While tracing dendrites on EyeWire I encountered the following situation. They belong to an Orientation Selective Ganglion Cell.

Dendrite hole

It's the first dendrite I've noticed with such a clear 'hole' in it, though which another dendrite passes.

Passing dendrite

Is it possible for one dendrite to pass through another? Perhaps I shouldn't have 'connected' the top part with the bottom part?


1 Answer 1


I can't think of any reason that would explicitly prohibit a dendrite from travelling through a space created by another dendrite's branching pattern. In non-EM images of dendritic branching patterns I don't think I have ever observed dendritic loops, but of course, that has nothing to do with whether or not they occur. Nonetheless, I think there are more likely explanations for what you are seeing in these particular images.

As shown, the upper and lower parts of the "dendrite" in the first image do not look like a continuous structure. I would guess that you are looking at a synapse or some other contact between two different processes in the first image. The 3-4'o'clock connection in the loop suggests this interpretation. Can you see vesicles in the images around those connections?

The other dendrite that shows up in your second image might be a glial cell or some other cell's dendrite that participates in the synapse. Alternatively, the other dendrite may just be a chance passing of two dendrites that don't interact functionally.

Of course, it could be that you and the computer have just completely misunderstood the data. My understanding of the Eyewire interface is that the computer guides a lot of the identification and doesn't easily allow you to see the surrounding areas so you can't track structures all the way to their sources. In such a situation, it seems probable that you would end up with ambiguous instances like this that you would have a hard time resolving without more information/expertise. But maybe Seung or one his colleagues can come weigh in on that?

  • $\begingroup$ "I don't think I have ever observed dendritic loops" ... On the images it did look like a loop, but I've noticed big staining gaps before. "do not look like a continuous structure" I don't know how the 3d model algorithm handles additions, and how much you can rely on it visually, although it does look like the structure is rounded prior to touching, which could indicate it's just a staining gap. $\endgroup$ Apr 27, 2012 at 12:09
  • $\begingroup$ "doesn't easily allow you to see the surrounding areas so you can't track structures all the way to their sources", ... correct. It's mainly just tracing neurons for now, but they are planning to work on other possible tasks as well. I agree that in many cases I am uncertain about what to do, ... hence this question, and your help being greatly appreciated. I'm still very curious to see how the project will evolve. $\endgroup$ Apr 27, 2012 at 12:12

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .