Action recognition of point-light displays presented with semantically (in)congruent auditory stimuli.