The common fruit fly, Drosophila melanogaster, long a model organism for genetics and behavior, now has a fully proofread wiring diagram covering the male fly’s brain and ventral nerve cord (the structure analogous to a spinal cord). The project, led by the Howard Hughes Medical Institute (HHMI) Janelia Research Campus with collaborators, is published in Cell under the title “Sexual dimorphism in the complete connectome of the Drosophila male central nervous system”.
The male connectome contains more than 166,000 neurons and about 125 million synaptic connections, making it the largest brain map by neuron count published to date. The dataset was proofread by human experts at HHMI Janelia and can be viewed, explored and downloaded via Neuroglancer, an open-source visualization tool developed by the team for very large multidimensional datasets.
Why this matters
This male connectome complements the recently released complete female fruit fly brain and ventral nerve cord maps. Having both sexes’ complete connectomes enables direct comparisons where neurons differ — useful for studying biological mechanisms of courtship, aggression and other sexually dimorphic behaviours — and allows researchers to examine individual variability in regions that are similar between sexes.
How a full connectome is made
Connectomics typically starts by sectioning a brain into millions of thin slices, imaging each slice with electron microscopy, and using computers and AI to stitch the images into 3D reconstructions. The Janelia teams build systems that use AI to turn flat electron microscope images into accurate, cell-scale 3D neural shapes.
Their AI tools include flood-filling networks, convolutional neural networks that begin at a single pixel and identify all other pixels belonging to the same object. In 2019 the Connectomics team released an initial, fully automated reconstruction of a female fruit fly brain. By 2020 they published a human-verified map of half a female fly brain containing about 25,000 neurons and 21 million connections, a record at that time. Work then continued toward the full, verified male brain map now completed.
Methods keep improving: recently the team incorporated synthetic neurons into training data, which enhanced speed and accuracy of their PATHFINDER reconstruction system. They are also developing new techniques to label and annotate specific neuron types. Currently, producing a fruit fly connectome still requires years of human effort for verification and annotation; reducing the need for manual correction will let research groups tackle even larger brain-mapping projects within reasonable budgets and timelines.
Extending connectomics to vertebrates
Connectomics is already moving into vertebrates, organisms with a spinal cord that are anatomically and functionally closer to humans. In a Columbia University-led study published this week in Nature, the team helped map a part of the elephantnose fish’s hindbrain involved in signal processing. That paper, “Connectome analysis of a cerebellum-like circuit for sensory prediction”, demonstrates how a static connectome can be combined with other data to study neural plasticity and learning, producing the most complete mechanistic model of learning in a vertebrate brain so far.
Larval zebrafish are among the few vertebrates whose brains are small enough to be mapped end-to-end with current techniques and are transparent at that stage, which allows measurement of neural activity during experiments. The ZAPBench dataset captures such combined activity and structural data. An upcoming paper with Harvard, “A connectomic resource for neural cataloguing and circuit dissection of the larval zebrafish brain”, is presented as the first whole-brain vertebrate dataset that includes both neural structure and molecular cell-type information across an entire brain. The team also released a preliminary dataset combining neural activity and structure in the same larval zebrafish brain as an open resource for researchers.
Conclusion
The male Drosophila connectome is a foundational resource that will support new experimental neuroscience. Three companion papers released alongside the connectome show its immediate use in studies of visual systems, taste and social behaviour. The methods developed will also aid projects such as the forthcoming fully proofread zebrafish map and efforts to map portions of the mouse brain. Although a full map of the human brain’s 86 billion neurons remains out of reach, these advances move the field closer to understanding how brains work and to uncovering mechanisms underlying disorders such as Alzheimer’s disease, depression and schizophrenia. The long-term hope is that this work will contribute to new treatments, improved brain health and approaches to neural repair.
Acknowledgements
The team thanks academic collaborators at HHMI Janelia and elsewhere and acknowledges core contributions from the Connectomics Team at Google. Special thanks are given to Hannah Hickey and Elise Kleeman for assistance, and to Lizzie Dorfman, Michael Brenner, John Platt and Yossi Matias for support, coordination and leadership.
