Abstract
Traditional models of face recognition view identity and
recognition of expression as implemented in separate systems, a view reinforced by the notion that prosopagnosia is a selective loss of the ability to recognize facial identity and is due to a loss of configuralprocessing
ability. Yet, not much attention has been devoted to investigating the role of configural processing in recognition of facial expressions either in normal viewers or in prosopagnosics. We investigated the influence of facial organization on discrimination of a local facial
feature, using a feature-matching task and comparing matching performance for either upright or inverted target presentation. Stimuli consisted of faces and face parts with or without expression, but task requirements were identical in both conditions. Normal viewers are equally sensitive to the face configuration whether the stimuli are neutral faces of face expressions. In contrast, prosopagnosic patients present an effect of the configuration only when there is a facial expression
present.
