The brain of an adult fruit fly has been mapped – human brains could follow suit | Mint

FRUIT FLIES are smart. For starters – the clue is in the name – they can fly. They can also flirt; fight; form complex long-term memories of their environment; and even warn each other of the presence of invisible dangers, such as parasitic wasps.

They do each of these things based on sophisticated processing of sound, smell, touch and vision, organized and controlled by a brain made up of about 140,000 neurons – more than the 300 or so found in a nematode worm, but far fewer than the 86 billion human brains, or even the 70 million in a mouse. This manageable but non-trivial level of complexity has made fruit flies an attractive target for those who want to build a “connectome” of the animal brain – a three-dimensional map of all its neurons and the connections between them. This appeal is further enhanced by the fact that fruit flies are already among the most studied and best understood animals on earth.

For years, it seemed likely that the race to assemble an adult fly connectome would be won by the FlyEM project at the Howard Hughes Medical Institute's Janelia Research Campus in Virginia. In 2020, FlyEM researchers, led by Gerry Rubin, an experienced flight biologist, published a connectome of an adult fruit fly “hemibrain”, an array of 27,000 neurons in the center of the organ. This was followed in 2023 by a connectome of the 3,016 neurons of a first-stage fly larva – the tiny caterpillar that hatches from an egg. But Janelia has been tasked with creating a connectome of an entire brain by a group called FlyWire, based at Princeton University. Ironically, Flywire used data collected by Janelia, but abandoned in 2018 because it was too difficult to analyze with the artificial intelligence (AI) software available at the time.

However, Mala Murthy and Sebastian Seung, the creators of FlyWire, had different AI software. They started the project in 2018 with support from the BRAIN Initiative (a US government effort to do for neuroscience what the Human Genome Project did for genetics) to analyze Janelia's now-abandoned data. The result, published this week in Nature, is a model that paints a detailed picture of a female fly's brain with 139,255 neurons, and locates some 54.5 million synaptic connections between them.

Creating a connectome means taking things apart and putting them back together. The dissection uses an electron microscope to capture the brain as a series of slices. Reassembly uses AI software to track the neurons' multiple projections across segments, recognizing and recording connections.

Janelia's researchers had developed two ways to do these things. The FlyEM team used a beam of gallium atoms to blast away nanometers of tissue from a brain sample, then captured an image of each newly exposed surface with a scanning electron microscope (which fires a beam of electrons at a surface and detects any radiation that is then emitted ). Their own fruit fly connectome, from a male, should be ready within a year.

Janelia's second method involved shaving off layers of a sample with a diamond blade and recording them using a transmission electron microscope (which directs its beam through the target rather than scanning the surface). This is the data that FlyWire uses. Using Janelia's library of 21 million images created in this way, Dr. Murthy and Dr. Seung, ably assisted by 622 researchers from 146 labs around the world (as well as 15 enthusiastic “citizen scientist” video gamers, who helped proofreading and annotating the results), betting their credibility in writing software on the ability to merge the images into a connectome. What they did.

In addition to the number of neurons and synapses in the fly brain, FlyWire researchers also counted the number of types of neurons (8,577) and calculated the combined length (149.2 meters) of the message-carrying axons that connect cells. . importantly, they have allowed the elucidation not only of a neuron's connections to its nearest neighbors, but also of the connections these neurons have to those more distant. Neural circuits can therefore be studied in their entirety. The project's researchers have more than doubled the number of known cell types in the fly's major optic lobes, showing how the new cell types are connected in circuits related to different elements of vision, including movement, objects and colors.

These kinds of things are scientifically interesting. But to justify the dollars spent on them, projects like FlyEM and FlyWire would also have to serve two practical purposes. One of these is to improve the technology of connectome construction so that it can be used on increasingly larger targets – perhaps eventually on the brains of Homo sapiens. The other is to discover the extent to which non-human brains can function as models for human brains (particularly models that can be experimented on in ways that will be approved by ethics committees).

This is where evolutionary biology comes into play. Fruit flies and humans are on opposite sides of a 670-meter-old division that splits bilaterally symmetrical animals into two groups: protostomes and deuterostomes. This separation almost certainly predates the evolution of the brain, meaning that the brains of insects (which are protostomes) and those of vertebrates (deuterostomes) have different origins. Drawing conclusions about one thing from another is therefore a risky undertaking.

This shouldn't matter for long. Several groups are currently working on mouse connectomes, pieces of which have already been put together. Although Janelia has no plans to move in this direction, Dr. Rubin (who, along with several other Janelia researchers, co-authored part of the package of nine Nature articles) thinks that within ten years there will be a complete mouse connectome could be created. if someone were willing to put aside $1 billion to pay for it. Analogous to the Human Genome Project, where the technology got cheaper as things got bigger, this would also reduce costs to a point where smaller connectomes, like those of flies, could be mass-produced.

The deuterostome-protostome division, together with more recent evolutionary shifts, also offers the possibility of a new science of comparative connectomology. In some cases it is already clear that giving multiple bites to the cherry through natural selection has resulted in more than one solution to the same problem. For example, the overarching organization of the neurons in fly brains and vertebrate brains is completely different. In other cases, however, both brains appear to work in the same way, suggesting that this may be the optimal way to do things.

These natural experiments, whose circuit diagrams will make available connectomes, could even help human computer scientists. After all, brains are quite successful information processors, so reproducing them in silicon could be a good idea. Since it is AI models that have made connectomics possible, it would be poetic if connectomics could in turn help develop better AI models.