2024 Kavli
Prize in
Neuroscience

Citation from the Committee

Recognizing faces is important for social interaction in many animals. Previous work in human psychology, clinical studies of brain-injured patients, positron emission tomography studies, and isolated face-selective neurons in non-human primates, had suggested the existence of a functionally specialized system for face recognition. However, face recognition had not been localized to any specific area of the brain. The present laureates used functional magnetic resonance imaging (fMRI) to localize different areas in neocortex specialized for face processing.

Nancy Kanwisher pioneered the establishment of the functional region of interest (fROI) approach to localize the fusiform face area (FFA) in humans using fMRI. Kanwisher was the first to develop and employ a paradigm to identify a region sensitive to faces in each person. This finding strongly supported the idea of modular localization of cognitive function in the neocortex. The use of such functional localizers is now widespread and applied to domains beyond the face recognition system.

Inspired by Kanwisher´s findings, Winrich Freiwald and Doris Tsao together used fMRI to localize similar face patches in macaque monkeys. Having localized these, they recorded from single neurons in each patch. They showed that the overwhelming majority of visually responsive neurons in the largest such region were face-selective. They proceeded to outline a system of multiple face patches, detailing their interconnections and functional specialization. Face recognition in the earliest patches was dependent on viewpoint, but later became viewpoint-independent through a series of processing stages. Winrich Freiwald in further work characterized populations of cells selectively responsive to faces familiar to the viewer. Doris Tsao identified different features of the face that make up a code enabling single cells to identify faces.

Together, the laureates, with their work on neocortical specialization for face recognition, have provided basic principles of neural organization which will further our understanding of recognition of objects and scenes.

Laureates bring the brain’s response to faces into focus

In a collective effort spanning decades, this year’s laureates have uncovered the neural mechanisms of one of the brain’s most complex tasks: responding to faces. In the early days of functional brain imaging, Nancy Kanwisher pinpointed the brain’s center for face processing, answering a longstanding question about whether some brain regions specialize in specific tasks. Doris Tsao and Winrich Freiwald then expertly combined functional imaging and recording from individual brain cells in macaque monkeys to reveal a six-region system that compiles facial information into a complete picture. Taken together, the laureates' work allows a glimpse of the neural architecture of the human mind.

By Rachel Zamzow, science writer

The biggest leaps in science often spring forth from a collective effort. Researchers build upon each other’s work, adding their own expertise, to produce a deeper understanding. This collaboration encapsulates the work of this year’s laureates for the 2024 Kavli Prize in Neuroscience: Nancy Kanwisher, Doris Tsao, and Winrich Freiwald. “No one of them could have taken it as far,” says awards committee member and neuroscientist John O’Keefe. A combined effort also defines their landmark findings of how the brain processes faces—by compiling information across distinct regions to create a complex picture.

In the early days of functional brain imaging in the 1990s, Kanwisher was fascinated by the possibility of probing the human mind by measuring blood flow to an active brain region. But it wasn’t easy to break into the emerging field, and Kanwisher’s first string of studies failed. As a last-ditch effort, she tested how the brain responds to faces, capitalizing on its privileged role in social interactions.

Kanwisher scanned her own brain while she viewed a series of images, some human faces and some other objects, such as tools and food. To her surprise, a small region on the underside of her brain on the right side responded much more strongly to faces than the other images.

But when Kanwisher scanned the brains of more people to replicate her results, she realized that this face-selective region, called the fusiform face area, was in slightly different locations in each individual. To correct for this, Kanwisher devised an analysis technique that first localizes that region in each person’s brain and then queries the region's response to different images. The approach, published with her face area findings in 1997, is now widely used in brain imaging studies.

Over several years of experiments, Kanwisher and her colleagues explored just how selective the fusiform face area is to faces. They found that the region responds strongly to different kinds of faces, such as those of animals, cartoon faces, and abstract drawings of faces, and much less so to images of body parts or the backs of heads. The researchers showed in 2020 that the region even responds in blind people while they run their fingers over models of human faces, suggesting that vision isn’t required for the effect.

Kanwisher’s findings addressed a longstanding debate among neuroscientists about how the brain works, O’Keefe says. Some believed it functioned as a distributed processing network like a computer, while others thought it was divided into localized modules, each responsible for a specific task. Though research would later prove that the brain has both characteristics, the fusiform face area was a clear example of the latter.

As such, Kanwisher and her colleagues have continued to discover many other functionally specialized brain regions, such as those that respond to places, bodies, sentence meaning, and music.

But when it comes to capturing the brain’s intricate circuitry, brain imaging is a rather blunt instrument. Digging deeper requires the help of animal models, in which researchers can track the activity of individual brain cells in real-time.

Patching Faces Together
Like Kanwisher, Tsao didn’t start out exploring how the brain responds to faces. Instead, she was interested in how the brain’s visual areas conjure a three-dimensional picture of the world. But after reading Kanwisher’s seminal paper describing a human brain region devoted to faces, Tsao saw a unique opportunity to ask a similar question with a lot more precision: How does a specific brain area decode an image as complex as a face?

Tsao teamed up with Freiwald, who was completing a postdoctoral fellowship under Kanwisher’s mentorship. Freiwald was pursuing an interest in how the brain focuses attention on objects when a colleague put him touch with Tsao.

Together, the researchers used brain imaging to test how macaque monkeys’ brains respond to faces. In 2008, they mapped what turned out to be six distinct brain regions, known as the face patch system, which respond selectively to faces more so than other objects.

The pair then recorded the activity of individual brain cells within these face patches by inserting targeted electrodes. They found that these cells had a remarkable proclivity for faces—firing with an explosion of sound when their activity was converted to an audio signal—and staying relatively silent in response to other images.

With further experiments, Tsao and her team untangled the connections of the monkey face patch system by tracing how dyes and electrical stimulation in one face area travel to others. Tsao and Freiwald revealed in 2010 how cells in some face patches specialize in faces with particular views, such as those looking straight ahead or to the left or right.

Tsao also identified how the face patches work together, assembling information across 50 or so dimensions, to identify a face. Some cells, for example, respond preferentially to the presence of hair or the distance between the eyes. In 2017, Tsao and a post doc used this neural code to predict which faces monkeys were viewing, just by analyzing the activity of around 200 brain cells. Tsao’s team is now exploring how the brain might perceive other objects in a similar way and ultimately create comprehensive visual scenes.

In his own work, Freiwald uncovered in 2021 that a separate brain region, called the temporal pole, accelerates our recognition of familiar faces. And in subsequent studies, he has started to unravel the complex brain circuitry underlying social interactions.

Taken together, the laureates’ work is “a beautiful model of how you go about analyzing the way in which the brain represents a particular feature of the world,” O’Keefe says. By carefully charting the brain’s face system and beyond, they’ve allowed a glimpse into the complex architecture of the human mind.