Two’s company!

EyeWirers complete J Cell #2, "EyeWire"We’ve mapped our second J cell!

(See this earlier post to find out how we judge completion.) The EyeWire community continues to amaze me. We’ve made another contribution to neuroscience, and that’s exciting.  Now we are ready to map cells that are connected to J cells.  This is even more exciting, since mapping connections is what EyeWire is all about.  We will start with starburst amacrine cell (SAC) #1, which is already available on EyeWire.  The plan is to map its branches, see whether they contact the J cells, and check whether the contact points are synapses.

SAC #1 was already partially mapped by EyeWirers during our pre-launch beta testing period.  Unfortunately, the mapping stalled, because SAC branches are so thin and difficult to trace. Now we are going to try again, but be warned that it will be much more challenging than the J cells.  In the next few days, we will post tutorial videos that should be helpful.  Starting today, everyone will be automatically directed to SAC #1.  If you are ever curious to explore other cells, then click on the blue “change cell” button in the Overview and manually select one.

how to change cells in EyeWire

But before you rush to map more cells, it’s worth pausing to examine the two J cells that we have so far. In science, as in many other endeavors, it’s important—and fun—to observe and discuss.

I like to think of our J’s as two puppy dogs from the same litter. Their likeness is lovable, but it’s also fun to look for differences, like a spot on one puppy’s ear. So switch back and forth between J Cell #1 and #2 in the EyeWire Overview by clicking on the blue “change cell” button.  How similar are our J’s? How are they different? Do you agree with me that they should be considered the same type of cell? (That’s what I told you, but any statement in science can be contested with data and logical arguments.)

J Cell 1 and J Cell 2 "EyeWire"


J Cell 1 and 2 EyeWire Of course, we can’t expect two cells of the same type to look perfectly identical. These are biological objects, not automobiles mass produced in a factory. To get some idea about the natural variations of J cells, it’s helpful to look at the many examples in the image below, which was produced using a light microscope. (EyeWire images come from an electron microscope, which has much higher resolution.)

This is a view of the retina from the top, when it’s laid flat.  Overall, the J dendrites point downward. At the same time, some dendrites veer to the right, and others to the left.

How was this remarkable image produced? A mouse was genetically engineered so that its J cells produce a fluorescent protein.  That’s why the J cells are visible in the image, while other cells remain invisible.  (The branches of our J cells have been colored by EyeWirers, rather than by a fluorescent protein.)  J cells were originally defined by this genetic method, and turned out to have a distinctive shape.

The images in EyeWire come from the retina of a regular mouse, not the special genetically engineered one mentioned above.   I claimed that our neurons are J cells, simply because they resemble the neurons in the image above.

Do you agree with me that our two J cells indeed deserve to be called J cells, i.e., are they the same type as appear in the image above?  And are the differences between our two J cells consistent with the natural variations evident in the image above? If so, then it’s correct to regard our two cells as being the same type. If not, then perhaps they should be classified as different types. What do you think?  Let’s discuss in the forum.  Rachel from Seung Lab created this post to get us started.

In our discussion, it may be helpful to refer to another image from the same paper, which shows two J cells from a different viewpoint.  Here the retina is viewed from the side, so you can see the J dendrites curving in the same direction.   Both images come from the original paper on the J cell, which was published in 2008.


Leave a Reply

Your email address will not be published.