The Neurocritic: February 2015

What do schizophrenia, bipolar disorder, major depression, addiction, obsessive compulsive disorder, and anxiety have in common? A loss of gray matter in the dorsal anterior cingulate cortex (dACC) and bilateral anterior insula, according to a recent review of the structural neuroimaging literature (Goodkind et al., 2015). These two brain regions are important for executive functions, the top-down cognitive processes that allow us to maintain goals and flexibly alter our behavior in response to changing circumstances. The authors modestly concluded they had identified a “Common Neurobiological Substrate for Mental Illness.”

One problem with this view is that the specific pattern of deficits in executive functions, and their severity, differ across these diverse psychiatric disorders. For instance, students with anxiety perform worse than controls in verbal selection tasks, while those with depression actually perform better (Snyder et al., 2014). Another problem is that gray matter volume in the dorsolateral prefrontal cortex, a key region for working memory (a core impairment in schizophrenia and to a lesser extent, in major depression and non-psychotic bipolar disorder), was oddly unaffected in the meta-analysis.

The NIMH RDoC movement (Research Domain Criteria) aims to explain the biological basis of psychiatric symptoms that cut across traditional DSM diagnostic categories. But I think some of the recent research that uses this framework may carry the approach too far (Goodkind et al., 2015):

Our findings ... provide an organizing model that emphasizes the import of shared endophenotypes across psychopathology, which is not currently an explicit component of psychiatric nosology. This transdiagnostic perspective is consistent...with newer dimensional models such as the NIMH’s RDoC Project.

However, not even the Director of NIMH believes this is true:

"The idea that these disorders share some common brain architecture and that some functions could be abnormal across so many of them is intriguing," said Thomas Insel, MD...

[BUT]

"I wouldn't have expected these results. I've been working under the assumption that we can use neuroimaging to help classify the different forms of mental illness," Insel said. "This makes it harder."

Anterior Cingulate and Anterior Insula and Everyone We Know

The dACC and anterior insula are ubiquitously activated ¹ in human neuroimaging studies (leading Micah Allen to dub it the ‘everything’ network), and comprise either a “salience network” or “task-set network” (or even two separate cingulo-opercular systems) in resting state functional connectivity studies. But the changes reported in the newly published work were structural in nature. They were based on a meta-analysis of 193 voxel-based morphometry (VBM) studies that quantified gray matter volume across the entire brain in psychiatric patient groups, and compared this to controls.

Goodkind et al., (2015) included a handy flow chart for how they selected the papers for their review.

I could be wrong, but it looks like 34 papers were excluded because they found no differences between patients and controls. This would of course bias the results towards greater differences between patients and controls. And we don't know which of the six psychiatric diagnoses were included in the excluded batch. Was there an over-representation of null results in OCD? Anxiety? Depression?

What Does VBM Measure, Anyway?

Typically, VBM measures gray matter volume, which in the cortex is determined by surface area (which can vary due to differences in folding patterns) and by thickness (Kanai & Rees, 2011). These can be differentially related to some ability or characteristic. For example, Song et al. (2015) found that having a larger surface area in early visual cortex (V1 and V2) was correlated with better performance in a perceptual discrimination task, while larger cortical thickness was actually correlated with worse performance. Other investigators warn that volume really isn't the best measure of structural differences between patients and controls, and that cortical thickness is better (Ehrlich et al., 2012):

Cortical thickness is assumed to reflect the arrangement and density of neuronal and glial cells, synaptic spines, as well as passing axons. Postmortem studies in patients with schizophrenia showed reduced neuronal size and a decrease in interneuronal neuropil, dendritic trees, cortical afferents, and synaptic spines, while no reduction in the number of neurons or signs of gliosis could be demonstrated.

This leads us to the huge gap between dysfunction in cortical and subcortical microcircuits and gross changes in gray matter volume.

Psychiatric Disorders Are Circuit Disorders

This motto tells us that mental illnesses are disorder of neural circuits, in line with the funding priorities of NIMH and the BRAIN Initiative. But structural MRI studies tell us nothing about the types of neurons that are affected. Or how their size, shape, and synaptic connections might be altered. Basically, volume loss in dACC and anterior insula could be caused by any number of reasons, and by different mechanisms across the disorders under consideration. Goodkind et al., (2015) state:

Our connection of executive functioning to integrity of a well-established brain network that is perturbed across a broad range of psychiatric diagnoses helps ground a transdiagnostic understanding of mental illness in a context suggestive of common neural mechanisms for disease etiology and/or expression.

But actually, we might find a reduction in the density of von Economo neurons in the dACC of individuals with early-onset schizophrenia (Brüne et al., 2010), but not in persons with other disorders. Or a reduction in the density of GAD₆₇ mRNA-expressing neurons in ACC cortical layer 5 in schizophrenia, but not in bipolar disorder. On the other hand, we could see something like an alteration in the synapses onto parvalbumin inhibitory interneurons (due to stress) that cuts across multiple diagnoses.

And it's not always the case that bigger is better: smaller cortical volumes can also be associated with better performance (Kanai & Rees, 2011).

As Kanai and Rees (2011) noted in their review:

...a direct link between microstructures and macrostructures has not been established in the human brain. A histological study directly compared whether histopathological measurements of resected temporal lobe tissue correlated with grey matter density as used in typical VBM studies. However, none of the histological measures — including neuronal density — showed a clear relationship with the grey matter volume.

So where do we go from here? Bridging the technological gulf between exceptionally invasive methods (like optogenetics and chemogenetics in animals) and non-invasive ones (TMS, MRI in humans) is a minor funding priority of the BRAIN Initiative. Another – more manageable – strategy for the present would be a comprehensive review of imaging, genetic, and post-mortem neuroanatomical studies of brains from people who lived with schizophrenia, bipolar disorder, major depression, addiction, obsessive compulsive disorder, and anxiety. This has been done most extensively (perhaps) for schizophrenia (e.g., Meyer-Lindenberg, 2010; Arnsten, 2011). Certain types of electrophysiological studies in primate prefrontal cortex may provide another bridge, although this has been disputed.

Goodkind and colleagues have indeed uncovered some “biological commonalities that may have been underappreciated in prior work,” but it's also clear there are “some fairly obvious distinctions between schizophrenia and bipolar disorder” at a clinical level (to give one example). In the rush to cut up psychiatric nosology along the RDoC dotted lines, let's not forget the limitations of current methods that are designed to do the carving.

Other comprehensive reviews:

Large-scale brain networks and psychopathology: a unifying triple network model

Does the salience network play a cardinal role in psychosis? An emerging hypothesis of insular dysfunction

Salience processing and insular cortical function and dysfunction

Critiques of phrenology-like VBM studies:

Now Is That Gratitude?

Should Policy Makers and Financial Institutions Have Access to Billions of Brain Scans?

Anthropomorphic Neuroscience Driven by Researchers with Large TPJs

Liberals Are Conflicted and Conservatives Are Afraid

Great discussion of a failure to replicate VBM studies (at Neuroskeptic):

Failed Replications: A Reality Check for Neuroscience?

Footnotes

¹ To quote Russ Poldrack:

In Tal Yarkoni's recent paper in Nature Methods, we found that the anterior insula was one of the most highly activated part of the brain, showing activation in nearly 1/3 of all imaging studies!

² Links to recent J Neurosci articles via @prerana123 and @MyCousinAmygdala.

References

Brüne M, Schöbel A, Karau R, Benali A, Faustmann PM, Juckel G, Petrasch-Parwez E. (2010). Von Economo neuron density in the anterior cingulate cortex is reduced inearly onset schizophrenia. Acta Neuropathol. 119(6):771-8.

Ehrlich S, Brauns S, Yendiki A, Ho BC, Calhoun V, Schulz SC, Gollub RL, Sponheim SR. (2012). Associations of cortical thickness and cognition in patients with schizophrenia and healthy controls. Schizophr Bull. 38(5):1050-62.

Goodkind, M., Eickhoff, S., Oathes, D., Jiang, Y., Chang, A., Jones-Hagata, L., Ortega, B., Zaiko, Y., Roach, E., Korgaonkar, M., Grieve, S., Galatzer-Levy, I., Fox, P., & Etkin, A. (2015). Identification of a Common Neurobiological Substrate for Mental Illness. JAMA Psychiatry DOI: 10.1001/jamapsychiatry.2014.2206

Kanai, R., & Rees, G. (2011). The structural basis of inter-individual differences in human behaviour and cognition. Nature Reviews Neuroscience, 12 (4), 231-242. DOI: 10.1038/nrn3000

Song C, Schwarzkopf DS, Kanai R, Rees G. (2015). Neural population tuning links visualcortical anatomy to human visual perception. Neuron 85(3):641-56.

Snyder HR, Kaiser RH, Whisman MA, Turner AE, Guild RM, Munakata Y. (2014). Opposite effects of anxiety and depressive symptoms on executive function: the case of selecting among competing options. Cogn Emot. 28(5):893-902.