Is it technologically feasible to create recording electrodes that can be sunk right through the skull and deep into the brain without breaking? Such a technology would greatly accelerate our ability to understand the brain. What kinds of issues do you see (i.e., are we going to kill some cells by piercing them?) or why is this flat out not conceivable? I’m thinking along the lines of carbon nanotubes with a diameter of 10nm if that helps get things started.
How does it do that?
There have been years of research into neural interfaces, and the issues relating to bio-compatibility and infection control have only just started to be reliably addressed. Longterm implantation of deep brain stimulation electrodes is only just becoming more than experimental. Small (100x100) neural grid arrays are being tested. There is a problem with rigidity - the brain flexes, and a rigid array will cause damage to the very neurons it is interfacing with as this happens. Materials science will move on, and flexible arrays will be developed, as will 3D penetrating arrays using nanowires of some sort. However, I suspect that some damage will always occur during insertion - the brain can cope with some damage, but avoiding scarring will be a factor.
However, accuracy introduces it’s own problems - monitoring a single neuron is exciting, but brain function is an emergent behaviour based on larger structures. The larger scale of fMRI identifies structural operation (with the loss of temporal detail, as fMRI is relatively slow).
Also, non-invasive techniques (Trans Cranial Magnetic Stimulation and the like) will improve and allow more specific targeting. There has already been research where fMRI identifies overactive neural paths and TMS is used to reset the specific neural area for intractable pain/dystonia. This is probably a better and ultimately safer approach than physically penetrating the skull.
The concept of neural interfacing is a staple of SciFi - Julian May describes this as part of her Galactic Milieu/Pliocene series, where Cerebrocybernetic interfaces are used to enhance metapsychic abilities. However, the protagonist is self-healing, so such interfacing is physically safe for him - the emotional/spiritual impact of the enhanced powers is a big part of how the series plays out.
Si
You are going to damage the brain and not get any useful information out of it.
Start reading up on artifical neural networks and you’ll quickly see what the problem is. We can simulate neurons fairly easily. If you compare the brain to a computer, a computer has a lot of dedicated parts that all interact on very small levels with a very limited number of other very specific parts. Computers are therefore very easy to break down into individual pieces and understand how they work.
The brain isn’t like that at all. What you see very easily with neural networks is that when you start adding more “neurons” they all start interacting with each other, and these interactions start to get really complex really quickly. Instead of three or four variables, you may suddenly be dealing with hundreds or thousands of simultaneous variables. And this is all before your artificial network really starts to get interesting. Make your artificial neural network even more complex, and you find that your artificial network starts to mimic the signals in the human brain. The problem is that by that point the thing is so ungodly complex that you can’t figure out what the hell it is actually doing. You can see the signals, but there are so many things all interacting with each other at once that it’s just not possible to break it down into simple steps and understand it. When you break it down into simple small parts, the interesting behavior goes away. When you scale it up to the point where it starts to get interesting, it is so complex that no one can understand it.
At this point, looking at signals doesn’t buy you much. We know what the signals from neurons look like. The hard part is figuring out how many billions of neurons with trillions of interconnections all work together.
Sorry but you guys seem to have missed the point. Seizures don’t occur in the places we’d like them to so we don’t get access to the parts of the human brain we’d most like to record from. What we really need is the ability to measure the activity of fairly arbitrary neurons in the brains of awake behaving humans. Such an ability would allow us to quickly play 20 questions with the brain and figure out how it works.
So…back to my question - the physics of a recording electrode that can simply be inserted through the skull. I’ve read that carbon nanotubes can have diameters on the order of angstroms, if that helps get things started.
Recordings electrodes that can go through the skull and target any part of the brain have existed for decades. The only problem is that they would indeed break going through the skull… so one simply has to drill a hole to put them through. Obviously you don’t want to do that to anyone, and therefore this kind of technique is restricted to animals.
There are exceptions, for instance electrodes are used to locate the site of the seizure in human patients suffering from major epilepsy trouble.
It is done, literally, but not on humans.
I suspect that cells would actually survive being pierced by such a small tube. But anything that thin would just break.
As a matter of fact, you get to understand quite a lot of interesting behaviours by doing exactly that. Of course we can’t get near to explaining the human mind, but we understand fairly well how the neurons in visual cortex work (and interact with each other), we know about motor neurons, decision making, reward through the dopaminergic system, etc.
My question is, specifically, what would the physical properties of an electrode that could go through the skull and not break be. In particular, could this be done with carbon nanotubes.
However, I give up on asking this question. For the most part, people just want to digress.
Carbon nanotubes are flexible, so you aren’t going to be making needles out of them. It might be theoretically possible to make some type of needle structure out of carbon. It’s going to have the same problems as a conventional electrode - you’ll need to actually make a hole in the skull to get to the brain, even if you’re using some kind of tiny carbon needle.