Hello! I am running a ‘cross validated’ RSA using fMRI data —i.e., examining the similarity of multivoxel patterns from a certain brain region from each trial with every other trial (except for trials from the same run) and I would like to examine these data using multidimensional scaling (MDS). However, it seems like MDS expects a square (n x n) matrix—which mine isn’t because of this ‘cross validated’ setup. Is there an easy solution for this beyond averaging across subjects?