Visual contrast drives face recognition, study finds

The answer to a long-standing mystery in visual neuroscience may also help explain how people with autism perceive faces, according to a study published in March in the Proceedings of the National Academy of Sciences¹.

People generally excel at recognizing faces and facial expressions, and scientists have long wondered whether there are distinguishing visual or structural features in a face that make this process so reliable.

The new study explored one of the few instances in which people can fail to recognize a face: in the negative of a photograph. Despite the fact that all of the visual information — such as the shapes of facial features and the locations of edges — are intact in a negative photo, most people find it difficult to identify faces in negatives.

“Our new study suggests that a large part of the answer might lie in the brain’s reliance on a certain kind of image feature,” says lead researcher Pawan Sinha, associate professor in the department of brain and cognitive sciences at the Massachusetts Institute of Technology.

For most people, the researchers found, that image feature is the contrast between the brightness of the eyes and the darkness of the areas surrounding them.

The practical applications of the work range from computerized face recognition systems in high-security settings, such as airports, to new experimental tools for characterizing face perception in people with autism, Sinha says.

Face perception is an important thrust of autism research because people with autism are known to avoid eye contact² and focus more on the mouth region than on the eyes³ to pick up emotional cues.

Understanding how people recognize faces might help researchers develop behavioral interventions for children with autism to shift their perceptual emphasis back to the eyes.

In the study, the researchers showed each of 15 participants 24 images of well-known celebrity faces. One group of five participants viewed the faces as photographic negatives; a second set saw the faces as silhouettes with only the eyes visible; and a third group viewed them as ‘contrast chimeras,’ or photographic negatives in which only the eye regions are reversed to positive and look as they normally would in a picture.

Only 54 percent of the participants who looked at full negatives and 13 percent of those viewing the silhouettes were able to recognize the celebrity faces. In contrast, more than 92 percent of the contrast chimera group recognized the faces correctly, suggesting that the eye region of the face plays an important role in recognition.

In the second part of the study, ten new participants viewed all the images in every experimental condition. Again, researchers found that the ability to recognize faces is highest when the images are contrast chimeras.

Contrast chimeras:

Functional magnetic resonance imaging of eight participants showed that brain areas normally associated with face processing — known as the fusiform gyri — are more active when subjects look at the contrast chimeras than when they look at the photographic negatives.

Similar imaging experiments, such as reversing only the mouth region of a negative photograph to positive, could help confirm whether children with autism focus more on the mouth than on the eyes, Sinha notes. For instance, if the fusiform gyri light up when children with autism look at mouth-based chimeras, that would suggest that the children primarily use the mouth to read faces.

Scientists could then conceivably design interventions that artificially enhance the salience of the eye region relative to the mouth — by boosting image contrast, color or sharpness, for example — in order to slowly shift a child’s emphasis back to the eyes. In gradually becoming more comfortable with gazing at the eye region, a child might become more adept at picking up on emotions that are better expressed by the eyes, Sinha says.

Researchers have used other techniques to show that people with autism tend to process the mouth rather than the eyes when reading facial expressions⁴, but Sinha’s contrast chimeras could still prove useful to the field, says David Simmons, lecturer in psychology at the University of Glasgow.

“It would certainly be interesting to see if people with autism spectrum disorder were more sensitive to mouth than eye contrast chimeras, even if it is only confirmatory,” Simmons says.

Others are more cautious, saying that the new work may inform autism research, but only to a point.

“While extending these kinds of experiments to autism will give us insight into how faces and other social signals are processed in autism, it is unlikely to really explain the cause of abnormal face processing, or directly lead to interventional strategies,” says Ralph Adolphs, professor of psychology and neuroscience at the California Institute of Technology.

“The reason is that we need to go a step back and ask why people with autism do not make normal eye contact in the first place,” Adolphs adds. “We need to understand the source of abnormal social motivation during early development.”

References: