It’s been exciting to see the J cell and the EyeWire community grow so quickly over the past few days. Yesterday we were almost finished with the J cell. Now it’s today. Are we done yet?
Before answering that question, let’s ask a more fundamental one: How do we know when we’re done? Lab member Jinseop Kim made a graph to visualize our progress:
The points on the graph show how the J cell has grown in size over time. To estimate future growth, Jinseop drew a curve through the points. The curve approaches the horizontal dashed line but never actually reaches it. This dashed line is Jinseop’s projection for the total size of the cell.
Now you can see that the answer to the question, “Are we done yet?” depends on what fraction of the cell we need to discover. Are we satisfied with 95%? Or do we insist on 99%? What we already have is probably accurate enough for the purposes of our research. In other words, the answers to our scientific questions (more on those in future blog posts) most probably will not depend on whether we go further. However, we have decided to let the community map the J cell for one more night, to see how many more branches will be discovered. In the image below, captured after lunch today, the newly discovered branches are colored in red. By tomorrow we will see more of these.
Why is the growth of the J cell slowing down? Looking for undiscovered J branches is kind of like prospecting for gold. The total amount of gold recovered from a mine increases according to a graph like the one above. In the beginning, you find gold frequently. Near the end, it becomes almost impossible to find gold. Economists call this the “law of diminishing returns.” At some point, you give up and say you’ve found all the gold, and move on. On the other hand, if you suddenly started employing a metal detector, that might speed up your search again. The EyeWire devs are working on the equivalent of that for neural branches. Stay tuned…