Wow!
The October 3rd, 2024 issue of Nature Magazine published a complete map—a wiring diagram a.k.a. connectome—of the nerve connections in a fruit fly brain. A fly brain? Tinier than the head of a pin; that’s not much, you may think. But I’m here to tell you that the nine research articles appearing in this issue demand admiration; and they demand careful scrutiny.
The fly brain has almost 140,000 nerve cells (neurons) that make 50 million connections (synapses). This complex structure took five years, 50 laboratories and hundreds of people to create and verify its wiring diagram. This is the era of big science. No longer is new science created by a singular creative mind. Many hands, many minds, many countries and many institutions cooperate to make new knowledge. This is a feature, not a bug (oops! Not a fly?) of how science works.
It’s been a while since scientists published the first complete connectome of an animal—a round worm called Caenorhabditis elegans.1 That tiny worm has 302 neurons and 10,000 synapses. It was an accomplishment, to be sure, but the fly brain is orders of magnitude more complex.
To map all its connections researchers had to mark each individual cell and nerve fiber in the intact brain, take apart the brain, and then reassemble it. First they fixed in place all the cells and connections in the intact brain. Then they stained the whole brain to make each cell visible in an electron microscope. Finally, they sliced it like a salami on a deli meat slicer set to ultra-thin and took a photographic image of each slice.
Back to the salami for a second. You know how there are peppercorns that get sliced in two on the slicer? Part goes to one slice, part to the next one, and maybe a small fragment to the third? If you want to reconstruct the whole peppercorn after slicing, and see which meat fragments it touches, you have to line up the salami pieces so that the sliced peppercorn fits back together. This was the problem our fly people faced. They made thousands of ultrathin sections, then used a high-speed, automated electron microscope to produce millions of images, and finally, they had to align the images, (like reconnecting the sliced peppercorns), so that they could trace each cell and its connections. This alignment allowed them to reassemble a 3D image of cells and their connections. The one reproduced above represents only 50 of the 140,000 cells and their connections. In principle the process is simple. Take it apart, fiddle around with it, and put it back together. In practice it is a lot harder.
This is a second feature (not a bug) of big science. No human or group of humans could have done this by hand. They had to automate everything, the slicing, the imaging and the reassembly. They had to use Artificial Intelligence to map the peppercorns. But as every student who has used ChatGTP to write a history assignment knows, computers and AI make mistakes. To ferret out and correct the mapping errors, the main research group recruited (and trained) volunteer proof-readers from fly labs around the world. I find this aspect of the project—the third feature (not a bug) especially moving. For sheer love of the project, excitement with knowing more about how the world works, for devotion to factual accuracy (imagine that!), scientists from around the world gave freely of their time.
The complexity of connections in that pin-head sized brain, with its 54.5 million synapses, is hard enough to visualize. But there’s more. As they mapped, the fly scientists found almost 8500 different types of nerve cells, of which 4500 were previously unknown. It is like having a parts list for a car, but not knowing what over half of the listed items do. Clearly the publication of the fly connectome is only the starting point for decades more research needed to learn the function of all these newy discovered cells.
Scientists can start digging into these newly posed mysteries because—and this is a fourth feature (not a bug) of good scientific practice—the elements of which I will call share, collaborate and teach. The researchers offered their data on a website called https://flywire.ai/ which they describe as “a human-AI collaboration for reconstructing the full brain connectome of Drosophila”. Using this website, neuroscientists can ask questions of interest to their own research projects. Other researchers can share apps to help in the analysis; and Flywire Academy provides high school and college level lessons on the material, as well as training videos for the teachers.
For all the excitement, though, certain questions nag. The fly brain map is derived from a single (female) brain. If I tried to publish a research paper with an N=1 it would never be assigned for review. So is this legit? “Have we” the authors of one of the Nature papers wrote, “collected a snowflake?”, a single brain unlike any other?2 This has long been a concern for connectome studies, a field in which researchers have to sacrifice frequentist statistics on the alter of complex, dense data. The debate centers around the question of how much all that connectivity varies from one fly to the next. If no two flies are alike, then the connectome map can only be used to explain the behavior of the (now deceased) fly whose brain they mapped. But scientists, of course, hope to generalize their findings to other flies.
In this example, our fly scientists checked their results against previous work on four different partial brains (bringing the sample size to 5). The answer to the snowflake problem, they decided, was: it depends. Some circuits are invariant in terms of cell number and connectivity, while others are not. These fly scientists don’t think they have spent the last five years collecting a snowflake; still, there is enough variability at the fine-tuning level that for every new behavior circuit they map, they will need to assess the biological variability, and variation introduced by their cell preparation and maooing methods, before they can generalize their conclusions to all flies.
They closed the snowflake section and the paper with a somewhat jarring final note by introducing gender/sex into the discussion. Some brain scientists love to hunt for sex differences in the brain while others think the search is theoretically wrong-headed (so to speak).3 The fly people worried that looking for differences in brain circuits between males and females might be prone to snowflakism. So they urged others to be circumspect when trying to identify “sexually dimorphic circuits..” .4 I, of course, agree with this caution. I also use it to point out that even when studying fly brains, gender and sex may be lurking beneath the surface. This is not the last you will hear from me on the topic of gender/sex.
White, J. G., Southgate, E., Thomson, J. N. & Brenner, S. The structure of the nervous system of the nematode Caenorhabditis elegans. Phil. Trans. R. Soc. Lond. B 314, 1–340 (1986).
Schlegel, P., Yin, Y., Bates, A.S. et al. Whole-brain annotation and multi-connectome cell typing of Drosophila. Nature 634, 139–152 (2024).
Rippon, Gina. Gender and Our Brains: How New Neuroscience Explodes the Myths of the Male and Female Minds. Pantheon (2019)
Schlegel, P., Yin, Y., Bates, A.S. et al. Whole-brain annotation and multi-connectome cell typing of Drosophila. Nature 634, 139–152 (2024). https://doi.org/10.1038/s41586-024-07686-5 p. 150