Why does Neuralink place place sensors in the neocortex instead of the old brain?

https://psychology.stackexchange.com/questions/27351/why-does-neuralink-place-place-sensors-in-the-neocortex-instead-of-the-old-brain

I think, old brain regulates most of our primary function like reflexes, heartbeat, breathing, fight or flight reaction, fear anger etc, it’s deeper in the structure and also it doesn’t hold any useful memory or creativity I think, its our monkey brain, so may be that’s why Elon didn’t want to mess with that I guess.

In vivo studies on monkey and chimp brains have pretty much defined ‘old brain’ behavior. Keep in mind, no pun intended, that chimps have a significant neocortex. What chimps do not have is language nor a consciousness executive process that manages the inner voice and our sense of self. Current research has implicated cerebral torque as the structure that enabled human language and this suggests that if a chimp could be genetically altered to develop a opposing right-left asymmetry of frontal and parieto-occipital regions they could learn to speak, although the chimp vocal apparatus and smaller neo-cortex, would limit verbal capabilities. Yes, ‘Planet of the Apes’, who knew?

So sensors on the neocortex makes a good deal of sense, because that is where the human specific thinking is happening.

I agree that if your goal is motor control for paralyzed people, reading brainstem or spinal neurons is plausibly a good way to do that. My best-guess explanation why they’re not doing that is: (1) the cortex is a lot easier to access, being on the outside, near the skull; (2) I hear the neuralink longer-term goal is futuristic high-bandwidth two-way brain-computer interfaces even for healthy people, and in that case (I assume) connections to cortex are better.

It’s probably also easier to work with because it has so many more neurons. For classifying neural states into movements based on old brain especially spinal cord, you might need to stick electrodes into specific neurons. That sounds like it’d kill the neurons, having a hole for decades. Whereas in cortex you can record outside cells to listen to a bunch of randomly picked neurons, and classify based on correlations or motor maps or whatnot.

1 Like

This article seems relevant.