by SM Fitzpatrick · Cited by 12 — The detailed technical discussion found below provides a brief intro- duction to the technical methodology because with functional imaging the.
10 pages
292 KB – 10 Pages
PAGE – 1 ============
Chapter I 0 Functional Brain Imaging Neuro-Turn o1· Ul1·ong Tzwn? Susan M. Fitzpatrick Research questions of interest to neuroscientists share a natural overlap with those pursued by scholars studying philosophy, art, music, history, or literature. The common ground is a shared desire to understand the ings of the human mind. What initially attracts someone to study ence, regardless of what aspects of nervous system function an individual career may become focused on (e.g., basic functions of the synapse), is the allure of contributing knowledge that deepens our understanding of our minds. Many neuroscientists want to know how it is that the activities of the cells of the nervous system, individually and collectively, contribute to our ability to see beauty, take pleasure from reading and writing, appreciate music, mull over past experiences, make plans for the future, and plate “big questions” such as the purpose of life. Of course these are also questions motivating scholars in the humanities and social sciences. From this shared perspective it is easy to see the appeal the neuroscientific turn holds for those interested in melding collaborations among researchers from neuroscience and scholars from the humanities and social sciences. This essay does not address what is gained or lost by more closely linking theories, methods, and findings from neuroscience with the ries, methods, or findings in more humanist disciplines. Rather, this essay focuses on a narrower question: do functional neuroimaging gies offer a constructive path for the neuroscientific turn in the humanities and social sciences? By reviewing some common misunderstandings and misconceptions about what neuroimaging1 can and cannot reveal about the mind, particularly concerning BOLD functional Magnetic Resonance r8o Functional Brain Imaging Ł r8r Imaging (BOLD fMRI) or Positron Emission Tomography (PET), I hope to offer one answer to the question. Although techniques for monitoring brain function by monitoring physiological parameters have been available for more than a centu1y general appeal of functional brain-imaging techniques as a way of studymg the mind2 is a fairly recent phenomenon. As I will argue later in this essay, I believe the appeal and to some extent the misconceptions of functioning brain-imaging studies is largely due to how the results of such studies have been popularly portrayed both by scientists and journalists. I have closely followed the development of brain imaging since the late 1970s, first as a scientist in laboratories using both traditional laboratory techniques and PET and later while a member of one of the pioneering in vivo Magnetic Resonance Spectroscopy (MRS) ries studying brain metabolism. I initially joined the James. S. Foundation OSMF), a philanthropic organization supportmg mmd/bram research, in 1993 and was surprised at the enthusiasm for and rapid tion of brain-imaging technologies by cognitive psychologists. firsthand how difficult it is to bridge the gap from measurements of bram metabolism to inferences about brain function, I thought it premature to use PET and BOLD fMRI to make inferences about complex cognitive and behavioral functions. Little did I suspect that by the end of the 1990s the highly complex physiological data obtained by would morph into the brightly colored “brain scans” descnbed as showmg our minds in action and now ubiquitous in newspapers, magazines, and other media. Joe Dumit (who appears in the afterword to this volume) vides an interesting perspective on how functional brain images morphed from a scientific tool for visualizing highly complex neurophysiological clinical and research data to a familiar cultural icon in Picturing Personhood: Brain Scans and Biomedical Identity (2003). Dumit’s book focuses on ies using PET imaging, but the conceptual and methodological issues he discusses are equally applicable to studies carried out with the now more widely available BOLD fMRI. . . Functional brain-imaging studies about the neurological underpm-nings of psychopathic behavior, desire, maternal love, the belief in God, why we like luxury goods, and myriad other topics are a natural draw for scientists, scholars, journalists and the general public.3 But are these ics actually well suited for study by functional brain imaging? I hope this essay will convince readers that the answer to this question.is ‘no.’ their seeming accessibility, functional brain-imaging stud1e.s are to design, execute, and interpret. In my experience, functwnal-1magmg
PAGE – 2 ============
182 Ł THE NEUROSCIENTIFIC TURN studies posing questions about aspects of human cognition that are poorly understood at the psychological and neurobiological levels contribute to general misconceptions about brain and mind and, even more generally, the links between biology and behavior. Too often, the finding that changes in the brain account for some behavior of interest is often pitched as though the brain-behavior link is tbe new, interesting finding. There has been little doubt in either psychology or neuroscience since the nineteenth century that behavior arises from the brain and body. In fact, the role of the brain in behavior was acknowledged at the time of the Hippocratic writings: “From nothing else but from the brain came joys, delights, laughter and jests, and sorrow, griefs, despondency and lamentations” (Plum and Posner 19So, r). The results from BOLD fMRI studies are often discussed in the ular media as though differences in observable behavior reveal intrinsic, hardwired differences in brain structure. This common misinterpretation is largely due to a misunderstanding of the capabilities of function brain imaging. Much like the way the results of genomic ·screens are thought of as more convincing than mere family histmy (see Marcus zo10), brain scans seem more convincing with respect to how our minds work than psychological data even when the primary nature of the results in a study is behavioral (McCabe and Castel zoo7; vVeisberg et al. zooS). The idea that brain imaging reveals something previously hidden about the “real mind” has led to a rush to use the widely available technology of BOLD fMRI to “neuro-fy” a number of academic pursuits, resulting in emergent hybrid disciplines including neuroeconomics, neurolaw, neuroethics, keting, neuroeducation, and neuroaesthetics. What primarily strikes me when reading publications from these attempts is that the hybrids could be more accurately hyphenated with cognitive in the place of neuTo-as the findings rarely reveal novel insights about brain structure-function tionships (for a discussion of this point see van Eijsden et al. zoo9) but are usually reinforcing psychological findings. I will leave it to others in this volume to demonstrate to what degree the knowledge gained by such hybrid pursuits is beneficial, or not, to the ing evolution of various academic disciplines. The concern of this essay is how the embrace of functional brain imaging in the development of the emergent neurohybrid disciplines rests on misunderstandings concerning the design and interpretation of functional brain-imaging studies and what information (if any) these visually seductive brain scans actually depict. A number of thoughtful reviews discuss some of the problems ent in functional neuroimaging studies in the areas of neuromarketing and social neuroscience, including some discussion of the point I make above, Functional Brain Imaging Ł r83 that oftentimes brain imaging is used to merely add “scientific weight” to what is already known from behavioral findings (Ariely and Berns zoro; Cacioppo et al. 2003). If one takes tl1e critical discussions seriously it becomes clear that the primary problems with using functional brain ing in the “neuroscientific turn” are not technical (although there are nical concerns, to be sure) as much as conceptual. It is not simply a ter of doing such studies better. Most neuroscientific-turn questions are framed assuming that we already understand how neural substrates serve cognitive functions, but these are questions that still need much cal and experimental work (for a discussion of this point see Haxby zo10). My reading of papers from the emergent neurodisciplines suggests we learn very little of substance beyond what vvas already known from the nitive psychological or behavioral studies-except for perhaps some verging evidence that behavior is accompanied by brain activity (I return to this point in greater detail below). It is not quite clear to me what other organs in the body were/are hypothesized to be responsible for mental functions. Beyond showing that the brain activity is involved in ing, deciding, reading, feeling, and so on, what else is gained by intensive brain-imaging studies? One reason given for incorporating tional brain imaging into disciplines such as economics or law is that brain imaging can provide neural explrmations for complex behaviors. The ficulty of constructing explanations that cross levels of analysis from brain function to cognitive functions to behavior are rarely acknowledged4 but are central to the difficulties of designing and interpreting brain-imaging studies addressing questions of interest to the humanities and social ences. What Does Brain Imaging Image? Before turning to some examples concerning in what ways brain ing can and cannot tell us about the mind, it is worthwhile to review the scientific basis of functional brain imaging. Most of the functional brain images we have become accustomed to seeing are provided to us devoid of the experimental details. The lack of details makes it is easy to overlook the technical assumptions and caveats that are required to meaningfully interpret the images. The rapid surge in fMRI technology during the last fifteen years builds on a century of research on brain metabolism and blood flow (Raichle 199S) coupled with a long history of advances in magnetic resonance physics (Logothetis zooS). At one time, most of tl1e individuals using the tools required for functional brain-imaging studies were part of
PAGE – 3 ============
184 Ł THE NEUROSCJENTJFJC TURN or well-versed in these research traditions. Today, it is more important for newcomers to actively keep in mind that functional brain imaging, be it with BOLD fMRI or PET, does not dh-ectly measure what is commonly referred to as “brain activity” (Fitzpatrick and Rothman 1999). Functional brain-imaging techniques rely on the measurement of changes in brain blood flow or metabolism and depend on the obsenration that brain energy metabolism and neuronal activity are coupled. This pling is what allows researchers to noninvasively monitor brain function (Roy and Sherrington r89o; Siesjo 1973; Sokoloff 1981). A breakthrough in this storied research effort, and one that made human studies routine, was the development thirty years ago of PET (Raichle 1998). In combination with experimental paradigms and models developed in cognitive ogy, PET allowed the first high-resolution metabolic maps of functionally specialized regions of the human brain. A drawback of the PET technology was its reliance on cyclotron-generated short-lived radioisotopes. The sequent development of functional magnetic resonance imaging (fMRI), which is also a way to detect changes in brain metabolism and blood flow but does not require the use of radio-labeled compounds, made functional brain mapping widely available (Kwong et a!. 1992; Ogawa et a!. 1992). It is important to note that in the early days of PET and BOLD fMRI the collaborations typically involved researchers with strong experimental backgrounds in cerebral blood flow and metabolism, neurophysiology, and neuroanatomy working with experts in PET and MR physics. The ies were a direct outgrowth of the traditional attempts to understand the 1·egionnl couplings among blood flmv, metabolism, and function. The rise of cognitive neuroscience brought the addition of cognitive psychologists in the 1990s who added expertise in experimental tasks designed to use functional brain-imaging studies to further probe brain structure/function relationships. Given the current enthusiasm for fMRI in fields like cognitive roscience and social neuroscience it is easy to forget that the basis of the signal, the statistical analysis of imaging data, the design of the cal or behavioral components, and the interpretation of the results remain active and intense areas of research and controversy in the field of tive neuroscience even after twenty years of effort. The lessons from tive neuroscience should send a note of caution to emerging neurohybrid disciplines.5 The detailed technical discussion found below provides a brief duction to the technical methodology because with functional imaging the devil really is in the details. Prior to crossing too many disciplinary bound-Functional B rain Imaging Ł r 85 aries (remember-functional imaging already combines aspects of science, biophysics, statistics, and experimental psychology) it is important to have enough local knowledge to ask the right questions. The technical foundations of functional neuroimaging should be kept in mind even if one’s only interest is as a casual observer of functional-imaging results. The application of PET and fMRI to localize cognitive processes is based on observations that functional neuronal activity increases when a region is involved in performing a cognitive task (see Posner and Raichle 1994 for a description of imaging study design). These functional neuronal ties are responsible for information processing by neurons and other brain cells and include neurotransmitter metabolism and the generation of ronal action potentials. The energy required for these and other brain cesses is provided primarily by oxidative glucose metabolism (Siesjo 1978). Functional-imaging techniques typically measure glucose metabolism or neurophysiological parameters coupled to glucose metabolism, such as changes in blood flow, changes in regional blood volume, or changes in the rate of oxygen utilization (Sokoloff 1981). The most popular method for performing functional-imaging studies, BOLD fMRI, is sensitive to these physiological parameters but does not directly measure them. BOLD is sensitive to changes in the ratio of oxygenated to deoxygenated bin in the brain’s blood supply (Ogawa eta!. I 998). The ratio of oxygenated to deoxygenated hemoglobin is determined by the delivery of oxygen to brain cells and the rate tl1at oxygen is consumed by cellular metabolism. In a typical functional-imaging study, a subject performs experimental tasks while the signal (the nature of which depends on the technique used) is acquired. The fimctionnl image depicts the incremental change in nal intensity during a task relative to a baseline state in which the subject rests in tl1e scanner (or relative to a control task). Cognitive processes are localized by functional imaging using experimental paradigms and analyses based upon theories of cognitive neuroscience. As an illustrative example, consider a study designed to assess whetl1er a subject’s frontal brain regions are involved in the general cognitive skill of verbal working memory. The subject would perform tasks requiring tl1is cognitive skill, such as bering lists of words, while being “scanned.” In one strategy tl1e degree of involvement of verbal working memmy in each task would be varied, but the requirements for other cognitive skills (such as visual recognition) are held constant. The relative intensity of the functional-imaging signal in the frontal region during each task would be statistically correlated with the verbal working memmy component (Posner and Raichle 1994). During an fMRI experiment the “images” obtained, representing the
PAGE – 4 ============
186 Ł THE NEUROSCIENTIFIC TURN spatial distribution of MR signal intensity changes, are acquired rapidly (on the order of one second each). The large data sets so acquired require the experimenters to use statistical analysis packages to compare the images obtained during different cognitive tasks. The statistical son searches for brain regions showing different signal intensities during the different tasks. The final presentation, often called the brain activity map, is usually an image of these statistically identified regions, color coded to represent the level of statistical significance, overlaid on an anatomical MR brain image. From these maps, researchers make inferences about the neural correlates of cognitive processes (for further discussion of this point see Fitzpatrick and Rothman 2002). If, during conversations with researchers primarily interested in using fMRI as a tool to interrogate cognitive behaviors (and not actively engaged in the technical development of brain-imaging science), I pose the tion, “vVhat does the fMRI signal measure?” the answers (in decreasing order of frequency and increasing order of accmacy) tend to be: regional neuronal activity incremental changes in regional neuronal activity incremental changes in regional cerebral blood flow and/or metabolism As is evident from the technical discussion provided above, none of these descriptions is completely accurate. An MR physicist would describe the most popular fMRI method, blood oxygen level-dependent imaging (BOLD), as measuring the change in the intensity of the nuclear MR signal due to changes in the transverse relaxation time and phase of the protons of water molecules in the blood and brain tissue as a result of changes in hemoglobin oxygenation and blood volume. There are strong links between the BOLD signal and the underlying activity shifts of ensembles of neurons, and these links are active areas of research (Logothetis 2oo8), but at present none of the functional-imaging methods commonly used directly measures neuronal activity. The take-home message from the technical details above is that despite the language used to discuss them, the brain images displayed in scientific publications and in tl1e popular media are not representations of changes in brain neuronal activity, or areas of “activation,” or the brain “lighting up” or “switching on.” Brain scans acquired with fMRI do not even cally depict the magnitude of the BOLD signal. Rather, the images are computer-generated, color-coded “maps” of statistically significant parisons among data sets. It must be stressed that the finding of statistically significant differences and a measured change in the actual magnitude of Functional B rain Imaging Ł 187 the signal acquired are not necessarily interchangeable. The same BOLD signal may be statistically significant in one subject and not distinguished from noise in another subject. In addition to measurement sensitivity issues, identifying a brain region as having a statistically significant BOLD signal depends on how well tl1e time course of the signal agrees with the assumed time course of tl1e cognitive task. Depending on the nature of the assumptions made in the statistical analysis package, the same data could yield statistical maps identifying different patterns of activation. Another caveat in directly relating functional images to regional ronal activity is that even the fiery red spots in the statistical map image overlays from a typical fMRI study may represent changes in neuronal activity of less than I percent relative to the brain’s regional neuronal activity at rest (Shulman and Rothman I998; Gusnard and Raichle 2ooi), and differences between tasks are commonly on the order of o. I% of the brain’s activity when not engaged in an imaging experiment. The tion, largely from other imaging methods such as PET and MRS, that the changes in regional neuronal activity during tasks are quite small has led to the emergence of studies uying to understand the “resting state” through examining changes in the fMRI signal when the subject is not engaging in directed activity (Raichle 2010). It is crucial to note tl1at the lack of a signal change (no “lighting up”) does not mean that a region is not involved in the performance of the task studied. It may be tempting for someone interested in using fMRI to answer questions about the neural correlates of viewing impressionist paintings or to discover what brain regions “light up” when reading Chaucer to dismiss concerns about the missing links in the chain of reasoning from the fMRI signal to neuronal activity and instead concenu·ate on making inferences about cognition and behavior from patterns of brain activity. However, considering the extent to which tl1e methodology influences the results, it behooves researchers and scholars interested in using fMRI or other brain-imaging modalities to gain a deeper understanding of the theoretical and practical strengths and limitations of the techniques. During a cal BOLD fMRI experiment, for example, over Io million measurements will be made of which only a small percentage are actually relevant to the experimental question (Haxby 20ro; Vul and Kanwisher 2010; Logothetis 2008), and the risk of making interpretation errors is large. Furthermore, even if the imaging methodologies perfectly measured changes in neuronal activity my general concerns about applying functional imaging to ticular kinds of questions would still hold. Most important, the critical but often overlooked component of any meaningful fMRI study is that inter-
PAGE – 5 ============
r88 Ł THE NEUROSCIENTIFIC TURN pretation of changes in neuronal activity is meaningless without a themy of the cognitive operations involved in performing the experimental tasks, and a well designed set of tasks to test the themy. There should also be a testable hypothesis about the neurobiological underpinnings of the tive operations under investigation. I expand on the role of understanding the cognitive operations necessary for performing a task in the next section of this essay. In the absence of the framework of a cognitive theoty, it is not possible to make inferences about the relationships between an observed behavior and the patterns of brain activity as revealed by functional brain imaging (Poldrack zooS; Henson zoos). Functional Imaging and Cognitive Neuroscience The burgeoning interest in the use of functional-imaging studies to rogate the neural correlates of behavioral and social phenomena is due in part to the role functional imaging played in the development of cognitive neuroscience over the past two decades. The ability to use measurements of brain metabolism and hemodynamic properties to “map” psychological concepts onto physical neurobiological substrates seemed tailor-made for a ne\v academic discipline intending to uncover the neurological pinnings of cognition (Gazzaniga, Ivty, and Mangun zooz). As already discussed in this essay, the appropriate use of functional brain-imaging methodologies in cognitive neuroscience is still evolving (Haxby zo1o; Logothetis zooS) and serves as a cautionary tale for other interdisciplinary endeavors. A good example of how functional neuroimaging can add to our standing of the neural substrates of cognition is provided by a now sic paper by the Washington University group known for pioneering applications of brain-imaging methodologies (Petersen et al. 199S). The experimental task design for the study demonstrates how neuroimaging data, when combined with behavioral data and a cognitive theoty of task performance, can inform our understanding of the underlying neural tems. The paper summarizes a rather complicated series of experiments investigating a type of learning called procedural learning or skill tion. This type of learning is labeled by psychologists as “nondeclarative” in the sense it is difficult for a person to explain or describe what has been learned-although it is clear that performance improves over time. vVe have all had experience learning a task that seems effortful when new but with practice and time becomes easier and automatic, even though we not fully articulate exactly what it is that has been learned. Learning to ride Functional Brain Imaging Ł r 89 a bicycle or to float in water are familiar examples of how a task feels ferent” as we progress from novice to skilled performer. The Petersen et al. study was motivated by the existence of two sible proposals for how neural modifications could accompany skill sition. One explanation was that as a skill became more practiced and expert over time, the neural circuitty involved in the task became more “efficient.” The second explanation suggested that as a skill is acquired the nature of the “task” undergoes changes in the way the task is sented and processed in the brain as the novice performance becomes more expert. It was hypothesized that in the “different tasks” case the functional-imaging data would reveal the accompanying changes in ral substrates. Of course, Petersen eta!. were quick to point out that it was most likely that these two explanations need not be mutually exclusive and that skill acquisition could involve both mechanisms. In this essay I will only discuss one of the tasks used in the functional neuroimaging studies, a type of maze learning. At first, the performance of an individual engaged in maze learning is slow with many errors. With practice formance speeds up and errors decrease. Together with the behavioral data, the imaging findings supported the idea that as an individual became more expert at the maze task, the relative pattern of activity of different areas of the brain changed consistent with the hypothesis of a change in representation with experience. The study serves as a good model for, among other characteristics, its intricate and thoughtful experimental design and by today’s standards rather modest conclusions. Readers of this essay are encouraged to consult the original paper as it demonstrates how careful one must be when tying out and interpreting results from a functional-imaging study even for seemingly simple questions about how the brain supports mental cesses. Furthermore it built on a strong foundation of prior results. ropsychological findings from individuals with discrete brain lesions had previously indicated that procedural learning, unlike learning and memoty for “declarative” types of knowledge (e.g., where you went on your last vacation), is not dependent on the medial temporal lobe. Conversely, cedural learning can be impaired by damage to other brain regions, such as the basal ganglia and the cerebellum that have less impact on declarative types of knowledge. It is important to emphasize that the Petersen et al. study was preceded by a robust tradition of studying procedural learning in neuroscience and in cognitive psychology. Generally, functional-imaging data are more likely to advance questions in cognitive neuroscience when the experimental design can call on well-articulated cognitive theories
PAGE – 6 ============
190 Ł THE NEUROSCIENTIF!C TURN about the nature of the task, solid behavioral data, and neuropsychological evidence about the possible neural underpinnings. Going Beyond Cognitive Neuroscience vVhen functioning imaging tools migrate from cognitive neuroscience into other disciplines, even closely related fields such as social neuroscience, the likelihood that the experimental criteria required to meaningfully interpret functional-imaging data can be met is less certain. A recent imaging study, “The Power of Charisma-Perceived Charisma Inhibits the Frontal Executive Network of Believers in Intercessory Prayer,” ried out by Uffe Schjoedt and colleagues at Aarhus University in Denmark (Schjoedt et al. 2010) and profiled in a Sciemist news piece (Coghlan 20ro), provides a good counterexample to the Petersen et al. study. The studies from the Schjoedt lab are, in my view, typical of attempts to fy topics of research (social networking, influence) that are traditionally of interest in the social sciences. For the purposes of this essay, Schjoedt et al. 2010 provides an example of the use of fMRI that needs to be pursued with caution. Schjoedt and colleagues compared BOLD fMRI data acquired from self-described devoted Christians (primarily individuals belonging to charismatic Christian denominations) and self-described “secular” pants (the publication states these were mostly BA students) with no prior experience with prayer. The volunteers completed a questionnaire on the · strength of their belief in God and the pmver of prayer prior to the ing study. After the experiments the participants completed a questionnaire on perceived qualities of the speakers and the feelings they had about God’s presence. Brain images were acquired while each participant listened to eighteen different prayers recorded by three different male readers fied to the research subjects as a non-Christian, an ordinary Christian, and a Christian known for healing powers. The main hypothesis of the paper, that “participants’ assumptions about the speaker would affect the evoked BOLD response,” was tested by contrasting the scans obtained from the Christians and from the ists. The secular group showed similar BOLD responses across the three conditions in the study. Analysis of the scan acquired from the Christians “only revealed significant activations in one of the contrasts, namely, in the contrast ‘non-Christian’ relative to ‘Christian known for his healing ers.”‘ The main finding reported was the inhibition of a frontal executive network in the Christian participants. The paper speculates the tion” (technically a decrease in the measured BOLD signal relative to that Functional Brain Imaging Ł 191 found in the secular participants) contributes to the facilitation of matic influence. Like the Petersen et al. study described above, the Schjoedt et al. paper is rather detailed, and I recommend that interested readers consult the nal publication. Unlike the Petersen et al. study, which made an important contribution to our understanding of the neural mechanisms underpinning procedural learning, it is not quite clear, at least to me, what we learn from the Schjoedt et al. study. In light of all the caveats I have discussed that can alter the findings in a functional-imaging study, what new contributions to our understanding of behavior, cognition, or neural function does this study provide us? We do not have a well-characterized behavioral or nitive themy of charisma as we do for learning as measured by improved performance on the maze task and the other tasks used in the Petersen et al. study. There is not, to my knowledge, a testable hypothesis, based on cognitive science, neuropsychological studies, or findings from patients with brain injuries that could suggest how “charisma” is represented in the brain. The findings of the Petersen et al. study also sound a warning bell that should be heeded for any functional-imaging study-experience changes the cognitive representation and the neural underpinnings of tl1e task. The self-selected, self-identifying devout Christian subjects willingly admitted to strong faith and a deep belief in the power of intercessmy prayer, while the secular participants denied such beliefs. Listening to prayers being read is, at its most fundamental, a different behavioral and cognitive task for each of these subject groups. For Christians praying is familiar (practiced), salient, and rewarding, and it may evoke an array of rich associations. Prayers by other Christians can have a power that words recited by an identified nonbeliever may not. For non-Christians much of the rich context for prayer is missing. Is it surprising to see different patterns of BOLD responses as the two subject groups listened to prayer? What the study cannot tell us is 1vby certain individuals become devout Christians and others do not, despite the report of a fundamental ence between the brain scan patterns of the two groups. The Schjoedt et al. study does not consider a common confound with functional-imaging studies comparing two groups selected for their ences-it is not possible to control for experience, beliefs, prior knowledge, attentional effects, affect, or the many other factors we encounter in life that influence cognition and behavior. The criticisms I am making here are not dependent on the teclmical and methodological difficulties discussed above. To me, the Schoedjt et al. studies are conceptually problematic; they derive from misguided attempts to use BOLD fMRI to directly map mind
292 KB – 10 Pages