Hi
I computed several classical GLM at the first level with the SPM option to perform a bayesian estimation allowing to get log-model evidences for each model within each voxel of each participant. Then, I want to compare them with the Bayesian Model Selection (BMS) tool at the second level. However, I realized that the models with the highest number of regressors were always winning (capturing the variance in all the grey matter). Therefore, I wanted to know if this is because they are better indeed to explain the variance of the whole brain or that a correction for the number of regressors is not implemented in the BMS tool?
Thanks in advance for any help on this!
Best
Nicolas
|