The RDoC Controversy: Alternate Paradigm or Dominant Paradigm?

While DSM and ICD are diagnostic manuals, RDoC is not. What RDoC is, instead, is a set of principles for deriving and studying biobehavioral constructs implicated in psychiatric disorders. Current diagnostic categories are valid for many purposes and represent the best effort to relate science to nosology. There is no better alternative at present for clinical application. However, there are very significant limitations in the scientific application of heterogeneous syndromes where the defining psychopathologies are not unique to the diagnostic class and vary within categories between cases. Consider the limitations of schizophrenia as the phenotype in genome-wide association studies, where individuals in a study cohort differ on virtually every clinical variable. Hence, results from such studies yield mostly weak findings with unknown relation to specific pathology and false negative exclusion of genes based on failure to apply robust phenotypes to select relevant individuals.

RDoC provides one of several paradigms addressing syndrome heterogeneity. DSM-5 provides eight symptom dimensions for psychotic disorders in section 3. These psychopathology domains require clinical attention and are targets for research. Translational research involving animals is used to generate knowledge hypothesized to relate to specific pathologies relevant to diagnostic class but not to represent the range of pathologies associated with the class. The Consortium on the Genetics of Schizophrenia (4, 5) and the Bipolar-Schizophrenia Network on Intermediate Phenotypes (6) endophenotype programs are outstanding examples of the deconstruction of clinical syndromes.

Science resulting in advances in clinical care has been modest. In this context, NIMH gave emphasis to RDoC as a paradigm with the hypothesis that research based on already defined behavioral constructs with known neural circuits will accelerate the development of fundamental knowledge applicable to psychopathology while reducing problems associated with heterogeneous clinical syndromes. The RDoC initiative neither replaces nor excludes other paradigms, but it is an NIMH strategic plan priority and competes for resources. In this regard it will eventually be judged by whether knowledge generated within this paradigm is validated in relation to human pathology and whether clinically relevant innovations are produced.

The issue is not whether RDoC is an alternative diagnostic approach but whether the paradigm will promote information that informs nosology and facilitates application of brain science in clinical diagnostic concepts. In addition to freeing investigators from the approach of diagnostic class versus normal control, possible outcomes relevant to DSM and ICD are that:

In the above examples, new information from RDoC and from other sources would be used to revise classification (think DSM-5.1, DSM-5.2, etc.), and brain science may be more closely related to clinical constructs over the long term. Of immediate interest is whether RDoC-generated knowledge will identify novel targets for therapeutic and prevention discovery. The current RDoC model can be viewed at https://www.nimh.nih.gov/research-priorities/rdoc/index.shtml and includes self-reports to facilitate translation to clinically assessed psychopathology.

In the context of frustration with the slow pace of advancing knowledge on major mental disorders, there is a concern that NIMH is moving away from clinical trials and psychopathology research in favor of more basic science. RDoC may be misunderstood in this regard. What the RDoC paradigm endeavors to do is to alert investigators to the fallacy of methods that equate diagnosis with a specific disease (7). Investigators are encouraged to address psychopathology in the context of hypothesized neural circuits and behavioral constructs with the freedom, but not the requirement, to work across diagnostic boundaries. It is hoped that selected behavioral constructs will increase the validity of mapping phenotypes from human to rodent and that exploring cellular and molecular mechanisms will be more informative when related to specific neural circuits and behavioral constructs. RDoC principles also include the view that psychopathology is dimensional and that a sharp line dividing pathology and normalcy is often not present. Studies embracing these principles have been conducted before the RDoC initiative and presently both within and apart from the formal RDoC initiative. The categorical methods common when contrasting normal volunteers with a diagnosis cohort or between diagnostic cohorts are outside the RDoC framework but are not excluded from NIMH support. However, if heterogeneity undermines hypothesis testing, it needs to be addressed. Confusion early in the RDoC initiative at the application and review stages has created misunderstandings in this regard and has contributed to negativity in the field.

The NIMH strategic plan calls for psychopathology more informed by brain science. RDoC is a tactical approach. It is not intended to increase basic science and decrease clinical science. Rather, the hypothesis is that translational research will be enhanced and that clinically relevant knowledge will be accelerated. It is too early to determine the effectiveness of the program, and it is good for the field to debate the best paradigms for advancing knowledge. But RDoC is compatible with the public health mission of NIMH and is motivated by the limited success of traditional paradigms.

RDoC is not intended to be a comprehensive approach to all the several hundred diagnostic categories described in DSM-5. The chapter on Schizophrenia Spectrum and Other Psychotic Disorders gives emphasis to the five symptom criteria for schizophrenia and additional dimensions on depression, mania, and cognition. The five RDoC domains are substantially relevant to cognition, depression, and negative symptoms but are less developed in relation to reality distortion, thought pathology observed in speech, and psychomotor abnormalities. Examining dimensions across diagnoses may identify critical similarities or differences between schizophrenia and bipolar disorder (8), for example. RDoC is a work in progress, and methods with relevance for these pathology dimensions may be addressed within the five RDoC domains or by establishing further domains (9–11).