Further evidence for a non-cortical origin of mirror movements after stroke.