>which are known to produce predictable fMRI signal changes in distributed brain regions.

Wondering how they created that baseline. Was it with fMRI data (which has deviance from actual data, as pointed out)? Or was it through other means?