The areas of the cortex that were first pinpointed to be important for cognition were concerned with language. These come from studies of aphasia, a language disorder that most often occurs when certain areas of brain tissue are destroyed by a stroke, the occlusion or rupture of a blood vessel to a portion of a cerebral hemisphere. Many of the important discoveries in the study of aphasia occurred in rapid succession during the last half of the 19th century. Taken together, these advances form one of the most exciting and important chapters in the neural science of human behavior.
Pierre Paul Broca, a French neurologist, was the first to identify specific areas of the brain concerned with language. Broca was influenced by Gall's efforts to map higher functions in the brain, but instead of correlating behavior with bumps on the skull he correlated clinical evidence of aphasia with brain lesions discovered post mortem. In 1861 he wrote, "I had thought that if there were ever a phrenological science, it would be the phrenology of convolutions (in the cortex), and not the phrenology of bumps (on the head)." Based on this insight Broca founded neuropsychology, a science of mental processes that he distinguished from the phrenology of Gall.
In 1861 Broca described a patient, Leborgne, who as a result of a stroke could not speak, although he could understand language perfectly well. This patient had no motor deficits of the tongue, mouth, or vocal cords that would affect his ability to speak. In fact, he could utter isolated words, whistle, and sing a melody without difficulty. But he could not speak grammatically or create complete sentences, nor could he express ideas in writing. Postmortem examination of this patient's brain showed a lesion in the posterior region of the frontal lobe, now called Broca's area (Figure 1–4B). Broca studied eight similar patients, all with lesions in this region, and in each case the lesion was located in the left cerebral hemisphere. This discovery led Broca to announce in 1864: "Nous parlons avec l'hémisphère gauche!" (We speak with the left hemisphere!)
Broca's work stimulated a search for cortical sites associated with other specific behaviors—a search soon rewarded. In 1870 Gustav Fritsch and Eduard Hitzig galvanized the scientific community when they showed that characteristic limb movements of dogs, such as extending a paw, could be produced by electrically stimulating discrete regions of the precentral gyrus. These regions were invariably located in the contralateral motor cortex. Thus the right hand, the one most used for writing and skilled movements, is controlled by the left hemisphere, the same hemisphere that controls speech. In most people, therefore, the left hemisphere is regarded as dominant.
The next step was taken in 1876 by Karl Wernicke, who at age 26 published a now-classic paper, "The Symptom-Complex of Aphasia: A Psychological Study on an Anatomical Basis." In it he described another type of aphasia, a failure of comprehension rather than speech: a receptive as opposed to an expressive malfunction. Whereas Broca's patients could understand language but not speak, Wernicke's patient could form words but could not understand language. Moreover, the locus of this new type of aphasia was different from that described by Broca: The lesion occurred in the posterior part of the cortex where the temporal lobe meets the parietal and occipital lobes (Figure 1–4B).
On the basis of this discovery, and the work of Broca, Fritsch, and Hitzig, Wernicke formulated a neural model of language that attempted to reconcile and extend the two predominant theories of brain function at that time. Phrenologists and cellular connectionists argued that the cortex was a mosaic of functionally specific areas, whereas the holistic aggregate-field school claimed that every mental function involved the entire cerebral cortex. Wernicke proposed that only the most basic mental functions, those concerned with simple perceptual and motor activities, are mediated by neurons in discrete local areas of the cortex. More complex cognitive functions, he argued, result from interconnections between several functional sites. In placing the principle of localized function within a connectionist framework, Wernicke realized that different components of a single behavior are likely to be processed in several regions of the brain. He was thus the first to advance the idea of distributed processing, now a central tenet of neural science.
Wernicke postulated that language involves separate motor and sensory programs, each governed by distinct regions of cortex. He proposed that the motor program that governs the mouth movements for speech is located in Broca's area, suitably situated in front of that region of the motor area that controls the mouth, tongue, palate, and vocal cords (Figure 1–4B). He next assigned the sensory program that governs word perception to the temporal-lobe area that he had discovered, which is now called Wernicke's area. This region is conveniently surrounded by the auditory cortex and by areas now known collectively as association cortex, a region of cortex that integrates auditory, visual, and somatic sensations.
Thus Wernicke formulated the first coherent neural model for language that—with important modifications and elaborations we shall encounter in Chapter 60— is still useful today. According to this model, the initial steps in neural processing of spoken or written words occur in separate sensory areas of the cortex specialized for auditory or visual information. This information is then conveyed to a cortical association area, the angular gyrus, specialized for processing both auditory and visual information. Here, according to Wernicke, spoken or written words are transformed into a neural sensory code shared by both speech and writing. This representation is conveyed to Wernicke's area, where it is recognized as language and associated with meaning. It is also relayed to Broca's area, which contains the rules, or grammar, for transforming the sensory representation into a motor representation that can be realized as spoken or written language. When this transformation from sensory to motor representation cannot take place, the patient loses the ability to speak and write.
The power of Wernicke's model was not only its completeness but also its predictive utility. This model correctly predicted a third type of aphasia, one that results from disconnection. Here the receptive and expressive zones for speech are intact, but the neuronal fibers that connect them are destroyed. This conduction aphasia, as it is now called, is characterized by an incorrect use of words (paraphasia). Patients with conduction aphasia understand words that they hear and read and have no motor difficulties when they speak. Yet they cannot speak coherently; they omit parts of words or substitute incorrect sounds and in particular they have difficulties repeating phrases. Although painfully aware of their own errors, they are unable to put them right.
Inspired in part by Wernicke's advances and led by the anatomist Korbinian Brodmann, a new school of cortical localization arose in Germany at the beginning of the 20th century, one that distinguished functional areas of the cortex based on the shapes of cells and variations in their layered arrangement. Using this cytoarchitectonic method, Brodmann distinguished 52 anatomically and functionally distinct areas in the human cerebral cortex (Figure 1–5).
Brodmann's division of the human cerebral cortex into 52 discrete functional areas. Brodmann identified these areas on the basis of distinctive nerve cell structures and characteristic arrangements of cell layers.
This scheme is still widely used today and is continually updated. Several areas defined by Brodmann have been found to control specific brain functions. For instance, area 4 is the motor cortex, responsible for voluntary movement. Areas 1, 2, and 3 constitute the primary somatosensory cortex, which receives sensory information primarily from the skin and joints. Area 17 is the primary visual cortex, which receives sensory signals from the eyes and relays them to other areas for further processing. Areas 41 and 42 constitute the primary auditory cortex. The drawing shows only areas visible on the outer surface of the cortex.
Even though the biological evidence for functionally discrete areas in the cortex was compelling, the aggregate-field view of the brain, not cellular connectionism, continued to dominate experimental thinking and clinical practice until 1950. This surprising state of affairs owed much to several prominent neural scientists who advocated for the aggregate-field view, among them the British neurologist Henry Head, the German neuropsychologist Kurt Goldstein, the Russian behavioral physiologist Ivan Pavlov, and the American psychologist Karl Lashley.
Most influential was Lashley, who was deeply skeptical of the cytoarchitectonic approach to functional mapping of the cortex. "The 'ideal' architectonic map is nearly worthless," Lashley wrote. "The area subdivisions are in large part anatomically meaningless, and misleading as to the presumptive functional divisions of the cortex." His skepticism was reinforced by his studies of the effects of various brain lesions on the ability of rats to learn to run a maze. From these studies Lashley concluded that the severity of a learning defect depended on the size of the lesion, not on its precise location. Disillusioned, Lashley—and after him many other psychologists—concluded that learning and other higher mental functions have no special locus in the brain and consequently cannot be attributed to specific collections of neurons.
On the basis of his observations Lashley reformulated the aggregate-field view by further minimizing the role of individual neurons, specific neuronal connections, and even specific brain regions in the production of specific behavior. According to Lashley's theory of mass action, it is the full mass of the brain, not its regional components, that is crucial to function. Applying this idea to aphasia, Head and Goldstein asserted based on their clinical studies that language disorders can result from injury to almost any cortical area.
Lashley's experiments with rats and Head's clinical observations have now been reinterpreted. A variety of studies have shown that the maze-learning used by Lashley is unsuited to the search for local cortical functions because it involves so many motor and sensory capabilities. Deprived of one sensory capability, say vision, a rat can still learn to run a maze using touch or smell. Besides, as we shall see later in the book, many mental functions are mediated by more than one region or neuronal pathway. Thus a given function may show anatomical redundancy and not be eliminated by a single lesion.
Soon the evidence for localization of function became overwhelming. Beginning in the late 1930s, Edgar Adrian in England and Wade Marshall and Philip Bard in the United States discovered that touching different parts of a cat's body elicits electrical activity in distinct regions of the cerebral cortex. By systematically probing the body surface they established a precise map of the body surface in specific areas of the cerebral cortex described by Brodmann. This result showed that functionally distinct areas of cortex can be defined unambiguously according to anatomical criteria such as cell type and cell layering, connections of cells, and—most importantly—behavioral function. As we shall see in later chapters, functional specialization is a key organizing principle in the cerebral cortex, extending even to individual columns of cells within a functional area. Indeed, the brain is divided into many more functional regions than Brodmann envisaged.
More refined methods have now made it possible to learn even more about the function of different brain regions involved in language. In the late 1950s Wilder Penfield, and later George Ojemann, reinvestigated the cortical areas that produce language. While locally anesthetized during brain surgery for epilepsy, awake patients were asked to name objects (or use language in other ways) while different areas of the exposed cortex were stimulated with small electrodes. If an area of the cortex was critical for language, application of the electrical stimulus blocked the patient's ability to name objects. In this way Penfield and Ojemann were able to confirm—in the living, awake, and conscious brain—the language areas of the cortex described by Broca and Wernicke. In addition, Ojemann discovered other sites essential for language, in particular the insula, a region that lies deep to Broca's area. As we shall learn in Chapter 60 the neural networks for language are far more extensive and complex than those described by Broca and Wernicke.
Initially almost everything known about the anatomical organization of language came from studies of patients with brain lesions. Today positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) allow anatomical analysis to be conducted on healthy people engaged in reading, speaking, and thinking (Chapter 20). Functional MRI, a noninvasive imaging technique for visualizing activity in the brain, has not only confirmed that reading and speaking activate different brain areas but has also revealed that the act of thinking about a word's meaning in the absence of sensory inputs activates a still different area in the left frontal cortex (Figure 1–6).
Specific regions of the cortex involved in the recognition of a spoken or written word can be identified with positron emission tomography (PET).
Each of the four images of the human brain shown here (from the left side of the cerebrum) actually represents the averaged brain activity of several normal subjects. In these PET images white indicates areas of highest activity, red and yellow quite high activity, and blue and gray the areas of minimal activity. The "input" component of language (reading or hearing a word) activates the regions of the brain shown in A and B. The "output" component of language (speech or thought) activates the regions shown in C and D. (Reproduced, with permission, from Cathy Price.)
A. The reading of a single word produces a response both in the primary visual cortex and in the visual association cortex (see Figure 1–5).
B. Hearing a word activates the temporal cortex and the junction of the temporal-parietal cortex (see Figure 1–2). The same list of words used in the reading test (A) was used in the listening test. The results of the reading and listening tests show that the brain does not use the auditory pathway to convey a transformed visual signal.
C. Subjects were asked to repeat a word presented through earphones or on a screen. The spoken word activates the supplementary motor area of the medial frontal cortex. Broca's area is activated whether the word is heard or read. Thus both visual and auditory pathways converge on Broca's area, the common site for the motor articulation of speech.
D. Subjects were asked to respond to the word "brain" with an appropriate verb, for example, "to think." This type of task activates the frontal cortex as well as Broca's and Wernicke's areas. These areas play a role in all cognition and abstract representation.
In separate studies, Joy Hirsch and her colleagues, and Mariacristina Musso, Andrea Moro, and their colleagues used fMRI to explore more deeply Wernicke's idea that Broca's area contains the grammatical rules of language. Hirsch and her colleagues made the interesting discovery that processing of one's native language and processing of a second language occur in distinct regions within Broca's area. If the second language is acquired in adulthood, it is represented in a region separate from that which represents the native language. If the second language is acquired early, however, both the native language and the second language are represented in a common region in Broca's area. These studies indicate that the age at which a language is acquired is a significant factor in determining the functional organization of Broca's area. In contrast, there is no evidence of such separate processing of different languages in Wernicke's area (Figure 1–7).
Functional magnetic resonance images of the brains of bilingual subjects during generation of narratives in two languages.
These bilingual subjects were either "early" or "late" bilingual speakers; "early" bilinguals had learned two languages together prior to the age of 7 years, whereas "late" bilinguals acquired a second language after age 11 years. Axial slices of brain that intersect Broca's area are shown for one representative "early" bilingual subject and one representative "late" bilingual subject. Regions of the brain that responded during the narrative tasks are shown in red (native language) and yellow (native and second languages). In high-resolution views of these areas centroids of activity associated with each language are indicated by (+) and areas of overlap between the two areas are shown in orange. (Reproduced, with permission, from Kim et al. 1997.)
A. In the early bilingual subject the locations of the centers and spread of activity for both languages are indistinguishable at the resolution of functional magnetic resonance imaging (fMRI) (1.5 × 1.5 mm) as indicated by the close proximity of the two (+) and the orange region indicating sensitivity to both languages.
B. In the late bilingual subject both the locations of the centers (+) and the spread of the native and second language are distinguishable at the same resolution.
Further evidence for the fundamental role of Broca's area in processing grammatical rules emerges from the fMRI studies of Musso and Moro on the language instinct. Because language is a uniquely human capability, Charles Darwin suggested that the acquisition of language is an inborn instinct comparable to that for upright posture. Children acquire the grammar of their native language simply by listening to their parents speak. They do not have to be taught the specific rules of grammar. In 1960 the linguist Noam Chomsky elaborated on Darwin's notion. He proposed that children acquire a language so easily and naturally because humans, unlike other primates, have the innate capability of generalizing to a complete and coherent language from a limited sample of sentences. Based on an analysis of the structure of sentences in various languages, Chomsky argued that all natural languages share a common design, which he called universal grammar. The existence of universal grammar, he argued, implies that there is an innate system in the human brain that evolved to mediate this grammatical design of language.
This, of course, raised the question: Where in the brain does such a system reside? Is it in Broca's area, as Wernicke's model would suggest? Musso, Moro, and their colleagues asked this question and found that the region of Broca's area concerned with second language becomes established and increases in activity only when an individual learns a second language that is "natural," that is, one that shares the universal grammar. If the second language is an artificial language, a language that violates the rules of universal grammar, activity in Broca's area does not increase. Thus Broca's area must contain some kind of constraints that determine the structure of all natural languages.
Studies of patients with brain damage continue to afford important insight into how the brain is organized for language. One of the most impressive results comes from a study of deaf people who have lost their ability to communicate through American Sign Language (ASL) after suffering cerebral damage. ASL uses hand gestures rather than sound and is perceived by sight rather than sound but has the same structural complexity as spoken languages. Signing is also localized to the left hemisphere; deaf people can become aphasic for sign language as a result of lesions in the left hemisphere, but not as a result of lesions in the right hemisphere. Damage to the left hemisphere can have quite specific consequences for signing just as for spoken language, affecting sign comprehension (following damage in Wernicke's area), grammar (following damage in Broca's area), or fluency.
These observations illustrate three points. First, the cognitive processing for language occurs in the left hemisphere and is independent of pathways that process the sensory and motor modalities used in language. Second, fully functional auditory and motor systems are not necessary conditions for the emergence and operation of language capabilities in the left hemisphere. Third, spoken language represents only one of a family of language skills mediated by the left hemisphere.
Similar conclusions that the brain has distinct cognitive systems have been reached from investigations of behaviors other than language. These studies demonstrate that complex information processing requires many distinct but interconnected cortical and sub-cortical areas, each concerned with processing some particular aspects of sensory stimuli or motor movement and not others. For example, in the visual system, a dorsal cortical pathway is concerned with where an object is located in the external world while a ventral pathway is concerned with what that object is.