WASHINGTON, D.C. September 15 -- Some hair is missing. Nevertheless, you recognize John McEnroe, the TV announcer who once ruled tennis with his intense serve and tantrums. New research indicates that ability to identify a face occurs in a special brain system. And surprisingly, the system is separate from the general purpose visual processing system that identifies objects such as tennis balls, sneakers, and stadiums.
"The results provide one of the more startling demonstrations that the brain is comprised of highly specialized processing areas, even for the perception of complex stimuli such as faces," says the study's lead author, Morris Moscovitch, of Rotman Research Institute at Baycrest Centre for Geriatric Care and the University of Toronto at Mississauga, Ontario.
Moscovitch's study, funded primarily by the Medical Research Council of Canada, is published in the September issue of the Journal of Cognitive Neuroscience.
"The study presents such a clear and striking dissociation between face and object recognition," says Daniel Schacter, a memory expert at Harvard University. "This dissociation was suspected before, but this research is an exceptionally clear example."
In the study, researchers performed 19 tests on a man known as CK who sustained brain damage in a road accident. CK can no longer read and has great difficulty recognizing common objects such as animals, flowers, trees, cars, furniture and utensils. "What is remarkable is that despite these difficulties, his ability to recognize faces seems to be completely normal, even though recognizing faces is much more difficult than recognizing that a small, straight, yellow object with a dark, pointy tip is a pencil," says Moscovitch. "The stark contrast between CK's normal face recognition and poor object recognition suggests that faces are special and that there are different brain systems devoted to recognizing faces and objects."
In the experiments, CK reviewed groups of pictures and revealed that he could recognize a face under a variety of circumstances. The researchers found that CK could recognize faces in photos even when they were altered with wigs, facial hair, glasses and hats, or age. In addition, he could identify family resemblance between faces. "We also learned that the face-recognition system is not specialized for dealing only with the full human face, but can detect any reasonable facsimile of it," says Moscovitch. For example, CK could recognize caricatures of famous people and the faces of cartoon characters.
Additional results indicate that the face-recognition system requires a visual code for activation. "The code is the upright configuration of facial features, especially the mouth, nose and eyes -- a facial template of sorts," says Moscovitch. "Anything that conforms to the code, even an arrangement of objects, will trigger the face-recognition system." For example, researchers found that CK could recognize faces comprised of different foods by the Italian painter Giuseppe Arcimboldo. While CK often was not aware that the pictures contained foods (objects), he knew that a large mushroom posing as a mouth, a turnip as a nose, and garlic as eyes, together depicted a face. On the other hand, CK's recognition of faces plummeted when the code was disturbed. He had difficulty making an identification if the face was inverted, fractured or if the top and bottom halves were misaligned.
The researchers plan to identify more precisely the visual code of a face and locate the brain regions that constitute the face-recognition system. In addition, they plan to determine how the face- and object-recognition systems interact to produce a unified, seamless visual perception of the world.
Moscovitch's co-authors were Gordon Winocur of Rotman Research Institute and Trent University and Marlene Behrmann of Carnegie-Mellon University. Moscovitch is a member of the Society for Neuroscience, an organization of more than 27,000 basic scientists and clinicians who study the brain and nervous system.
Materials provided by Society For Neuroscience. Note: Content may be edited for style and length.
Cite This Page: