RDoC: We should look before we leap

We begin our investigations with patients displaying symptoms. We categorize and study them. We clarify the nature of their underlying disorders. Our efforts are not complete until we have returned to where we started and can explain to our patients how their symptoms arose

Kenneth Kendler

In 2013, the director of the US National Institutes of Mental Health (NIMH), Dr Tom Insel, announced that their funding model would no longer fund projects investigating Diagnostic and Statistical Manual of Mental Disorders (DSM) diagnoses. Instead, he said, they would seek projects linking supposedly basic neurobiological constructs to neurocognitive correlates, and eventually to mental illness. These neurobiological constructs are described as ‘brain circuits’ (Insel, 2013). These ‘brain circuits’ form the crux of a partially complete matrix called the Research Domain Criteria (RDoC). Mental ‘systems’, such as ‘acute threat’ and ‘reward’, are situated in the rows of the matrix, and they intersect with ‘units of analysis’ in the columns. These ‘units of analysis’ are genes, molecules, cells, circuits, physiology, behaviour, self-reports and paradigms (neurocognitive tests). The cells of the matrix are supposed to categorise and prompt new research.

It is informative to read Insel’s editorial writing in 2013 and 2014, and find it similar to Robert Spitzer’s, regarding the Diagnostic and Statistical Manual of Mental Disorders–Third Edition (DSM-III) in the 1970s. Spitzer believed the ‘two approaches, structuring the interview and specifying all diagnostic criteria … sponsored by the NIMH … will result not only in improved reliability but in improved validity which is, after all our ultimate goal’ (Spitzer and Fleiss, 1974). Insel (2013) believes the NIMH sponsored ‘Research Domain Criteria project to reformulate psychiatric diagnosis according, in part, to emerging biology rather than the current approach, which is limited to clinical consensus (will lead to increased) diagnostic validity’.

The DSM has a number of problems: It is too complex to be used routinely by clinicians (Gornall, 2013); it takes vague and heterogeneous symptom clusters and declares them to be coherent clinical entities known as ‘mental illnesses’ or the euphemistic ‘mental disorders’ (reification); it demands circular diagnostic research as inclusion in trials is determined by DSM criteria; it appears stringent at the same time as being actually loose so that diagnoses are highly heterogeneous and encompass too many non-forms or weak-forms (Galatzer-Levy and Bryant, 2013); many diagnoses have shown a reduction in diagnostic reliability in supposedly improved versions (Lieblich et al., in press); it does not incorporate neurobiological markers and its structure is not informed by pathophysiology. The problem of reification is immense because it confuses thinking about mental problems at all levels. The DSM used the language of the positivist (restricted to the consideration of empirical evidence) medical model to describe all the various kinds of distress and divergence from normative patterns of thinking, feeling and behaving that the pre-DSM-III mind was still trying to understand and in doing so delegitimised other forms of understanding.

In its current form, the RDoC makes clear that taxonomies of mental illness are provisional, and so refutes the reification of mental illnesses that has been a side effect of the DSM, but it remains an ostensibly positivist project, inasmuch as the patient’s subjective experience is objectified and made generic in its codification as a symptom, or constrained by itemised questionnaires, and is subordinate to measures of behaviour. It is only ostensibly positivist because although the means of enquiry are empirical, the bedrock of the discipline is in fact metaphysical, relying on ontological claims that certain patterns of thinking and behaviour represent entities called ‘mental illnesses’.

The RDoC encodes a project of global reductionism asserting that brain circuits are the fundamental units of the mind, making it hard for other philosophical approaches to contribute, and it does not have explicit outcomes for the phenomenological understanding of mental illnesses. That is, the RDoC does not seek to return our understanding of dysfunction in brain circuits to an understanding of the subjective experience of patients, therefore preventing us from explaining to them how their symptoms arose in the context of their own experience.

The RDoC and supporting editorial content (Insel, 2013) are couched in language that subordinates the medical model of mental illness to the neuroscientific model of mental illness. One may agree with the project of ‘starting from scratch’ with the diagnosis of mental illnesses, but nevertheless disagree with this change in the philosophical approach to the tools of the trade. In this way, the RDoC is a decidedly non-medical document. It applies the concepts and jargon of neuroscience and psychology. The terms used in the matrix align more readily with clinical psychology than with clinical psychiatry. The columns of the matrix pertaining to contact with human beings, the ‘self-report’ and ‘paradigm’ columns imply the practice of psychology that concerns itself with neurocognitive testing. The ‘self-report’ cell is in fact the only one that a clinical psychiatrist would ordinarily have the tools to interrogate. However, such a clinician would be unlikely to refer to this practice as ‘gathering self-report’. A rigorous phenomenologist may also take issue with the idea that the stipulated phenomenological method is to gather a self-report. In fact, self-reporting is associated with significant bias and therefore does not necessarily represent the subjective experience faithfully. If self-report is taken too credulously, then conscious and unconscious forces in the subject and psychiatrist that co-opt the self-report to obscure the subjective experience are not counteracted.

The notion that we can find ‘evidence based’ diagnoses without first engaging in a concerted project to decide on the appropriate type of evidence is troubling. Evidence comes in so many different forms, none of which are suitable for every application required of a system of classification of mental illnesses. Robert Spitzer in the 1970s sought valid forms in the patterns of speech and behaviour that arise in psychiatric patients. Tom Insel in the 2010s is seeking valid forms in the patterns emerging from genetic testing, neuroimaging and neurocognitive assessments. Whether we choose degree of agreement on the phenotype or degree of agreement on the endophenotype as our benchmark, we are nevertheless depending on reliability in both instances – that is, on a consensus of learned colleagues as to what does and doesn’t constitute a mental illness or a faulty brain circuit – while validity remains elusive. It seems unlikely that there will be a static entity in the anatomical unit of the brain that drives all of the changes notable in the patients treated by psychiatrists.

The RDoC is a clear vote of no confidence in the Diagnostic and Statistical Manual of Mental Disorders–Fifth Edition (DSM-5) by a major funder of mental health research, as much as it is a plan of attack for the eventual discovery and classification of valid clinical forms. The most important innovation of the RDoC is the return of all levels of the hierarchy (genotype, phenotype, endophenotype, aetiology, diagnosis) to provisional status and the commitment to an iterative approach that has been long considered essential to the gradual improvement of mental health taxonomies. However, Insel (2013) is too enamoured of the idea of ‘Faulty Circuits’, and this project seems to be an explicit declaration of global reductionism along materialist monist lines, that is, it declares that the mind and brain are made of the same stuff, declares that stuff to be the anatomical unit of the brain and tacitly favours anatomical descriptive terms over descriptions of the subjective experience. This declaration implicitly undervalues the empathic, imaginative and phenomenological aspects of psychiatric science. We suggest that a hybrid taxonomy of ‘brain diseases (and) problems in living’ must include as an essential component a phenomenological understanding of psychiatric patients. The RDoC might therefore be taken as an opportunity for a sustained iterative approach to psychiatric taxonomies without foreclosing prematurely or too narrowly on the nature or identity of its basic subunits. We support the notion that ‘brain circuits’ should be integrated into our understanding of mental illnesses, but declare there is a lot more to it than that.