Sunday, October 16, 2016

The Balanced Rhetoric Against Neuroscience

The New York Times editorial pages continue to be a place where anti-neuroscience rhetoric can be expressed primarily as decreased funding or more accurately portion of the available NIMH funding.  Maybe there has been some pro-neuroscience opinion expressed there and if there was I have missed it.  I recently posted an exciting development in neuroscience teaching for psychiatrists and psychiatric residents.  In that post I reference an opinion piece by Richard Friedman, MD a psychiatrist (1).  Dr. Friedman makes several arguments for psychotherapy as if it is unrelated to neuroscience and based on that premise concludes that there is no substitute for psychotherapy, that people are more than a brain in a jar, and that anyone benefiting from psychotherapy seems to prove  that.  I found that to be an incredible statement considering that (according to Koch in above graphic):  "The brain is the single most complex object in the universe." There is also the fact that with 7.4 billion people on earth - there are 7.4 billion unique conscious states - the vast majority of which are not accurately described by any DSM or psychodynamic diagnosis/formulation.  All the time that Dr. Friedman is mounting this critique he also discusses the importance of clinical research and suggests shifting the funding balance away from neuroscience.

In the recent case John C. Markowitz a professor of clinical psychiatry at Columbia has a more subtle form of the argument.  In this case and the previous opinion piece the authors both endorse the importance of neuroscience to a point.  In this case the argument is - yes neuroscience is important but let's reestablish balance between neuroscience and clinical studies such as looking at the efficacy of psychotherapies.  Breaking it down, Dr. Markowitz makes the following points:

 1.  Under the directorship of Thomas Insel, the NIMH clinical research budget was "strangled" and the resources were diverted to neuroscience research.  The author acknowledges both the need for neuroscience research and the primitive stage of psychiatric diagnostics based on clusters of signs and symptoms.  This was really the basis for Insel's RDoC initiative looking at more reliable markers of psychiatric syndromes.  Any practicing psychiatrist who has seen all of the iterations of the DSM realizes that we are as far as we can go with this manual.  That includes from the standpoint of validity but also in terms of the clinical examination by psychiatrists.  As long as we are all contained by this manual, the clinical method of psychiatry will remain stuck somewhere in the 1940s.  That should be extremely disconcerting to the profession and future psychiatrists.

DSM technology is extremely limiting in terms of the usual clinical trials.  The NIMH sponsored Star*D study is a decade and a half old at this point.  It has defined the response rates for both antidepressant therapies and provided a discussion point for psychotherapy trials of depression.  Clinical trials of antidepressants provide an equally varied result.  Any practicing clinician knows that these studies are all seriously flawed out of the gate by using DSM diagnoses and also an intent-to-treat analysis that does not resemble clinical practice.  The variation in diagnoses from depression to anxiety to depression plus anxiety as seen in clinical practice should point to the fact, that patient selection into clinical trials currently results in very heterogenous patient populations in terms of both therapeutic effects and medication tolerability.  We can continue to spend large sums of money on these trials of mixtures of patient populations and post modest positive results or we can attempt to identify patients who will respond specifically and not experience side effects from a particular therapy.  That is the real promise of neuroscience based research.

2.  The patients who need help are poorly served by current neuroscience research.  The helpful psychotherapies listed by the author like interpersonal psychotherapy (IPT), cognitive behavioral therapy (CBT), and other psychotherapies have been around for decades.  I happen to have copies of Interpersonal Psychotherapy by Klerman, Weissman, Rounsaville, and Chevron and Cognitive Therapy of Depression by Beck, Rush, Shaw, and Emery.  The publication date of the former is 1984 and the latter is 1979.  Both therapies have been out there for over 30 years.  At this point both have been studied hundreds of times.   Looking at clinical trials on Medline yields 1711 for CBT and 261 for IPT.  Not only that but some of the clinical trials that were successful (like IPT for cocaine use) have never made  it into clinical practice.  In fact, in most places getting a therapist who actually practices any of the specific research proven psychotherapies is impossible.  The problem does not seem to be a lack of psychotherapy research but a lack of access to practitioners who use research proven psychotherapies.  Mental health treatment is the most highly rationed treatment resource and additional studies that continue to prove that existing psychotherapies work seems superfluous at this point.  Any current studies are often compared to existing therapies and with the DSM problem contributing to diagnostic heterogeneity.  Any new trials should only be funded for serious conditions where the therapy might be useful.  There is no reason to expect that a new therapy applied using the current diagnostic system or clinical trials technology will lead to any enhanced treatment effects.

3.  Existing treatments are not "good enough".  The author attributes this "good enough" statement to Insel himself.  I understand the point he is trying to make.  The author points to continued suffering, treatment failures and suicides as evidence that more is needed now.  The problem is that there is no assurance that clinical research will add any more at this time.  Certainly a focus on suicide as a stand alone problem (not suggested at all by DSM) and on serious disorders with no treatment like adult anorexia nervosa is warranted.  But even then we are left with a clinical trials technology that consistently produces modest results at best.  More multimillion dollar trials of psychotherapy that we already know is somewhat effective when patients have no chance of ever receiving it against a backdrop of "is this really depression or anxiety" seems like a waste of time and money to me.  It seems like a much better idea to develop a neuroscience method to determine who needs psychotherapy and who might benefit from medications.  But even then, the only treatments that will be readily available will be the medications and even then less than half of the affected patients will get access to treatment.  Good luck trying to find a psychotherapist and an insurance company willing to cover the cost of the number of sessions used in the psychotherapy research. Research proven therapies are only as good as the number of practitioners using them and access to those practitioners.

4.  The placing all of your eggs in one basket argument.  This is basically saying that if the ratio of clinical to neuroscience funding is 10% to 90% the risk is missing something big in the clinical research and not getting any useful results from neuroscience.  Given the history that I have provided, there needs to be a clear advance on the clinical side in order to fund large trials.  It does not make any sense to continue to  fund more of the same  or slight modifications of treatment for common disorders.  Our eggs have been all in one basket and I would call that treatment as usual.  In the 30 years that I have been in practice, there is nothing that I would call a major breakthrough.  Clinical research results come and go.  Effective psychiatrists are effective psychiatrists not based on breakthroughs but how they approach clinical practice.  Even that mode of treatment is threatened by widespread support for "collaborative care" that is being justified using the same kind of research that justified managed care in the first place.  In the end there has been nothing more destructive in terms of access to care for mental disorders than managed care.

In many ways these ongoing arguments resemble the arguments of the biological psychiatrists and psychotherapy psychiatrists that I trained under in the 1980s.  Many programs were split under this artificial division with the residents left to identify with biological or psychotherapy faculty.  It is interesting to note that this division occurred at a time when Kandel wrote a paper on how psychotherapy is neuroscience in action (3).  That may have been missed because the biologically based psychiatrists at the time were really focused on pharmacology and neuroendocrinology rather than a comprehensive neuroscience.  Neuroscience and the old diagnostic technology and clinical methods seem to be the current points of division.

A lot of the criticism is directed at Insel.  I have heard him talk about the initiatives and the rationale sounded clear to me.  I think that rationale is very similar to what I have discussed so far, but for clinical psychiatrists it is also the realization that as long as we live in an approximate world - we will get approximate results.  The inertia to stay in that place is always puzzling to me.

But - it is time to move out of the 1950s.

Clinical psychiatry the way it is currently researched and practiced holds no promise for understanding the most complex known object in the universe.  Neuroscience is one of the big ways out of that predicament.

George Dawson, MD, DFAPA      


1:  Friedman RA. Psychiatry's Identity Crisis. New York Times July 17, 2015. p SR5.

2:  Markowitz JC.  There’s Such a Thing as Too Much Neuroscience.  New York Times October 14, 2016. p A21.

3:  Kandel ER. Psychotherapy and the single synapse. The impact of psychiatric thought on neurobiologic research. N Engl J Med. 1979 Nov 8;301(19):1028-37. PubMed PMID: 40128.



  1. There's a whole lot of neuroscience that points to the importance of the social environment and attachment in shaping both brain structure and function. Most mental health clinicians of all stripes are totally in the dark about this stuff.

    1. Good point.

      Attachment was one of the early paradigms that led to some of the first biological models. At one of the places I trained several faculty were members or the primate lab and did active neuroendocrine research on attachment behaviors. I gave a seminar on psychosocial dwarfism that was observed in children in orphanages in WWII and the possible neuroendocrine basis. Since the neuroscience of attachment has become more refined. Attachment has also been a significant component of psychotherapy and psychoanalysis since Bowlby and Spitz. In medical school we had an Infant Psychiatry group that studied the works of Robert Emde - a pioneer in the field. Another good example of the relationship between the clinical work and neuroscience.

  2. I'm curious what you think about the common factors/context model. That's what I think about when I read that Op-Ed, even though I think you are right that they will just do more studies trying to establish the world dominance of CBT. That aside, I have to believe that studying psychotherapy more effectively would lead to improvements in outcomes. I always think of the example of 'facilitative interpersonal skills' which, if accepted as an important ingredient of psychotherapy, could lead to modifications in approaches to training, or even in selection of candidates. Put too simply, if the therapist/candidate is 'not a very good listener' that probably is related to their effectiveness as a therapist. Seems actionable, at least superficially.

    1. I think that the common factors/contextual model is the basis of a lot of treatment is psychiatry beyond psychotherapy per se. For example, a focus on the relationship and the therapeutic alliance is what I consider to be the basis for treatment and it has to be a reference point for how seeing a psychiatrist differed from seeing a physician who is not a psychiatrist.

      When I was trained, manual driven psychotherapies for research including the authors mentioned in the post but many others were just arriving on the scene. Most psychiatrists were trained in supportive psychotherapy and psychodynamic based psychotherapy. The early supportive psychotherapy contained elements of most subsequent therapies but it was not standardized in a manual.

      I agree that specific elements of therapy should be actionable to a point, but there are some people who don't get it or for other reasons are unable to do psychotherapy. Thankfully they are very rare in the field.

      I don't think that further psychotherapy research will necessarily lead to better outcomes for the reasons I mention in the post. I don't think it is possible to treat a heterogenous population (as in any DSM diagnostic group) with a medication or psychotherapy and get better results. I think that the current psychotherapies need to be applied to groups that are selected with more precision than the DSM.

      My speculation is that will be the true value of neuroscience. A common error is that people often think that neuroscience equates to medications or physical interventions and not psychotherapy.

      For anyone interested in what the author of this post is referring to - a brief review is available at: