Testing key predictions of the associative account of mirror neurons in humans using multivariate pattern analysis: Commentary on Cook et al