The TED Model of Mental Illness: Thomas Insel

In my abortive attempt to start a blog last year, I appeared to be very optimistic about some popular talks from the Technology, Entertainment and Design conference (widely known as TED, or the “Youtube for smart people”). More recently, I’ve come to recognize some major caveats with this model of education and public outreach.

Many of the most popular TED talks attempt to shed light on major issues of social and individual conscience in our modern society. The hallmark signature of these talks is that they are short, engaging and optimistic–which often tends to conflict rather jarringly with the reality of many of the topics being addressed. Whether it comes to solving climate change, soil erosion, water shortages, world poverty, economic woes or major illness, TED probably has a happy solution to share. But if the problem is big enough or complex enough and solutions are difficult, onerous and perhaps entirely beyond reach, then it almost seems disingenuous to paint the ubiquitously optimistic picture that TED proffers up to its viewers.

This morning I stumbled upon a recent TED talk about mental illness which really rankled. Thomas Insel, director of the National Institute of Mental Health (NIMH), proposes that we adopt a “new understanding of mental illness” to address the many challenges that we currently face.

The problem

Mental illness is a common, frequently early-onset condition which is considered the most common cause of major disability and is frequently implicated in premature death (suicide).

Insel likens mental illness to various other medical ailments such as stroke, heart disease, AIDS and leukemia which have been (somewhat) successfully controlled through early identification and treatment. If it is possible to identify and intervene early on, then perhaps success rates will increase dramatically.

I have to agree on this point–engaging individuals early on to prevent subsequent episodes of mental illness is a commendable objective. The big question is how do we identify individuals who are at-risk and more importantly, what is the most appropriate treatment to help prevent illness once these people have been identified?

The solution

Insel’s bright new model of the mind?

Mental illness is a “brain disorder”. Moreover, we need to see mental illness as a malady of the brain in order to solve it.

I stand in awe.

But given that the “brain” or neuroscientific model of mental illness is now into its third or fourth decade (hardly novel or inspiring) with little success in eliminating the high levels of disability due to mental illness, it is even more surprising that Insel believes the chief criticism of this model is that it is reductionist.

In fact, the problem is not one of reductionism but of incompleteness. Our understanding of the brain and what influences its development and current function is most certainly incomplete. Moreover, what we know about the external factors that can influence this brain (generally considered “environmental” factors such as societal, inter-personal, lifestyle factors and exposures) and how all of these things interact is necessarily limited.

Can Insel address this (obvious) point?

Well he counters the challenge of neuroscientific reductionism by arguing that the brain is complex (it has millions of neurons and synaptic connections!), and that current models of mental illness as a disorder of the brain are appropriately nuanced. Therefore, not reductionist.

So what is this new model?

Apparently, most of the popular mental disorders are disorders of connections, which is probably true in a generic sense. This includes major depression, obsessive compulsive disorder (OCD), post-traumatic stress disorder (PTSD) and schizophrenia.

In fleshing out this perspective in 2010, Insel notes in Scientific American that “over-activity” of Brodmann area 25 is frequently noted in Depression and appears to normalize with successful treatment. In PTSD, he notes “abnormal” activity in the ventral medial prefrontal cortex (vmPFC) and that rodent models of “fear conditioning” recapitulate this abnormal function in the rodent infralimbic cortex (IL) a possibly analogous brain structure. (Actually, Brodmann area 25 is located squarely within the vmPFC, and may very well represent the closest region comparable to the rodent IL, a curious coincidence.) In OCD, there is a loop between the thalamus, caudate and orbitofrontal cortex which is abnormally active and in schizophrenia abnormal development of the frontal cortex has been implicated.

In the current TED talk, Insel presents decade-old research on childhood-onset schizophrenia as a profound example. Children who developed schizophrenia at a young age had notably lower levels of grey matter throughout their neocortices than control subjects of the same age group. Moreover, Insel argues, we can use this sort of developmental trajectory to predict who will be at risk for schizophrenia years before symptoms develop.

Where Insel goes wrong

But it’s not at all clear how any of this will come to pass. Childhood schizophrenia is a rare form of a rare disorder, whereas many of the most common and disabling ailments may not even have an obvious developmental hallmark.

Moreover, Insel is still caught up on the most basic charge of reductionism: in fact, he has reinforced it. When he claims that depression, PTSD and OCD can be explained by “over-activity” at single neural connections and that this over-activity is a core feature of the disorder, this is inherently reductionist.

In reality, these features are potential biomarkers, not necessarily critical causal factors in a disorder’s progression. The brain activation studies indicate that “over-activation” more likely in groups of people who have current illness than those without, but in order to be useful for treating these disorders, biomarkers need to identify at-risk individuals. Very few neuroimaging studies have taken this step to identifying specific individuals (rather than average differences between groups) thus far. Not only is this correlation without causation, it’s not even necessarily very useful correlation just yet.

A second core issue is that Insel is disingenuous about how he discusses mental illness. He compares schizophrenia, and particularly a rare childhood-onset schizophrenia, to major depression, OCD and PTSD–disorders that are as diverse as they are dissimilar (also another form of reductionism). Critically, the research Insel presents from individuals with childhood-onset schizophrenia is collected over decades, and so might provide a useful measure to identify at-risk individuals early on and thus improve treatment (although, I am not familiar with any work to suggest that this technique can pick out individual at-risk kids before they develop either childhood or adulthood schizophrenia, nor whether this is even practical).

In contrast, the studies of individuals with depression and other more common mental disorders don’t allow us to predict who will develop a disorder before it comes about. In fact, these studies exclusively look at individuals who already have the disorder. In fact, they differentiate between people who have current symptoms and those who do not. A person with major depressive disorder but who is not currently symptomatic (currently feels well, but still “has depression”) would not be identified with the brain imaging methods available today.

In other words, Insel’s “early identification and prevention” premise is completely unsupported by the reality. In fact, the “evidence” he uses to support this strategy for going forward (from a very special case of childhood-onset schizophrenia) may have no correlate in very common and disabling disorders like depression. Even if we could implement society-wide brain scans at an early age, there’s a very good chance that there won’t be anything useful to discover there for predicting the course of common mental disorders.

Insel states that we “aren’t there yet, we don’t know what the tools will be, nor what to precisely look for in every case to get there before the behavior emerges”. He’s right about this. In fact, I would take it one step further: we don’t even know if there will be something to find when we arrive.


2 thoughts on “The TED Model of Mental Illness: Thomas Insel

  1. Sean Blackwell

    Very sharp article!
    Did you hear that Insel just threw out the upcoming DSM 5 at the NIMH, and in its place will be new criteria for mental ‘diseases’ which will be based on ‘biomarkers’ which, of course, psychiatry has been searching for for a century!!!!

    I’m going to enjoy this!
    Sean Blackwell

    1. zstansfi Post author

      I agree Sean. I just heard about this most recent push yesterday afternoon (see my comment here). Despite my above criticisms of Insel’s talk at TED (which I feel still stand) I think that the planned expansion of the Research Domain Criteria Project to define biomarkers without following specific diagnostic criteria offers quite a bit of promise. I think it’s almost without doubt that there are specific and measurable features of biology that will closely map on to some set of (currently unknown) symptoms or clusters of symptoms. Hopefully, these biomarkers will offer some insight into the causes and solutions for mental illness as well.

      This implies a major open question for current initiatives: can we identify enough valid biomarkers using the proposed new technologies to usefully influence treatment outcomes?

      Regardless, I think that even if this sort of “data-driven” approach fails to powerfully impact valid diagnosis and treatment, it will at least provide a strong means to test of the hypothesis that the most meaningful features of mental illness can be understood through an understanding of neurobiology and neurophysiology. A counter option might be that mental illness emerges between multiple constructs including measurable biological states, emergent psychological states and external/social influences. If a strong version of this latter position is true, it may require a good understand of not just the biology, but also its interactions at many different levels to adequately treat these disorders.


Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>