I've struggled with this issue as well. My "solution" was to either just use a spherical seed placed according to prior study coordinates or to use other parts of the Freesurfer-based Destrieux atlas (like the circular insula ROIs for dorsal anterior insula). Of you have AFNI you can always play with some single subject data using their instacorr option in the GUI and see if things look right. Another option would be Thomas Yeo's 17 network parcellation map.
Since those designations (especially dorsal and ventral anterior insula) are more so based on functional data, my guess is we aren't likely to get an "atlas" that incorporates those subregions.
Sent from my iPhone
> On Sep 25, 2014, at 11:25 PM, Ethan <[log in to unmask]> wrote:
> Hi there,
> Does anyone know how to create insula subregion masks? (dorsal anterior, ventral anterior, posterior)
> It seems that AAL and Harvard-Oxford atlas only have whole insula segmentation.
> Freesurfer can segment insula into anterior and posterior, but I still want the three subregion version.
> Many thanks!