3 ms·
> Octopuses are famously smart, but their central brain is only a small part of their nervous systems. Two-thirds of the roughly 500 million neurons in an octop
by thaumasiotes 6y ago
> Octopuses are famously smart, but their central brain is only a small part of their nervous systems. Two-thirds of the roughly 500 million neurons in an octopus are found in its arms.
OK...
> For the octopus, with thousands of suckers studding symmetric arms, each of which can bend at any point, building a central mental representation of how to move seems like a computational nightmare. But experiments show that the octopus doesn’t do that. “The brain doesn’t have to know how to move this floppy arm,” Cheng said. Rather, the arm knows how to move the arm.
How is this supposed to explain anything? The computational problem is the same whether you solve it with neurons located in the head or neurons located in the arm...?
- Closi 6y agoI think they are kinda saying that the function of the arm is abstracted away from the main brain. Like the main brain can say “move leg” and doesn’t need to worry about how exactly that leg achieves its goal.
- eganist 6y agoNot sure if the analogy I'm about to put forward is oversimplified, but it feels akin to "display Thing on Screen" being instructed to the GPU and the GPU doing all the heavy lifting to calculate what Thing is and ultimately displaying it on Screen.
- ipv6ipv4 6y agoIf this is the case, then ants have a similar characteristic. A newly decapitated ant body will stand upright. It will stand steady against a breeze, stick to an inverted surface and right itself if tipped over. It essentially stands waiting for instruction to walk.
- asdff 6y agoIt's beneficial to do it like this for the octopus given the constraints of its environment. Have a close run in with a seal or a shark or some other predator, and lose an arm, and you aren't wasting energy maintaining resources for the missing arm in the central head brain. Two octopus being equal aside from a central brain versus a limb based system, and you chop off an arm, the limb based octopus will survive and eventually out compete the centralized brain octopus due to lower total metabolic costs from their more modular brain system. The central brain octopus needs to eat more food to pay for a brain it isn't using to its fullest extent. Humans are like the central brain octopus. We have experiences like the phantom limb phenomenon after becoming paraplegic; there is still the signalling network set up for the missing limbs and this is still a metabolic cost to keep up these connections in the brain even with the limb no longer present on the body.
- thaumasiotes 6y ago> It's beneficial to do it like this for the octopus given the constraints of its environment. Have a close run in with a seal or a shark or some other predator, and lose an arm, and you aren't wasting energy maintaining resources for the missing arm in the central head brain. This is a perfectly reasonable thing to say, but it tends to conflict with the article's own interpretation of what it's saying. For example, the image captioned "Octopus movements are too complex to be centrally coordinated." Their point does not appear to be that you can save on energy budget by storing the arm software in the arm, so that it goes away when the arm does. Rather, they seem to be saying that it's possible for the arm software to solve problems that would be too computationally complex to solve, if you were trying to solve them inside the head instead of the arm.
- asdff 6y agoI think the article doesn't make very many conclusive points to be fair, it's more just probing an interesting question about offloading biological compute, but this is certainly an energy question. Everything in biology is due to entropy. The article even concludes talking about the ideal experiment would be to measure the hypothetical metabolic costs of web building vs having more brain tissue and no web building, if such a setup of otherwise equivalent spiders were possible. I think its just an example of how it might be simpler to distribute the load. It's usually simpler to build multiple cheap things than one big expensive thing. We see this with CPUs: our supercomputers aren't single core monsters, but instead thousands of desktop class CPUs strung up in parallel and we distribute our computing jobs across an array of them because this is much cheaper.