Read on Twitter
A report of my recent dig in the tree shrew visual system, https://biorxiv.org/cgi/content/short/2020.12.08.416651v1, where we found a new type of retinotopic map that expands our understanding of topographic representations in the cortex. Thread:
1st, tree shrew? Tree shrews are fast-moving diurnal creatures that look like a squirrel but visual system (more) like a primate. They have orientation columns in V1, a single area V2, and importantly (for imaging) a smooth brain without area-hiding folds. 1/n
We checked that retinotopic maps in tree shrew V1 were conventional: the two axes of the visual field were mapped on the two, albeit squeezed and squished, perpendicular axes of the cortical area. This is what most reported retinotopic maps look like: nice and simple. 2/n
The map in V2? not simple. Here I’ve separated the up/down axis of the visual field (elevation) and the left/right axis (azimuth). Instead of the simple maps we expected, we found that the mapping of one axis of the visual field was striped, while the other axis was smooth. 3/n
What sort of transform of the visual field could produce such maps? E.g. how would you create the striped map in V2 from the simple map of visual field? A sinusoidal transform would do it. Here's a video since it’s a non-trivial mental exercise: https://bit.ly/3na8TXm 4/n
Now because V2 gets its retinal information mostly through V1, this implies an intricate pattern of connectivity between V1 and V2. We did tracer injections and yep, the V1 axon projections in V2 were consistent with sinusoidal hypothesis. See Fig 4! 5/n
Why does V2 exhibit this transform? Maybe because it is quite elongated. Inspired by prior work, we used self-organizing models used to simulate formation of topographic maps to show that elongation of a simulated area can produce the sinusoidal transform. 6/n
More in paper, including striped functional feature maps. In sum: simple topographic maps are just one instantiation of possible sensory surface transforms in the cortex. More imaging in the extrastriate will likely reveal more transforms. 7/n
This was fun work with David Fitzpatrick, co-authors Kuo-Sheng Lee, Rachel Satterfield, and @NicoleShultz16. Thanks to them and others at MPFI. 8/8.
