i]Setup.Options [/i]the default '[b]Volume: same as mask[/b]' option in the'[i]Analysis space (voxel-level)' [/i]menu, and the '[b]Explicit mask (mask.volume.brainmask.nii)[/b][i]' [/i]option in the '[i]Analysis mask (voxel-level)' [/i]menu, and re-run the [i]Setup&Denoising [/i]steps (that will have CONN automatically resample your data to 2mm voxels jointly across all subjects as part of the [i]Setup [/i]step).[/color]
[color=#000000]method 2) if, on the other hand, your functional data is not yet in MNI-space, then normalize your data first (e.g. run a "functional normalization" preprocessing step, in [i]Setup -> Preprocessing[/i]), and then follow method (1) above. [/color]
[color=#000000]Best[/color]
[color=#000000]Alfonso [/color]
[i]Originally posted by kaija:[/i][quote]Hello,
I am currently trying to perform voxel-to-voxel analyses on subjects from two different datasets. When I try to run any analysis in the voxel-to-voxel category, I get the following error:
"Unequal analysis voxels across subjects. Modify 'spatial resolution' and/or 'analysis mask' fields in Setup->Options to match analysis voxels across subjects"
I have gotten around this by running the ICA and selecting only the new subjects I want to add, and now I assume that in the 2nd level results I am looking at both datasets combined, but I don't know if this is the correct way of doing things.
Otherwise I will have to try and fix this voxel number but I am not sure how to go about doing that?
Sorry if the answer to this if fairly obvious, but any help would be appreciated!
Thank you very much,
Kaija[/quote]
Not exactly. Since group-ICA is basically a group-level analysis, running ICA twice, each on different groups of subjects, is not the same running it once on the entire dataset. I would recommend to first resample your data to the same resolution using one of the methods below, and then re-run ICA on the entire dataset.
[color=#000000]method 1) if your data is already in MNI-space, then select in [i]Setup.Options [/i]the default '[b]Volume: same as mask[/b]' option in the'[i]Analysis space (voxel-level)' [/i]menu, and the '[b]Explicit mask (mask.volume.brainmask.nii)[/b][i]' [/i]option in the '[i]Analysis mask (voxel-level)' [/i]menu, and re-run the [i]Setup&Denoising [/i]steps (that will have CONN automatically resample your data to 2mm voxels jointly across all subjects as part of the [i]Setup [/i]step).[/color]
[color=#000000]method 2) if, on the other hand, your functional data is not yet in MNI-space, then normalize your data first (e.g. run a "functional normalization" preprocessing step, in [i]Setup -> Preprocessing[/i]), and then follow method (1) above. [/color]
[color=#000000]Best[/color]
[color=#000000]Alfonso [/color]
[i]Originally posted by kaija:[/i][quote]Hello,
I am currently trying to perform voxel-to-voxel analyses on subjects from two different datasets. When I try to run any analysis in the voxel-to-voxel category, I get the following error:
"Unequal analysis voxels across subjects. Modify 'spatial resolution' and/or 'analysis mask' fields in Setup->Options to match analysis voxels across subjects"
I have gotten around this by running the ICA and selecting only the new subjects I want to add, and now I assume that in the 2nd level results I am looking at both datasets combined, but I don't know if this is the correct way of doing things.
Otherwise I will have to try and fix this voxel number but I am not sure how to go about doing that?
Sorry if the answer to this if fairly obvious, but any help would be appreciated!
Thank you very much,
Kaija[/quote]