Connectomics: Mapping the Brain's Wiring

In 1986, scientists published the complete wiring diagram of a brain.

It was the brain of a roundworm called Caenorhabditis elegans—302 neurons, about 7,000 synaptic connections. It took over a decade of painstaking work, slicing the worm into thousands of thin sections, imaging each one with electron microscopy, and tracing every connection by hand.

The result was the first connectome: a complete map of every neuron and every connection in a nervous system.

Thirty years later, we've done it for a fruit fly. About 140,000 neurons. Over 50 million connections. The map took years to produce, involved thousands of researcher-hours, and required machine learning just to trace the wires.

The human brain has 86 billion neurons and roughly 100 trillion connections.

We're going to map that too.

Connectomics is the quest to create complete wiring diagrams of brains—and the bet that knowing the wiring will help us understand how minds work.

Why Wiring Matters

Imagine trying to understand how a computer works by measuring the electricity flowing through it. You'd see patterns. You'd notice that certain power signatures correlate with certain programs running. But you'd never really understand what's happening—not until you had the circuit diagram.

That's roughly where neuroscience has been. We can measure brain activity with fMRI and EEG. We can see which regions are active during which tasks. But we don't have the schematic. We don't know exactly how the 86 billion neurons are connected, which pathways information flows through, which connections are strong and which are weak.

The connectome is that schematic.

Sebastian Seung, the Princeton neuroscientist who popularized the term, puts the bet plainly: "I am my connectome." The pattern of connections in your brain—shaped by genetics, sculpted by experience—is what makes you you. Your memories, your skills, your personality: all encoded in how your neurons wire together.

If that's true, then mapping the connectome isn't just an engineering challenge. It's a path to understanding what a mind actually is.

The C. elegans Proof of Concept

Let's start with the worm.

C. elegans is a one-millimeter-long nematode that lives in soil. It's transparent, reproduces fast, and every individual has exactly 302 neurons wired in exactly the same pattern. It's the fruit fly of neuroscience—a model organism simple enough to study completely.

Sydney Brenner, John White, and their colleagues at the MRC Laboratory of Molecular Biology in Cambridge spent over a decade mapping the worm's nervous system. They sliced thousands of worms into ultrathin sections (about 50 nanometers thick), photographed each section with electron microscopy, and manually traced every neuron through every slice.

In 1986, they published the result: the complete connectome of C. elegans. Every neuron. Every synapse. The full wiring diagram.

This was a monumental achievement. And it raised an uncomfortable question: even with the complete wiring diagram, we still couldn't predict the worm's behavior from first principles.

We knew exactly how the neurons were connected. But neurons aren't simple on-off switches—they have complex biophysics, neuromodulation, plasticity. The wiring diagram tells you the structure, not the dynamics. It's like having the schematic of a radio but not knowing what stations it's tuned to.

The C. elegans connectome proved that connectomics was possible. It also proved that the connectome alone isn't sufficient to understand a brain. You need the wiring, but you also need to know what's flowing through it.

The Fruit Fly: Scaling Up

Fast forward three decades.

In 2023, researchers published the complete connectome of an adult fruit fly (Drosophila melanogaster). The FlyWire project, a collaboration between Princeton, Cambridge, and other institutions, mapped roughly 140,000 neurons and over 50 million synaptic connections.

This wasn't just a bigger worm. This was qualitatively different—a brain with complex behaviors: flight, navigation, courtship, learning, memory. Flies have mushroom bodies (critical for learning), optic lobes (processing visual information), and a central complex (navigation and action selection). The circuitry is sophisticated.

Mapping it required new technology:

Serial section electron microscopy at scale. The entire fly brain was sliced into about 7,000 ultrathin sections, each imaged at nanometer resolution. The resulting dataset was 120 terabytes—just the raw images.

Machine learning for segmentation. Humans can't trace 140,000 neurons by hand. AI systems, trained on human-labeled examples, did the initial segmentation—identifying which pixels belonged to which neuron. But the AI made errors, so...

Crowdsourced proofreading. Thousands of citizen scientists, using a platform called Eyewire (and later FlyWire), reviewed and corrected the automated segmentations. It was like a video game where the goal was fixing neuron traces—and the prize was contributing to science.

The result is stunning: a complete map of how a fly's brain is wired. You can explore it online. You can trace the pathway from a photoreceptor in the eye, through visual processing regions, to motor neurons that control the wings. The whole circuit, synapse by synapse.

For the first time, we can see how a complex brain is actually constructed.

What the Fly Map Revealed

Having the connectome immediately enabled discoveries that weren't possible before.

Central complex circuitry. The central complex is a structure in the fly brain that integrates sensory information and controls navigation—like an internal compass. With the connectome, researchers could trace exactly how visual information flows into the compass, how the compass updates the fly's heading, and how heading signals drive motor output. The circuit was more elegant than anyone expected—a ring attractor network that mathematically resembles models theorists had proposed decades earlier.

Descending neurons. Flies have about 350 neurons that descend from the brain to the nerve cord, controlling the body. The connectome revealed which brain circuits connect to which descending neurons—essentially mapping the "action channels" that translate decisions into movements.

Unexpected connections. Some neurons that seemed unrelated turned out to be directly connected. Some pathways were shorter than expected; others were more convoluted. The real wiring often differed from the wiring we'd hypothesized.

The fly connectome doesn't answer every question about fly behavior. But it provides a ground truth that experiments can now test against. When you manipulate a neuron, you know exactly what it's connected to. When you model a circuit, you can use the real topology, not a guess.

The map doesn't explain everything. But it makes everything more precise.

The Human Frontier

Okay. 302 neurons for a worm. 140,000 for a fly. How about 86 billion for a human?

The short answer: not yet. Maybe not for decades.

The long answer: progress is happening faster than most people realize.

The cubic millimeter milestone. In 2021, a collaboration between Google and Harvard published the largest brain connectome yet: a cubic millimeter of mouse cortex. That's about 100,000 neurons and 500 million synapses. It took petabytes of data and years of computation. One cubic millimeter is a grain of sand. The human brain is about 1.4 million cubic millimeters.

Human tissue samples. The same collaboration has mapped small volumes of human cortex—tissue removed during epilepsy surgery. These aren't whole-brain maps, but they're real human neurons, with real human wiring. Early findings suggest human neurons have more complex dendritic structures and more diverse connection patterns than mouse neurons. Evolution made human brains different at the cellular level, not just bigger.

Post-mortem challenges. To get nanometer-resolution images of an entire human brain, you'd need to slice and image it after death. Current technology could do this in principle—but it would take decades of imaging and exabytes of storage. The practical timeline depends on improvements in imaging speed, automation, and data processing.

In-vivo approximations. You can't slice a living brain, but you can estimate connectivity from other signals. Diffusion MRI tracks the paths of white matter fibers—the long-range connections between brain regions. It gives you a coarse wiring diagram, showing which regions connect to which, but not at single-synapse resolution.

Current estimates suggest a nanometer-resolution human connectome might be achievable by the 2040s or 2050s—if technology continues improving exponentially. That's speculative. But the fly connectome was speculative too, until it wasn't.

The Limitations of Wiring Diagrams

Let's be honest about what a connectome can't tell you.

Static vs. dynamic. The connectome is a snapshot—it shows which neurons are physically connected, not which synapses are currently active or how strong they are. A synapse can be anatomically present but functionally silent. Activity depends on neuromodulators, recent history, and ongoing plasticity. The wiring diagram is like a road map that doesn't show traffic.

Electrical vs. chemical. Neurons communicate through synapses, but they also communicate through gap junctions (direct electrical connections) and through volume transmission (chemicals diffusing through extracellular space). The standard connectome maps synapses. It often misses the other channels.

Individual variation. The C. elegans connectome is the same in every individual. The human connectome isn't. Your brain's wiring is partly genetic, but it's also shaped by everything you've experienced. Two humans have similar large-scale architecture but different fine-grained connections. Mapping one human brain gives you one human brain—not a universal template.

Interpretation. Even with the complete wiring, you need to understand what each neuron does—what information it encodes, what computation it performs. The connectome shows structure; function requires additional experiments. Having the schematic of a radio doesn't tell you what song is playing.

Sebastian Seung anticipated this critique. He argues that the connectome is necessary but not sufficient—you need the wiring plus the activity plus the biophysics plus the learning rules to understand a brain. The connectome is one layer of the answer, not the whole answer.

It's still a big layer.

The Allen Brain Atlas and Mouse Atlases

While nanometer-resolution connectomics advances slowly, coarser maps are already transforming neuroscience.

The Allen Brain Atlas provides gene expression maps for entire mouse brains—which genes are active in which regions. It also provides connectivity maps at the mesoscale (region-to-region, not neuron-to-neuron). Researchers can see which areas project to which other areas, based on tracer injections.

The Mouse Brain Connectivity Atlas maps long-range axonal projections across the entire mouse brain. It's not synapse-level resolution, but it shows the major highways—which regions send information to which other regions.

These resources have become standard tools. When a researcher wants to know what a brain region connects to, they don't have to trace the connections themselves—they can look it up. The infrastructure of brain science is catching up with the infrastructure of genomics.

The Human Connectome Project, funded by the NIH, has mapped coarse connectivity in over a thousand human subjects using diffusion MRI. It's revealed systematic variation—how connectivity patterns correlate with cognitive abilities, personality, and clinical conditions. Not synapse-level maps, but population-scale insight.

From Maps to Simulations

The ultimate dream of connectomics isn't just to have a map. It's to simulate a brain.

The Blue Brain Project at EPFL has been building biologically detailed simulations of cortical tissue—not just the wiring, but the biophysics of each neuron type, the dynamics of each synapse. Their models of a cubic millimeter of rat cortex contain digital neurons that fire like real ones.

The idea: if you get the structure right and the biophysics right, emergent behavior should follow. You don't program cognition—you simulate neurons, and cognition emerges.

This is controversial. Critics argue that simulation doesn't equal understanding—you can simulate a system without knowing why it works. Supporters argue that simulation is understanding, just a different kind: you understand something when you can build it.

Whether or not full brain simulation leads to artificial general intelligence, it's already useful for testing hypotheses. You can lesion a simulated brain, manipulate a connection, change a parameter, and see what happens—experiments that are impossible or unethical in real brains.

The connectome is the foundation. Simulation is the building you construct on top.

What the Wiring Diagram Tells Us About Ourselves

Here's the philosophical bet underneath connectomics.

If "you" are your memories, your skills, your personality, your patterns of thought—and if those are encoded in synaptic connections—then the connectome is the physical substrate of identity. Damage the connectome, and you change. Upload the connectome (if that were ever possible), and you could, in principle, persist without the original neurons.

Sebastian Seung calls this "neural Darwinism"—the idea that your neurons compete and cooperate, strengthening some connections and weakening others, and that this selection process writes your experience into your wiring.

The connectome is where nature meets nurture. Your genes specify the initial wiring rules. Your experience sculpts the connections within those rules. The result is a structure that's both universal (all human brains have similar architecture) and unique (no two brains have identical fine-grained wiring).

You are your connectome. And your connectome is the most complex structure in the known universe.

Mapping it is the project of the century.

Further Reading

- Seung, S. (2012). Connectome: How the Brain's Wiring Makes Us Who We Are. Houghton Mifflin Harcourt. - White, J.G. et al. (1986). "The structure of the nervous system of the nematode Caenorhabditis elegans." Philosophical Transactions of the Royal Society B. - Dorkenwald, S. et al. (2023). "FlyWire: Online community for whole-brain connectomics." Nature Methods. - Shapson-Coe, A. et al. (2021). "A connectomic study of a petascale fragment of human cerebral cortex." bioRxiv.

This is Part 2 of the New Neuroscience series. Previous: "Every Neuroscience Headline Lies." Next: "Optogenetics: Controlling Neurons with Light"—Karl Deisseroth's revolution that let us turn neurons on and off with beams of light.