Covert person recognition: its fadeout in a case of temporal lobe degeneration.

Covert person recognition was investigated longitudinally over a three-year period in a patient suffering from "Crossmodal Familiar Person Agnosia", possibly due to a fronto-temporal dementia in its right temporal variant (Gentileschi et al., 2001). The progressive neuronal degeneration in the cortical regions critical for face recognition (viz., right infero-temporal areas) presented us with the opportunity to check Burton et al.'s (1991) and Farah et al.'s (1993) hypothesis on the dissociation between overt and covert face recognition in a neuropsychological condition which, however, is neurologically and cognitively different from that of focal "associative prosopagnosia". Covert person recognition starting from overtly unrecognised faces was assessed by means of learning tasks of face/name association involving celebrities. It was assumed that some unconsciously spared information would selectively enhance the relearning rates when famous faces were paired with their true names. In fact, the true-name advantage (i.e., selective saving for experimental relearning of true name pairings) reached significance at first assessment, carried out five years from clinical onset. Effect faded away two and three years later on, thus abolishing the overt/covert dissociation in face recognition. These findings support Burton et al.'s (1991) and Farah et al.'s (1993) hypothesis of covert face recognition as the consequence of partial and incomplete activation of person semantics, due, in the present case, to the impoverishment of Gentileschi et al.'s (2001) "exemplar semantics" storehouse. Moreover, it turned out that covert recognition does not imply a different learning slope, but an overall different level of the learning profile.