Films · Developmental Neuroscience
A Label Is Not a Wire
Read the transcript of “A Label Is Not a Wire”
Film transcript: A Label Is Not a Wire
How does a brain wire itself, partner by partner? Start with one small circuit. Tap the tendon below the knee, and the leg kicks. A sensory neuron in the thigh muscle excites the motor neurons of that same muscle. Through an inhibitory interneuron, it also quiets the motor neurons of the opposing muscle. The right partners, with the right signs. Development built every one of those connections. Now scale up. A human brain has about 86 billion neurons. They are joined by something like 100 trillion synapses. The genome that builds them has about 20,000 protein-coding genes. Could the genome simply list the connections? Even using every letter of the genome, the list would be about a million times too long.
So is the trouble giving every neuron its own name? No. Names are cheap. Each extra on-off gene doubles the number of combinations. A few dozen genes could give every neuron its own name. But a name says which cell you are. It does not say whom to connect to.
A label is not a wire. So the genome stores rules. The first is a graded label. Retinal ganglion cells carry a receptor whose amount depends on where they sit in the eye. Their target in the brain carries the matching cue, in an opposing gradient. The second is a local rule. Each growing tip, the growth cone, reads only what is right next to it. Axons with more receptor are kept away from places with more cue.
In our toy model, each axon settles where receptor times cue reaches a set threshold. That rule and the opposing gradients give the right overall order. Real maps also use competition, and more cues. Now take 200 evenly spaced axons and a tenfold gradient. Suppose each reading has an independent random error of about 10 percent. In this toy model, overall order stays high, but neighbouring pairs land in the right order only about 53 percent of the time.
A third rule sharpens that: overproduce, then select. Each axon branches more widely than it needs. Neighbouring cells in the retina tend to fire together. A branch among its retinal neighbours fires in step with them, and together they drive the same target cells. Such branches tend to stay. Branches among strangers tend to retract. The tissue measures this after the connections form, so the genome never has to store it. Here is a test. A goldfish can regrow a cut optic nerve. Block every spike from its eye with tetrodotoxin while the axons grow back. Which part of the map survives: the rough order, or the fine precision?
The axons still reached their target, in roughly the right order. But their branches stayed spread over a wide area, instead of clustering tightly. Molecular cues probably placed them. Activity sharpened them. Now block spikes in zebrafish embryos, as their first map forms. The map formed in order, and single arbors stayed small and tight. So the line can fall in different places. But these fish also differ in age, and in regrowing a map versus building one.
The takeaway: the genome stores rules, and the tissue measures the rest. The course follows those rules in three acts: where cells go, how axons find their targets, and who survives.
About 20,000 genes, about 86 billion neurons. The genome cannot list the wiring, so what does it store instead?
Continue in the lesson: The Wiring Problem: How a Brain Gets Built →
4:26 · Free, with captions and a transcript. The lesson has the full text, the questions and the sources.
How this film was made. Narration: an AI narrator voice (Kokoro, af_heart) reading narration written from this lesson's text. Animation: AI-generated in code (Manim) by Claude, a schematic with no drawn molecular structures. Captions come from the narration script. See something wrong? Report it.