Dear all,

I am creating new ROIs from MNI coordinates in Conn Toolbox. When defining the paramenters of the ROI, the last gap to fill is the ROI-file resolution. By default I have 2mm, but I would like to know

(1) which measure is usually recommended for this parameter; and 

(2) where this parameter for the ROI-file resolution come from, why is 2mm?

Thank you in advance!

Oihana

Dear Oihana,

Regarding (1) the recommendation is to have ROI files that have at least as high spatial resolution as your functional data. In general, when extracting BOLD timeseries from an ROI, CONN will look at each individual voxel in your ROI definition file, and for each of those voxels it will extract the BOLD data using nearest neighbor interpolation (i.e. extracting from the nearest voxels in the functional data), so having ROI resolutions that are equal or higher than your functional data makes sure that all functional data voxels within your mask will be included when computing ROI-level BOLD timeseries. 

Regarding (2) the 2mm value is the default voxel-size that your functional data will have been resampled after spatial normalization, so that makes it a reasonable default value for the spatial resolution of your ROI files.

Best

Alfonso

Originally posted by oihana:

Dear all,

I am creating new ROIs from MNI coordinates in Conn Toolbox. When defining the paramenters of the ROI, the last gap to fill is the ROI-file resolution. By default I have 2mm, but I would like to know

(1) which measure is usually recommended for this parameter; and 

(2) where this parameter for the ROI-file resolution come from, why is 2mm?

Thank you in advance!

Oihana