The role of depth of encoding in attentional capture