Scientists from Harvard and MIT have discovered that human brains tend to perceive a face as either male or female depending on where it appears in our field of view, according to a recently published study.

On a computer, the team generated a number of faces ranging from extremely female, to androgynous, to extremely male. They then flashed these faces at various locations on a computer screen and recorded which gender the subjects observed.

The researchers discovered that more androgynous-looking faces, when shown in certain parts of the visual field, would sometimes look male and sometimes look female.

But the researchers also found that the regions that recognize a given face as male or female differ from person to person. That is, a face that looks male in a certain portion of the visual field might not appear male to another person when displayed in the same visual field.

“You expect the same face to look the same everywhere,” said Arash Afraz, a post-doctoral fellow at MIT and the lead author of the study. “When you examine it carefully you notice a huge difference.”

The study’s findings challenge the previously accepted notion that “visual perception is homogenous” or that an object always looks the same to a brain no matter where it is located in the visual field.

The researchers also concluded that the ability of a person to accurately identify an object depends on the size of that object.

It is much easier to recognize the gender of a large face than of a small one. For simpler objects like a line, the object need not be as large for a person to recognize it as a line.

In addition to gender, the study tested items like perceived age and weight of the faces. It found that visual field biases existed in these areas as well, although they were independent of gender bias.

According to co-author Patrick Cavanagh, a professor at the University of Paris and a research professor at Harvard, the study raises questions about the accuracy of eyewitness testimony.

“If, in a study, we present a man, and people are saying it’s a woman, that’s not very reassuring in terms of eyewitness testimony,” Cavanagh said.

According to Afraz, people have maintained these biases because they have no evolutionary need to get rid of them.

Because there is no reason behind their existence, Afraz said that such biases “can be taken as more evidence against intelligent design.”

Cavanagh said that as a result of this study, scientists should be able to measure the area of the brain within which analysis is made.

The findings will appear in a forthcoming issue of the journal Current Biology.