For much of human history, the source of human intelligence and individual character was thought to have been the heart, the liver, or the spleen—not the brain. Before and after the mind was linked to the brain, the supposed significance of the organ has shaped how it is represented—both as a body part and as the locus of the self. Images of the brain have for the most part been, and still are, speculative, thanks to the opaque relationship between the organ and its functions. The subjective experience of consciousness—dynamic, diachronic and synchronic—cannot readily be transposed onto the brain's physiology. The kidney, by contrast, filters and secretes a fluid with properties that can be correlated and classified according to smell, color, and sedimentation, leaving traces of a time-based process with a clear beginning and end. Diagnosis from urine, practiced for many centuries, is a deductive process based on the commonsensical intelligibility of this process: intake, excretion, repeat.
We tend to think of contemporary, digital images of the brain—the beguiling, arresting concoctions derived from functional magnetic resonance imaging (fMRI) machines—as evidence of “activity,” regardless of the complicated mathematical operations involved in their production. We encounter them in reports about “your brain on poker” or “your brain on sex.” The use of brain scans to trumpet what are often insignificant and sensational studies has demystified the discipline and helped provoke a backlash against so-called neurophrenology, and against the use of neuroscience as an explanatory panacea. You cannot blame neuroscientists for hyperbole: The search for the site of consciousness has preoccupied humanity for a long time. And bright, easily readable brain scans have liberated neuroscientists from the tedium of slicing tissue samples in lab coats, giving them a sexy, public voice. I don’t intend to scold neuroscientists for making rash claims. After all, as much as they might be expected to note the speculative, notional character of an fMRI scan, those of us who “consume” such images must curtail easy assumptions about what “mapping” the brain actually means.
The ubiquity of brain scans creates what philosopher of science Gaston Bachelard calls an “epistemological obstacle” to the development of scientific knowledge. In The Formation of the Scientific Mind (1938), he writes, “it can very clearly be seen that an over-familiar scientific idea becomes weighed down by too much psychological concreteness, amassing too many analogies, images, and metaphors, and gradually losing its vector of abstraction, its sharp abstract point.” He warns of complacency, advocating a constant renewal of scientific thought by remaining vigilant about accepted truths.
Looking at historical representations of the brain and its function, from medieval diagrams to rare Japanese watercolors of dissected brains, from the perspective of medical methodology and visual strategy, might provide a corrective for the complacency of which Bachelard warns. Many of these historical representations seem quite foreign to the contemporary viewer, and compel us to acknowledge both the continuity and contingency of the inquiry into the nature and origin of thought; they challenge us to see contemporary images of the brain as if they, too, were unfamiliar. In doing so, we might ask what the digital images of multicolored floating brains we so often confront today really show us and what we really want to see: a tangle of neural activity that approximates the mystery of consciousness, or a piece of flesh that anchors consciousness in material processes? And has it always been this way?
The brain is a fragile organ, and extricating it without making a mess requires considerable anatomical knowledge and surgical skill. Once removed from its bony vessel, the brain immediately quits functioning and loses its shape: The ventricles collapse, the gyri dent and deform. Like custard, the brain can scarcely be touched without being dimpled, ripped, squished. This fragility of the brain helps explain why there was apparently no attempt to accurately depict the organ intact before 1500, since there is documentation of basic brain surgery (trepanation) going back to the Neolithic period. Brain surgery is described in medical treatises from Hippocratic times onward, and representations of medieval medical practice often include images of physicians attending to head wounds and removing parts of the skull. But the distinctive image of the brain that we have become used to—shaped roughly like a walnut and bifurcated, also like that nut, with a twisting surface of hills and valleys, or gyri and sulci—is relatively new.
The images that dominated the medieval era are not based on the morphology of the brain, but rather on a schematic representation of the ventricular system that consists of a human head divided into three or four “cells,” which are often connected to suggest some kind of mental process. These shapes are a kind of shorthand for the four fluid-filled spaces deep in the brain, which comprised a metaphoric model of the mind that remained virtually uncontested in the West or East for a thousand years.1 The classification of these cells varies among authors, but generally they stand for perception (sensus communis and imaginatio), thought (ratio, fantasia), higher thought (cogitativa), and memory (vis aestimativa and memoria).
The apparent unwillingness to show the brain as a material organ is sometimes considered indicative of a lack of medical sophistication; for a long time, medieval depictions were negatively compared to those of the Italian Renaissance, which were championed as pinnacles of rational humanism. But neither the visual conventions of the Renaissance nor those of the medieval era provide a clear correlation between aesthetic strategy and medical knowledge. And representing the brain metaphorically, with its functions abstracted, conforms closely to medieval philosophical ideas. Scholastic thought considered observation deceptive, open to misinterpretation, and therefore altogether unreliable. “Medieval thinkers concerned themselves with large systems, asking not if something—be it knowledge received from antiquity, a phenomenon observed, an event described in the Bible—were true but how to explain that it must be true,” Melanie Holcomb writes in Pen and Parchment. “They sought elegance in their explanations because they understood the cosmos to be an orderly, harmonious enterprise created by God.” Knowledge was to be deduced logically from a general principle.
Several of the most famous medieval illustrations of brain function integrate the cell doctrine into a system that also includes references to the five senses. And sometimes they collate layers of knowledge from diverse sources—astrology, cosmology, Christian doctrine, humoral theory, myth—into geometric systems, organizing diverse kinds of information into one pictorial system. Often these sketches were made on empty pages within manuscripts, probably as memory aides.
The fifteenth-century schematic drawing known as MS Canon Misc. 366 is relatively simple: The brain is a circle separated into five rows of boxes, which form cells that correspond to the main tenets of the popular theory (though in this case regarding animals). The image is abstract, notational, with a minimal amount of anatomical information—the words “nervi otici” mark two lines as crossing optic nerves, thus identifying the circular shape as the brain. Its simplicity highlights the advantage schematic images can have over mimetic drawings: Such diagrams may be abstract, symbolic, simple, but they are powerful enough to subsume large quantities of data if necessary. Their very openness points to the limitations of what we know about the workings of the mind, and foregrounds the graphic transposition of knowledge between disciplines or epistemic media (writing, mathematics, drawing). In this sense, the medieval cell models are not unlike contemporary brain scans, which are composites of different kinds of knowledge and different image-making systems.
This image, taken from a later copy of a thirteenth-century Islamic medical treatise, Khalīfah ibn Abī al-Maḥāsin al-Ḥalabī’s The Sufficient Book on Ophthalmology, combines detailed anatomical description with oblique, abstracted references to physical shapes, complicating the relationship between anatomy and metaphor. Two circular figures (the eyes) and an incomplete ellipse (the skull) are connected by an X-shaped structure (bisecting optic nerves). Although the image is a diagram, the references to the body are specific: The concentric circles refer to parts of the eye, the contour line of the ellipse refers to the pericranium and the skull, the twin lines just inside the ellipse refer to the dura mater and pia mater (membranes covering the brain), and the vertical lines and the triangular shape refer to the skull’s sutures (fibrous joints that allow for minimal movement in adults). The writing within the image and around its border refers to anatomical details and to “the five powers” ostensibly located in the ventricles.
Although the drawing is highly stylized, the anatomical information is precise and intricate in comparison with other medieval images. Physicians in the medieval Islamic world may have conducted surgery and dissection, as evidenced in their corrections and annotations of Galen’s work. Yet their depictions of the brain—like those of their Western contemporaries—remained abstract. In surveys of Islamic medicine, Western historians have often belittled medieval Islamic texts as derivative of lost Greek manuscripts. Despite the absence of Western medieval images of greater or equivalent anatomical accuracy, Karl Sudhoff (who wrote about the drawing in 1914) singled out geometric abstraction as a symptom of Islamic prohibitions against dissection and image-making. Sudhoff compounded his cultural bias with a visual bias, which led him to ignore the originality and sophistication of this image.
Only in the final stage of my writing did I receive a scan of Khalīfah's manuscript from a Turkish library. Much to my surprise, upon opening the file I realized that the diagram has been reproduced upside down ever since first being published in Julius Hirschberg’s 1905 German treatise on the history of Arab ophthalmology. (This may be the first time the orientation is correct.) Why was this remarkable reversal, whether intentional or accidental, accepted for so long? I haven't been able to find an answer; all other diagrams of this kind, as Hirschberg and Sudhoff undoubtedly knew, show the eyes below the head or cells.
By the end of the fifteenth century, Renaissance visual practice began to inflect medical illustration. Concurrently, humanist scholars and physicians became increasingly committed to empirical research. Together with technological advances in printing, this led to a surge in the publication of anatomical treatises. The representation of the brain as a morphological object, cradled in the skull, reflected the high regard in which observation, verisimilitude, and perspective were held. At the same time, these depictions ceased alluding to any physiological process or to mental function. The human body was either depicted as manifestly dead object of study or mockingly shown as a sentient being evidently unaware of its wounds.
Nevertheless, observation was still influenced by expectation. By the second century Galen had identified the rete mirabile, a web-like network of blood vessels found in some ungulates at the base of the brain, while dissecting pigs and cows. Later physicians inferred that humans must also possess the rete mirabile and instilled it with a pseudo-mystical aura and the job of extracting “psychic pneuma” from the “vital pneuma,” or life force, produced in the heart. Even after its existence in humans was disputed by Berengario da Carpi in 1521, Vesalius—who of course had never seen the rete mirabile in humans—went on to depict the structure in a 1538 anatomical treatise, deferring to the general consensus that such a thing existed. Vesalius corrected this mistake in his next treatise, the groundbreaking De Humani Corporis Fabrica (1543), but the rete mirabile continued to appear in other anatomical illustrations for some time, even as the convolutions of the cortex and the brain stem made their way into manuscripts.
Standards are especially conspicuous when they shift dramatically, as they did following the 1664 publication of Thomas Willis’s Cerebri Anatome, illustrated by the famed architect (and astronomer) Christopher Wren. Willis removed the brain from the skull and studied it in the round, especially from below, then sliced through the organ to reveal its inner structures. He hypothesized that the brain stem was more important than the ventricles—which had recently been discovered to contain fluid, making them unlikely to house the mind—and so Wren’s drawings afford uncommon views of the base of the brain. As an architect, Wren was fond of classicizing shapes and elegant domes, and the fourteen drawings of the brain in Cerebri Anatome display how his sense of proportion informed his understanding of the organ’s spatial and structural logic.2 Willis’s medical advances are evident in Wren’s drawings, and yet Wren’s technical innovations—he discovered that he could harden and fix the brain matter by injecting alcohol into the soft tissue—and artistic talent allowed him to create images imbued with such spatial coherence and aesthetic authority that they in turn reinforced the credibility of the anatomist’s claims.
We treat technical images, especially those intended for teaching, differently from images created for other purposes; there is an inherent assumption that they are more objective, and perhaps more strenuously avoid subjective style or composition. But such a notion of objectivity is tied to changing conventions of seeing and constructing an image.3 This is particularly evident looking at medical illustrations, which have always tended to clean up the messiness of flesh in order to clearly depict anatomy. In the case of the brain, prior assumptions about texture, composition, and connectivity determine how the organ is pictured. Today’s brain scans claim objectivity by showing a seemingly straightforward anatomical view, evoking the tacit association between mimesis and scientific impartiality. But the actual standard of what appears naturalistic at any moment is not based on observation but convention; it is specific to historical, professional, and cultural contexts and always serves a purpose—even in its alleged absence, naturalism.
“I sawed across the skullbone, laid open the brain and then studied it carefully,” wrote the Japanese physician Kawaguchi Shinnin in Kaishihen (Complete Notes on the Dissection of Cadavers, 1172). “Further I enucleated the eye balls and could thereby discover that the optic nerve continues into the brain.” Two years earlier, after getting permission from his feudal lord, Shinnin had been given a severed head and two headless corpses from an execution to perform the second officially sanctioned dissection in Japan. The resulting book detailing his observations contains twenty-three hand-tinted woodcuts by Shunmei Yo, including four images of the head and brain.
Kawaguchi had trouble reconciling medical doctrine with what he saw: “It is written in old books that the heart is connected to the ear,” he writes. “This can be believed. Judging from the results of my dissection, one cannot understand however that the kidney connects to the ear and the liver to the eye.” Though European physicians had arrived in Japan with anatomy books, this knowledge was not considered relevant for the treatment of illness in Japan. This may explain why Kawaguchi’s teacher, Gengai Ogino, discouraged him from publishing the book, arguing the anatomical findings might simply embarrass earlier physicians but fail to produce changes in therapeutic practice.
One of the images in Kaishihen shows the head opened to reveal brain matter, with a narrow bifurcating channel running through the middle but no other regular delineation; instead, there is a tangle of lines reminiscent of a ball of thick yarn, or a squirming mass of worms. In another, brain matter appears to float on the page, connected to an eyeball that is depicted quite neatly as a circle with a tail, the optic nerve. There is no connotation of reason or authority here; in fact, the apparent disorder of the brain matter appears in stark contrast to the orderly depiction of the eye and the precisely drawn line leading to the caption. These images diverge from Western conventions of medical illustration and seem to indicate the insignificance of the brain in Japanese medicine of the time. But we know too little to draw conclusions about these images that are not overly determined by our own conventions; the image may reflect stylistic traditions of Japanese printmaking rather than anatomical concepts. Even so, the vivid illustrations of Kawaguchi’s laconic description of the shape of the brain as “similar to that of a chicken’s intestines” provide an alluringly different, but also unsettling, view of where the self presumably resides. Examining the unease produced by the images in Kaishihen helps explain the lure of the orderly and reassuringly immaterial brain scans we favor today.
The cerebral cortex only came to dominate the attention of anatomists around 1800, with the advent of Franz Joseph Gall’s theory of cranioscopy, or phrenology. Gall believed that character traits could be mapped onto parts of the cortex; quantities of virtues and vices could be gauged by charting elevations or depressions of the overlying cranium. The mapping of “temperaments” in pictures and corresponding busts enjoyed a great vogue, especially in the US. The popular phrenological images showing these correlations are fantastical combinations of diagrammatic and metaphoric projection. And the illustrative models, which offer an intuitive understanding of the self, have survived in today’s images of the brain.
Nineteenth-century science was preoccupied with disease and alienation, and therefore interested in identifying and classifying all manner of criminal, racial, and sexual aberrations. The examination of the brain with new investigative tools—result of microscope, electricity, tissue staining and cortical stimulation—promised answers. By the middle of the century, scientists had produced intricate maps of the brain, charting the convolutions of the cortex in orderly patterns. They proceeded from speculation about the functions of specific areas of the brain, sometimes obtained by studying brain-damaged patients, to experimentation: stimulating certain regions of the cortex and registering sensory and motor responses. Eventually, this combination allowed scientists to deduce localization, which led to a new kind of topography of the brain.
Concurrently with the mapping of the brain, anatomists studied its histology. But brain tissue presented a challenge: At the time, nerve cells were not visualized as discrete entities. That changed in 1873 when Camillo Golgi developed a selective staining technique that made it possible to observe isolated neurons; soon after, Santiago Ramon y Cajal refined Golgi’s technique and illuminated discrete nerve cells, dendrites, and axons, paving the way for an understanding of “communication” between cells.4 The drawings Ramon y Cajal created are stunningly beautiful, yet he was disdainful of what he considered the aestheticization of scientific images; he insisted that he only represented what he saw. However, the life-like—almost excitable—quality of neurons in his drawings, and the degree to which he imbues them with a material presence, seems to foreshadow the discovery of their crucial but elusive role as functional signaling units.
In 1924, following the earlier discovery that electrical currents from the surface of living animal brains stem from synchronous activity of nerve cells, Hans Berger developed electroencephalography to record so-called brain waves. The study of electrical impulses opened an entirely new realm, eventually enabling scientists to “animate” the brain. Thus an effect of the functional brain portrait, which linked measurable physical actions to the workings of the brain, was to give the impression of agency: Whereas the brain had been depicted from the outside before, now the organ appeared to leave indexical traces of its activity. Though fMRIs are functional inasmuch as they track neurological processes, they are also topographical—they characterize individual sections of the brain based on the premise that increased blood flow indicates activity. Which is to say they localize based on the quantification of physiological processes and assumptions of causality.
The information provided by such time-based images always depends on a particular framework; just as an EEG only has expressive value in relation to a baseline—established parameters that give relevance to the results—today’s brain scans are interpreted using conventional sets of data. Reducing the “noise” to “smooth” data is only one of the externally imposed conventions that order results so as to produce coherent, seamless images. Like all scientific experiments, the experimental setup of the scans determines the results; the questions delineate the range of answers.
Digital imaging is particularly vulnerable in this regard because its representations are almost entirely based on computational analysis and the mathematical manipulation of information. Articulating the blood-flow data within images derived from other sources—X-ray, photography, computer animation—provides the appearance of an action occurring within the brain. But in fact the morphological references to the brain are mere ornamentation.
What is wrong with making such captivating images? First, their power—and tendency—to signify knowledge far exceeds the findings, which are so often speculative and unspectacular. The current passion for brain scans expresses an unrealistic positivism in regard to the explanatory power of empirical science and the ability of images to transparently communicate knowledge. Second, these images exploit conventional signifiers of knowledge—graphs, grids, blueprint-like 3-D structures—while employing sophisticated techniques of visual manipulation to make the images “easier” to read and more expressive. The images tend to connote “science,” and yet they dumb down what should be more overtly complex; they are methodologically incoherent.
While we cannot in real life untether ourselves from the material world, the ubiquitous floating brain achieves a measure of transcendence. Such images contribute to what the philosopher Colin McGinn, in a recent critique of language that humanizes neurons, calls “surreptitious homunculus talk” that “generates an illusion of theoretical understanding”—not only of the way the brain works, but of how it creates the self. If we want to avoid the lulling effects of Bachelard's epistemological obstacle, we need to disavow that unproductive illusion.
The Human Brain Project was recently selected as one of two Future and Emerging Technologies flagship projects by the European Commission. The goal is to use supercomputers to build the equivalent of a human brain using models and simulations—a network that mimics the interactions of 100 billion neurons and of the 100 trillion connections between them. “We now have a blueprint to understand the molecular architecture that defines the development and organization of the human brain,” one scientist confidently says in an HBP promotional video. “It’s like building a giant telescope to peer into deep space, only we’ll be able to look deep into the brain.” Fittingly, the analogy makes a misleading association between empirical knowledge and the ability to understand the overall dynamics of a complex system: The greater the level of magnification, the greater the epistemic reward.
Over a hundred years ago, in The Interpretation of Dreams, Sigmund Freud also used the analogy of a telescope to make roughly the opposite conclusion:
Ideas, thoughts and psychical structures in general must never be regarded as localized in organic elements of the nervous systems but rather, as one might say, between them, where resistances and facilitations provide the corresponding correlates. Everything that can be an object of our internal perception is virtual, like the image produced in a telescope by the passage of light-rays.
Freud works to dispel the spatial and topographic connotations of the language used to describe the systems of conscious and unconscious thought, and thus to discourage us from imagining the psychic apparatus in terms of anatomical or metaphorical localization. He insists that thinking is a dynamic process; a thought does not move from one place (conscious) to the other (unconscious), but undergoes Besetzung, or a change in energy (which Freud translator James Strachey unfortunately labels “cathexis”).
Freud made diagrams of this psychic apparatus—a dynamic system superimposed on a whimsical brain-like shape, reminiscent of the entanglement of anatomy and metaphor found in illustrations of medieval cell doctrine. (His compositions are collected in From Neurology to Psychoanalysis: Sigmund Freud’s Neurological Drawings and Diagrams of the Mind, edited by Mark Solms and Lynn Gamwell and published in 2006.) The drawing that accompanies one of Freud’s most influential metapsychological works, “The Ego and the Id” (1923), is peculiarly ambiguous. Freud inscribes elements of the psychic apparatus on a faint, almost quaint brain-shaped figure. While the description is abstract, Freud employs concrete references: The perception of the outside world enters through W–Bw, passes through the Ego into the Id.
Freud alludes to the anatomy of the brain and explicitly refers to a “cap of hearing”; he jokes about the Ego wearing that cap cocked. The strange “tunnel” on the right side of the drawing, where the repressed ("Vdgt," meaning Verdrängtes) enters the Id but is kept apart from the Ego, might be a nod to the vermis (the “worm” connecting two cells) evident in those medieval representations. Although Freud explicitly notes that the pictorial representation should not be interpreted apart from its depiction of relationships, his whimsical references encourage the viewer to do just that.
But one does not have to go as far as invoking the bête noire of current neuroscience for a cautionary note on neuroanatomical determinism. In Mind, Language and Society: Philosophy in the Real World, after provisionally defining consciousness as a biological phenomenon, John Searle observes, “In the case of consciousness,” unlike other biological phenomena such as mitosis, meiosis, or digestion, “we have an irreducible subjective element left after we have given a complete causal account of the neurobiological basis.” But scientists are undeterred, and continue to seek answers to all questions in the most elemental units, reducing biology to chemistry, chemistry to physics, and so on. Paradoxically, the explanation of Searle’s “irreducible subjective element” eludes these efforts; even as the magnification level increases, the answer remains just out of reach.
More than anything, looking at historical representations of the brain reveals their absolute contingency: Every period has its own tools and instruments—intellectual and physical—and their limitations never seem to be fully grasped. Only when the next tool comes along, or the paradigm shifts, do changes in the questions we ask also make possible other answers.