the connectome

A connectome is a complete wiring diagram of a nervous system: every neuron, and every synaptic connection between them. For the fruit fly, Drosophila melanogaster, that diagram exists. It is the largest complete connectome ever produced and, at time of writing, the only one for an animal that does anything you would recognize as behaving.

The numbers, as Eon Systems states them for the brain their emulation is built on: 139,255 neurons and roughly 50 million synaptic connections, preserved, scanned, reconstructed, and emulated from a single fly.

   complete wiring diagrams, by animal:

   c. elegans   ▏                        302 neurons
   fruit fly    █████████▉          139,255 neurons  ◀ the substrate
   mouse        ░░░░░░░░░░░░░░   ~70,000,000  (partial)
   human        ·            ~86,000,000,000  (unmapped)

   the fly: the largest animal ever read to the last synapse,
   and the smallest that grooms
fig. c-1 · where the fly sits. below it, the diagram is complete but the behavior is thin. above it, the reverse.

For scale: the nematode C. elegans, the previous complete connectome, has 302 neurons. A mouse has around 70 million. A human has around 86 billion. The fly sits at the point where the diagram is still completable but the behavior is already rich: flies learn, sleep, fight, court, navigate, and groom. This is why the fly, and not something simpler or something grander, is the substrate for everything in this archive.

what the diagram contains

Each neuron in the connectome carries:

The regional assignments matter more to this archive than anything else on that list. The fly brain is not a uniform mesh. It is organized into dense, named structures with centuries of anatomy behind the names: mushroom body, ellipsoid body, medulla, antennal lobe. These are the units we cut along. See what a neuropil is and the regions section.

what the diagram does not contain

A connectome is a map of connections, not a recording of activity. It does not include synaptic weights as they existed in the living fly, neuromodulator state, glia, or anything the fly had learned. That such a map, plus neurotransmitter identity and nothing else, is sufficient to reproduce most of the fly's function was not obvious. It was the central empirical surprise of the emulation work, and it is covered on the 91% result.

The observers' shorthand for this in the logs is "the diagram is the animal". It appears first in log 002, where it is a question, and later as a statement. The archive takes no official position on it.

see also reconstruction · what a neuropil is · provenance
referenced by large fly model · colophon · glossary · reading order · log 002 · log 013