On 2 October 2024, a consortium led by Princeton’s Sebastian Seung and Mala Murthy published the first complete connectome of an adult fruit fly brain in a special collection of nine papers in Nature. The map traced all 139,255 neurons of Drosophila melanogaster and the roughly 54.5 million synaptic connections between them, packed into a brain about the size of a poppy seed. The lead paper, by Sven Dorkenwald and colleagues, made it the largest complete brain wiring diagram of its time.
The scale jump from the 302-neuron worm connectome of 1986 was made possible by combining machine learning with human effort. Automated AI segmentation traced neurons through a vast stack of electron-microscope images, then a community of researchers and citizen scientists in the FlyWire project proofread and annotated the result. The organizers estimated this collaboration represented about 33 person-years of work; doing it by hand without the AI would have taken on the order of 50,000 person-years.
Beyond the wiring, the team annotated cell types, identifying 8,453 distinct types, of which 4,581 were newly described. The full dataset was released openly so that any lab could simultaneously explore, analyze, and build on it, a deliberate break from the tradition of guarding data until publication.
The fly connectome sits at the meeting point of neuroscience and AI in two directions. AI made the map possible, and the map in turn gives AI researchers a fully specified biological neural network - a real brain’s complete graph - to study, model, and compare against artificial networks.