We compared both models to fMRI responses while participants watched animated social interactions
Surprising result: BOTH models explained neural responses in BOTH pSTS and TPJ, even after controlling for the variance explained by the other model.
⇒Computations not strictly segregated by region.
