Dear SPMers,
We ran source reconstruction on the same MEG dataset with different versions of SPM8 (v4010, v5236) and got drastically different results. The main step of difference seems to be group inversion (calling the function "spm_eeg_inv_group"). We are wondering whether SPM users have encountered such issues, whether there are underlying changes in the algorithm that causes such big differences in estimated sources, and what the recommendations will be regarding version differences.
Thanks so much!
Best,
Lily
|