Hello all,

I am running a group-ICA to look at f-connectivity differences between two patient groups. However, our patient and control groups are of unequal size (patient=65, control=25). In SPM you are able to specify "assume unequal variances" but I do not see this option in CONN- I'm wondering whether this is already accounted for within the software or not?

Thanks in advance for any help,
Breanne

Hi Breanne,

Sorry, the "unequal variances" option is not available in CONN's second-level GLM analyses (you would have to run the analyses within SPM in order to use that option). That said, this option is unrelated to the two groups having different sample sizes, GLM between-group comparisons do not assume equal sample sizes and work perfectly fine even in the extreme case of one group having just a single subject (GLM analyses do assume equal variances within each group, though, which is what the 'unequal variances' option extends)

Best
Alfonso
Originally posted by Breanne Kearney: