Recognizing faces, a critical social skill, is among the most challenging perceptual and memory tasks the human visual system performs. Some patients with cerebral lesions suffer from prosopagnosia, a relatively selective failure to recognize faces. The overall aim of our work is to advance our understanding of the basis of this unusual failure in different patients, both in its functional and anatomic aspects, information that is important also for cognitive theories of normal face processing. We focus on several key issues. First, in some patients the recognition problem is due to a failure to perceive the structure of a face. By using facial stimuli with carefully manipulated changes, we will characterize the type of structural encoding that is lacking. These focus upon theorized distinctions between features, external contour, internal facial geometry, and normative rules of facial structure. We compare these data to parallel studies of normal subjects viewing inverted faces, a maneuver said to disable the normal adult human expertise with faces. Our protocols will address not only what is seen in faces but also how these data are processed in both normal subjects and patients. We will also use non-facial stimuli to determine if the encoding defect is selective for faces or involves other objects, a point of theoretical importance debated in both the patient and functional imaging literature. To clarify the specificity of the perceptual encoding defects in prosopagnosia that we uncover, we perform parallel studies in similar patients with medial occipitotemporal damage but not prosopagnosia. We use both behavioral studies to characterize their function, and functional imaging to characterize the structure-function correlations in both prosopagnosic and non-prosopagnosic controls. Second, we focus on another stage in cognitive models of face processing, facial memories. We use imagery to examine not only the status of facial memories, but the type of facial memories present (feature- or configuration-based). We hypothesize that anterior temporal structures may be critical to this function, and propose a study of patients with temporal lobectomy to address the question of the anatomic correlate of facial memories. Last, we integrate the perceptual and imagery data in prosopagnosia in a study of covert recognition. From our prior work, we hypothesize that covert or unconscious recognition is the residual product of a partially damaged face network in the brain. We propose a functional imaging experiment to test this hypothesis.
Showing the most recent 10 out of 35 publications
