RDoC Frequently Asked Questions (FAQ)
RDoC: its constructs, conceptual bases, the matrix
- Are the constructs currently detailed in the RDoC matrix the only ones that should be used in studies?
No. In keeping with the idea of RDoC as a framework for research, NIMH encourages the study of new constructs (or refinements to the current constructs). The constructs in the matrix are intended to represent particularly good exemplars and to serve as a starting point for research, but new (and modified) constructs are essential to the growth of the overall framework. We encourage discussion about constructs in the RDoC Forum, and have specific guidelines for proposing new constructs (see next question). Although RDoC RFAs to date have required applications to use constructs currently in the RDoC matrix, this has been done to ensure an initial set of projects devoted to the extant matrix elements; investigator-initiated projects are always welcome.
- How do I modify existing RDoC constructs, or propose new ones?
It is an expectation of the RDoC project that new and refined constructs, and the experimental paradigms designed to assess them, will be developed as more researchers use RDoC and provide data regarding validity and utility.
To propose a new construct, a researcher must meet the criteria used in the initial workshops. These are: (1) evidence for a functional behavioral or psychological construct, (2) evidence for a neural system or circuit that plays a major role in implementing the function, and (3) a putative relationship to some clinically significant problem or symptom (e.g., fear, anhedonia, hallucinations, memory, social cognition). In addition, the construct should have an appropriate “grain size”, i.e., a granularity that facilitates the study of relationships among measures from the various units of analysis; in other words, the functional aspect of the construct should be neither too broad nor too narrow to find meaningful relationships with its biological measures. The current RDoC constructs were devised with all of these criteria in mind, and so can serve as useful exemplars.
- How does the RDoC Matrix get filled? Would NIMH think, “we have these cells filled in the matrix, but we need more research in these other cells,” and how would that be communicated to investigators?
The RDoC Matrix is a compendium of tasks and measures that were gathered from topic experts during the RDoC Workshops. Curating and annotating the matrix is a work in progress. The matrix elements are good examples of what can be used and studied. They are not meant to be exhaustive or prescriptive; they do not dictate the measures or tasks that can be used in RDoC grant applications. Investigators may, however, find the matrix to be a convenient repository of tasks and measures when they are writing grants, or a useful resource when they are approaching a new topic of research. One way that the RDoC matrix may help to facilitate the scientific review culture is to provide a place for common terms and approaches.
Blank or sparsely populated cells in the matrix represent clear gaps in the field. They offer investigators an opportunity to conduct new studies to develop and/or validate appropriate measures for that area. As with many aspects of the RDoC framework, NIMH is not targeting particular cells at this time, leaving it to the investigator community to determine the high priority areas for grant applications. Occasionally, certain constructs may be highlighted in RDoC RFAs and on the RDoC website as requiring further study. Investigators are encouraged to submit applications to study any part of the matrix that they think deserves further research.
- Where does RDoC take development into account? And what about the effects of environment?
Two major dimensions of the RDoC framework involve environmental factors and neurodevelopment, and the interactions between them. These two factors might be considered as an overall context in which the constructs, as studied across multiple units of analysis, are explored. To date, funded RDoC grants have been distributed approximately equally between child and adult samples, and studies of developmental trajectories are strongly encouraged. Studies spanning the normal and abnormal development of neural circuits and behavior are of interest. RDoC does not specify particular points in development, nor particular types of environmental inputs, for study; this stance is intentional, since the goal is to liberate investigators to focus upon those aspects that they consider most relevant rather than imposing new constraints that might hamper their research designs.
- Why is RDoC reducing all behaviors to brain circuits and activity?
RDoC’s goal is not to reduce the understanding of disorders to circuits (or to molecular/cellular processes), but rather to integrate and synthesize multiple measures to lead to a more comprehensive understanding of the constructs in the matrix, and in turn to the symptoms to which they relate.
- Isn’t RDoC focusing on the biological at the expense of the psychological?
Mental disorders involve both psychological and biological components. That is why subjective experience and the questionnaires that measure them are a significant part of the RDoC matrix. In fact, the five major domains are based on prior work that has tried to understand human behavior using a variety of methods such as temperament and personality measures, and structural analyses of the patterning among multiple mental disorders (e.g., internalizing vs. externalizing, Big Five or Big Three, etc.). RDoC is an attempt to explore the biological systems relating to these psychological constructs starting with the knowledge we already have. By emphasizing quantified measures of behavior and cognition, as opposed to relatively subjective and vaguely defined symptom reports, RDoC aims to elevate the role of both psychology and biology in understanding and classifying mental disorders.
- I’m wondering what the current thinking is about the relevance of human neuroimaging in the RDoC initiative.
Human neuroimaging is very much relevant to RDoC. RDoC’s goal is to link up self-report and behavior to underlying neural circuitry and activity, where possible. RDoC’s focus is on translational and clinical research in human populations. As neuroimaging is one of the few tools that provide valuable insight into brain activity in humans, studies with experimental paradigms centered around neuroimaging will be relevant to RDoC.
- How much should an investigator be trained in neuroscience to do RDoC research?
While neuroscience training would be useful for RDoC research, it is not essential. As mentioned on the website, and as can be seen in the matrix, RDoC attempts to take into account relationships between different units of analyses such as self-reports, behavioral tasks, EEG, MRI, genetics, psychometric measures, etc. As such, research teams that use an RDoC perspective in their studies would likely be multidisciplinary in nature as well. Members of such a team should include investigators with training and expertise appropriate to the units of analysis included in any particular project. No one team member or domain would be singularly essential for all RDoC grants.
Putting together RDoC grant applications
- Imagine developing a biomarker to look at predicting treatment response to SSRIs in depression and anxiety. RDoC would encourage looking at the overall SSRI response, as opposed to just depression or just anxiety. However, rating scales for improvement in depression symptoms and anxiety symptoms are not the same, so it would seem like combining the clinical trial results of depression and anxiety would be like combining apples and oranges. How would an investigator finesse this?
We would encourage investigators to look at SSRI response in relation to RDoC domains (e.g., positive and negative valence systems and other relevant constructs from other Domains). Measures of depression and anxiety may also be helpful to link results to constructs that have been studied with reliable and valid measures.
- For a clinical trial, say of 'depression' treatment, one implication is for larger sample sizes in order to look at sub-groups. Is that right?
This would depend on how the experimental paradigm was designed. If the idea behind the study is that there are numerous subgroups in a DSM condition such as “depression” that could vary in different ways on treatment response in relation to some construct, then a larger sample size would indeed be necessary. If, on the other hand, you are studying one or two RDoC constructs (e.g., such as loss and threat) in relation to individuals who have a range of depressive symptomatology, then a smaller sample may be sufficient for statistical power and thus hasten relevant clinical findings.
Additionally, it is worth noting that as more researchers contribute to RDoCdb, it will become possible to combine data across studies from different sites for analyses that would have previously required large-scale studies.
- If I am interested in studying emotion regulation issues, but work in an area where threat constructs are not employed, how can I propose a study that addresses emotion regulation processes?
In this case, it would be helpful to study emotion regulation in relation to a construct that is relevant to your area of research. As we have outlined on the web site, regulatory or modulatory processes that can be viewed as inherently related to a construct of interest are heuristically considered to be part of that construct – e.g., as noted for threat regulation. Thus, as another example, regulatory processes that modulate certain types of aggressive behavior might be considered in terms of the Frustrative Nonreward construct. Self- or emotion-regulation will also be relevant to constructs within the Positive Valence (e.g. regulation of approach) and Social Processes Domain (e.g. regulation of social communication). The Cognitive Control construct may provide yet another conceptual approach.
Additionally, as mentioned elsewhere in the FAQ, new constructs that are not part of the RDoC matrix can be proposed, and this would apply to a regulating process involved with that construct as well. In such cases, the grant application should outline how the new construct fits RDoC principles – i.e., there should be (1) evidence for a functional behavioral or psychological construct, (2) evidence for a neural system or circuit that plays a major role in implementing the function, and (3) evidence of the significance for mental disorders.
- Could you speak briefly about best practices for approaching problems from an RDoC angle when you believe that your particular problem (e.g., aggression) involves dysfunction in two or more systems (e.g., negative valence systems and systems for social processes).
RDoC encourages the study of multiple constructs to examine complex clinical phenomena such as aggression (itself a heterogeneous concept; see the Proceedings of the Negative Valence workshop). RDoC does not specify a particular “best practice” since there are a variety of research designs and statistical analyses that can be used to study such issues. The most appropriate approach may depend upon detailed aspects of the research plan. Investigators are encouraged to discuss specific research plans with NIMH program officers prior to submission of a grant application.
- If an investigator is interested in social engagement as a mechanism that is related to positive and negative valence systems, how transdiagnostic should she be, as this affects many disorders?
The answer involves two related parts. First, RDoC domains and constructs cut across several disorders, and it may not be feasible to include subjects with all kinds of symptomatology. We would suggest starting in a relatively circumscribed way (e.g., subjects with varying degrees of symptoms for a particular class of disorder such as those in the schizophrenia spectrum chapter in the DSM-5). Based on results, including construct validation within that context, larger studies including subjects across various spectra of psychopathology could then be planned. Second, RDoC welcomes studies that examine multiple domains and/or constructs, such how Positive and/or Negative Valence affect social engagement.
RDoC grant reviews
- Do “RDoCian” grant applications go to different study sections at NIH?
Study section assignment depends upon the type of grant application submission. Applications that are received in response to an RDoC RFA will go to an NIMH convened study section; investigator-initiated grant applications (including those in response to Program Announcements (PAs or PARs) go to various CSR study sections according to the usual CSR referral guidelines.
- RDoC is a great idea, but my experience is that the study sections are ignoring it in evaluating proposals, with basic scientists not that interested in translation and clinical scientists still wanting to see DSM-5 criteria to define samples. How is NIMH going to address this issue?
While some anecdotal comments have noted lower enthusiasm for RDoC-themed applications in clinically oriented study sections, other investigators report hearing the opposite -- that applications planning to use DSM categories are downgraded for not involving RDoC dimensions. RDoC Unit staff are aware of these concerns, and are working to provide guidelines for review that will promote a consistent approach to evaluating applications with either RDoC or DSM orientations. It seems clear that reviewers in CSR study sections are becoming much more conversant with the RDoC framework and RDoC-oriented grant applications.
Basic scientists are welcome to propose research involving further study of the RDoC constructs at a pre-clinical level. These investigators are also encouraged to conduct studies leading to new knowledge about the neural mechanisms underlying complex behaviors which may relate to clinical symptoms when dysfunctional. Among other positive contributions, such studies may provide the basis for new RDoC constructs (see basic science FAQ).
- Can DSM disorder categories still be included in grant applications to NIMH?
While it has been stated that NIMH will re-orient its translational research toward RDoC, this does not mean that research with DSM categories is no longer accepted. Moving forward, it is important to enable the research community to build upon the research databases and literatures that already exist using DSM categories. Large GWAS studies represent one notable example where studies using DSM diagnoses are likely to continue for some time, in order to build upon extant data sets. Even for studies with clinical samples that emphasize RDoC constructs, it is reasonable to state in publications the numbers in various diagnostic groups, and to report the means and confidence intervals of the various measures for each group; this assists readers in understanding how RDoC data relate to traditional DSM analyses.
For studies designed around DSM categories, a variety of research designs can combine RDoC approaches with DSM samples – for instance, examining RDoC dimensions across two or three DSM groups (e.g., schizophrenia and bipolar disorder or multiple anxiety disorders). Grant applications that focus solely on comparing a single DSM category to controls may be useful if the major aims are to explore dimensions or subgroups within the category. Studies adopting the traditional design in which patients given a single DSM diagnosis (considered as a unitary disease entity) are contrasted to healthy controls, however, will be given lower priority. In all cases, it is important to consider strongly the dimensional aspects of the construct under study, such as patients not quite meeting criteria for a DSM diagnosis, controls who may have some history of past disorder, etc.
RDoC in basic science research and clinical services research
- Is RDoC interested in animal studies?
RDoC is a translational research project, so studies conducted under the RDoC aegis would be confined to humans. However, projects employing animals can play an important role in the overall RDoC project. See Simmons & Quinn (2013) for a thorough discussion of this aspect.
- Are basic researchers now required to study only RDoC domains and constructs?
No. To the contrary, NIMH continues to support a wide variety of basic research relevant to its mission. The Institute looks to basic researchers for new discoveries in behavioral and brain research that could potentially lead to new RDoC constructs, thus opening up novel studies in the clinical realm.
Insofar as RDoC represents a research framework involving the translation of basic-science constructs to clinical phenomena, basic researchers are also welcome to submit pre-clinical applications intended to further the understanding of the current constructs.
More information regarding NIMH priorities in basic research can be found through the Division of Neuroscience and Basic Behavioral Science websites and program officers.
- When can the RDoC approach be applied to the general population?
RDoC studies are encouraged to include the full range of functioning along the dimensions of interest, from “normal/healthy” through “sub-clinical” through various degrees of clinically significant impairment. The goal is to have constructs validated and applicable in such a way as to improve targeting and treatment effectiveness at different levels of impairment, as defined on an empirical basis. Studies focused solely on healthy human subjects fall within NIMH’s basic science programs/priorities.
- Can RDoC replace the DSM/ICD?
The RDoC project is still in an early stage, and cannot yet replace the current diagnostic systems. Because it is still an experimental framework only, the RDoC matrix should not now be used for clinical diagnostic purposes. The aim of the RDoC initiative is to stimulate a different approach to translational/clinical research, with the goal that the knowledge gained will inform future clinical diagnostic systems. We expect that such changes to diagnostic manuals would be phased in gradually as more investigators adopt RDoC principles in their studies and more constructs prove to be clinically valid and useful.
- Can RDoC be used for clinical care or everyday problems at the current time?
RDoC’s long-term goal is to provide more effective diagnosis for individuals, and thus is consistent with precision medicine approaches to diagnosis in other areas of medicine. While it may not be immediately applicable for clinical purposes, it is possible to envision how research on basic neural circuitry and functional systems involved in mental disorders can readily translate to both therapeutic and preventive interventions as RDoC increasingly builds upon its knowledge base in the years to come. As has been said, the road to better therapeutics starts with better diagnostics.
- Much current intervention research focuses on improving functional outcomes (e.g., employment and relationships) among people with specific disorders. Does this type of research relate to RDoC at all?
Yes. RDoC emphasizes specific behavioral or cognitive functions (such as working memory or social processes) that often have very specific implications for examining functional outcomes (as opposed to outcomes examining aggregate change across the multiple symptoms that comprise most DSM disorders). This research may play an increasingly important role as the best methods for implementing RDoC constructs in practice settings are developed. Such outcome variables may be located in the “Behavior” and “Self Reports” columns.
- Investigators in healthcare services and implementation science have found it difficult to frame research questions and methods using RDoC. Will there be additional guidance on this? For example, when studying healthcare services delivery for anxiety-spectrum, how does RDoC relate?
For the most part, services and implementation research at the current time involve traditional disorder categories since the use of RDoC constructs in clinical contexts is still so new. However, a number of investigators are developing innovative ways to examine outcome measures related to RDoC constructs in electronic health care records, ambulatory monitoring, and treatment evaluation metrics. Interested investigators are encouraged to contact NIMH program staff to discuss examples and possible applications.
RDoC and existing datasets
- How does RDoC impact non-RDoC studies that are already in progress? For example, are there any ways in which those studies should be adapting data collection and analysis plans to incorporate the RDoC framework?
It depends on where the study is in its timeline. If still early in the data collection phase, investigators may consider adding relevant RDoC measures to their protocol. However, if the study is already further along, investigators may decide that it is prudent to stick to their proposed data collection protocol. All are encouraged to analyze data on a dimensional basis alongside their original design factors.
- Is there any way that I can analyze already existing data sets in terms of RDoC domains and constructs?
Yes. See PAR 14-008 , “Secondary Data Analyses to Explore NIMH Research Domain Criteria.”
- When will RDoCdb be ready for querying data for analysis?
RDoCdb is part of the NIH/NIMH Data Repositories (NIMHDR). As with other NIMH Data Repositories, investigators upload data to RDoCdb every 6 months in January and July. Descriptive data (those data that simply describe the characteristics of the sample, and are not related to the primary aims of the study) remain in a private state for up to four months after submission, at which time they are then shared. Analyzed data (those data related to the primary aims of the study) remain in a private state until such time as they are published (see rdocdb.nimh.nih.gov ). Since the first projects submitted to RDoC RFAs were funded in 2012, it will be some time before data from these projects are shared and available for querying. However, NIMH has provided supplemental funding for some relevant ongoing studies whose data are closer to publication, so these may be available in six to 12 months. Also, there are some data from projects that have been entirely completed that are available for querying and analysis; check with RDoCdb staff .