The Tell-Tale Brain

fusiform area (visual processing) and auditory cortex mediate the bouba-kiki effect. The cross-modal abstraction required for this probably requires initial passage through the angular gyrus. 2, interactions between the posterior language areas (including Wernicke’s area) and motor areas in or near Broca’s area. These connections (the arcuate fasciculus) are involved in cross-domain mapping between sound contours and motor maps (mediated partly by neurons with mirror-neuron-like properties) in Broca’s area. 3, cortical motor-to-motor mappings (synkinesia) caused by links between hand gestures and tongue, lip, and mouth movements in Penfield’s motor map. For example, the oral gestures for “diminutive,” “little,” “teeny-weeny,” and the French phrase “en peau” synkinetically mimic the small pincer gesture made by opposing thumb and index finger (as opposed to “large” or “enormous”). Similarly, pouting your lips outward to say “you” or (in French) “vous” mimic pointing outward.

There was an accelerated development of the left IPL in primate evolution culminating in humans. In addition, the front part of the lobule in humans (and humans alone), split into two gyri called the supramarginal gyrus and the angular gyrus. It doesn’t require deep insight to suggest therefore that the IPL and its subsequent splitting must have played a pivotal role in the emergence of functions unique to humans. Those functions, I suggest, include high-level types of abstraction.

The IPL (including the angular gyrus)—strategically located between the touch, vision, and hearing parts of the brain—evolved originally for cross-modal abstraction. But once this happened, cross-modal abstraction served as an exaptation for more high-level abstraction of the kind we humans take great pride in. And since we have two angular gyri (one in each hemisphere), they may have evolved different styles of abstraction: the right for visuospatial and body-based metaphors and abstraction, and the left for more language-based metaphors, including puns. This evolutionary framework may give neuroscience a distinct advantage over classical cognitive psychology and linguistics because it allows us to embark on a whole new program of research on the representation of language and thought in the brain.

The upper part of the IPL, the supramarginal gyrus, is also unique to humans, and is directly involved in the production, comprehension, and imitation of complex skills. Once again, these abilities are especially well developed in us compared with the great apes. When the left supramarginal gyrus is damaged, the result is apraxia, which is a fascinating disorder. A patient with apraxia is mentally normal in most respects, including his ability to understand and produce language. Yet when you ask him to mime a simple action—“pretend you are hammering a nail”—he will make a fist and bang it on the table instead of holding a “pretend” handle as you or I might. If asked to pretend he is combing his hair, he might stroke his hair with his palm or wiggle his fingers in his hair instead of “holding” and moving an imaginary comb through his hair. If requested to pretend waving goodbye, he may stare at his hand intently trying to figure out what to do or flail it around near his face. But if questioned, “What does ‘waving goodbye’ mean?” he might say, “Well, it’s what you do when you are parting company,” so obviously he clearly understands at a conceptual level what’s expected. Furthermore, his hands are not paralyzed or clumsy: He can move individual fingers as gracefully and independently as any of us. What’s missing is the ability to conjure up a vibrant, dynamic internal picture of the required action which can be used to guide the orchestration of muscle twitches to mime the action. Not surprisingly, putting the actual hammer in his hand may (as it does in some patients) lead to accurate performance since it doesn’t require him to rely on an internal image of the hammer.

Three additional points about these patients. First, they cannot judge whether someone else is performing the requested action correctly or not, reminding us that their problem lies in neither motor ability nor perception but in linking the two. Second, some patients with apraxia have difficulty imitating novel gestures produced by the examining physician. Third and most surprisingly, they are completely unaware that they themselves are miming incorrectly; there is no sign of frustration. All of these missing abilities sound compellingly reminiscent of the abilities traditionally attributed to mirror neurons. Surely it can’t be a coincidence that the IPL in monkeys is rich in mirror neurons. Based on this reasoning my postdoctoral colleague Paul McGeoch and I suggested in 2007 that apraxia is fundamentally a disorder of mirror-neuron function. Intriguingly, many autistic children also have apraxia, an unexpected link that lends support to our idea that a mirror-neuron deficit might underlie both disorders. Paul and I opened a bottle to celebrate having clinched the diagnosis.

But what caused the accelerated evolution of the IPL—and the angular gyrus part of it—in the first place? Did the selection pressure come from the need for higher forms of abstraction? Probably not. The most likely cause of its explosive development in primates was the need to achieve an exquisitely refined, fine-grained interaction between vision and muscle and joint position sense while negotiating branches on treetops. This resulted in the capacity of cross-modal abstraction, for example, when a branch is signaled as being horizontal both by the image falling on the retina and the dynamic stimulation of touch, joint, and muscle receptors in the hands.

The next step was critical: The lower part of the IPL split accidentally, possibly as a result of gene duplication, a frequent occurrence in evolution. The upper part, the supramarginal gyrus, retained the old function of its ancestral lobule—hand-eye coordination—elaborating it to the new levels of sophistication required for skilled tool use and imitation in humans. In the angular gyrus the very same computational ability set the stage (became an exaptation) for other types of abstraction as well: the ability to extract the common denominator among superficially dissimilar entities. A weeping willow looks sad because you project sadness on to it. Juliet is the sun because you can abstract certain things they have in common. Five donkeys and five apples have “fiveness” in common.

A tangential piece of evidence for this idea comes from my examination of patients who have damage to the IPL of the left hemisphere. These patients usually have anomia. (difficulty finding words), but I found that some of them failed the bouba-kiki test and were also abysmal at interpreting proverbs, often interpreting them literally instead of metaphorically. One patient I saw in India recently got 14 out of 15 proverbs wrong even though he was perfectly intelligent in other respects. Obviously this study needs to be repeated on additional patients but it promises to be a fruitful line of enquiry.

The angular gyrus is also involved in naming objects, even common objects such as comb or pig. This reminds us that a word, too, is a form of abstraction from multiple instances (for example, multiple views of a comb seen in different contexts but always serving the function of hairdressing). Sometimes they will substitute a related word (“cow” for “pig”) or try to define the word in absurdly comical ways. (One patient said “eye medicine” when I pointed to my glasses.) Even more intriguing was an observation I made in India on a fifty- year-old physician with anomia. Every Indian child learns about many gods in Indian mythology, but two great favorites are Ganesha (the elephant-headed god) and Hanuman (the monkey god) and each has an elaborate family history. When I showed him a sculpture of Hanuman, he picked it up, scrutinized it, and misidentified it as Ganesha, which belongs to the same category, namely god. But when I asked him to tell me more about the sculpture, which he continued to inspect, he said it was the son of Shiva and Parvati—a statement that is true for Ganesha, not Hanuman. It’s as if the mere act of mislabeling the sculpture vetoed its visual appearance, causing him to give incorrect attributes to Hanuman! Thus the name of an object, far from being just any other attribute of the object, seems to be a magic key that opens a whole treasury of meanings associated with the object. I can’t think of a simpler explanation for this phenomenon, but the existence of such unsolved mysteries fuels my interest in neurology just as much as the explanations for which we can generate and test specific hypotheses.

LET US TURN now to the aspect of language that is most unequivocally human: syntax. The so-called syntactic structure, which I mentioned earlier, gives human language its enormous range and flexibility. It seems to have evolved rules that are intrinsic to this system, rules that no ape has been able to master but every human language has. How did this particular aspect of language evolve? The answer comes, once again, from the exaptation principle—the notion that adaptation to one specific function becomes assimilated into another, entirely different function. One intriguing possibility is that the hierarchical tree structure of syntax may have evolved from a more