4 ms·

The foldable subdural ecog step feels like a "then draw the whole owl" situation. Foreign body reactions are a major problem that I don't see a satisfying solut

The foldable subdural ecog step feels like a "then draw the whole owl" situation. Foreign body reactions are a major problem that I don't see a satisfying solution for yet. Is the gradual signal degradation from the tissue reaction solved? Are we going to need to be tinkering around everyone's calvariums every 5 years to replace these? In my brief search online I'm finding these low powered studies on the safety of these devices. I'd like to see what the rates of meningitis or neoplasia look like once the implanted population climbs into the thousands. I strongly suspect that invasive measurement of brain activity will eventually be regarded as a foolish endeavor.

konical 3y ago
Concur with all points sans belief there will never be a method without inflammation, scar tissue or other deleterious long-term response to invasive measure activity. It could be as simple as patching some adipose stem cells with a bit of extra code, onto the ends of our tech. Or perhaps a coating of proteins? Ultimately, all of these direct interfaces requires this anti-scarring biologic integration technology before being able to be mainstream.
- TaylorAlexander 3y ago
  > It could be as simple as patching some adipose stem cells with a bit of extra code, onto the ends of our tech. Is that simple? Genuinely curious, do you have any links to research about this? That sounds very not-simple to me but I don't actually know anything about this field.
  - XorNot 3y ago
    Yes and no: adding the right protein coat solves a lot of problems. But proteins aren't durable - they degrade, and your body is designed to continuously replace them. Functionally when you want to coat something with a layer of biological <thing X> to fix a compatibility problem, the issue isn't that it won't work, it's that it won't be long term stable - you're not needing to build infrastructure, you're needing to build a whole little self-regulating system that will replace and turn itself over stably.
    - Zenzero 3y ago
      Right. It's a problem I don't see any near term solution for. It is not my area of expertise, but my expectation is that we will ultimately see some sort of externally applied miniaturized EEG/MR combo with improved spacial resolution for feeding into deep learning models. Normalizing electrode implantation sounds like lunacy to me.

konical 3y ago

Concur with all points sans belief there will never be a method without inflammation, scar tissue or other deleterious long-term response to invasive measure activity. It could be as simple as patching some adipose stem cells with a bit of extra code, onto the ends of our tech. Or perhaps a coating of proteins? Ultimately, all of these direct interfaces requires this anti-scarring biologic integration technology before being able to be mainstream.

TaylorAlexander 3y ago
> It could be as simple as patching some adipose stem cells with a bit of extra code, onto the ends of our tech. Is that simple? Genuinely curious, do you have any links to research about this? That sounds very not-simple to me but I don't actually know anything about this field.
- XorNot 3y ago
  Yes and no: adding the right protein coat solves a lot of problems. But proteins aren't durable - they degrade, and your body is designed to continuously replace them. Functionally when you want to coat something with a layer of biological <thing X> to fix a compatibility problem, the issue isn't that it won't work, it's that it won't be long term stable - you're not needing to build infrastructure, you're needing to build a whole little self-regulating system that will replace and turn itself over stably.
  - Zenzero 3y ago
    Right. It's a problem I don't see any near term solution for. It is not my area of expertise, but my expectation is that we will ultimately see some sort of externally applied miniaturized EEG/MR combo with improved spacial resolution for feeding into deep learning models. Normalizing electrode implantation sounds like lunacy to me.