Scientists Finished Wiring Up a Fly Brain, and AI Did the Tracing
Researchers at HHMI Janelia, Google Research and the Cambridge Connectomics Group published the complete wiring diagram of a male fruit fly's nervous system: 166,000 neurons and 125 million connections, the largest such map yet.
A decade long project has produced the complete connectome of a male fruit fly, meaning a map of every neuron in its brain and nerve cord and the connections between them. The numbers are more than 166,000 neurons and around 125 million connections, which makes it the largest brain map by neuron count anyone has finished. The work was led by the FlyEM team at the Howard Hughes Medical Institute’s Janelia Research Campus, with the Cambridge Connectomics Group and Google Research, and was published in the journal Cell under the title “Sexual dimorphism in the complete connectome of the Drosophila male central nervous system.”
A connectome is a wiring diagram, and it is built the hard way. You slice a brain into sections thinner than a wavelength of light, photograph them with an electron microscope, and then trace each individual neuron through thousands of images to see where it goes and what it touches. Doing that by hand for 125 million connections is not merely slow, it is impossible, which is where the machine learning comes in: models trained to follow the outline of a single cell across image after image do the first pass, and humans check and correct. That partnership is the whole reason a complete map exists now rather than in fifty years. Anyone can open the result in Neuroglancer, the free tool Google built for exploring enormous image datasets, and follow a neuron around. 404 Media reports that people have already wired the simulated fly circuitry up to novelty projects, including getting it to play Doom, which tells you more about the internet than about neuroscience.
What is behind this
The reason a male map matters when a female fly brain was already mapped is in the paper’s title: sexual dimorphism, the differences between the sexes. Having both lets researchers ask which circuits differ and how those differences produce different behaviour, which is a much sharper question than anything a single map allows. It is also worth being precise about what a connectome is not. A wiring diagram tells you what connects to what, not how strongly each connection fires, what chemicals are involved, or how any of it changes while the animal learns. Knowing the circuit is a starting point for understanding behaviour, not the finish line, and researchers in the field are consistently clearer about that than the headlines tend to be.
What this means for you: For most of us this is not a tool, it is a good example of what AI is actually doing well right now, and it looks nothing like a chatbot. There is no clever reasoning here: the models do one narrow perception task, follow a thin grey line through a stack of photographs, at a scale and consistency no team of people could match. That pattern, a narrow model doing an enormous amount of careful looking while scientists ask the questions, is quietly behind a lot of recent progress in biology and materials. If you are curious, the data is public and free to explore in the browser, which is a rather lovely thing to be able to say about a decade of research.
Sources
Source: https://research.google/blog/a-connectomics-milestone-mapping-the-complete-male-fruit-fly-brain/
Zuckerberg Says Safety Is a Competitive Edge, Not a Reason to Slow Down
Mark Zuckerberg declined to join the call for frontier labs to pace themselves, arguing that any lab ignoring alignment will fall behind on its own. He said Meta delayed Muse for months over safety, and did not ask rivals to do the same.
