Ask a scientist about the link between facial expressions, language and cognition

COLUMBUS, Ohio—On Friday, April 29, Aleix Martinez will answer questions in a live online Q&A session on the social media site reddit. He'll discuss how he identified the "not face"—a universal expression of disapproval—as well as the implications of his research for human cognition and the origin of language.

Go to the reddit science AMA page on Friday morning to submit your questions, and tune in at 1 p.m. ET to see the answers and join the discussion.

Aleix Martinez

As a cognitive scientist, Martinez wants to understand how the human brain works. But his methods are different than you might think.

"I take a computational view," said the professor of electrical and computer engineering at The Ohio State University. "This means I assume the brain performs a series of computations, similar to those executed by a computer, to decide how to interpret and respond to some input or situation."

In computational terms, a facial expression encodes information, and researchers have long wondered how exactly the brain is so good at decoding that information. 

"You can think about my work as identifying the algorithm the brain uses to solve a problem," he added.

Previously, Martinez showed that facial expressions of emotion can be universally understood, and the brain may interpret them in a way similar to spoken words. In his latest research, he and his team have created a machine learning algorithm that can identify the facial expression a person is looking at 60 percent of the time, based solely on the image of his or her brain in a functional magnetic resonance imaging (fMRI) machine. The fMRI images enabled them to pinpoint, for the first time, the site in the brain that decodes facial expressions.