Researchers mapped 166,000 neurons in a fly brain, the biggest brain map yet made

Scientists have finished mapping every neuron and synaptic connection in the brain and central nervous system of an adult male fruit fly, producing the largest brain map ever assembled by neuron count. The project, a decade-long collaboration between HHMI Janelia Research Campus and Google Research, reconstructed more than 166,000 neurons and roughly 125 million synaptic connections in three dimensions. It gives neuroscience a complete wiring diagram of an entire animal nervous system for the first time, not just an isolated brain region.

What the male fruit fly connectome actually contains

The finished map, formally described in the journal Cell as “Sexual dimorphism in the complete connectome of the Drosophila male central nervous system,” covers every neuron in the brain, both optic lobes and the ventral nerve cord that runs down the length of the fly’s body. That combination means the map does not stop at the brain the way earlier partial reconstructions did; it traces circuits from sensory input all the way through to the nerve pathways that drive movement and behavior.

Researchers describe the achievement as a milestone for the field known as connectomics, the effort to chart every neuron and connection in a nervous system with enough precision to model how signals actually travel through it, according to Google Research’s account of the project. A finished connectome functions less like a static picture and more like a circuit diagram, letting scientists trace a specific sensory input, such as a puff of odor detected by an antenna, through a chain of neurons all the way to the muscle movements it eventually triggers.

How AI stitched together millions of microscope images

Building the map required combining millions of two-dimensional electron microscope images of thin brain slices into a single, continuous three-dimensional reconstruction, then tracing each neuron’s shape and following its synaptic connections one by one. Machine learning models trained specifically for this kind of image segmentation did much of that reconstruction work, a task that would take human anatomists working alone many years longer to complete by hand. The project also drew on collaborators at the MRC Laboratory of Molecular Biology and the University of Cambridge, extending a partnership that has run for roughly ten years.

Even with automation handling the bulk of the reconstruction, human proofreaders still had to check and correct the computer-generated tracings, since a single misidentified connection can distort how a downstream circuit is interpreted. That mix of large-scale machine labor and careful manual verification is part of why a project of this size took roughly a decade to finish, even with two well-resourced research organizations working in parallel.

Building on a smaller 2020 female fly map

The same broader research effort previously released a human-verified map of about half of a female fruit fly’s brain in 2020, covering roughly 25,000 neurons and 21 million connections. The new male connectome is both larger in scope, since it includes the full nervous system rather than half a brain, and more than six times larger by neuron count, reflecting several additional years of imaging and computational refinement. Having that earlier female reference map already in hand made the new male map immediately useful for direct comparison rather than a standalone dataset.

The gap between the two projects also illustrates how quickly the underlying technology improved. Reconstructing half of one female brain took years of dedicated effort in the late 2010s, while the team was able to complete an entire male nervous system, several times larger, within a comparable stretch of time once better microscopy and more capable image-segmentation software became available.

What comparing male and female wiring can reveal

With both a male and a female nervous-system map now available, researchers can compare the two circuit by circuit to study the biological mechanisms behind behaviors that differ sharply between the sexes, including courtship rituals and aggression. Fruit flies have long served as a model for these behaviors because their courtship and fighting patterns are stereotyped and well studied, which makes it easier to trace a specific behavioral difference back to a specific difference in neural wiring rather than guessing at the connection. The HHMI Janelia announcement frames that comparison as one of the most immediate scientific uses of the new dataset.

Some of the clearest differences researchers expect to examine involve circuits tied to pheromone detection and courtship song production, both of which behave differently in male and female flies despite the two sexes sharing most of the same basic neural architecture. Mapping exactly where those circuits diverge gives researchers a concrete anatomical basis for behavior that previously could only be described in general terms.

Why a fly, and how far this is from a human brain

Fruit flies remain a preferred organism for this kind of exhaustive mapping because their nervous systems are small enough to reconstruct completely with current imaging and computing power, while still sharing many of the same basic neural building blocks and signaling chemistry found in more complex animals. A human brain, by contrast, contains roughly 86 billion neurons, a scale that remains far beyond what any current connectomics project can fully map. Researchers involved in the fly work, along with the National Institutes of Health’s own summary of the project, have pointed to the techniques developed for this map as a foundation that could eventually help interpret smaller circuits within damaged or diseased human brain tissue, even though a complete human connectome remains a distant goal.

In the near term, the more realistic payoff is expected to come from other small model organisms and from isolated circuits within larger brains, such as mouse visual-processing regions, where the imaging and computational methods refined on the fly connectome can be applied at a more manageable scale. Each successful project of this kind also improves the underlying software, gradually closing the gap toward nervous systems far larger than a fruit fly’s.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview