A Framework for Mental Health Research (RDoC)

Although the Research Domain Criteria project (RDoC)  is not particularly new (the description of it on the NIMH website is dated June 2011 and it’s been on my reading list for at least a year), there has been a lot of attention drawn to it recently. This is partly because the DSM-5, the new psychiatric diagnostic handbook, is due to be published on May 18th, prompting Thomas Insel, director of the National Institute for Mental Health (NIMH), to recently write about the necessary next steps in mental health research (see here and here). I’m sure many others (for example The Neurocritic) have written about this recently. Having finally gotten around to reading about RDoC on the NIMH website  myself, I wanted to briefly summarise what it is and why it’s brilliant. If you haven’t heard of RDoC or keep meaning to look it up, this is for you.

The RDoC project is a framework for thinking about and researching all aspects of human psychopathology/mental health, without confining the research to existing diagnostic labels. The DSM-5 is the best thing we currently have for the purpose of clinical assessments and diagnosis, in the hope of trying to treat and improve the lives of people with mental health difficulties. But research suggests that it isn’t good enough.

In reality, mental health conditions overlap greatly, both in terms of clinical presentation, associated features (e.g. cognitive difficulties) and in terms of apparently non-specific risk factors (e.g. many genetic variants have now been shown to play a role in more than one condition (e.g. schizophrenia, autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD), all previously thought to be quite distinct conditions)). Even within a diagnostic category, there is a lot of variability in severity and also, you don’t need a diagnosis to have problems; symptoms below diagnostic thresholds often cause difficulties for affected people. It has become clear that mental health problems are not binary. Instead, there appears to be a continuous distribution of mental health difficulties, ranging from none to very severe.

The RDoC framework is about cutting across these diagnostic labels and looking at the underlying dimensions of behaviours and measures of neurobiology. The idea is that recruiting participants to a study on mental health based on their diagnoses and then trying to determine how they differ from “healthy” controls in order to inform research on diagnoses, is actually a little circular. The alternative approach, suggested by the RDoC framework, is to recruit participants with a range of related problems (e.g. all types of mood disorders) and look within this group.


(Example of the 2 main dimensions of the RDoC Framework)

The framework divides up mental health into a list of constructs (such as ‘Responses to potential harm (Anxiety)’, ‘Reward Learning’, ‘Cognitive Control’ or ‘Social Communication’) within more general domains (e.g. Cognitive Systems). It then divides research approaches into units of analysis (e.g. Genes, Molecules, Cells, Observed Behaviour etc.). Two other dimensions are ‘Developmental Aspects’ (how these constructs change over time) and ‘Environmental Aspects’ (how the environment affects and interacts with the constructs).

The hope is that considering mental health in terms of these dimensions rather than diagnoses, will serve as a research framework for improving our understanding of mental health and creating better diagnostic categories for the future. This framework seems to me a much more valuable way of doing research in this area. It reminds me of a great blog post by Dorothy Bishop from 2010, in which she argued that neurodevelopmental problems should be considered on a number of developmental dimensions, rather than as discrete clusters of difficulties (i.e. diagnostic categories). I was very inspired by this way of thinking when I first started my PhD and so I think it’s great to see that the NIMH is encouraging researchers in mental health to adopt this way of thinking.

Concerns about genetic testing for autism/ASD

Autistic spectrum disorder (ASD) is a collective term for neurodevelopmental problems involving social & communication deficits as well as restricted and repetitive behaviours. Scores of studies over the past couple of decades consistently agree that genetics play a very important role in this condition (with estimates of heritability around about 80%).

Naturally, this observation has opened up a floodgate of research trying to hunt down the genetic factors that may be involved. The genetics of such a complex and variable condition, however, are (unsurprisingly) incredibly complicated. It’s quite clear that there is no “gene for autism/ASD” – there is no one, single genetic factor which can explain every single case of ASD. There are also environmental factors involved, as well as epigenetic changes (where the structure of DNA is unaffected but what genes actually get expressed (i.e. used) in the body is changed). But the story is still not that simple. There are likely to be a good deal of genetic factors involved (some estimate over 1000), most likely interacting together within biological systems (known as gene x gene interaction). What’s more, the exact pattern of which of these genetic factors are involved is likely to differ from individual to individual. This is complementary to the variability of symptoms of ASD and associated difficulties, from each autistic person to the next. As the saying goes, “if you’ve met one individual with autism, you’ve met one individual with autism”.

I’ve read a fair amount on the subject, but I’m relatively new to this field and I’m certainly no expert. Nonetheless, I was mildly concerned when I read a recent post on the SFARI website about a “new diagnostic” test for ASD. The test apparently involves Next Generation Sequencing (NGS) in an attempt to examine an autistic individual’s DNA for point mutations of single nucleotides within 62 genes that have some small level of support in possibly being implicated in some cases of ASD. NGS sounds fancy and expensive. That’s partly because it is a very new (and exciting) technique and as such, it genuinely is quite costly – though getting cheaper by the moment.

However, all the reading & literature searching I’ve been doing over the last few months on the genetics of ASD makes me a little worried about what possible benefit individuals with ASD and their parents could actually get out of this particular “test”. More established cytogenetic tests for gross chromosomal abnormalities, such as fragile X, which have been around for a while, already have a body of robust scientific research and seem to have real-life implications and clinical applications. However, NGS is quite different and as it is a pretty new technique, there is not enough robust evidence out there yet to definitively implicate any particular factors in ASD, much less definitively implicate all of those 62 genes. 3 very recent exciting studies published in Nature (discussed quite well here, here and here) are the biggest ones to date using NGS to investigate ASD; yet still the specific genetic mutations they implicate are far from being definitely confirmed as risks for the condition. Their effect sizes are quite small and some of them are likely to be false positives – see Dorothy Bishop’s blog post for a good explanation of the meaning of effect sizes in genetics. Moreover, scientists don’t even know what exactly most of these genes do in our bodies, where they are expressed or what other genes they interact with. We are still a very, very long way from practical, clinical applications.

So why is this company pretending otherwise? What good will it do for a given family to find out that their child has a rare mutation at a single point (or even at multiple points) in these 62 genes? Every single human being, with or without a neurodevelopmental problem, has the odd rare mutation. It takes a lot more than such a tiny blip in our DNA to overthrow the entire brain and create as complex a set of symptoms as are involved in ASD. Telling a family such a thing seems to me to be a little meaningless and unhelpful, until more research is done.

I feel a little bit mollified having scanned through the website of the company who provide this test, as they claim to be not-for-profit and provide families with a package of support involving genetic counselling, which is an important step up from just handing people a genetic test report without any help in interpretation. I think they may be genuinely trying to do something useful and they even appear to be involved in research, which is surely a good sign. I can only assume that some of the other services they offer on their website have a direct benefit to people. I’m just a little worried that there will be other companies who will charge these extortionate amounts of money and just won’t be able deliver anything of actual tangible help or even much in the way of reassurance to families, given the current level of knowledge and understanding on the subject.