News Release

Car fanciers’ experience mental traffic jams when viewing autos and faces simultaneously

Peer-Reviewed Publication

Vanderbilt University

Aficionados may not only treat their automobiles as if they are people, but it now appears that they recognize their cars with the special part of the brain that is also used to identify faces. And, when they try to identify cars and faces at the same time, they are likely to experience a kind of perceptual traffic jam.

Those are some of the implications of research conducted at Vanderbilt University and the University of Colorado at Boulder. Researchers there compared how the brains of auto experts and novices process pictures of cars and faces. They found that viewing cars elicits signals from the brains of car experts that are just like the signals evoked by viewing faces in other brains. Moreover, the experts' skill interfered with their ability to identify faces when they were forced to process cars and faces simultaneously.

The findings, reported online on March 10 in the journal Nature Neuroscience, directly challenge the widely held view that a small, specialized area in the brain is specially hardwired to recognize faces. When confronted with a novel object, people use different parts of the brain to identify it by breaking it down into pieces. By contrast, the special facial recognition area appears to recognize faces holistically, all at one time, and does so more quickly than the piecemeal approach.

Some researchers, including Isabel Gauthier, assistant professor of psychology at Vanderbilt who co-authored the current paper, have argued that faces are not recognized in a special-purpose module but rather by a general purpose visual processor that can be trained to identify other objects holistically, not just faces.

Three years ago Gauthier published a study that showed car fanciers and bird watchers both used the facial recognition area in the brain to identify the objects of their interest in addition to faces. Last year, she published work showing that as people are trained as experts on identifying novel, computer-generated objects, they begin to recognize them holistically.

But these studies left unanswered the question of whether the same neural circuitry was involved in processing faces, birds and automobiles or whether the faces and objects were processed by different neural networks that are intermingled in the same small area in the brain. So Gauthier, working with Tim Curran, assistant professor of psychology at CU Boulder, designed a study to address this issue.

"With this study, we show that the holistic identification process takes place very early in the sequence of visual processing and that at least some of the same neural circuitry must be involved in identifying faces and other objects of extreme interest," says Gauthier.

The researchers recruited 40 men for the study, 20 car fanciers and 20 car novices. They had the subjects view alternating sequences of faces and cars and asked them to compare each car to the previous car they saw and each face to the previous face they saw. In this fashion, the person had an image of a car in his mind when he was looking at the faces. A trick the researchers used was to cut both the images of the faces and cars in half and ask the subjects to ignore the top parts of the images. By modifying the top halves of the images, they were able to measure whether the subjects looked at both the cars and faces in a holistic or piecemeal fashion.

Gauthier and Curran found that individuals with the greatest degree of car savvy recognized the cars in a holistic fashion, but this came at a cost. It reduced their ability to process faces holistically at the same time. By contrast, auto novices used the piecemeal approach to identify the cars and that didn't interfere with their ability to recognize faces in a holistically.

"This indicates that the two holistic processes are not independent," Gauthier maintains.

In order to determine the timing of the interference between holistic car and face recognition, the researchers had their subjects put on a net intermeshed with electric sensors that measured their brain waves. They took the readings from all the subjects and averaged them together to reduce individual variability using a technique called event-related potential (ERP). This allowed them to identify the timing and general location of the processing associated with both car and face recognition.

The ERP analysis found the difference in a brain wave labeled N170 that has been associated with facial recognition in previous studies. It also established that the conflict between face and car recognition in the auto experts takes place shortly after a person views an image. "This indicates that it is a basic perceptual process, not something that happens because auto experts attend to, or reason about, cars in a different way," says Gauthier.

It also located this activity in the right hemisphere in the same area where functional MRI brain scans have located the facial recognition area, known as the fusiform face area. The fMRI brain scanning technique provides higher-resolution mapping of brain activity than ERP, but does not provide information about how this activity varies over the short time periods involved in visual processing.

"The ERP results indicate that holistic processing of faces and cars by experts both involve fast-acting visual recognition processes that occur less than one-fifth of a second after faces or cars are seen," Curran explains.

If the brain's holistic processing capability can be applied to automobiles, which are about as visually distinct from faces as possible, then it should be possible to train it to identify almost any type of object, the researchers argue.

Kim Curby, a doctoral student at Vanderbilt and Daniel Collins, a research assistant at CU-Boulder, also contributed to the study.

###

The research was funded by the James S. McDonnell Foundation, the National Institutes of Health and the National Science Foundation.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.