Descending neuron · Lobeworks/17

A descending neuron is a neuron whose cell body lies in the brain and whose axon runs down into the nerve cord, carrying the brain's commands to the circuits that move the body; in the fruit fly about 1,300 of them form the brain's whole output to the legs and wings.

Descending neuron. A descending neuron is a neuron whose cell body lies in the brain and whose axon runs down into the nerve cord, carrying the brain's commands to the circuits that move the body; in the fruit fly about 1,300 of them form the brain's whole output to the legs and wings.

Each descending neuron reaches circuits of the ventral nerve cord that already know how to produce a movement, so its message is closer to a choice than to a muscle command. Some act as switches for a whole behaviour: activating the moonwalker descending neurons makes a fly walk backwards, and silencing them stops it from backing away. Their counterparts in a vertebrate are the tracts from the brain to the spinal cord.

They are the readout of whole-brain models. The fly simulations of 2026 read the firing of descending neurons and translated it into commands for a body model, because the circuits below them were not modelled.

They mark where a brain map ends. In a map of the brain alone they are cut at the neck, so their targets, and everything that turns their choice into a gait, are absent.

Questions: What do the moonwalker descending neurons do? They make a fly walk backwards. Activating them switches the fly from forward to backward walking, and silencing them impairs backing away, so a few neurons carry a command for a whole change of behaviour. They are an example of why the descending neurons are the natural readout of a brain model: their activity is already close to a decision. What does the fly's nerve cord do that the brain does not? It turns a choice into movement. Descending neurons from the brain tell the nerve cord what to do, walk, groom or take off, and the circuits of the nerve cord produce the timed contractions of the muscles in each leg and wing. A brain alone can select a behaviour but cannot perform it, which is why a model of the brain without the nerve cord needs a body model to move. Why do whole-brain fly models read out descending neurons? Because they are the brain's only output to the body. Every command to the legs and wings leaves the brain through about 1,300 descending neurons, so a simulation of the brain alone can do nothing with its activity except read those neurons and translate their firing into commands for a body model, which is what the virtual flies of 2026 did.