Divided attention to grasp-relevant locations