Sunday, November 29, 2015

Carving Up Brain Disorders

Neurology and Psychiatry are two distinct specialties within medicine, both of which treat disorders of the brain. It's completely uncontroversial to say that neurologists treat patients with brain disorders like Alzheimer's disease and Parkinson's disease. These two diseases produce distinct patterns of neurodegeneration that are visible on brain scans. For example, Parkinson's disease (PD) is a movement disorder caused by the loss of dopamine neurons in the midbrain.

Fig. 3 (modified from Goldstein et al., 2007). Brain PET scans superimposed on MRI scans. Note decreased dopamine signal in the putamen and substantia nigra (S.N.) bilaterally in the patient.

It's also uncontroversial to say that drugs like L-DOPA and invasive neurosurgical interventions like deep brain stimulation (DBS) are used to treat PD.

On the other hand, some people will balk when you say that psychiatric illnesses like bipolar disorder and depression are brain disorders, and that drugs and DBS (in severe intractable cases) may be used to treat them. You can't always point to clear cut differences in the MRI or PET scans of psychiatric patients, as you can with PD (which is a particularly obvious example).

The diagnostic methods used in neurology and psychiatry are quite different as well. The standard neurological exam assesses sensory and motor responses (e.g., reflexes) and basic mental status. PD has sharply defined motor symptoms including tremor, rigidity, impaired balance, and slowness of movement. There are definitely cases where the symptoms of PD should be attributed to another disease (most notably Lewy body dementia)1, and other examples where neurological diagnosis is not immediately possible. But by and large, no one questions the existence of a brain disorder.

Things are different in psychiatry. Diagnosis is not based on a physical exam. Psychiatrists and psychologists give clinical interviews based on the Diagnostic and Statistical Manual (DSM-5), a handbook of mental disorders defined by a panel of experts with opinions that are not universally accepted. The update from DSM-IV to DSM-5 was highly controversial (and widely discussed).

The causes of mental disorders are not only biological, but often include important social and interpersonal factors. And their manifestations can vary across cultures.

Shortly before the release of DSM-5, the former director of NIMH (Dr. Tom Insel) famously dissed the new manual:
The strength of each of the editions of DSM has been “reliability” – each edition has ensured that clinicians use the same terms in the same ways. The weakness is its lack of validity. Unlike our definitions of ischemic heart disease, lymphoma, or AIDS, the DSM diagnoses are based on a consensus about clusters of clinical symptoms, not any objective laboratory measure.

In other words, where are the clinical tests for psychiatric disorders?

For years, NIMH has been working on an alternate classification scheme, the Research Domain Criteria (RDoC) project, which treats mental illnesses as brain disorders that should be studied according to domains of functioning (e.g., negative valence). Dimensional constructs such as acute threat (“fear”) are key, rather than categorical DSM diagnosis. RDoC has been widely discussed on this blog and elsewhere it's the best thing since sliced bread, it's necessary but very oversold, or it's ill-advised.

What does this have to do with neurology, you might ask? In 2007, Insel called for the merger of neurology and psychiatry:
Just as research during the Decade of the Brain (1990-2000) forged the bridge between the mind and the brain, research in the current decade is helping us to understand mental illnesses as brain disorders. As a result, the distinction between disorders of neurology (e.g., Parkinson's and Alzheimer's diseases) and disorders of psychiatry (e.g., schizophrenia and depression) may turn out to be increasingly subtle. That is, the former may result from focal lesions in the brain, whereas the latter arise from abnormal activity in specific brain circuits in the absence of a detectable lesion. As we become more adept at detecting lesions that lead to abnormal function, it is even possible that the distinction between neurological and psychiatric disorders will vanish, leading to a combined discipline of clinical neuroscience.

Actually, Insel's view dates back to 2005 (Insel & Quirion, 2005)....2
Future training might begin with two post-graduate years of clinical neuroscience shared by the disciplines we now call neurology and psychiatry, followed by two or three years of specialty training in one of several sub-disciplines (ranging from peripheral neuropathies to public sector and transcultural psychiatry). This model recognizes that the clinical neurosciences have matured sufficiently to resemble internal medicine, with core training required prior to specializing.

...and was expressed earlier by Dr. Joseph P. Martin, Dean of Harvard Medical School (Martin, 2002):
Neurology and psychiatry have, for much of the past century, been separated by an artificial wall created by the divergence of their philosophical approaches and research and treatment methods. Scientific advances in recent decades have made it clear that this separation is arbitrary and counterproductive. .... Further progress in understanding brain diseases and behavior demands fuller collaboration and integration of these fields. Leaders in academic medicine and science must work to break down the barriers between disciplines.

Contemporary leaders and observers of academic medicine are not all equally ecstatic about this prospect, however. Taylor et al. (2015) are enthusiastic advocates of a move beyond “Neural Cubism”, to increased integration of neurology and psychiatry. Dr. Sheldon Benjamin agrees that greater cross-discipline training is needed, but wants the two fields to remain separate. But Dr. Jose de Leon thinks the psychiatry/neurology integration is a big mistake that revives early 20th century debates (see table below, in the footnotes).3

I think a distinction can (and should) be made between the research agenda of neuroscience and the current practice of psychiatry. Neuroscientists who work on such questions assume that mental illnesses are brain disorders and act accordingly, by studying the brain. They study animal models and brain slices and genes and humans with implanted or attached electrodes and humans in scanners. And they study the holy grail of neural circuits using DREDDs and optogenetics. This doesn't invalidate the existence of social, cultural, and interpersonal factors that affect the development and manifestation of mental illnesses. As an non-clinician, I have less to say about medical practice. I'm not grandiose enough to claim that neuroscience research (or RDoC, for that matter) will transform the practice of psychiatry (or neurology) in the near future. [Though you might think differently if you read Public Health Relevance Statements or articles in high profile journals.]

Basic researchers may not even think about the distinction between neurology and psychiatry. Is the abnormal deposition of amyloid-β peptide in Alzheimer's disease (AD) an appropriate target for treatment? Are metabotropic glutamate receptors an appropriate target in schizophrenia? These are similar questions, despite the fact that one disease is neurological and the other psychiatric. There are defined behavioral endpoints that mark treatment-related improvements in either case. It's very useful to measure a change in amyloid burden4 using florbetapir PET imaging in AD [there's nothing similar in schizophrenia], but the most important measure is cognitive improvement (or a flattening of cognitive decline).

Does Location Matter?

In response to the pro-merger cavalcade, a recent meta-analysis asked whether the entire category of neurological disorders affects different brain regions than the entire category of psychiatric disorders (Crossley et al., 2015). The answer was why yes, the two categories affect different brain areas, and for this reason neurology and psychiatry should remain separate.

I thought this was an odd question to begin with, and an even odder conclusion. It's not surprising that disorders of movement, for example, involve different brain regions than disorders of mood or disorders of thought. From my perspective, it's more interesting to look at where the two categories overlap, with an eye to specific comparisons (not global lumping). For instance, are compulsive and repetitive behaviors in OCD associated with alterations in some of the subcortical circuits implicated in movement disorders? Why yes.

But let's take a closer look at the technical details of the study.

Crossley et al. (2015) searched for structural MRI articles that observed decreases in gray matter in patients compared to controls. The papers used voxel-based morphometry (VBM) to quantify regional gray matter volumes across the entire brain. For inclusion, disorders needed to have at least seven published studies to be entered into the analysis. A weighted method was used to control for number of published studies (e.g., AD and schizophrenia were way over-represented in their respective categories), and 7 papers were chosen at random for each disorder. The papers were either in the VBM database or found via electronic searches. The x y z peak coordinates were extracted from each paper and entered into the GingerALE program, which performed a meta-analysis via the activation likelihood estimation (ALE) method (see these references: [pdf], [pdf], [pdf] ).

They found that the basal ganglia, insula, lateral and medial temporal cortex, and sensorimotor areas were affected to a greater extent in neurological disorders. Meanwhile, anterior and posterior cingulate, medial frontal cortex, superior frontal gyrus, and occipital cortex were more affected in psychiatric disorders.

- click on image for a larger view -

The authors also looked at network differences, with networks based on previous resting state fMRI studies. Some of these results were uninformative. For example, psychiatric disorders affect visual networks more than neurological disorders do. That was because neurological disorders affect visual regions much less than expected (based on the total number of affected voxels).

Another finding was that abnormalities in the cerebellum occurred less often than expected in neurological disorders. But this is obviously not the case in cerebellar ataxia, which affects (you guessed it) THE CEREBELLUM. So I'm not sure how useful it is to make global statements about cerebellar involvement in neurological disorders.

ALE map (FDR pN < 0.05) from 16 VBM studies of ataxia.

ALE map above was based on 16 papers in the BrainMap database (from a search including 'Ataxia', 'Friedreich ataxia', or 'Spinocerebellar Ataxia'). Gray matter decreases are seen in the cerebellum.

It was sort of interesting to see all the neurological disorders lumped together and compared to all the psychiatric disorders (the coarsest carving imaginable), but I guess I'm more of a splitter. But an integrative one who also looks for commonalities and overlap. The intersection of neurology and psychiatry is a fascinating topic that could fill many future blog posts.


1 Comedian Robin Williams, who died by suicide, was initially thought to have depression and/or PD. However, an autopsy ultimately diagnosed Lewy body dementia (‘diffuse Lewy body disease’). PD isn't purely a motor disorder, either. Symptoms can include cognitive changes, depression, and dementia.

2 That's a fascinating history that may be covered at another time. For now, here's the table from de Leon (2015).

3 It's interesting to see the prediction for 2015: we should be in the age of diagnostic biomarkers by now...

4 That article came to a surprising conclusion:
If these data support a regional association between amyloid plaque burden and metabolism, it is for the somewhat heretical inversion of the amyloid hypothesis. That is, regional amyloid plaque deposition is protective, possibly by pulling the more toxic amyloid oligomers out of circulation and binding them up in inert plaques, or via other mechanisms...


Benjamin S. (2015). Neuropsychiatry and neural cubism. Acad Med. 90(5):556-8.

Crossley, N., Scott, J., Ellison-Wright, I., & Mechelli, A. (2015). Neuroimaging distinction between neurological and psychiatric disorders.. The British Journal of Psychiatry, 207 (5), 429-434 DOI: 10.1192/bjp.bp.114.154393

David, A., & Nicholson, T. (2015). Are neurological and psychiatric disorders different? The British Journal of Psychiatry, 207 (5), 373-374. DOI: 10.1192/bjp.bp.114.158550

de Leon J. (2015) Is psychiatry only neurology? Or only abnormal psychology? Déjà vu after 100 years. Acta Neuropsychiatr. 27(2):69-81.

Insel TR, & Quirion R (2005). Psychiatry as a clinical neuroscience discipline. JAMA, 294 (17), 2221-4 PMID: 16264165

Martin JB. (2002). The integration of neurology, psychiatry, and neuroscience in the21st century. Am J Psychiatry 159(5):695-704.

Taylor JJ, Williams NR, George MS. (2015). Beyond neural cubism: promoting a multidimensional view of brain disorders by enhancing the integration of neurology and psychiatry in education. Acad Med. 90(5):581-6.

Subscribe to Post Comments [Atom]


At December 01, 2015 7:23 AM, Anonymous herb wiggins, MD said...

It's very simple, actually, to predict what's going to happen to psychiatry versus clinical neurology/neuroscience. The former will be absorbed and explicated by the latter. This is true simply because of the structure/function relationship methods we use to understand how brain structure creates brain outputs, or functions. The so called clinico-pathological localization, now widely more achieved with MRI and fMRI of brain. This is the structure/ function relationship. Where function is damaged, or a deficit in neurological terms, there is structure damaged, as well. Where there is structural damage, there is functional loss.

Thus, the "Astonishing Hypothesis" of Sir Francis Crick. That brain functions arise from brain structures, solely, and of course their interaction via the senses with the surrounding events in existence providing the input.

I have developed an effective, fruitful model based upon this approach which directly, and clearly creates a model of mind/brain with mind's higher functions/abstractions being generated from the 100K's of cortical cell columns. It's quite simple, elegant, least energy, and explains much with little.

These two articles give the basics of it, with the rest of the articles written and still coming detailing how the brain/mind works.

Herb Wiggins, M.D.; Diplomat Am. Bd. of Psychiatry and Neuro

At December 07, 2015 8:52 AM, Anonymous Anonymous said...

As a intern in clinical psychology in 1967 at the University of Minesota Medical Center, I was in a grand rounds run by Starke Hathaway, the author of the MMPI, and he asked us how many of us believed that biology would eventually be the source of information regarding the etiology of mental illness. I was the only one who raised my hand. I wonder if the number has really changed.

At December 08, 2015 5:37 PM, Anonymous Anonymous said...

They should have included a group diagnosed with conversion disorder. That's where you might find a bridge between neurology and psychiatry/psychology.

At December 08, 2015 6:37 PM, Blogger The Neurocritic said...

That's a very good point. I was thinking the same thing about Tourette's. At the time the authors ran their analysis, however [probably in 2013, judging by the initial submission date], they didn't find 7 studies of conversion disorder.

Coincidentally, a Trends in Cognitive Sciences article published several days after this post touched on some similar topics (Downar et al., 2015). In that article, they searched 4 top neurology journals and top 4 psychiatry journals for a list of 32 disorders, then determined the percentage for each disorder that were published in neurology journals.

Mental retardation 78.8% Neuro
Developmental delay 77.0%
Tourette 56.3%
Conversion disorder 54.5%
Delirium 34.7%
Autism 19.9% Neuro (so mostly Psych)

Tourette's and conversion disorder were the ones most in the middle, with about 55% neuro and 45% psychiatry.

At December 08, 2015 6:40 PM, Blogger The Neurocritic said...

Anonymous of December 07, 2015 8:52 AM -- I imagine that would vary by program / department / institution?

At December 20, 2015 3:51 AM, Blogger Mark p.s.2 said...

Regarding your question "Where are the clinical tests for psychiatric disorders?"
For testing , you omit the distinction between the voluntary patient seeking medical help from the involuntary "patient" who doesn't want to take any test.


Post a Comment

<< Home

eXTReMe Tracker