The challenge to define cortex (By Ashley Juavinett)

“Our fragmentary and rapidly evolving understanding is reminiscent of the situation faced by cartographers of the earth’s surface many centuries ago, when maps were replete with uncertainties and divergent portrayals of most of the planet’s surface.”

– David Van Essen, 2003

We love talking about cortex. It’s bumpy, it’s got layers, and it’s probably the brain structure that makes us the very verbal, skilled primates that we are.

We also love all of the different areas of cortex—there’s one for face recognition, another for motion detection, and many for decision-making. Often, labs stake claims on their cortical area of interest, diving deep into how that particular patch gets its job done.

But how well can we really divvy up that important sheet of tissue that makes us human? Can we confidently say we’ve left one area, and moved into the next? And how well can we translate these borders to smaller animal models, such as mice?

Tiny brains with big aspirations

Mice are super important to neuroscientists. Sure, they’re quite small and not exactly the most brilliant animals, but we’ve been able to engineer them to mark specific cell types, express glowing proteins, and more. As a result of this powerful murine toolbox, mice have gained a lot of attention from scientists who want to understand circuits and cell types in the brain. In particular, the visual cortex of the mouse has been the site of a lot of discussion, with many researchers hoping that we could use our extensive knowledge about the coarse organization of the primate visual system to ask detailed questions in the mouse brain.

However, if we want to use powerful genetic and recording tools in mice, we first need to understand how their cortex is organized. So, many neuroscientists have been working to combine textbook knowledge about primate brain organization with novel techniques designed for the tiny mouse brain.

After a few landmark papers showing that mice have multiple visual areas like primates, the field immediately grabbed its two-photon microscopes and optical fibers and got to work. But as it turns out, mouse visual cortex isn’t exactly arranged like primate cortex, and many mouse researchers are left wondering if they have indeed isolated a meaningful cortical “area,” or a dubious patch of cortex. This challenge to disentangle the mouse brain has unearthed several fundamental questions about defining cortex.

An age-old problem

Fortunately for the new generation of mouse vision researchers, many before us have given careful thought to how we should consider the cortex. Primate researchers are quite familiar with the problem of defining cortical areas as well as the difficulties that arise when comparing across species. Multiple cortical partitioning schemes exist for macaque visual cortex and mapping these areas onto humans is quite contentious. In primates, there are 10-12 areas with clear organization, but there is anatomical and functional evidence for more than 40 distinct visual areas.

In an elegant book chapter on the topic, cortical map guru David Van Essen lays out four ways to divide primate visual cortex: architecture, connectivity, retinotopy (an ordered map of the visual world), and functional specialization. Still, when we compare maps from these types of data they’re often conflicting, even in the relatively large primate brain.

Usefully, Van Essen also defines three frequent problems encountered when defining a cortical area. First, there may be subtle (or imprecise) boundaries between areas, and therefore ground truth is not easily defined. Second, areas may be internally heterogeneous, either because they have repetitive structures, asymmetric organization, or gradients of function. Last, areas may differ between individual animals. Even in primates, primary visual cortex and the middle temporal area can vary two-fold or more in size between monkeys. And while anatomical boundaries are often used to locate these areas in primates, these landmarks are quite variable.

The current state of cortex

 In the past year or so, multiple studies have been released that are challenging our current conceptions of mouse and human cortex. Research from the Allen Brain Institute suggests that mice have even more visual areas than we previously thought—something on the order of sixteen. As in primates, many of these areas don’t clearly receive input from primary visual cortex, and some of them have really weak and/or partial maps of the visual world.

Meanwhile, the research on human cortex is also getting a bit complicated. Using an enormous dataset from the Human Connectome Project, Van Essen and his team identified 97 new cortical areas. That’s more than double the number of areas previously reported. They were able to do this by using multiple sources of information about organization, such as the thickness of cortex, or the activity in a given area during resting state. Like all good big data papers, they used machine learning to train a classifier to determine area boundaries.

Both of these projects represent a new wave of approaches to integrate many types of data in order to accurately draw cortex area borders. Ultimately—as Van Essen himself once proposed—we may need to settle on probabilistic boundaries: area #1 likely ends here, and area #2 is very likely next door.

So as it turns out, dividing up the cortex isn’t as simple as separating yellow and red M&Ms. However, with multiple types of anatomical and functional information across species, we’re hopefully moving towards a common framework for the widely-adored cortex.

References

Glasser, M.F., Coalson, T.S., Robinson, E.C., Hacker, C.D., Harwell, J., Yacoub, E., Ugurbil, K., Andersson, J., Beckmann, C.F., Jenkinson, M., Smith, S.M., Van Essen, D.C. (2016). A multi-modal parcellation of human cerebral cortex. Nature. 536: 171-178. doi:10.1038/nature18933

Maunsell, J.H., Van Essen, D.C. (1987). Topographic organization of the middle temporal visual area in the macaque monkey: representational biases and the relationship to callosal connections and myeloarchitectonic boundaries. J Comp Neuro. 266(4): 535-55. doi: 10.1002/cne.902660407

Van Essen, D.C., Newsome, M.T, Maunsell, J.H. (1984). The visual field representation in striate cortex of the macaque monkey: asymmetries, anisotropies, and individual variability. Vision Res. 24(5): 429-48.

Van Essen DC: Organization of Visual Areas in Macaque and Human Cerebral Cortex. In Visual Neurosciences (L. Chalupa and J Werner, eds.). 2004.

Zhuang, J., Ng, L., Williams, D., Valley, M., Li, Y., Garrett, M., Waters, J. (2016). An extended map of mouse cortex. eLife. 6. doi: 10.7554/eLife.18372

Image credit https://www.flickr.com/photos/38703275@N06

Any views expressed are those of the author, and do not necessarily reflect those of PLOS.

Ashley Juavinett is a neurosciences postdoctoral fellow at Cold Spring Harbor Laboratory, a lover of cortex, and a writer of both prose and music. Her exploits be followed at @analog_ashley.