quote:
Dear PLS experts,
I know that it would be possible to do repeated measures for structural/functional data in PLS if one knows which factors are within subject and which are between. However I am not sure how one can do it for two sessions longitudinal structural data with considering only GM condition? Is it like one should make a datamat (from GM) per subject and per session and then include all subject's session 1 as group 1 and all subjects session 2 as group 2 for permutation?
Thanks
Klas
Hi Klaas,
The configuration would depend on how close the two time points you measure are. Strictly speaking, two time points would be a within-subjects factor only and you would enter both sessions into a single data matrix - essentially a two-condition study. You could also treat it as two groups if the time points are several years apart (I have seen this in other fields), in which case you would create two session files per subject and treat them as different 'groups'. Interestingly (sort of), in the case of this sort of design (two conditions, one group), the permutation rules are the same regardless of whether you treat it as one group measured twice or two groups. It would make a difference if you had two distinct groups with more than one within-subj factor.
I hope that made sense.
cheers
Randy