Specialised structural descriptions for human body parts: Evidence from autotopagnosia

Previous accounts of autotopagnosia (e.g., Ogden, 1985; Pick, 1908; Semenza, 1988) propose that the disorder is attributable to deficits in "mental images," visual body schema, or semantic representations. A recent account (Sirigu, Grafman, Bressler, & Sunderland, 1991b) posits deficits in visual structural descriptions of the human body and its parts, in the context of spared semantic and proprioceptivespatio-motor body representations, but provides no evidence bearing on the nature or format of the putatively damaged representation. We report data from a man with autotopagnosia consequent to lefthemisphere brain damage which bear directly on the nature of the representation impaired in the disorder. The subject, GL, is unable to localise body parts on himself or others, whether cued by verbal or visual input. In contrast, he uses body parts precisely in reaching and grasping tasks, correctly matches items of clothing to body parts, and localises the parts of animals and man-made objects without error. We also demonstrate that GL is unable to match pictured or real human body parts across shifts in orientation or changes in visual appearance, but can perform analogous matching tasks with animal body parts and man-made object parts. The data extend the account of Sirigu et al. (1991b) in suggesting that human body part localisation depends upon structural descriptions of human (but not animal) bodies that enable viewpoint-independent body part recognition and participate in the calculation of equivalence between the body parts of self and others across transformations in orientation.