^-- Up

A Minimally-biased Philosophy of Life: Physics/Resource distribution for directed spines

The primary task of this ideal brain (in both importance and bandwidth) is to control a simplified spine. The brain's modules are its canonical computational resources; they take up space and consume energy. Efficiency dictates the two should be well-matched: portions of space and scale modules represent should be distributed approximately equally along the spine.

But approximately does not mean exactly, because boundary conditions matter too. For one thing, a real spine has two ends, which form discontinuities in sensation, acceleration, and control, meaning that both head- and tail-ends require more computational resources (modules) than does the middle. Furthermore, the two ends are not the same. The tail-end is lightweight like a whip, easily accelerated and thus ideal for signaling. The head-end, on the other hand, contains the brain and certain sensory organs; it accelerates less easily, but is physically closer to the sensors it does contain, so it can communicate with them better. So already, based only on the difference between head and tail, we expect the brain-space dedicated to head control and tail control to specialize differently.

The main difference is communications bandwidth. Both latency and the timing jitter of sensory signals scale linearly with distance travelled. Even if proprioception were the only kind of sensory signal, and spinal control its only function, we would expect modules controlling the head (at the high-bandwidth end) would be more numerous and more fine-grained than tail modules.

So what of communication, heart, lungs, digestion, and all the other metabolic and sensory processes which must be controlled as well? Those all very over seconds, not milliseconds--and hence require less bandwidth than the spine--so they can piggyback on the baseline control arrangement outlined above, with higher-bandwidth tasks toward the head and lower ones toward the tail.

But that division requires not only that the bandwidths of a metabolic system match the bandwidth of its corresponding section of the spine, but that the system physically reside there. In other words, that the basic physical layout of vertebrate organs originally appeared (500 million years ago) primarily to satisfy computational constraints. In this unorthodox view, the high-bandwidth eyes and ears reside in the head in order to have the tightest computational link with the brain. The next-highest consumers of bandwidth, mouth and throat, are placed physically by the neck. Further down comes breathing, then digestion and visceral sensation. Finally the slowest-changing tasks, elimination and reproduction, are processed in the lowest-bandwidth brainspace representing the tail-end.