The Inheritance, Power and Predicaments of the “Brain-Reading” Metaphor
- First Online:
- Cite this article as:
- Gilbert, F., Burns, L. & Krahn, T. Medicine Studies (2011) 2: 229. doi:10.1007/s12376-010-0054-0
With the increasing sophistication of neuroimaging technologies in medicine, new language is being sought to make sense of the findings. The aim of this paper is to explore whether the “brain-reading” metaphor used to convey current medical or neurobiological findings imports unintended significations that do not necessarily reflect the genuine findings made by physicians and neuroscientists.
First, the paper surveys the ambiguities of the readability metaphor, drawing from the history of science and medicine, paying special attention to the sixteenth through nineteenth centuries. Next, the paper addresses more closely the issue of how metaphors may be confusing when used in medicine in general, and neuroscience in particular. The paper then explores the possible misleading effects associated with the contemporary use of the “brain-reading” metaphor in neuroimaging research.
Rather than breaking new ground, what we see in current scientific language is a persistence of both a constraining and expansive set of language practices forming a relatively continuous tradition linking current neuroimaging to past scientific investigations into the brain.
The use of the readability metaphor thus carries with it both positive and negative effects. Physicians and neuroscientists must resort to the use of terms already laden with abstracted meanings, and often burdened by tradition, at the risk of importing through these words connotations that do not tally with the sought-after objectivity of empirical science.
KeywordsBrainMedical imagingMetaphorNeuroimagingBrain reading
Textual evidence of how various authors write about “reading the brain”, “reading the mind”
The following examples are the results of an internet search of journals from the Nature Publishing Group archives (January 2006–April 2009). The table is meant only as a sampling of evidence of the brain/mind-reading metaphor. We have chosen Nature Publishing Group simply because it is one of the highest impact factor publishers in the field of science in general as well as neuroscience in particular. As such, it can be regarded as a very influential media that both reflects and shapes the relevant discourse(s) on this subject. This table is simply provided as a set of examples and is not meant to be understood as representative of the discourse of the “brain-reading” metaphor. The locutions searched include “reading the brain”, “reading mind”, “neural code”, and “neural signature”. Any occurrence of mind-reading referred to as “empathy” has not been included in this list.
Part 1: examples (not exhaustive) of metaphors referring to “brain reading”
Year of publication
“Neurotechnologies such as brain reading” (Conti and Corbellini 2008)
“Developments in neuroimaging (…) including the reading of brain states” (De Charms 2008)
“Possibility of directly reading aspects of a person’s brain-activation state” (De Charms 2008)
“General brain-reading device that could reconstruct a picture” (Kay et al. 2008)
“Useful in reading the brain states” (De Charms 2008)
“rtfMRI can potentially read complex brain states” (De Charms 2008)
“Applications for reading brain signals” (De Charms 2008)
“‘Brain-reading’ algorithms” (Op de Beeck et al. 2008)
“Decode a person’s conscious experience based on (…) ‘brain reading’“(Haynes and Rees 2006)
Part 2: examples (not exhaustive) of metaphors referring to “mind reading”
Year of publication
“Reading out a person’s thoughts does exist” (Anonymous 2009)
“Mind-reading with a brain scan” (Smith 2008)
“rtfMRI increases (…) our ability to ‘read’ mental states by decoding this information” (De Charms 2008)
“Directly reading a person’s ongoing mental images” (De Charms, 2008)
“Related methods, with the goal of mind reading” (De Charms, 2008)
“Such feats of rudimentary ‘mind-reading’“(Owen and Coleman 2008)
“Explicitly explores ‘mind-reading’ but with respect to tracking microsaccades” (Martinez-Conde and Macknik 2007)
“New ways of enhancing, controlling and reading the mind.” (Stefansson 2007)
“The possibility of “reading” a person’s thoughts (…) with only his or her mind.” (Schrock 2007)
“The power of fMRI to read and predict human experience” (Anonymous 2006)
“Some even liken it to mind-reading” (Pearson 2006)
“Reading the private intentions of a person” (Amodio and Frith 2006)
Number of articles
With the increasing sophistication of neuroimaging technologies in neuroscience research and medicine, new language is being sought to make sense of the findings. This paper demonstrates that the current project of understanding and explaining the brain through neuroimaging that claims to make it “readable” has significant precedents throughout the history of science in the West. It is our view that the “brain-reading” metaphor, across its uses in science and medicine, has had and still has both a restrictive and expansive effect on our understanding of this most mysterious organ, the brain. The aim of this paper is to explore whether the “brain-reading” metaphor used to convey current medical or neurobiological findings imports unintended significations that do not necessarily reflect the genuine findings made by physicians and neuroscientists.
To assess the value of the “brain-reading metaphor” in its relevant contemporary application for making sense of investigations into the brain through neuroimaging, we do well to consider what grounds it. First, the paper surveys the ambiguities of the readability metaphor, drawing from the history of science and medicine, paying special attention to the sixteenth though nineteenth centuries. Next, the paper addresses more closely the issue of how metaphors may be confusing when used in medicine in general, and neuroscience in particular. The paper then explores the possible misleading effects associated with the contemporary use of the “brain-reading” metaphor in neuroimaging research and its relevant applications. The paper concludes that rather than breaking new ground, what we see in current scientific language is a persistence of both a constraining and expansive set of language practices forming a relatively continuous tradition linking current neuroimaging to past medical and scientific investigations into the brain. The use of the readability metaphor thus carries with it both positive and negative effects.
Historical Precedents of the Reading Metaphor
Reading the Universe
In general terms, reading is a complex cognitive process by which the ordered presentation of linguistic materials (e.g. printed characters, visual symbols, works of art) is decoded for the purposes of deriving and/or constructing meaning (cf. Goodman 2000). The intelligibility of any given text depends on the logic or system by means of which the signs relate to one another or refer beyond themselves. While a written text may serve as the paradigm for reading, non-linguistic and naturally occurring signs may also be read as texts in the relevant sense (Peirce 1998). The “readability” of the universe (or “legibility”, as Blumenberg (1981) calls it in his extensive elucidation of the metaphor) carries with it connotations of a grand unveiling of a hidden truth about the natural world. It thus invokes the tradition of the Book of Secrets going back to the Middle Ages, if not further (cf. Eamon 1994).
The consequence of this view is that scientists would have to develop a new rationalist discourse in order to interpret the book of the universe (writ large) as well as its various component systems (as per the natural sciences).4
Philosophy is written in this grand book, the universe, which stands continually open to our gaze. But the book cannot be understood unless one first learns to comprehend the language and read the letters in which it is composed. It is written in the language of mathematics, and its characters are triangles, circles, and other geometric figures without which it is humanly impossible to understand a single word of it; without these, one wanders about in a dark labyrinth. (Galilei 1957: 237–238)
Reading the Body
The drive to make sense of our corporeal existence through science has across various epochs also propelled a view of the “readability” of the body. In this regard, the human body is certainly not a conventional text; however, novel developments in anatomy in the sixteenth century helped to re-conceptualize the body as a text that could be read.
Thus, the readability metaphor played a key role in establishing the scientific legitimacy of anatomy in the century before Galileo, albeit in much less direct fashion. The most exemplary and ground-breaking anatomical work in this regard is Andreas Vesalius’ De Humani Corporis Fabrica, the publication of which in 1543 constituted a scathing critique of Church dogma and helped to clear the way for a new approach to anatomy and medicine. For almost a millennium and a half, the Christian Church had not strayed from the traditional model of human anatomy derived from Galen of Pergamum’s animal dissections and his theory of the bodily humours (i.e., blood, phlegm, yellow bile and black bile) (Ackerknecht 1968: 53). Although Galen was certainly a profoundly innovative and successful physician, experimental physiologist, writer and public orator in second-century Rome (Von Staden 1995), his anatomy work was problematic on at least two accounts. First, Galen did not use human cadavers in his research. Instead, he dissected animals such as pigs and apes and then tried to apply his insights via analogy to human bodies. One of the most egregious examples of the errors caused by this approach was the claim that humans, like oxen, had a set of vessels called the rete mirabile at the base of the brain (Martensen 2004: 61). Second, even when the use of human cadavers was no longer prohibited in medieval universities, the authoritarian use of Galen’s texts inhibited innovation and discovery. That is to say, anatomy instruction was hopelessly mired in tradition, particularly where the strict dichotomy between the priest’s role as “educated commentator” and the dissector’s role as “manual labourer” was concerned (Ackerknecht 1968: 90). Vesalius heaps much scorn on this practice in his De Humani Corporis Fabrica when he compares the priests to “jackdaws” who sit in their “high chairs” telling the illiterate surgeons below what to cut, neither party really knowing what they were doing (Vesalius 1998: lv). Vesalius’ text corrected this schism by drawing attention to the intelligibility of the actual human body (rather than Galen’s defective animal-based texts) and by intending his anatomy atlas to be used as a practical aid to dissection—to look inside human cadavers—rather than a substitute for first-hand experience. Early modern science thereafter sought to find what Vesalius referred to as the “fabric” of the human body (Zwart 1998: 116). But, as Hub Zwart, notes, the “fabric” of the human body was not thought of as “something which could be recognized immediately” (114): in other words, this fabric was not thought to be simply observable.
It should be apparent that the preceding description of Vesalius’ work does not make direct reference to the reading metaphor. In contrast to the explicit use of the metaphor by Galileo, the readability metaphor nevertheless functions as a guiding principle or underlying paradigm for anatomy during the Renaissance. To see this, it is necessary to engage in a more extensive analysis of the sociocultural context within which anatomy developed, drawing on some recent work in science and technology studies (STS). Pioneered by Thomas Kuhn and Michel Foucault in the 1960s, STS has become a rich source of insight into how scientific inquiry is socially situated or constructed (cf. Golinski 2005; Hacking 2007; Sismondo 2010). Daston and Galison illustrate this attitude to science in their book Objectivity (2007) by elucidating how the content of scientific inquiry is shaped in part by the methodology or standpoint of the inquirer. This is not to say that scientific facts are arbitrary or wholly dependent on conceptual schemes. Rather (as the discussion of Foucault below will demonstrate), new configurations of social institutions (e.g., the hospital and the university) may be productive and stimulate discovery. In the case of anatomy, Catherine Waldby (2000) draws on the work of Daston and Galison, Jonathan Sawday and Bruno Latour to document the influence of technology and the reading metaphor on the science of anatomy.
Reflecting on the differences between modern computer-generated anatomical models and Renaissance anatomy atlases, Waldby notes that in each case, the form of representation helps to construct the object represented (i.e., the content). In particular, the two-dimensional medium of the printed page and the mass production afforded by the printing press created new opportunities for Vesalius at the same time as they placed new constraints on the representation of the human body. Some of these constraints included the need to clearly demarcate tissues, to hide fluids and to add colour coding (Waldby 2000: 98), but the most significant was the requirement that the body be represented spatially as a series of two-dimensional maps that could be overlaid upon one another (rather than as a volume of matter subject to putrefaction and decay).
The extent to which this interpretive anatomization of the body as text is a creative act (but no less objective for all that) is explored in greater detail in the discussion of the difference between reading and seeing in the context of Foucault’s STS work.
In this sense the “anatomical” body and the anatomical text were “co-emergent”, coming into being only in relationship to each other. Moreover, the anatomical body is produced only in the act of dis-integration, becoming fully anatomical only in the acts of surgery or anatomical dissection, that is, in the act of being anatomically “written”. (Waldby 2000: 91–92)
Reading the Brain
[These imprints] are definite physical modifications of the acting fibre […]. Every man’s brain contains the history of his life, interior as well as exterior, spelled out in real letters, but [these letters] belong to a specific language that nature uses to address us all. Thus, an anatomist, with either very good eyes or a perfect microscope capable of seeing5 these letters, and with an ability to comprehend their signification, might be able to read a dead person’s brain, the thoughts he or she might have had during his or her lifetime. (Godart 1755: 209–210. Translation by F. Gilbert)6
Godart’s eighteenth-century theory explicitly suggests that a perfect seeing of the brain with “good” (presumably well-trained) eyes or vision-enhancing technologies will allow reading the very essence of the brain. On Godart’s view, the laws of nature work through the fibres in the same way for all people making the imprinting process on the fibres standard across different brains (Godart 1755: 359)7; however, the actual “print out” of sensations as cerebral letters (by way of the rustling of fibres in the brain (Godart 1755: 174–175)) is distinct for each individual as dictated by subjective experience (Godart 1755: 359). Although Godart’s use of the “readability-of-the-brain” metaphor represented a new way of talking about the brain that had the potential to be revelatory, Godart himself held that with the state of human “industry” (presumably, technology) at the time, there would be almost no chance that anyone could ever perfectly understand the subtle neural anatomy enough to distinguish the brains of markedly different human subjects. As he writes: “At the point where things are now, we are very far from being able to draw such distinctions; even Malpighi8 was unable, despite all his dexterity, to discover the difference between the brains of the great Lancisius,9 and one of the biggest idiots” (Godart 1755: 210). In the end, Godart adopted a cautious skepticism, noting that even if one were able to see the cerebral fibres as letters of the alphabet of the mind, this would not yet constitute knowledge of their meaning as sensible expressions of the brain-fibre language which remained for him practically hermetic, no more than a theoretical postulate (Godart 1755: 210).
Nevertheless, five years later, the Parisian doctor Antoine Le Camus explained in his Memoire on the Brain that, with the help of his scalpel, he had searched for connections between the structure of the brain and the organization of thought. But, he noted, “I can only see a greyish mass, furrowed by smooth rays that converge into a very white mass. Unsatisfied, I therefore no longer trust my eyes only, as they are unable to let me see more organization within the Brain’s substance” (Le Camus et al. 1760: 3–4). Feeling helpless, he turned to the available technological support of that epoch, the magnifying glass as well as the microscope (see Wilson 1995; Smith 2010). Even so, he met with disappointment as he was unable to find any trace of the sought after neural fibres from which he had hoped to read the brain and in so doing discern its nature (Le Camus et al. 1760: 3–4).
Thus, we see in the historical examples given previously a belief in, or hope for, the readability of the brain as a kind of text: hence, the metaphor, “reading the brain”. But as Rabinowitz aptly notes, “a reader can only make sense of a text in the same way he or she makes sense of anything else in the world: by applying a series of strategies to simplify it—by highlighting, by making symbolic, and by otherwise patterning it” (Rabinowitz 1998: 19). Treating the brain as a readable entity in and of itself is not yet misleading, but we do well to investigate not only from whence this metaphor has come to us (as detailed earlier), but more generally how it, like metaphors more generally in medicine, has the capacity not only to reveal, but also to conceal or distort, the reported findings.
Metaphors that Sometimes Confuse: Seeing and Reading as Metaphors in Medicine
According to the Oxford English Dictionary, metaphor is defined as “a figure of speech in which a name or descriptive word or phrase is transferred to an object or action different from, but analogous to, that to which it is literally applicable” (OED 2010).10 Metaphors play a very important role in medicine, as in many other fields, but we do not always appreciate their implications and their potentially distorting consequences. After all, language has many metaphorical dimensions that are sometimes hidden from—or at least not explicitly apparent to—us but whose pervasiveness and guiding power help us to make sense of our experience of the world (Lakoff and Johnson 1980). In medicine, metaphors such as the “war on disease”, “pulling the plug”, and the search for a “miracle cure” help to guide our thinking about the nature of medicine. In Practical Reasoning in Bioethics, James Childress presents an extensive list of medical metaphors but cautions us to be aware of the implications of using them uncritically (Childress 1997). While metaphors can illuminate aspects of phenomena that remain hidden, they can also be extremely limiting and may distort reality.11 In this regard, we are not always in control of our metaphors, especially when they are engrained in the way we speak and experience the world. Nevertheless, metaphors are both useful and inescapable, so we should learn how to reflect on them and use them properly (or draw attention to their misleading tendencies) rather than aim to rid them from medical discourse.
In this paper, we are particularly interested in the potentially misleading consequences of the metaphor of “brain reading” or “reading X in the brain”, which is closely related to the metaphor of seeing. As stated previously, Godart explicitly suggests that a perfect seeing of the brain with “good” (presumably, well trained) eyes or vision-enhancing technologies (a “perfect microscope capable of seeing”) is part of what is necessary to read the brain (Godart 1755: 209–210). However, as Godart recognized, seeing is distinguishable from reading even if seeing is necessary for the process of reading the brain. This distinction, to which we now turn, needs further exploration to better understand the lineage and import of the “readability-of-the-brain” metaphor.
Analytically speaking, a doctor reads a patient’s charts and laboratory reports, but it would be wrong, except in a metaphorical sense, to speak of a doctor as reading the patient’s body. However, in the case of brain imaging, it seems fitting to say that the neurologist can read the thoughts or mental states in the brain of a patient with the aid of, for example, an fMRI machine. Accordingly, it will be useful to establish some basic distinctions between seeing and reading in the medical context, both to develop a clearer idea of why the metaphor of reading has been selected in recent scholarship and because of the degree of confusion in the current use of these metaphors. That is to say, several references to reading the brain borrow some of their rhetorical force from the metaphor of seeing. Moreover, as Michel Foucault shows, the metaphor of seeing often overlaps with that of reading, where the key point of differentiation is the degree to which interpretation or mediation is required. Yet, since every reading is also a seeing in a meaningful way and vice versa, there is a potential distortive element to both metaphors.
In his 1963 work The Birth of the Clinic, Michel Foucault analyses the role that the metaphors of seeing and reading have played in medicine. He describes his approach as archaeological because he sifts through layers of historical documents and artefacts in order to retrace the development of medical concepts and practices, paying special attention to the relationship between medical discourse and other social phenomena, discourses, and events. Foucault attributes the development of a privileged role for seeing (which he calls the “medical gaze”) to innovations of a scientific and social nature that occurred during and after the French Revolution. In this time of great turmoil, when the university teaching of medicine was abandoned in favour of hospital-based clinics, the body was reconceptualized as a three-dimensional space of tissues whose “darkness” could be “illuminated” by those who had the proper expertise and training (Foucault 1994: 165–166). In addition to hospital reforms and the intensification of autopsy studies, the cultivation of such an eye depended on a series of new developments, foremost among them being the discovery of different tissue morphologies by Xavier Bichat in 1802 (Bichat 1813), the rediscovery of the pathological anatomy of Giovanni Battista Morgagni (1761) and the refinement of clinical examination, as illustrated by René-Théophile-Hyacinthe Laennec’s introduction of the stethoscope in 1819 (Laennec 1979).
The glance is of the non-verbal order of contact, a purely ideal contact perhaps, but in fact a more striking contact, since it traverses more easily, and goes further beneath things. The clinical eye discovers a kinship with a new sense that prescribes its norm and epistemological structure; this is no longer the ear straining to catch a language, but the index finger palpating the depths. Hence the metaphor of ‘touch’ (le tact) by which doctors will ceaselessly define their glance. (Foucault 1994: 122)
The essential element of such a glance is that it is immediate, like touching (ignoring for the moment the fact that touching involves the rapid relay of neural signals from the finger to the brain and back). However, the tendency of this metaphor to distort lies in the fact that the eye of the physician must first be trained through extensive experience and training. The art of seeing must be cultivated, making seeing a kind of “instantaneous reading” based on past experience and training (i.e., reading the “signs” in the encounter). Nevertheless, the metaphor does accurately capture the idea that the exercise of judgment in context12 is required. That is to say, one cannot substitute an algorithmic calculation process (e.g., a computer program) for the physician who relies on instinct as well as experience when consulting a patient.
Although more closely related than it may initially seem, the metaphors of seeing and reading may be distinguished based on the degree of mediation that each requires. Reading requires the most mediation as it presupposes a syntax, grammar, code or algorithm against which what is seen must be interpreted. As a result of this necessary processing, there is a serious risk that unintended significations may be imported either through error (i.e., misreading or misapplication of the code) or as the result of the ambiguity inherent in language. For example, unless one knows English, one may see words but one cannot be said to read them. Yet, even if one speaks English fluently, the associations that attach to the words that are read will vary with each individual. In its idealized, rationalistic form, however, the metaphor of “reading nature” assumes that the secrets of nature are written out like a language that can be understood by anyone who reads it. Moreover, this natural grammar in turn corresponds to the grammar of our own conceptual language(s), which enables us to conduct an exhaustive reading of the natural world and to impose order and predictability onto it (or rather, to see the order that is already there) (Foucault 1994: 165).13
However, even though it reflected a commitment to an abstract model of disease, Foucault shows that the seeds for a new conception of pathology and a new organization of the clinic were already taking root in this project of nosography. That is to say, the difficulties encountered during the effort to write nosologies would ultimately redirect the attention of the physician to the patient and to the importance of seeing the effects of disease on the patient’s physiology.
At last, there emerges on the horizon of clinical experience the possibility of an exhaustive, clear, and complete reading: for a doctor whose skills would be carried ‘to the highest degree of perfection, all symptoms would become signs’, all pathological manifestations would speak a clear, ordered language. (Foucault 1994: 94–95)
The “new” clinics that were created in the wake of the French Revolution did not have a teaching function in the strict sense; they were sites of research and innovation. The vast scale of new hospitals such as the École de Santé in Paris meant that physicians were exposed to a wide range of diseases and to variations of each disease that did not lend themselves to neat categorization according to rigorous nosological categories. Instead, it became essential to pay closer attention to the unique features of individual cases and to employ probabilistic reasoning to manage the uncertainty that characterized the variations in the presentation and progress of symptoms in different patients (Foucault 1994: 97–98). The irony is that Bichat, the man who articulated this new model of pathology best by means of the differentiation of tissues, was actually guided by Pinel’s nosological perspective (Foucault 1994: 102). Nevertheless, Bichat’s metaphor of seeing into the depths of tissues provided a new “window” into physiology and pathology that ultimately replaced the reading metaphor (Foucault 1994: 102).
Although Foucault’s analysis appears to rigorously distinguish seeing and reading, he is also careful to show how these metaphors may also shift their meaning depending on the context in which they are used. There is a degree of flexibility in metaphors that may work to our advantage but may also lead to conceptual slippage or loss of meaning. For example, seeing cannot always be distinguished so easily from reading given that direct perception still presupposes a schema or framework within which what is seen needs to be interpreted. To revisit the former example, if one speaks English as a native speaker, one often just sees the meaning of words (e.g., as one approaches a STOP sign) and it is possible to forget that one is actually reading very quickly. Moreover, seeing and reading both involve the use of other senses that seem to deviate from the metaphor: for example, reading in Braille substitutes touch for sight and one can of course listen to a public reading or lecture. However, seeing “happens” in an instant when one opens one’s eyes, or at least it seems to if one ignores the series of interpretive acts that occur (not least of which is the inversion of the image transmitted from the eye to the brain). By contrast, reading takes longer and requires a code or language that is more explicitly dependent on interpretation than is the seeing of objects or colours, for example. Still, the dream of “reading nature” has not died out and its effects are still uncertain. That is to say, the pressure to restate novel medical possibilities in terms that are already at hand threatens to distort their meaning and limit the possibility of what may yet be achieved, notwithstanding the benefits it may also bring.
Possible Misleading Effects of the “Brain-Reading” Metaphor
Thus far, we have been concerned to survey the ambiguities of the readability metaphor, drawing from the history of science and medicine, paying special attention to the sixteenth through nineteenth centuries. Next, we investigated more closely how metaphors may be confusing when used in medicine in general. Now we will look at how the readability metaphor is employed more specifically in the context of contemporary neuroimaging research and its relevant applications, bearing in mind how the historical freight of this terminology both augments and constrains our understanding of the relevant findings.
According to the Merriam–Webster Dictionary—much like the definition from the Oxford English Dictionary given earlier—a metaphor is “a figure of speech in which a word or phrase literally denoting one kind of object or idea is used in place of another to suggest a likeness or analogy between them (as in drowning in money)” (Merriam-Webster 2010). Perelman and Olbrecthts-Tyteca have noted that in the context of argumentation, metaphors act as condensed analogies. Aristotle shows this explicitly in the Poetics (XXI, 1457b) with the following example: “As old age is to life, so is evening to day. One will accordingly describe evening as the ‘old age of the day’ and old age as the ‘evening of life’” (quoted in Perelman and Olbrechts-Tyteca 1969: 399, nt.179). Perelman and Olbrechts-Tyteca have analysed the structural relations of Aristotle’s example, noting that it demonstrates a logic whereby “‘A is to B as C is to D’ yields the expression ‘C of B’ to designate A” (Perelman and Olbrechts-Tyteca 1969: 399). So also, we can say that: “(A) Neuroimaging is to (B) the brain, as (C) reading is to (D) symbolic expression (e.g. written language)”. Accordingly, this yields the expression “reading of the brain” (‘C of B’) to designate (A) neuroimaging. It is important to remember, however, that metaphors are a kind of trope, and as Quintilian explains in The Institutio Oratoria (VIII, VI, 1), a trope involves “‘the artistic alteration of a word or phrase from its proper meaning to another’” (quoted in Perelman and Olbrechts-Tyteca 1969: 398–399, nt. 174). Since metaphors are commonly used as rhetorical figures of speech (that achieve their effects via analogy), we need to further probe how the analogies between reading and modern neuroimaging (as built into the “brain-reading” metaphor) rightfully explain or potentially obfuscate the meaning of reported findings achieved through neuroimaging. With this in mind, we turn now to a discussion of contemporary neuroimaging and the discourses surrounding it.
Modern neuroimaging technologies—for example, computed axial tomography (CT), diffuse optical imaging (DOI), event-related optical signal (EROS), magnetic resonance imaging (MRI), functional magnetic resonance imaging (fMRI), electroencephalography (EEG), magnetoencephalography (MEG), positron emission tomography (PET) and single photon emission computed tomography (SPECT)—hold out an unprecedented promise of medical innovation (Logothetis and Wandell 2004).14 Even if these imaging technologies are not restricted to neurological applications, some have been used to provide structural imaging of the brain, for the purposes of diagnosing larger-scale intracranial diseases, tumours, injuries and stroke. Others have been used to provide functional imaging of the brain and applied for various purposes including: i) the diagnosis of smaller tumours, lesions and diseases (e.g. metabolic diseases) on a finer scale as well as ii) brain mapping for research into neurological and cognitive psychology that correlates brain areas with specific functions (Haberfeld et al. 2010). With the latter expansive applications of the technology, we are developing new ways of investigating the brain that extend far beyond clinical purposes to correlate mental or psychological states to physical (i.e. brain) states or structures (cf. Zanzotto and Croce 2010).
As it stands, many neuroscientists and certain neurophilosophers are now talking as if the newly available technological powers for observing the brain will surely lead us to the mind—that the mind can be read from the brain (see Table 1, part 2). As Kamitani and Tong write: “The potential for human neuroimaging to read out the detailed contents of a person’s mental state has yet to be fully explored” (Kamitani and Tong 2005). Following this hope, Haynes and colleagues purport to use brain scans to “decod[e] mental states from brain activity in humans” (Haynes and Rees 2006). Friston further claims that “all demonstrations of functionally specialized responses—demonstrating a significant mapping between mental states and brain signals—represent an implicit mind reading” (Friston 2009; cf. De Charms 2008).
The arc of Godart’s eighteenth-century theory—that claims that the brain holds inscribed within itself the whole of an individual’s personal history—clearly traces through to present-day conceptions that apply correlation analyses to investigate the readability potential of the brain. We see evidence of this in talk of “decoding” the brain for mental contents (Kamitani and Tong 2005) and finding in it “neural signatures” or “neural codes” (see Table 1, part 3). Modern neuroimaging technologies also carry forward the impetus of Godart to open the brain for viewing, and like a text, read from it the language contained therein. What is new with these technologies, when compared with the instruments for investigation in previous epochs is the capacity to, as one textbook puts it, “see inside and ‘read’ a living human brain” (emphasis added), and to do so with increasing sophistication (Haberfeld et al. 2010). But as Watson points out, “Bridging the gap between understanding the brain and understanding the mind and behaviour continues to elude us” (Watson 1996: 544). Even so, the belief in the readability of the brain by way of modern neuroimaging techniques seems to trace forward Godart’s belief that we are closing this gap, first “with very good eyes”, then with the promise of “the perfect microscope”, and now with instruments of much greater accuracy as available through modern neuroimaging (see Table 1, part 2).
However, Godart’s “very good eyes” or, as the case may be, neuroimaging technology that can “perfect” our powers of vision count as examples of sight, and sight counts as a prerequisite for most kinds of reading (Braille, being an exception to this norm). But, to repeat, powers of vision, no matter how accurate, are insufficient by themselves to make reading possible: as Godart also noted, it is not enough to be able to see the imprinting of the “letters” of thought in the brain (cf. Willingham and Dunn 2003); also needed is “an ability to comprehend their signification” (Godart 1755: 208–210). Is neuroimaging, then, better described as a practice in “seeing”, as opposed to “reading”, the brain? Answering this question, as we shall see, can indirectly also answer our query as to the limitations of the “brain-reading” metaphor.
A proper answer to the question at hand hinges on what kind of reading is implied. On the one hand, the practice of reading as applied to texts like laws, works of art (e.g. plays, novels) or newspapers all involve a great deal of interpretation on the part of the reader to produce understanding. On the other hand, the term “reading” can also be used to describe “data indicated by an instrument” (Merriam-Webster 2010) as we would speak of, for example, “a meter reading” or “a temperature reading”. To read the latter sort of “read-outs”15 obviously involves comparatively little (thought arguably still some) interpretation. Reading of this kind would almost be akin to near “instantaneous reading” of a STOP sign—that is to say, reading practices that are so routine they have veritably become non-conscious or unconscious. As stated previously, Foucault has shown that the metaphor of “seeing” often overlaps with that of “reading”, where the key point of differentiation is the degree to which interpretation or mediation is required. So, where reading involves very little interpretation and therefore involves minimal cognitive processing, it becomes like seeing. Insofar as the metaphor “reading the brain” is used to designate modern neuroimaging practices, where the “reading” involved is understood as requiring a great deal of interpretation, the use of the trope seems valid. In the act of reading a sentence (as with reading a language), one observes letters but seeing them as letters that form words whose ordering delineate sentences that signify propositions requires seeing beyond and past the surface appearance of scratches on the page. So also, “reading the brain” with neuroimaging is much more than a matter of straightforward observation.
Michael Coles, in an article entitled “Modern Mind-Brain Reading” surmises that “[p]sychophysiological measures, particularly those of the event-related brain potential… can serve as ‘windows’16 on the mind and as ‘windows’ on the brain” (see Coles 1989). Yet, this idea does not reflect the real possibilities offered by new neurotechnologies. Moreover, the “window” metaphor adds to the confusion because it implies that one can just see the brain even though such “seeing” is in fact a reading. That is to say, a technological device is required to “open” that window using sophisticated computer programming to read the data. After all, the results of fMRI, for example, rely on computer processing based on algorithms derived from statistical comparisons that are used to generate images that approximate representations of brain functioning. On a popular and cultural understanding, neuroimaging technologies can be applied from the outside of a subject to represent (or mirror) the interior of the subject (Kevles 1997; cf. De Charms 2008). This commitment assumes that images stand in for something more fundamental of which they are technological copies (Van Dijck 2005). However, most modern imaging technologies do not operate as X-rays do (Doby and Alker 1997). The images are not to be confused with accurate representations, but rather approximate configurations, producing meaningful correlates rather than a mirror of the reality of what is imaged. As such, they are types of description that act, not as observations, but as ciphers requiring interpretation by physicians and neuroimagists.
In this regard, the “brain-reading” metaphor effectively captures the highly mediated nature of neuroimaging in relation to the relevant living brain that is being imaged. Insofar as the metaphor “reading the brain” is used to designate modern neuroimaging practices, where the sense of “reading” that is implied is much like that when reading a text of some sort, the metaphor seems to be valid. However, insofar as the “brain-reading” metaphor is taken to imply the kind of reading involved when removing information from an automatic device (e.g. water metre read-out, or computer read-out), it denies the interpretive aspects of the process and in this regard, the trope is potentially very misleading.17 Greater powers of “seeing” may seem available to us through modern neuroimaging, but the greater accuracy of representations have not obviated the interpretive element for determining findings of tumours, lesions and diseases in the brain, let alone findings for such things as “intentions”, “thoughts”, “the will” (Haggard 2008), or “the mind”.
As outlined earlier with regard to the power of metaphors, it is essential to reflect on the history or genealogy of metaphors in order to be able to use them wisely. The often implicit (or hidden) meaning(s) of metaphors may serve to distort reality or limit our capacity to chart new directions and to make sense of the research findings. Even the metaphor of reading, as distinct from seeing, admits of two different senses that are often run together in the neuroscientific literature and in popular literature. That is to say, the expression “brain reading” or “reading X in the brain” crystallizes relatively dissimilar expectations. The more restrictive scope of what in the brain presents itself for reading, to which actual neuroscientific findings correspond, should be rigorously distinguished from the more expansive scope of what the brain presents for reading, to which representations and fantasies (supposedly) correspond. In other words, the process of collecting brain data is often confused with reading as a cognitive process of decoding symbols (an interpretive process) that is effectively analogous to making sense of expressions of a language.
The “brain-reading” metaphor in itself is not responsible for this confusion. Indeed, it might be even worse if we were to routinely speak and write of “seeing X in the brain” as opposed to “reading X in the brain”.18 At the very least, the trope of reading conveys the important need for interpretation to make sense of the findings of (data produced by) neuroimaging technologies. But treating the brain as if it were a subject to be read—a veritable linguistic entity—as opposed to an object to be observed is potentially misleading. Doing so, and with the routine use and entrenchment of the metaphor, we sometimes forget that it is only a metaphor and come to believe that our observations of the brain through neuroimaging provide “something” more than data; instead, these “readings of the brain” are made to assume a certain symbolic status, entrenching a belief in metaphysical notions like the mind (see Table 1, part 2), just as scratches on a page signify words that stand in for concepts or ideas.
There is currently a discrepancy within the domain of neuroimaging technologies between what is induced by hyperbole (including talk of being able “to read X in the brain”) and what is concretely possible through science and medicine. The explicit reading of an individual’s accumulated experience (i.e. mental states, personality, intentions, memories, etc.), for instance, is still very far from being possible today. Quite the contrary, neuroscientists and physicians are among the first to admit the extent of what remains unexplained in brain functioning. For this reason, even the diagnostic use of brain scans remains problematic. Although researchers are able to see the brain’s bioelectrical activity in relation to specific goal-related tasks carried out in laboratories and have software that recognizes certain predefined paths of cerebral activity, this highly mediated and technological act of “reading” stands in stark contrast to some of the hyped promises for this technology: for example, promises to detect lying (Thompson 2005; cf. De Charms 2008: 724; Fenton et al. 2009; Ireland 2007; Kerr et al. 2008; Langleben et al. 2002; Langleben 2008; Spence et al. 2004), criminal intent (Barrie 2008; West 2007; cf. Khoshbin and Khoshbin, 2007; Krahn et al. 2009; Marks 2007; Morse 2006) or the lack thereof (Hartocollis 2007; cf. Appelbaum 2009).
New technology brings new ways of thinking about neuroscience and medicine. These new ways of thinking may be shaped by the way that we talk about new developments. Where we are forced to draw on older metaphors that are inappropriate or inaccurate, a distortion may result. This is especially true when the metaphor is used uncritically or without due appreciation of its scope and impact. Neuroscientists and physicians must resort to the use of terms already laden with abstracted meanings, and often burdened by tradition, at the risk of importing through these words connotations that do not tally with the sought-after objectivity of empirical science. The use of the readability metaphor carries with it both positive and negative consequences. We do not condemn the “brain-reading” metaphor per se as used to discuss the practice of neuroimaging. To be fair, the “brain-reading” metaphor as an explanation of neuroimaging is not without its merits: it makes accessible for a multiplicity of stakeholders what we otherwise find very difficult to describe in non-technical language and it speaks to the explicit interpretive elements germane to the technology. The distortion occurs when the metaphor is understood not as a figurative form of speech—“an artistic alteration of a word or phrase from its proper meaning”, to quote Quintilian (1920: VIII, vi, 1)—but is instead taken literally, not as a metaphor. In effect, this sets up the brain as a text—a veritable linguistic entity. All too commonly this presumption, when not acknowledged as the product of metaphorical thinking, then feeds into expectations to find, by way of neuroimaging technologies, evidence of metaphysical entities that are (not surprisingly) not empirically observable. Though historical precedents explain some of the sources and traditions that underwrite the sensibility of the “brain-reading” metaphor as it has its life in discussions of modern neuroimaging technologies, we do well to critically check certain distortions (even false expectations) carried forward from the past that are also bundled with the use of this way of speaking (and writing) about the brain.
This paper expands upon some arguments that were initially sketched in a commentary by first author Gilbert and Wenger (2010)
Norman et al. state that: “A key challenge for cognitive neuroscience is determining how mental representations map onto patterns of neural activity. Recently, researchers have started to address this question by applying sophisticated pattern-classification algorithms to distributed (multi-voxel) patterns of functional MRI data, with the goal of decoding the information that is represented in the subject’s brain at a particular point in time. This multi-voxel pattern analysis (MVPA) approach has led to several impressive feats of mind reading” (2006: 424). De Charms also writes: “fMRI data have increasingly been analysed using computer pattern-classification algorithms and related methods, with the goal of ‘mind reading’” (2008: 723). Cf. references 26–36 in De Charms 2008: 729.
“According to Lacan, the true break through of symbolization, the first truly consistent effort to mathematize the real, remains the accomplishment of Galilei. His genius is of another, more modern kind and resides in the systematic reliance on on a mathematical methodology, at the expense of appearance, the imaginary and the intuitive, at the expense of Anschauung. Galilei’s achievement results from his reliance on algebraic formula rather than visual experience, formula that can never completely realize themselves in the world of everyday existence because of the many interfering circumstances—the noises of the real. Galilei’s genius resided in the insight that the Book of nature was written in algebraic symbols rather than in primordial images or types” (Zwart 1998: 116).
It would be fair to say that modern physics has not deviated from this view in any fundamental sense, although a degree of uncertainty has crept into the mathematical representation of the universe through quantum physics and non-Euclidean geometry (Friedman 2002). Even so, the language of mathematics has been reinvented and refined rather than abandoned. Unlike modern physics, modern biology has not enjoyed the same kind of mathematical precision, although the project of reading genetic codes goes a long way to doing just that (cf. Kay 2000; Zwart 1998).
See Sect. 3 below on the distinction between “seeing” and “reading” in the medical sciences.
All subsequent English translations of Godart are provided by (first author) Frederic Gilbert.
Laura Balladur (2006: 163–164) explains that: “By the eighteenth century, the clockwork precision of the world now extended to the mind’s intimate mechanisms. Because mechanisms could fall under ideal laws, so did physiology: mind and machines shared the same language of physics, which could describe the great forces acting between planets or the imperceptible force of impressions in the mind. …One of the more ‘physicalist’ treatises of the period, Godart’s La Physique de l’âme, captures this trend to depict the mind in terms of physics. …[O]ne finds in Godart’s work on the physics of the soul an attention to concepts of time, frequency, liveliness, and depth of impressions—the latter suggesting that a potential energy was inherent to sense impressions. Like the stretched string, whose potential energy diminishes over time, the mind experiences a type of decay. Forms simply disappear: ‘[Les] empreintes sont d’autant plus profondes, qu’elles sont plus récentes […] que leur retracement a été vive’ (Godart, 287). Not only does the distinctness of the marks relate to the mechanics of the impression, it also relates to the quality of the sense impression itself, its ‘liveliness,’ which creates even deeper imprints. In turn, the memory of these ideas depends on the strength or force of their impressions. Through frequency and liveliness, the initial impressions are retained in the memory of the traces: ‘La durée des idées depend de la profondeur des empreintes et de la ténacité des fibres’ (Godart, 287)” (Balladur 2006: 163–164).
Marcello Malpighi was a seventieth-century Italian doctor famous for his contributions in physiology.
Lancisius was an Italian anatomist who wrote a monograph on the sympathetic nerves.
The OED also gives an alternative definition of metaphor as: “Something regarded as representative or suggestive of something else, esp. as a material emblem of an abstract quality, condition, notion, etc.”(OED 2010).
Susan Sherwin explains: “All metaphors function as models or analogies. As such, they emphasize similarities between the two realms in question while obscuring differences. That is, in the process of helping us to understand one set of relations in terms of another, metaphors highlight certain features of the domain to which they are applied and, inevitably, they hide or distort others” (Sherwin 1999: 199).
Laennec’s exhaustive account of the different sounds one may hear when listening to the patient’s chest through the stethoscope provides a good indication of how much discernment and training is required to conduct a physical exam and to make a diagnosis (Laennec 1979).
Even though the natural world as a whole may have a rational or intelligible structure (e.g., in Galileo’s sense of a mathematical composition), it is still possible to differentiate levels of intelligibility (much like the scholastic notion of a chain of being in which humans have more reality/perfection/goodness than inorganic matter (like rocks) but less than God (Lovejoy 2009)). On this logic, since the human brain would certainly be at the higher end of the scale, given that it is the organ that allows us to think, one could assume that it is more rational/complex in its composition or organization than a spleen or a heart may be. The affinity between the organ that supports thoughts and the thoughts themselves may be so great that the organ itself manifests the thoughts in a direct manner: i.e., the thoughts are visible in the tissue as if they are part of the very fabric of the brain. According to this line of thought, the “first-hand reading” of one’s own thoughts should in principle be accessible to a second-hand reading from an external vantage point (i.e., by someone who does not actually think those thoughts). To be readable, concepts must somehow be inscribed “in the flesh” or in the fibres of the brain (cf. Cheung 2010), or at least this is the hope expressed by the quotations above from Godart and Le Camus.
Our approach to the subject has emphasized historical precedents for contemporary neuroimaging techniques as following from our discussion of early modern understandings of human anatomy in general and neuroanatomical mechanisms in particular. In terms of contemporary neurotechnologies, our emphasis is on indirect imaging techniques that attempt to “read” the brain. Equally significant are direct imaging techniques that are used to represent the brain. These technologies (e.g., EEG) have also fostered the “brain-reading” metaphor. (For examples of the prevalence of the “brain-reading” metaphor in the reporting of some of the relevant neuroscience research and technology developments, see Anonymous, 2010; Caron, 2010; Max-Planck-Gessellschaft 2009). Neuroscientists employ technologies such as EEG as direct devices of measurement (i.e. direct recordings of neural activation), whereas functional imaging devices—more commonly associated with what we call "neuroimaging" devices, examples of which include: SPECT, PET and fMRI—are regarded as indirect forms of measurement because they “interpret” neural activation via processes of brain metabolism (e.g., blood oxygenation levels) (see Logothetis & Wandell, 2004). Historical precedents for technologies that measure direct recordings of neural activation could be found in models that would portray the brain as an electrical unit (see for example, Smith 2010: 111–113). We do acknowledge the importance of devices employed to measure brain activation directly and it is important to recognize their contribution to the development of the “brain-reading” metaphor in the brain sciences. However, for the purposes of this paper, we have focused primarily on functional imaging devices employed to indirectly measure brain activity and structures. We thank an anonymous reviewer for helping us to further clarify these distinctions.
The Oxford English Dictionary defines “read-out” as follows: “a: the process of removing information from an automatic device (as an electronic computer) and displaying it in an understandable form b: the information removed from such a device and displayed or recorded (as by magnetic tape or printing device) c: an electronic device that presents information in visual form”.
Note how the National Institute of Mental Health (NIMH 2010) also speaks of neuroimaging as providing a “window into the brain”. Cf. also De Charms for a discussion of how MEG and EEG “have historically been windows on the brain” (2008: 722).
Not uncommonly the brain is understood as a data device (e.g. like a computer) that produces a read-out which we are now able to register with modern neuroimaging technologies. See Haynes and Rees 2006 as well as Kamitani and Tong 2005. Cf. O’Toole et al. 2007.
Thanks to Alexandre Wenger for enlightening discussions which were invaluable to us in drafting the argument of this paper. Thanks to Françoise Baylis and the Novel Tech Ethics research team for feedback on earlier drafts. Research for this project has been funded by Canadian Institutes of Health Research, MOP 77670, Therapeutic Hopes and ethical concerns: Clinical research in the neurosciences and by Canadian Institutes of Health Research, NNF 80045, States of Mind: Emerging Issues in Neuroethics.