Nosological Reflections

Dimensionalization as the Path to Validity

Eventually, the Task Force’s ambition to achieve a paradigm shift, coupled with its desire to tackle validity, led it to view dimensionalization as the means by which it would correct for the failings of the DSM’s categorical system and the rigid, perhaps arbitrary, boundaries it drew (Whooley and Horwitz 2013). The Task Force announced that “one of the major—if not the major—differences between DSM-IV and DSM-5 will be the more prominent use of dimensional measures in DSM-5” (Regier et al. 2011:xxvii).

The logic of the Task Force’s plan for dimensionalization was as follows. Since DSM-III, every edition of the DSM has classified mental disorders as discrete categorical syndromes, composed of clusters of symptom manifestations. To meet a diagnosis, a patient must demonstrate a certain number of symptoms within the syndrome. Categories and symptom criteria are organized around an either/or logic: if a patient meets the proposed criteria, he or she is seen as having a clinically significant psychiatric disorder; if not, the patient does not receive a diagnosis. Dimensionalization would change this. Rather than viewing mental disorders as categorically distinct from normal states, dimensions represent mental illness and well-being as existing on a spectrum, with mental disorders conceived of as extreme variations along that spectrum. Patients would be assessed along a continuum, and diagnoses would be expressed numerically.

Dimensionalization was to be achieved by including two different types of numerical scales: an overarching “cross-cutting” scale and severity scales for each diagnostic category. The intent of the former was to move away from reification, of thinking of DSM categories as discrete things, and to encourage reasoning that looked across categories by assessing numerically some common psychiatric symptoms. Diagnostically, this would provide clinicians with a general psychiatric profile of each patient. The severity scales would foster dimensional reasoning within categories by measuring the severity of the identified diagnosis. Each work group was charged with creating its own severity scales. In relation to reification and comorbidity, the introduction of these two types of scales would make the boundaries between disorders more porous.

The promise of dimensionality was that in place of the crude either/or categories of DSMs past, DSM-5 would allow the construction of more sensitive measures that perhaps could better pick up the neurological and genetic variance and thus assist in capturing the “the biological basis of mental disorders” (interview, Task Force member). A Task Force member explains how dimensionality relates to validity:

I think the most important reason for introducing dimensions is that most areas of psychopathology present in a continuous dimensional manner. They do not present themselves in a dichotomous manner. So the more we can capture that variance the more we are representing the way they appear in nature. And after all that’s the goal of the diagnosis is to be able to plan treatment, determine outcomes and predict outcomes and understand the conditions. And so a continuous measure looks like it represents nature much more readily for almost all areas of psychopathology.

The previous categorical model has failed in achieving more valid categories because it misspecifies how we should be thinking about mental disorders. By introducing a dimensional logic, through the inclusion of numerical scales, the Task Force sought to reconfigure psychiatric nosology by making it more sensitive to capture genetic and neurological mechanisms: “I think the real driving issue [for dimensionalization] is the increased emphasis on the biological basis of psychiatric disorders. And I think there’s a growing realization that until we have dimensional measures we’re not going to figure that out” (interview, Task Force member). Thus within the process, dimensionalization came to signify the means by which DSM-5 would accomplish a paradigm shift in how psychiatry understands mental disorders.

Dimensionalization Unrealized

Although simple in theory, the construction of dimensional scales proved difficult for the work groups. The project of dimensionalization proposed by the Task Force would involve quantifying mental illness, transforming discrete symptom presentation into numerical values on a scale. Diagnoses, according to this plan, would be accompanied by a number. However, the process of commensuration and quantification is inexact (Espeland and Stevens 1998), as the work groups’ struggles in developing severity scales attest. Because of proprietary issues, the APA could not use existing scales in DSM-5, so the work groups were charged with coming up with new ones. “Developing scales is a very complicated process” (interview, work group member), because many of things that the scales intend to measure are hard to quantify. Furthermore, the Task Force adopted a hands-off approach to the work groups, giving them few instructions on how to construct scales. The problem was, lacking some overall guidance, “different work groups and different individuals have interpreted dimensionality differently” (interview, internal critic). Some groups suggested severity scales that simply counted the number of symptom criteria a patient met. Others were more ambitious and tried to quantify the criteria themselves, which could be weighted and added together to achieve some sort of composite score. Some of these latter proposals would involve computer-generated algorithms. The end result was a diversity of scales, with different logics, “a hodgepodge,” according to one critic, that threatened to undermine the DSM’s consistency.

The inherent difficulties in creating dimensional scales were compounded by the DSM-5 Task Force’s orientation toward the revision process. Despite scaling back its initial ambitions, the Task Force continued to promote the scales as a paradigm shift. The persistence of such rhetoric stretched the concept of paradigm shift rather thin, as the proposal sought to merely introduce dimensional logic into the manual by hitching scales to the existing categories, leaving the categorical system of DSM-III intact. But whether the talk of paradigm shift accurately reflected the new goals, it shaped the Task Force’s bottom-up, hands-off approach toward the work groups. In placing “no constraints on the degree of change” (Kupfer, quoted in Roan 2009) on the work groups, the Task Force hoped to encourage innovation; what they bred instead was confusion. One work group member who eventually resigned from the process described the situation: “I don’t think there was anything in the way of clear guidelines.” This led to fragmentation, most evident in the proliferation of diverse severity scales with different underlying logics. Absent “specific directions coming from the Task Force about how this should be done,” the work groups approached dimensionality in an “uncoordinated way,” with “some favoring existing instruments and others, you know, taking different approaches” (interview, work group member). Such disorganization and inconsistency provided fodder for DSM-5’s critics to argue that the process was being mismanaged.

While the Task Force pursued dimensionalization in a disorganized fashion, resistance to DSM-5 among psychiatrists grew. The rhetoric of “paradigm shift” raised red flags for many psychiatrists, particularly members of previous DSM Task Forces. Both Spitzer and Allen Frances, the chair of DSM-IV, began a campaign against the DSM-5 revisions that focused on the perceived disorganization and lack of transparency of the process (Frances 2009; Spitzer and Frances 2010), the Task Force’s perceived lack of attention to clinical utility (First 2005), and the construction of untested and unproven dimensional assessment tools using a process unsuited for such a task (Frances 2010a). At the core of these criticisms was a sense that the Task Force, in trying for a paradigm shift, was acting too rashly, courting public scorn and unintended consequences (First and Frances 2008; Frances 2009, 2010b). Whereas the Task Force felt that something needed to be done to address the lack of progress in validity, these critics urged patience, arguing that the DSM-III would yield the desired results if given more time. Moreover, radical changes would hinder research as it would introduce chaos and problems of continuity. Critics countered the Task Force’s desire for revolution with a commitment to iteration, the incremental accumulation of scientific evidence (Kendler and First 2010). Behind this relatively abstract debate over how psychiatric nosology should proceed was a concern for the unintended consequences for making changes too prematurely, a position expressed vehemently by Allan Frances, who has expressed concerns that revisions he oversaw in DSM-IV that led to diagnostic inflation (Bastra and Frances 2012) and false epidemics (Frances 2012b). Within psychiatry, the conflict between the DSM-5 Task Force and its critics turned vitriolic and personal, culminating in the publication of a New York Times op-ed piece by Frances (2012a) that made the stunning suggestion that psychiatry should give up its monopoly over writing the DSM, as it had proved itself not up to the task. With the debate becoming increasingly public, the Task Force’s proposals came under increasing scrutiny by APA officials, rank-and-file psychiatrists, and other mental health professionals.

In May 2012, these issues came to a head at the APA’s annual meeting. The APA Assembly, citing the undue burden the new scales placed upon clinicians, voted unanimously to relegate all dimensional scales to the appendix, crushing what little hope was left for a radical change to the DSM-5. They claimed that clinicians found the scales too cumbersome, reflecting the fact that they were created in a hurried, disorganized fashion: “There’s no rush to make this diagnostic system. It can wait until it’s justified. And it should be justified” (interview, assembly member). The Task Force reluctantly accepted the assembly’s ruling and, despite years of work on scales, left them out of the diagnostic criteria altogether. In a letter written by Dilip Jeste (2012), president of the APA, announcing the board of trustees’ approval of the final manuscript in December, dimensionalization was not even mentioned. And despite the manual’s endorsement of dimensional thinking in its introduction (American Psychiatric Association 2013:5–6, 12–13), DSM-5 includes only cross-cutting scales as suggested “techniques to enhance” clinical decision making in need of “further study” in section III, “Emerging Measures and Models” (American Psychiatric Association 2013:733); the severity scales are absent altogether, save for the inclusion of a three-level severity scale for autism spectrum disorder and an eight-item scale on psychosis symptom severity. With the last remaining innovations relegated to the section on further study, the final version of the DSM-5 displays only modest changes that belie the Task Force’s initial bluster.

The Emergence of RDoC as an Alternative

The debate over the revisions and the ultimate failure of the DSM-5 Task Force to achieve its ends raises real questions about the future of the DSM. DSM-5 will likely still reap tremendous profits for the APA, as it remains integral to the bureaucratic functioning of the mental health system. Its useful coding system is not going anywhere. But for decades, the DSM has been the authority in the conceptualization of mental distress. Although often maligned, its professional, scientific, and popular influence cannot be denied. However, the fallout from the revisions and subsequent developments indicate that the DSM may no longer be the platform through which major conceptual efforts are sorted out. For the first time since 1980, uncertainty proliferates in the world of psychiatric nosology.

One indicator of the DSM’s changing status comes from the Task Force itself, with its determination that the DSM will now become a “living document.” A member of the Task Force explains the reasoning behind this decision:

Essentially the idea is not to wait until 18 years go by, or 12 to 18 years [to] go by. If new discoveries come up that are relevant to diagnosis and treatment to not wait until the next DSM comes along to incorporate it.

The move to a living document is represented in a change to the naming convention of the DSM; the title will shift from roman numerals to Arabic numbers (DSM-5 to be followed by DSM-5.1 and so on). The implication of this decision is that knowledge of psychiatric disorders will emerge piecemeal and changes to the nosology should adjust accordingly. The revision of the DSM will no longer become a major event that involves a wholesale reflection on the part of the profession every decade or so.

Most telling, however, is the emergence of RDoC, an alternative nosological framework developed by the NIMH in a process separate from the DSM. On April 29, 2013, days before the release of DSM-5, the NIMH announced on its Web site that it was moving away from DSM by prioritizing research that eschews DSM categories altogether (Insel 2013). The announcement portrayed the DSM symptoms-based approach to classifying mental disorders as retrograde, a vestige of a more primitive era of medicine. Two geneticists cited in the announcement argue, “At the end of the 19th century, it was logical to use a simple diagnostic approach that offered reasonable prognostic validity. At the beginning of the 21st century, we must set our sights higher” (Craddock and Owen 2010:95). According to Bruce Cuthbert and NIMH director Tom Insel, the DSM system, based on presenting signs and symptoms, does “not adequately reflect relevant neurobiological and behavioral systems—impeding not only research on etiology and pathophysiology but also the development of new treatments” (Cuthbert and Insel 2013:126). Insel was more explicit in a New York Times article: “As long as the research community takes the D.S.M. to be a bible, we’ll never make progress. People think that everything has to match D.S.M. criteria, but you know what? Biology never read that book” (quoted in Belluck and Carey 2013:A13). Thus, the NIMH, reacting to the problems with the DSM-5 revision, has determined that the DSM is not an effective vehicle through which to address validity. For the Task Force, the rise of RDoC presented a challenge, for “the timing was not the best” (interview, Task Force member).

RDoC signals that the research community within psychiatry is moving on from the DSM. The infelicitous timing of the announcement, coupled with the DSM-5’s inability to make strides in addressing validity, signals a potentially radical change in psychiatric nosology. Indeed, existing concerns over the DSM, highlighted dramatically in the revision process, have led the NIMH to decouple its taxonomic efforts from the concerns of clinical practice. Arguing (like many on the Task Force) that artificial DSM categories have hindered its research objective of achieving a biological understanding of mental disorders or, at the very least, the development clinical tests for mental disorders (Kapur, Phillips, and Insel 2012), the NIMH is now promoting RDoC as the preferred framework for researchers seeking funding. The NIMH plans to develop this parallel nosology of mental disorders, which would advance research on the etiological underpinning of mental disorders that, when complete, would lead to valid diagnostic categories.

RDoC begins from three assumptions: that mental illnesses are brain disorders, that the tools of clinical neuroscience (i.e. functional neuroimaging) can identify the dysfunctions in neural circuits that cause mental illnesses, and that data from genetics and neuroscience will yield biosignatures that can be used to diagnosis mental disorders (Insel et al. 2010). As a Task Force member noted, its “basic proposition is that we need to understand pathophysiology at the neuro-circuit level.” The hope is that building a nosology from the bottom-up using “recent breakthroughs in genetics and molecular, cellular and systems neuroscience” (National Institute of Mental Health 2011) will lead to biological and neurological understandings of mental disorders and, in turn, produce more valid categories. In contrast to the DSM, RDoC encourages researchers not to think in terms of discrete categories. Rather than parsing distinctions into finer and finer categories, the goal is to develop research on “major domains of functioning” (National Institute of Mental Health 2011), by funding research that focuses on “fundamental biobehavioral dimensions that cut across current heterogeneous disorder categories” (Cuthbert and Insel 2013:126) to identify a common underlying biological, neurological, or genetic cause. For example “clinical trials might study all patients in a mood clinic rather than those meeting strict major depressive disorder criteria,” wrote Insel (2013). “Studies of biomarkers for ‘depression’ might begin by looking across many disorders with anhedonia or emotional appraisal bias or psychomotor retardation to understand the circuitry underlying these symptoms.” In constructing its research “matrix,” the program proposes seven classes of variables on which research should focus: genes, molecules, cells, neural circuits, physiology, behaviors, and self-reports (National Institute of Mental Health 2011). Although RDoC remains very much in the preliminary, abstract stages, it proffers an understanding of mental distress as a brain disorder, stressing that “core aspect” of all seven classes of variables are circuits. It is upon research on circuitry that RDoC’s promise as the new vehicle for paradigm shift rests. Unencumbered by clinical concerns—which, in the words of one work group member, leads to the “lowest common denominator” in the DSM—it is hoped that the new nosology will achieve a valid system for categorizing mental disorders rooted in neurology.

Decontextualization and the DSM

In maintaining the symptoms-based categories of DSM-III, DSM-5 perpetuates the decontextualization of previous editions. The DSM intentionally constructs its symptom criteria to avoid contextual concerns, seeking to achieve reliability by defining mental distress to manifest symptoms, often defined without consideration of their contextual appropriateness. Indeed, the goal of DSM-III was to construct transposable diagnostic categories to increase reliability across contexts. The whole point was to minimize idiosyncratic contextual considerations by standardizing criteria. Decontextualization has gained momentum over recent revisions, as the criteria have become less and less amenable to the appreciation of the social embeddedness of the symptoms themselves. For example, depressive and anxiety symptoms are measured as present and, in turn, pathological, if they meet a certain threshold, regardless of whether these symptoms are an appropriate response to distressing social factors (Horwitz and Wakefield 2007, 2012). The latest manifestation of DSM’s decontextualizing impulse is the elimination of the bereavement exclusion from major depressive disorder. In previous editions, doctors were instructed to exclude the diagnosis of major depressive disorder if a patient’s depressive symptoms resulted from the grief of losing a loved one. The logic behind the exclusion was that such symptoms were not indicative of a mental disorder, when contextualized in the experience of loss. However, the DSM-5 has eliminated the bereavement exclusion and, with it, one of the few explicit (and symbolic) bulwarks against the decontextualization of psychological distress in the manual.

Moreover, DSM-5’s focus on validity this time around promoted further decontextualization by explicitly embracing a neuroscientific and genetic model for identifying the etiologies of mental disorders. This represented a shift in DSM rhetoric around etiology. Previous editions claimed to be agnostic toward etiological theories. To minimize controversy, the DSM-III Task Force decided to avoid etiological theories by focusing narrowly on symptom manifestations (American Psychiatric Association 1980:7). Many have questioned the authenticity of this agnosticism (e.g. Horwitz 2002:73–74; Kirk and Kutchins 1992:111–16), but it did leave room for ambiguity and alternative understandings of the root causes of particular symptoms. However, with the DSM-5 Task Force’s commitment to tackling validity through neuroscience and genetics, the long-held, tacit assumptions of the creators of the DSM has been made explicit: mental disorders are problems of the brain.

It should be noted that even if the dimensional scales had been adopted in DSM-5, the problem of decontextualization would have remained. Dimensionality is not a solution to decontextualization. Indeed, in the discussions over dimensionality, the Task Force and work groups never considered the issue of decontextualization. The various scales, though diverse, typically hew closely to the existing criteria. In turn, they display an asocial character. Therefore, although questions revolved around how best to conceptualize mental disorders, never once did the Task Force consider the possibility that a more valid conceptualization of mental distress might involve an acknowledgement of contextual factors and the social embeddedness of mental distress. Valid diagnoses, in this case, were treated as synonymous with biological and neurological definitions of mental disorders.

RDoC: Decontextualization through Decoupling

RDoC is poised to accelerate the trend of decontextualization that began with DSM-III in two crucial ways. First, in its commitment to producing classifications for researchers only, RDoC has decoupled the construction of its nosology from the exigencies of clinical practice—exigencies that, in many ways, forced the DSM to retain (however reluctantly) an appreciation of social and environmental elements in mental disorders. By decoupling research and clinical concerns in developing its nosology, RDoC expressly brackets out the subtleties and complexities of the clinical encounter (in which contextual reasoning remains relevant).

One of the long-standing tensions at the heart of the DSM is that it proposes to serve many different aims and interests with its one-manual-fits-all approach. According to the DSM-IV-TR (American Psychiatric Association 1994), “The utility and credibility of DSM-IV require that it focus on its clinical, research, and educational purposes and be supported by an extensive empirical foundation” (p. xxiii). Although DSM-IV-TR claims that “fortunately, all these many uses are compatible with one another” (p. xxiii), this has not been the case. Researchers have sought to make the categories as scientific and accurate as possible, whereas clinicians have adopted a more practical orientation, seeking a system that can work within the confines of clinical practice.

The arguments over clinical utility during the DSM-5 revisions underscore the disconnect between the members of the revision process and clinical realities. The researcher-dominated Task Force made gestures to clinicians’ needs, but the nature of the debates over clinical utility indicate a basic lack of understanding of the exigencies of clinical practice. With the shifting emphasis on genetic, pharmacological, and neurological research in the past few decades, there are few resources for psychiatric researchers who study clinical practice. Moreover, the field trials, the only element in the revision process specifically gauged to analyze the use of the DSM, was narrowly focused on whether the proposed changes are reliable, not on the more complex dimensions of clinical utility. Given the paucity of existing research on clinical use, the discussion over clinical utility and dimensional scales never got beyond superficial talking points. Both sides, those for and against the revisions, marshaled clinicians to support their positions. Those in favor argued that clinicians want dimensional scales to assist in diagnostic decision making and assist patient monitoring (i.e., Kraemer 2007). Those against the scales countered that past efforts to introduce scales have been roundly ignored by clinicians (i.e., First 2005). The appeal to clinicians in support for diametrically opposed positions in the DSM-5 debate underscores the extent of the divide between clinicians and psychiatric researchers. Although most work groups proposed simpler scales, this made it “almost inevitable that those scales will not be used for research” claimed a critic of DSM-5. And even these simple scales were ultimately rejected altogether by the clinician dominated APA Assembly. This outcome led some to question the DSM’s long-standing commitment to providing a single manual:

So it gets back to my more fundamental question which is that the DSM is a single book that is not good for anybody. Like it’s not good enough for researcher for a number of reasons, and if you make it good for clinicians . . . it’s the lowest common denominator that creates a problem. (Interview, work group member)

Given these tensions, the NIMH has decided that it is time to decouple the interest of clinicians and researchers when developing a psychiatric nosology and has positioned RDoC as a “research classification system rather than one intended for routine clinical use” (Cuthbert and Insel 2013:127). This decision is driven by RDoC’s desire to capture the complexity of neuroscience and genetics, a project that would be difficult, if not impossible, to realize if its nosology had to also be useful for clinicians. Eventually, according to the NIMH’s proposals, the research findings that RDoC yields will form the basis of a clinical nosology, but there is no need to hamper research from the start.

However, there is a cost to decoupling, and that cost is further severing our conceptualizations of mental distress from considerations of its social embeddedness, because in many ways, the need to address the vicissitudes of clinical practice serves as a counterweight to the DSM’s tendency to decontextualize. Although the DSM has become explicitly decontextualized, implicitly, the commitment to clinical practice has meant that in its use, context is regularly applied to the various criteria in assessing particular patients, to avoid false positives. DSM-5 itself avows that a “careful clinical case history” should include social factors (American Psychiatric Association 2013:19), despite the fact that such factors are typically excluded from the criteria sets. Moreover, Wakefield and First (2012) demonstrate that because psychiatry lacks biological tests, it is incumbent upon clinicians to determine whether the symptoms experienced by patients are in fact dysfunctional and not responses to normal stressors. To do so, they gather additional information to situate these symptoms in their social milieus and the experiences of the patients. Wakefield and First further argue that an appreciation of context is not just a matter of clinical practice, but contextual reasoning is often implied in the DSM text itself. Therefore, although context is often written out of the explicit criteria, it reemerges tacitly in the diagnostic decision-making process.

The NIMH, on the other hand, is committed to ensuring that RDoC remains unburdened by the constraints of making its nosology useful for clinical practice. RDoC promises to allow psychiatric research to go wherever it needs, and it has committed to going into the brain. And if it is successful in decoupling, the understanding of mental disorders will no longer be kept honest by the vicissitudes of clinical practice and will be further removed from their social context. This has some involved with the DSM-5 concerned: “I think the distance between practitioners and researchers who are the ones who really focus on brain function is increasing and I don’t think that’s necessarily good for the field” (interview, work group chair). In its agenda, RDoC sacrifices the unofficial space for contextual reasoning allowed by the DSM.

RDoC: Decontextualization via Brain-Centrism

Freed from the exigencies of the clinical encounter, RDoC intends to pursue mental disorders primarily through the laboratory. The second source of RDoC’s increased decontextualization is its overwhelming focus on the brain. RDoC unambiguously embraces the view of mental distress as a problem of neurocircuitry and has constructed a research agenda that approaches them as such. Although such a program may not explicitly reject the relevance of social dimensions, the NIMH’s public statements about RDoC reveal an agenda that clearly demotes the relevance of the social. This focus raises long-held concerns by medical sociologists, especially the worry over medicalization (see Conrad 2007; Link and Phelan 1995)—the focus on the individual at the expense of the social, the failure to contextualize risk and health outcomes, and the resulting depoliticization of health problems.

RDoC’s brain-centrism does not preclude sociological investigation. Indeed, research in social epidemiology and health disparities demonstrates that the biomedical model of disease need not prohibit considerations of the social. Nevertheless, ontological positions—what the object of analysis is—are closely bound to epistemological positions—how said object can be known. Underlying concepts and definitions are not determinative, but they are constraining; as Ludwig Fleck (1979:8) observed, the “definition selected still determines some conclusions” as it becomes intertwined in the basic assumptions of a thought style. Assumptions about the phenomenon under study shapes how we seek explanations for it and what types of methodological approaches present the most promising venues to account for it. Although the brain does not have to be treated as isolated in the skull (and indeed, good neurology treats the brain as interactive with its surrounding environment), the assumptions and tendencies that come along with RDoC’s brain-centrism encourage research designs that tend toward reductionism.

RDoC’s brain-centrism raises two distinct but interrelated questions vis-à-vis the concern over decontextualization: (1) whether its ontological commitment (i.e., mental disorders as problems in brain circuitry) is warranted and accurate and (2) whether the construction of a research agenda along these lines not only allows for but promotes a nuanced appreciation of social factors acting upon the brain. Both questions must await final answers, but early indicators raise some red flags for those proffering a socially contextualized view of mental distress.

The ontological issue is the most complicated. RDoC is making a particular bet that mental disorders are problems with neural circuitry. This is not a critique but a banal observation culled from the history of science; all paradigms and research programs rest on a degree of faith (Kuhn 1996). Researchers need to make some assumptions about what it is they are studying in order to study it. However, recognizing this allows one to bring some historical perspective to the RDoC’s bet. Historically, understandings of mental distress have fluctuated between a biomedical disease model, built on assumptions of internal physiological or neurological dysfunction, and more environmental conceptions, which view it as a by-product of unhealthy environments or problematic social constraints. In viewing mental disorders as “biological disorders involving brain circuits” (Insel 2013), RDoC explicitly embraces the biomedical model of mental distress, joining a long tradition of seeking a physiological, neurological, and/or biological cause of mental disorder that has remained vexingly elusive. Many physiological bases for mental illness have been proposed in the past to great fanfare—disruptions in blood circulation, focal sepsis, overstimulation of nerves, faulty neurotransmitters, genetic markers—only to give way to disappointment. Indeed, the history of the search for psychiatry’s “holy grail,” the underlying mechanisms of mental disorders, is one of cyclical patterns of enthusiastic optimism and fatalistic pessimism, but ultimately a story of failure (Grob 1998). Pinning down mental distress has proven frustratingly difficult, and any new attempt to do so must be assessed in relation to this history. Of course, the historical record is not inherently damning to RDoC; its neural circuitry model may very well pan out. Nevertheless, the historical record is not promising, and given this, it is reasonable to question whether psychiatry is chasing “mythical entities” (Mirowsky and Ross 1989:17) that might be more accurately thought of as something altogether different. Therefore, it is important to point out that these ontological assumptions have epistemic ramifications; namely, given the nature of the object of analysis, how should one go about securing knowledge about it?

The more practical question is whether a research agenda built on such brain-centric ontological assumptions leaves room for the consideration of the social in its actual research designs. Although RDoC is very much in its preliminary stages, insight into its proposed research agenda can be gleaned through the NIMH’s public statements on RDoC, and these public comments are not reassuring when it comes to contextualization.

According to a member of the DSM-5 Task Force who is favorably disposed toward RDoC, the goal is “to use fundamental building blocks of brain structure and function to define meaningful clusters of either symptoms or pathology that will lend themselves to treatment and genetic and clinical neuroscience research” that ultimately will “be the building blocks for DSM-6.” As such, the NIMH proposes to harness “the power of modern research approaches in genetics, neuroscience, and behavioral science to the problem of mental illness” (Insel and Lieberman 2013).

Translated into research designs, the commitment to “basic behavioral neuroscience” means promoting laboratory-based research designs using imaging technologies such as functional magnetic resonance imaging (fMRI; National Institute of Mental Health 2011). In reaction to the problems with the discrete DSM categories (see above), RDoC will promote studies that analyze dimensions of function in the brain, defined as “behavioral entities tied to neural systems” (Cuthbert and Insel 2013:129), which cut across traditional diagnostic categories. The NIMH has outlined these “fundamental units of analysis,” or dimensions in a matrix, and has grouped them into five general “domains of functioning”: “Negative Valence Systems (i.e., systems for aversive motivation), Positive Valence Systems, Cognitive Systems, Systems for Social Processes, and Arousal/Regulatory Systems.” Studies will focus on a single dimension (e.g., “acute threat”) within one of these domains (e.g., “negative valence systems”) (National Institute of Mental Health 2011). Although schematic at this point, the NIMH has offered some examples of what this might look like in concrete terms. One such example taken from the NIMH’s Web site is as follows:

A design to study fear circuitry might thus have as inclusion criteria all patients presenting at an anxiety disorders clinic. Classification variables: The construct of interest is Fear/extinction, in the domain of Negative Affect. The independent variable for grouping would be the extent of responding to fearful stimuli using a measure such as amygdala response (from fMRI) or fear-potentiated startle (i.e., a circuit-level variable). Dependent variables would be symptom measures on various fear and distress measures, in order to test hypotheses about mechanisms by which hyper-reactivity and hypo-reactivity to threat cues affect the nature and severity of presenting symptoms. (National Institute of Mental Health 2011)

In this example, the abstract goals of RDoC become more evident. First, rather than focusing on a particular mood or anxiety disorder, this research “cuts across” DSM diagnostic categories by sampling patients with different types of anxiety disorders (e.g., general anxiety disorder, obsessive compulsive disorder, social anxiety disorder, phobia, panic disorder) to identify the underlying mechanisms of all these diagnoses. Second, to examine the “fear/extinction” dimension, it seeks to expose participants to a fearful stimulus in a laboratory setting and assess its effects on neural circuitry—amygdala response—using fMRI. With its emphasis on neural circuits as the “core aspects” (National Institute of Mental Health 2011) of the variables under consideration, RDoC espouses a commitment to designs that use, among other things, new imaging technologies in controlled laboratory settings.

Such research, born from RDoC’s brain-centrism, raises two issues related to decontextualization. The first relates to the level of analysis. RDoC encourages researchers to focus on seven classes of variables: genes, molecules, cells, neural circuits, physiology, behaviors, and self-reports. Excluded from this list are social and environmental variables. RDoC does acknowledge a place for additional “elements” that are “not represented in the matrix per se, but are considered to be critical elements in research fostered by RDoC” (National Institute of Mental Health 2011). But if this is the case, if the social and environmental factors are critical, why not include them in the formal matrix? Why not treat social and environmental variables on par with its other variables? The exclusion of social variables from RDoC’s guidelines for researchers suggests that social factors will be, at best, reduced to a residual concern and, at worst, ignored altogether.

The second issue revolves around the artificiality of the laboratory setting. The prioritized research designs lean heavily on the laboratory, adopting “behavioral neuroscience” as one of their models. However, experimental designs that seek to induce stimuli in the laboratory have a tendency to construct designs that display superficiality. Complex experiences that are operationalized to fit the laboratory context take on a certain “thinness” and, in the process, get reduced to shallow facsimiles of the experience outside the laboratory that is of interest. This problem is evident in behavioral neuroscience research on morality. Rather than addressing complex, deep issues of morality, research in this field is organized around thought experiments that use thin concepts that are “methodologically more tractable and theoretically more docile than thick ones” (Abend 2011:151). The use of thin variables raises questions as to how well findings might hold up in explaining thick phenomenon in the complex social world. To what extent do the anxious conditions, the fearful stimuli, to which subjects are exposed in the laboratory capture anxiety produced by the stressful social environments outside the lab? If we take seriously the idea that RDoC itself promotes, namely, that the “central nervous system is exquisitely sensitive” to its environment, we need critical appreciation and assessment of how the artificiality of the laboratory environment affects the brain and what shows up on functional magnetic resonance images. RDoC’s brain-centrism brings with it the potential to neglect the manner in which the brain interacts with its surroundings, by focusing solely on what happens in the brain. If we recognize the brain not as isolated but ever interacting with its surroundings, how much can findings produced in a highly controlled context inform us about mental distress in other contexts, especially when the relevant social variables are excluded from RDoC’s matrix?

The concerns regarding RDoC and decontextualization outlined here are offered tentatively and preliminarily, gleaned as they are from proposals and not actual research. The devil is in the research design details, and such research may ultimately incorporate social and environmental dimensions. But this does not mean that such concerns are without warrant. The science promoted by RDoC is built upon laboratory methods. The question, long prominent in the sociology of science, is an epistemological one: what insight can be achieved from the artifice of the laboratory setting? If mental distress is inherently related to social context, as many sociologists hold, what can be gained from research designed to eliminate context from consideration through laboratory controls?