From Scalpels to Scanners: Studying the Brains Language System


The methodical, incisive Paul Broca was an influential physician-scientist in mid-19th-century France. He made many discoveries about the mechanisms of human diseases like cancer, muscular dystrophy, and the illness he termed aphémie—the loss of speech.

On a spring day in 1861, a 51-year-old man was admitted to the ward of Paul Broca, a respected surgeon and anatomist, at Paris’s Bicêtre Hospital. The patient suffered serious gangrene in his right leg. His entire right side, in fact, had been paralyzed for years. Although the man could move his face and tongue, he was also unable to speak. Well, mostly.

 The methodical, incisive Paul Broca was an influential physician-scientist in mid-19th-century France. He made many discoveries about the mechanisms of human diseases like cancer, muscular dystrophy, and the illness he termed aphémie--the loss of speech.

“He could no longer produce but a single syllable, which he usually repeated twice in succession; regardless of the question asked him,” wrote Broca in a report to the Société d’Anatomie (Anatomical Society) four months later. “He always responded: tan, tan, combined with varied expressive gestures. This is why, throughout the hospital, he is known only by the name Tan.”

Tan seemed to register what he was told, but he could not articulate a reply. After Tan died, his mysterious affliction was settled via scalpel. When the skull was autopsied, Broca saw a lesiona diseased cavitylarge enough “to hold a chicken egg” in the lower left part of the frontal lobe of Tan’s brain. During the following two years, Broca carefully described seven more speech-impaired patients, all of whom had nearly identical brain damage. He concluded that this left-hemisphere region was the seat of language. It was duly named Broca’s area.

Broca’s cases settled a prickly argument among the intellectual elite at Paris’s scientific societies. Many believed that the spongy, seemingly uniform tissue of the brain performed uniform functions. With his evidence, Paul Broca challenged the establishment and laid a founding principle of neuroscience: that specific parts of the brain do specific things.  

Broca’s wild idea--called localization--has stood the test of time. It’s been upheld by every successive revolution in neuroscience, including the latest: the era of brain imaging. Today, high-powered scanners can peer directly into the complex brains of living, thinking, talking peopleno scalpel needed.



Above is the preserved brain of Tan, Paul Broca's first aphémie patient. The lesion is clearly visible on the bottom left side. It was so obvious, in fact, that Broca didn’t touch his scalpel to the interior of Tan’s brain. Instead, he preserved the organ intact in alcohol, and entrusted it to the collections of Paris’s medical museum, the Musée Dupuytren, for posterity.

If Broca had done a dissection, neuroscientists could have speculated in 1861 that brain regions beyond his namesake area also help humans speak. In 2007, a University of California research team did a virtual dissection of Tan’s preserved brain using an imaging technique called volumetric MRI. The high-resolution scans showed that the lesions extended far deeper than Broca had reported. The cavity even reached through a large bundle of neuron fibers that links language areas in the front and the back of the brain. Broca’s area, as it’s defined, may therefore play a smaller role in articulating speech than the surgeon imagined.

Photo: Dronkers, et al., 2007

Uniquely Human

Our species is distinguished by our capacity for language. Many animals can communicate elaborately. Cuttlefish, for example, send signals by changing the color patterns of their skin, and honeybees “dance” to publicize prime locations for feeding and nesting. Chimpanzees, our closest living relatives, have a complex system of gestures to exchange messages. Intriguingly, this body language stems from the counterpart of Broca’s area in the chimpanzee brain. Still, no other kind of animal can infinitely combine symbols--words, in our case--to relate abstract concepts such as plans and ideas.

“Humans are sort of like chimpanzee 2.0,” says Gary Marcus, a psychologist at New York University and an expert in childhood language development. “Our brains are built on some fairly complicated machinery that’s already there. It’s like the icing on a cake that was already in the oven for a while.” It’s not surprising that scientists have long been drawn to languagea swirl of icing, if you willto decipher how the human brain works in general.

Opening the Black Box

Until recently, scientists thought that our language control center had two parts: Broca’s area and Wernicke’s area, which helps us understand what we hear. Modern brain scans are showing, however, that language recruits many different regions. “It’s much more complicated than anybody realized,” says Marcus. “I think that’s because language itself is so complicated. It needs to have the full participation of the brain to be able to do this amazing feat.” 

One method neuroscientists use to unscramble this puzzle is functional magnetic resonance imaging, or fMRI. The technique shows what different brain areas do. It relates structure to function, as Paul Broca did. Had fMRI been around in his time, he surely would have put Tan in the scanner straightaway. (In fact, in 2007, neuroscientists did scan Tan: See sidebar.)

Before fMRI became widely available in the late 1990’s, scientistsBroca includedtreated the brain as a black box, says Joy Hirsch, the director of Columbia University’s brain imaging center, the Program in Imaging and Cognitive Sciences. “We knew what went in and we knew what came out,” Hirsch explains. “We could only infer properties of brain organization.”

How does fMRI map this organization? First, a subject lies down and places his or her head into the scanner, a large tube that emits a harmless magnetic field. To image the subject’s language system, the researcher might set up a simple experiment. One example is an “object-naming task”: the subject looks at pictures of objects and names them as the scanner runs.


An fMRI scanner at Columbia University's Program for Imaging and Cognitive Sciences.


During any mental effort, the brain regions involved recruit oxygenated blood as fuel. When the blood carries more oxygen, its magnetic properties change. This activity disrupts the scanner’s magnetic field. The regions that change in oxygenthus, activity“light up” on the scanner’s computer screen. Scientists like Hirsch can get a complete map of the language system handling a specific task in five minutes.

Beyond Broca

Studies using fMRI have led to new insights about our vast capacity for language that autopsy or analysis of language-impaired people, by themselves, cannot. With fMRI, neuroscientists have confirmed that Broca’s and Wernicke’s areas are indeed waypoints for language processing. But they don’t act alone. Other brain regions, such as the auditory cortex and the visual cortex, have proven critical to the language system as well. Furthermore, not all of the language system resides exclusively in the left hemisphere, as Broca believed.  

Neurologists have also discovered that the human language system works like a circuit. If one part of the circuit activates, the whole system “fires up,” ready to operate. Hirsch’s lab and others have noted this during scanner tasks where the subject listens to a story read aloud. Just listening, explains Hirsch, will not only activate the mental apparatus to process sounds, including Wernicke’s area, but it will also jump-start Broca’s areathe speech production centereven though the subject isn’t actually speaking.


An fMRI scan indicates what regions of the brain are active during a particular task. This scan shows activity in one "slice" of a person's brain as he calls out the names of everyday objects.

Columbia University PICS

Additionally, Hirsch’s team has learned that a child’s language system is fully formed long before he or she is verbal. This was clear when the team scanned babies with the fMRI to map their brains for surgery. During the scans, the parents calmed their babies with soothing words. “The process of listening to their parents speak actually activates an expected cascade of language areas as if language system was fully operational,” says Hirsch. “But the child hasn’t learned to speak yet.” These findings are reinforced by Marcus’s studies of how young children recognize the grammatical patterns of human language. His lab has found that even seven-month-old babies are attuned to the roots of grammar.

Mapping the Brain in Action

An fMRI scan can show a lot. But since the technique does not image neurons directly, it cannot reveal how the parts of the brain’s language system signal each other. It can’t outline the routes of the signals or the order in which they travel. “The technologies that we really want as brain scientists don’t exist yet,” says Marcus. “My fantasy brain technology would look at individual neurons in the course of us speaking sentences and look at how those neurons communicate together.” 

For the time being, researchers approximate the connections and timing of the language system by coupling fMRI with other technologies. One is an electroencephalogram, or EEG. This device records the brain’s electrical activity, a result of the firing of neurons. When neuroscientists apply EEG sensors to a subject’s head while in the scanner, the electrical patterns suggest which parts of the language system fire first, next, and last.

Diffusion imaging is a cutting-edge technology that shows the major thoroughfares of neurons that link disparate brain areas. A diffusion-imaging scanner delineates axonsthe long fibers that extend from the neuron’s cell body to transmit impulses to the next cells in line. It works by sensing how water molecules found naturally in brain tissue move along axons. Diffusion imaging cannot detect which neurons are firing. But when scientists compare a diffusion image with an fMRI scan, they can identify the location of the fibers that connect the clusters of activity that lit up during the language task.

The Next Brain Revolution

Above is a human brain scan from a diffusion spectrum imaging machine. It reveals the major bundles of nerve fibers. The colors indicate the general direction that the fibers follow.

Van Wedeen

As brain imaging continues to innovate, scientists may have a shot at answering one of neuroscience’s million-dollar questions. “If our brains look so much like the brains of our close nonprimate relatives, why do we speak and they do not?” asks Hirsch. “There are a number of hypotheses that we and others are exploring.” Presently, the strongest hypothesis is that our brains are more interconnected. Researchers eagerly await the day when neurotechnologies can produce complete, high-resolution signaling maps of animal brains to compare. Paul Broca, if he could, would no doubt be among the first in line to see them.