Neural representations of graspable objects: are tools special?