Distributed representations of dynamic facial expressions in the superior temporal sulcus