Baku, February 22, AZERTAC

The decoding of a brain region that distinguishes one face from another may offer clues to the very nature of consciousness, according to Nature. In the spring of 1995, Nancy Kanwisher was granted access to the functional magnetic resonance imaging (fMRI) machine at the Massachusetts General Hospital. After making little headway using the nascent technology to explore the visual perception of shape, she climbed inside the machine, gazed at photos of people’s faces, and had her coworkers scan her brain.

What Kanwisher discovered would launch a 30-year journey to unravel how humans perform a vital social task: making sense of each other’s faces. Those early experiments revealed a small patch of increased blood flow in a brain region that responded more vigorously to images of faces than random objects. That first scan showed “a promising blob on the bottom of my right hemisphere,” Kanwisher writes in a 2017 retrospective about the work. To be certain, she had her colleagues scan her again and again. “To our delight, the trusty little blob showed up in exactly the same place every time.”

Like other seminal discoveries, Kanwisher’s finding made sense in hindsight. Faces hold a special significance for us and for our primate relatives. “We are social creatures,” says Winrich Freiwald, of The Rockefeller University, who joined Kanwisher’s lab as a postdoc in 2001. “Faces are so important for those social interactions that finding a specialized circuitry for faces seemed a safe bet.”

In the decades since Kanwisher, now at MIT, spotted that first, promising blob, that bet has paid off in spectacular fashion. Discovery of these face-specific systems in humans and monkeys launched an intensive avenue of research that has deconstructed how the brain identifies and analyzes faces. In 2024, the groundbreaking work earned Kanwisher, Freiwald, and Doris Tsao of the University of California, Berkeley, the 2024 Kavli Prize in Neuroscience.

“Nancy Kanwisher is a real pioneer: Her original paper has about 7,000 citations, which is just astronomical,” says neuroscientist Bruno Rossion of the University of Lorraine in France, who studies face recognition in humans and did not participate in the work. “And she has since taken a leading role in defining the function of these regions in human face recognition.”

Tsao and Freiwald’s subsequent work localizing the equivalent face-specific regions in monkeys and then recording the activity of individual neurons in that network with electrodes “changed the field and allowed them to make real progress in understanding how these neurons were coding for various properties of faces”, Rossion says.

A neural code

Kanwisher’s findings were particularly exciting since they demonstrated that the brain’s architecture could be functionally compartmentalized. “This question had been debated heatedly in our field for nearly 200 years,” she writes in her 2017 retrospective. “And now here was a little piece of the brain that seemed to do just one thing: perceive faces.”

Deciphering how the cells within this patch perform this remarkable feat, however, was not easy to do in the human brain, where low-resolution fMRI imaging could only detect the collective response of hundreds of thousands of neurons. That’s where Tsao and Freiwald come in.

The pair started by performing fMRI imaging on macaques. Tsao even reached out to a friend at MIT to design a seat that would allow the monkeys to recline horizontally inside the scanner, “a chair I still use in my lab today”, she says. With their subjects in a comfortable sphinx position, it didn’t take long for Tsao and Freiwald to locate an area that lit up specifically when they showed macaques pictures of faces. The next step was to dip into that site with an electrode and locate the cells that were responsible.

“The very first time, the first cell we recorded from was face selective,” says Tsao. So was the next.

“Around the fifth or sixth cell,” Freiwald adds, ‘we looked at each other and said, ‘This is incredible!’ All the cells we've isolated so far are face selective.”

That discovery enabled them to do more sophisticated studies: determining which orientation or facial features individual cells were tuned to recognize and mapping the interconnections and functional specializations within the macaque face-patch system.

As a faculty member at Caltech, Tsao and one of her postdocs would go on to discover a neural code for faces, designing an algorithm that would allow them to predict with uncanny accuracy which face a macaque was looking at based on the responses of just 200 neurons. And Freiwald and his team at The Rockefeller University identified nearby brain regions that respond most strongly to faces that are familiar.

The rudiments of recognition

Despite this stunning progress, much about face recognition remains to be decoded. “Surprisingly basic questions remain unanswered about the information represented in each region, the computations it conducts, and the connectivity between that region and the rest of the brain,” writes Kanwisher.

For his part, Freiwald is fascinated by what the brain does with all the information we glean when we gaze at a face: “How do you go from the perception of a face to making inferences about the emotional state of others? Are they angry? Are they happy? What are they feeling and thinking—and what will they do next?” For that matter, how do we reliably recognize the same person when we see them out of context or after they get a haircut or remove their glasses, he asks.

“We're now in a much better place to tackle more complex questions about other aspects of face perception and processing because of the work that Nancy, Doris and Winrich have done over the past 20-plus years,” says Brad Duchaine, a neuroscientist who studies face recognition at Dartmouth College. Understanding how the brain unpacks facial expressions, identifies individuals, and recognizes the familiar smile of a loved one offers a way to explore the broader processes of cognition and memory—perhaps even consciousness.

“That’s a question we’re actively investigating in my lab,” Tsao says. “We want to understand how the brain takes visual information and uses it to generate a crystal-clear, coherent perception of the world. That will give us a toehold into understanding these high-level questions about cognition, language, consciousness and thought.”

And it all begins with the face. “The brain is a system for recognition—that’s its main function,” Rossion says. “The day that we fully understand how the human brain performs face identity recognition, we'll have made major progress in understanding how the how the human brain works.”