Encoding differences in visual and haptic face recognition