QnAs with David Van Essen

David Van Essen has long been interested in understanding the structure, function, development, and evolution of the cerebral cortex. From the early days of his career, when he manually generated cortical flat-maps of the macaque brain, to his current role as a principal investigator for the Human Connectome Project, Van Essen’s work has helped tease apart the complexity of cortical organization and connectivity. A neuroscientist at Washington University in St. Louis, Van Essen was elected to the National Academy of Sciences in 2017. In his Inaugural Article, Van Essen refines his 1997 model of how the cerebral cortex develops and folds (1).

David Van Essen. Image credit: Washington University School of Medicine.

PNAS: How did you become interested in cortical folding?

Van Essen: I have been obsessed with the issue of cortical folding since the mid-70s, when I started “undoing” the cortical folds by making cortical flat-maps and using them to study cortical function and organization of the visual system in macaque monkeys. In 1996, while studying primary and secondary visual areas, V1 and V2, in postmortem macaque brains, we found that the connections between V1 and V2 start to form around the same time that folding commences and a gyrus emerges along the V1–V2 border. That led to my favorite light bulb moment of my entire scientific career: hypothesizing that the timing between the formation of connections between two strongly connected areas and the fold that emerges between them may not be coincidental. Maybe it’s causal because the axons generate enough mechanical tension to bring strongly connected regions close together. That was intriguing, and I realized that it could be generalized to other folds in the cerebral cortex.

PNAS: How did this finding lead to your model for cortical morphogenesis?

Van Essen: That led to what I initially called the tension-based folding hypothesis, which, in a nutshell, said that strongly connected regions have a gyrus—an outward sticking fold—that forms between them. To keep things balanced, there also have to be inward folds, [or] sulci, [and] the deep grooves formed because the connections between the two sides of a sulcus weren't strong enough to prevent inward folding. In 1997 I published a more general tension-based morphogenesis hypothesis, which proposes that mechanical tension along axons and dendrites and also the elongated processes that we call radial glial cells could work in concert to try to keep connections short overall (2).

For the cerebral cortex, I realized that if the radially oriented dendrites and radial glial cells are also generating tension, just as I had originally invoked for axons, that would tend to pull the upper and lower layers close together and keep the cortex a thin sheet. Cortical folding occurs secondarily when the surface area of the thin sheet becomes so large that it needs to fold to keep in touch with the blob-like nuclei that it enshrouds. In contrast, I hypothesized that the dendrites of blob-like structures like the subcortical nuclei are also gently tugging, trying to pull things close together, but because these dendrites are randomly oriented, the nuclei end up as compact blobs rather than sheets.

PNAS: How have you revised the model since 1997?

Van Essen: In 2010, a colleague of mine, Phil Bayly, published a paper, which suggested that tension is indeed present in the brain but not in the right location to explain my tension-based morphogenesis hypothesis (3). So, for my Inaugural Article, I decided to revisit this hypothesis carefully and critically (1). It’s now crystal clear that mechanical tension and pressure do play a crucial role in many aspects of cellular morphogenesis. For the formation of sheet-like structures, the only plausible mechanism that I’m aware of is this idea of radial tension along dendrites, axons, and glial cells. With regard to how the sheet-like structures become folded, I realized that my original idea is unlikely to be the full story. There may be, in addition, a kind of buckling process that others have hypothesized, but I think a major part of that buckling has to do with a sandwich concept that gives rise to the hypothesis’s new moniker, the Differential Expansion Sandwich (DES).

PNAS: How does this new model differ from the original one?

Van Essen: The idea is that in addition to what’s going on in white matter and at the interface between the cortical gray matter and white matter, there’s a major role played in the folding process by a thin layer at the outer margins that has relatively few neuronal cell bodies and is dominated by dendrites and axons. So that’s the sandwich I invoke in this revised model: a thick layer underneath, a thin layer above, and in between a relatively thin sheet of cortical gray matter. But it’s still dominated by tension along elongated processes as the driving force for how the cortex gets its shape. I added a “+” to make it the DES+ model because there are other things going on that likely impact shape, such as the emergence of deep bulging structures like the subcortical nuclei inside. The DES+ model aims to bring the major players into a common conceptual framework and give them important but complementary roles.

PNAS: Where do you see this research going in the future?

Van Essen: My article proposes a variety of ways to examine the role of tension in morphogenesis using cutting-edge technology, such as photoablation of individual axons and dendrites. I also think the field would benefit tremendously from greater appreciation of the synergy between genetic and molecularly oriented approaches and the biomechanical realities of how these molecules generate and mediate forces to affect the shape of cells or tissues. There [are] now opportunities for rapid progress in figuring out in detail how these forces are generated, how they interact in healthy brains and nervous systems, and what goes wrong in the myriad ways that brains get miswired.