Structural encoding precludes recognition of face parts in prosopagnosia.

The extent and the impact of spared processing of facial stimuli in the prosopagnosic patient LH is examined using the inversion effect and the face context effect. Our study asked how the deficit in individual face recognition is related to two perceptual abilities that are spared in this patient but between which there is interference when both are applied to the face stimulus, i.e. structural encoding of the face and parts-based matching procedures. Three experiments studied this relationship with task demands and stimulus properties designed to trigger the parts-based processes. In the first experiment, human and animal faces are presented upright or inverted with good performance only for the inverted condition. In Experiment 2 normals show a clear face context effect (matching of upright faces easier than scrambled or inverted ones) in the full face matching task whereas in the parts matching task the face superiority effect disappears. In contrast, LH shows a face inferiority effect when matching full faces but also when matching an isolated face part to a face part in a full face context. The results show that structural encoding of the face overrules parts-based procedures that could otherwise be helpful to tell individual faces apart.