Advances in noninvasive medical imaging have opened new windows into the living brain. Like observatories pointed inward, modern brain scanners routinely capture breathtaking images of the gyral swirls and neuronal clusters that underlie human cerebral nature. These brain portraits are composed of three-dimensional voxels, or volume picture elements, digitally displayed in shades of grey.
Knowledge in neuroscience can be measured in degrees of resolution. Greater neuroimaging resolving power means more finely detailed representations of the human brain. As the brain is the physical correlate of the mind, its grey matter and intricate interconnections are subject to scientific investigation. Neuroimaging methods increasingly are able to map out, voxel by voxel, the neurobiological pathways underlying all aspects of thought and behavior, including those involved in moral judgment and ethical reasoning.[1],[2],[3]
Whether clarity in neuroimaging might help to resolve, not only clinical questions, but also ethical grey matters, is a prime question for neuroethics. The repertoire of voxels has so multiplied that it is now possible to speculate whether, from a science of the brain, one can derive a coherent and valid system of ethics. Psychologist Michael Gazzaniga has proposed that, “there could be a universal set of biological responses to moral dilemmas, a sort of ethics, built into our brains.”[4] Images have always profoundly influenced cultural perceptions of human nature. To the extent that neuroimaging informs a brain-based model of ethics, its fundamental unit of significance is the voxel.
To begin to explore the implications of a voxel-based paradigm for ethical theory, it is helpful to examine, not just the quality of images, but also the methods and presuppositions of neuroimaging. Implicit in every voxelous reconstruction of the brain is the idea that the brain is virtually, if not essentially, reducible to matter. Reductionism can clarify, but it can also mislead. Vibrant voxels may elucidate pertinent facts. Exclusive attention to them may overlook important truths.
Neuroimaging is a product of the last hundred or so years, with the greatest progress having occurred in the last three decades. Following Röntgen’s invention of the x-ray machine in 1895, for much of the 20th century, visualization of the diseased brain was possible only through pneumoencephalography, a painful procedure in which air was injected into the spine and allowed to rise to outline the contours of the brain as seen on a skull x-ray. The first computed tomography (CT) scan in clinical use at Mayo Clinic in 1973 supplanted pneumoencephalography with a hundred-fold increase in the resolution of brain imaging. Presenting anatomic images slice by slice, its spatial resolution was coarse by today’s standards with a field of view of 13 mm per voxel. By contrast, current CT technology achieves sharp spatial resolution with typical fields of view of 0.7 mm, and emerging methods of multidetector row high resolution CT achieve 0.4 mm per voxel. CT angiography now achieves resolution of the major intracranial blood vessels about as clearly as early CT resolved such larger structures as the brain’s lobes and ventricles.
Whereas CT utilizes ionizing radiation to measure tissue density, magnetic resonance imaging (MRI) utilizes radiofrequency pulses to define soft tissue molecular composition at resolutions of 1–3 mm per voxel. More powerful 8 Tesla research magnets can achieve high resolution images with voxels corresponding to just 0.2 mm. This level of imaging detail compares to the 0.01–0.05 mm size of most neurons.
Ever sharper shades of spatial resolution seem at first to suggest that, if only the brain were to be imaged in sufficient detail, then all the brain is and does might be explained. The structure of matter, however, is not all there is. Exact knowledge of the brain’s shape, its configuration, its density and spatial orientation, even the atoms that make up its grey matter, while necessary for accurate neuroscience, yet are insufficient. Also to be considered is what the brain does. As regards function, too, neuroimaging is yielding astonishing details. Anatomical correlations to specific neurological capacities are possible through functional MRI (fMRI), which takes advantage of the paramagnetic properties of oxygenated hemoglobin to detect real time changes in regional blood flow in response to increased neural metabolic activity. Other current methods of functional imaging technologies include electroencephalography (EEG), single photon emission computer tomography (SPECT), positron emission tomography (PET), and magnetoencephalography (MEG). The terminology at times appears to lengthen in inverse proportion to shrinking voxel size.
Voxels, which correspond to changing units of tissue volume, are represented mathematically in the four dimensions of height, width, depth, and time. Voxels corresponding to units of nature can also be viewed philosophically as the product of four independent causal categories. In the language of Aristotle, voxels have material, formal, efficient, and final cause. Each category has its proper sphere of explanation as well as its epistemological limits.
At the level of material cause, voxels represent units of matter. Voxels have material cause in the electrons or flickering light-emitting diodes that display brain images. Voxels correspond to material cause in the physical elements, molecules and cytoplasm that make up the brain probed by the scanner. Because voxels are arranged uniformly accordingly to an arbitrary grid, they only approximate the actual arrangements of matter. Material cause tells something of what the brain is without supplying an understanding of what the brain is doing, how the brain came to be, or why it exists. Voxels at the level of material cause are correlations of matter to matter. CT can distinguish blood from water within the brain, but analysis of material cause alone cannot distinguish life from inanimate mass.
At the level of formal cause, voxels represent shape and spatial relationships. While voxels themselves do not interact with one another, their corresponding molecular configurations have formal cause in such spatial arrangements as the DNA’s elegant double helix, neurons’ arborizing dendrites, and the cerebral cortex’s convoluted gyri. Analysis of formal cause leads from correlation to identification. Recognition of formal cause allows microscopic imaging to distinguish neurons from other types of cells. Recognition of formal cause allows neuroimaging to distinguish the specific cortical and subcortical brain regions involved in discrete cognitive functions.[5],[6],[7]
Exact knowledge of formal cause can never be exhaustive. There are physical limits to how much information about reality voxels can reveal. Imaging the submicroscopic realm, for example, is fraught with loss of detail because, with greater spatial resolutions come lower signal-to-noise ratios. Efforts to obtain extremely fine views of molecular structure encounter fuzziness as the resolution of the imaging modality approaches the wavelength of the light or other energy source used to obtain the image. Indeterminacy at the quantum level also imposes limits on what can be known about the position or behavior of a given molecule.
Nor does more detail necessarily lead to greater knowledge. One rightly suspects that there is more to the story of the ceiling of the Sistine Chapel than patterns painted. The mission of interpretation must look beyond the flecks, forms and symmetries of its frescos to the thoughts and imagination of the artist and to the source of his inspiration.
At the level of efficient cause, voxels represent mechanisms of action and reaction. Causation in the ordinary sense of the word usually refers to efficient causation. Voxels themselves have efficient cause in the human actions and machine processes that generate brain images. From scanner design to construction and operation, from magnetic field flux to software data processing and digital image reconstruction, voxels are the product of highly organized streams of efficient causation. Efficient causation is the very language of voxels. Their brilliance is the direct outcome of antecedent electromagnetic pulses. Consistently obedient to digital command, their obligatory arrays reconstruct reality in approximate black and white.
The brains that voxels represent consist of streams of efficient causation of exceedingly greater complexity. A cubic millimeter of cerebral cortex contains as many as a billion synapses by which neurons signal one another.[8] Accordingly, the brain must be understood at many levels of organization. Deciphering efficient causes and their effects leads to progressively coherent explanations of overall brain function. These explanations portray cerebral activity in digital brush strokes from a palette of scintillating voxels. Any one brain image, like a single frame from a movie, captures just a slice of meaning.
Available mathematics and computer software are, for the time being, remarkably inadequate to the task of tracing out the totality of neuronal paths of efficient causation underlying human thought. Whether Raymond Kurzweil is correct or not in predicting that computer processing power will one day surpass human cognitive capacity,[9] the belief among some forecasters that the evolutionary trajectory of artificial intelligence can, given sufficient time, approximate human intelligence already challenges traditional notions of free will. Computer-generated structural and functional images of the brain tend to reinforce the idea that the brain is much like a computer. A computer may be a complex machine, but it is essentially a mechanism. Deterministic models of the brain that regard conscious will as illusory seek to explain moral agency exclusively in terms of material efficient cause.[10],[11]
Conscious will may, perhaps, act from somewhere in between the voxels. Whatever lies outside or above efficient material causative connections remains undetectable to empirical investigation. What science can learn of brain behavior is, in fact, limited to the reproducible. Science is interested in events that repeat themselves predictably. Scientific method, therefore, is not well suited to the systematic study of unique phenomena, which may include diverse sorts of human thoughts, judgments and decisions.
At the level of final cause, voxels have aims. Voxels find their final cause in the purpose for which they were made, which for neuroimaging is primarily diagnosis. Voxels find their ultimate meaning in relation to something beyond the immediate chain of efficient causation. Their existence presupposes a motivated inventor of the neuroimaging equipment, an interpreter who reads the images with a goal in mind, all for a patient in need of medical attention.
Analysis of final cause in neuroscience leads from coherence to unifying explanations. The quest for understanding encounters in the domain of final cause signs pointing to purpose and destiny. The desire to explain phenomena at one level and then to predict phenomena at another level, the yearning to understand, and the belief that cerebral behavior is discoverable all are comprehensible within a framework of final cause. That brains can ponder their own depictions implies self-consciousness and a purpose to human thought. That purposeful brains exist and possess awesome intricacy inexplicable by any theory of origin based on blind chance implies an intelligent Designer. These are inferences that voxels cannot definitively spell out in the language of material, formal, or efficient cause, yet their truth is accessible to the mind that considers relationships of final cause.
Only by recognizing final causes can one judge whether it is good to image the brain. The neurologist who, like the fictional detective Sherlock Holmes, has a “passion for definite and exact knowledge,”[12] appreciates well what neuroimaging contributes to medical care and neuroscience research. The physician values brain images because, above all, they benefit the patient. The patient is the starting point for making sense out of grey voxels and the focal point for exploring the meaning of human nature. For, as Holmes elaborates, “One’s ideas must be as broad as Nature if they are to interpret Nature.”[13] Brain images and the neural arrangements they represent are not, after all, a complete portrait of the human being. The encounter at the bedside is a compelling reminder that the suffering patient is no mere collage of voxels. The patient, who comes to the physician in a state of illness or distress, is the central reason for pursuing more accurate, detailed, fast, and noninvasive methods of neuroimaging. All causal categories converge in attending to the interests of the patient, who experiences and chooses, who feels pain and welcomes comfort, who needs and is needed, all in ways that voxels cannot adequately capture and that caregivers cannot fully know.
Voxels in formation convey volumes of information. While precise and useful, voxels are an imperfect medium by which to resolve questions of neuroethics. Voxels can supply needed facts, and deductions drawn from their patterns can help to project potential consequences, but they cannot supply formulae to complete the ethical analysis. Vacuous voxels cannot by themselves reach a valid moral decision how to treat the patient. The reason is that voxels represent the shape of what is. There is simply no voxel the shape of an ought.
Imaging the realm of is differs from perceiving the realm of ought. The two are so incommensurable as to be invisible to one another. Questions framed in terms of should andshould not elude the most penetrating neuroimaging methods. X-rays, for example, cannot fathom the dimensions of moral reality any more than they can outline a whispered breath. The mind, aware of other persons and oriented to purpose, is able to perceive what the x-ray cannot, even though the eye cannot directly perceive the x-ray. Analogously, Wilhelm Röntgen observed that, “The retina of the eye is not sensitive to these rays. Even if the eye is brought close to the discharge tube, it observes nothing.”[14]
Finer precision of anatomic resolution and tests that trace out the brain’s functions have afforded remarkable improvements in diagnostic certainty. At the bedside, classical approaches to eliciting historical clues and interpreting uncertain physical signs are yielding some of their former prominence to the power of sophisticated neuroimaging. Clear pictures of brain activity provide rapid, relevant information to assist timely clinical decision-making. A brain scan, of course, is not all there is to a healing encounter. The physician’s knowledge, skills, experience, judgment, and human touch remain essential to framing the proper clinical question, reaching a diagnosis, interpreting and explaining test results, and implementing a treatment plan tailored to the individual patient. It is important not to lose sight of the person for the voxels. Likewise, depending on one’s perspective, a brain-based theory of ethics written in the language of voxels might supersede, or it might complement, traditional approaches to discerning the moral dimensions of medicine.
There is, finally, no color of voxel to signify awe. The sense of wonder that the study of the brain arouses suggests that there is meaning to the contents of the human cranium that surpasses what voxels can outline and more to human thought than be traced out by the ostensibly necessary paths of efficient causation. Sherlock Holmes considered it a matter of deduction that in the design of the rose rests the highest assurance of providential goodness transcending sheer necessity:
All other things, our powers, our desires, our food, are all really necessary for our existence in the first instance. But this rose is an extra. Its smell and its colour are an embellishment of life, not a condition of it. It is only goodness which gives extras, and so I say again that we have much to hope from the flowers.[15]
As regards the brain, we have much to learn from visible voxels, and much to hope from unseen superfluities.
[1] Moll J, de Oliveira-Souza R, Eslinger PJ. Morals and the human brain: a working model. Neuroreport 2003;14(3):299–305.
[2] Greene J, Nystrom LE, Engell AD, Darley JM, Cohen JD. The neural bases of cognitive conflict and control in moral judgment. Neuron 2004;44(2):389–400.
[3] Borg JS, Hynes C, Van Horn J, Grafton S, Sinnott-Armstrong W. Consequences, action, and intention as factors in moral judgments: an fMRI investigation. J Cogn Neurosci 2006;18(5):803–817.
[4] Gazzaniga MS. The Ethical Brain. Dana 2005, p. xix.
[5] Moll, de Oliveira-Souza, and Eslinger, Morals and the human brain.
[6] Greene et al., The neural bases of cognitive conflict and control in moral judgment.
[7] Borg et al., Consequences, action, and intention as factors in moral judgments.
[8] Cheshire WP. Glimpsing the grey marble. Ethics & Medicine 2007;23(2):119–121.
[9] Kurzweil R. The Singularity is Near: When Humans Transcend Biology. New York: Viking, 2005, p. 138.
[10] Wegner DM. The Illusion of Conscious Will. Cambridge, MA: Massachusetts Institute of Technology, 2002.
[11] Peterson GR. Minding God: Theology and the Cognitive Sciences. Minneapolis, MN: Augsburg Fortress Press, 2003, pp. 95–98.
[12] Doyle AC. A Study in Scarlet. In: The Complete Sherlock Holmes. New York: Gramercy, 2002, p. 10.
[13] Doyle, A Study in Scarlet, p. 18.
[14] Wilhelm Conrad Roentgen. On a New Kind of Rays. Eine Neue Art von Strahlen. Würzburg, 1895. In: Logan Clendening, Source Book of Medical History, New York: Dover, 1960, p. 670.
[15] Doyle AC. The Adventure of the Naval Treaty. In: The Complete Sherlock Holmes. New York: Gramercy, 2002, p. 194.
Editor’s Note: The views expressed herein are Dr. Cheshire’s own and do not necessarily reflect the position of Mayo Clinic or Mayo Foundation, USA. This article originally appeared in Ethics & Medicine: An International Journal of Bioethics 23, no. 3, (2007) and is used with permission.
