Day #142: So you want to record from the brain…
Today’s post was inspired by a conversation I was having yesterday in the comment section (you know who you are and thank you for the questions). I thought I would elaborate on how we record from the brain and why. There are a lot of different ways we can do this, some of them are super invasive and others are non-invasive. In the lab I work in now, we do things non-invasively there are good things about this and bad things about this, so let’s get into it!
The story begins a long time ago (spoiler not that long ago) back in the late 1930’s and again in the late 1940’s and starts with two men named Hodgkin and Huxley. Using squid giant axon they characterized how neurons worked. This would lay the groundwork for what we do in my lab, recording from the brain!
Like any good story though, there are multiple beginnings and while Hodgkin and Huxley were the fathers of how the neurons work (an important bit of information for sure) long before then back in 1875, we found a way to record non-invasively from the brain using something called Electroencephalography (EEG) which is something we still use today. While we have considerably refined the technique since then, the principle remains the same. The brain has changing electrical potentials and we assume the body is charge neutral, this means on the surface of the skin we can measure the changes in the electrical potential (how positively or negatively charged the skin is).
There are limitations, something called volume conduction limit how we record, so we cannot record from single neurons non-invasively, instead we record from populations of neurons (tens of thousands of neurons) and measure how they become more synchronized or unsynchronized. This is also why we use a lot of electrodes (typically 64 electrodes in our lab), because we can use multiple electrodes and models of how volume conduction causes the signal propagation to spread to back calculate where the signal originated. The other thing is that we also cannot record from deep structures of the brain, the skull is bad at conducting electricity so it does a semi good job of blocking the small amplitude signals that would originate from the deep structures of the brain (small amplitude after they’ve traveled to the skull).
So to this point we’ve talked about a single way to record, let’s look at some of the invasive ways. For example we have nanowires, which is what Elon Musk’s neuralink company is using. Here we eliminate the distance from the recording and get nice and close to the neurons. This means we can get as close as we can to the “ground truth” or the actual, ACTUAL signal from the neuron(s). This makes figuring out what is going on in that part of the brain slightly easier and it means the signal to noise ratio is a LOT higher (this is a good thing).
There is also neural dust, which is coming from a lab run out of Berkeley. I actually know the PI of the lab, we’ve spent a lot of time together recently at the Neurotech workshop. The idea is that instead of sticking wires in, we can inject tiny sensors into the brain. They are wireless, externally powered, and are (in my opinion) slightly less invasive when compared to nanowires.
One of the last big technologies that are invasive is electrocorticography (or ECoG). This is even less invasive (still very invasive mind you) and involves gently placing a flexible sensor array onto the brain, not into the brain, big difference. This has a lot of benefits, namely longevity, but also it is less invasive so less change of causing serious issues. We use this for a lot of different things, but the idea is that this is the same principles that we use for EEG, but instead of outside of the skull we are now recording inside the skull. This means that we are recording from a larger population of neurons again, but we can better localize the activity than we can using EEG. The downside is the area covered, ECoG is very small compared to the size of the brain, this is mostly a materials and invasiveness limitation, if we had the right materials and didn’t mind going crazy with how much of the skull we opened up, we could probably cover the entire brain… potentially.
This brings us all the way back to EEG. Everything we do with EEG can be done (mostly) with more invasive ways (and easier… again mostly). In EEG we have to deal with a lot of noise, if the sensor moves, noise. If we are anywhere near civilization we have electrical signal from the wall, noise (line noise in this case, 60 Hz here in the US and 50 Hz in europe). If you blink or move your eyes, you guessed it noise! We have ways to filter the eye movement/blink noise, remove the sensor movement noise, etc. We can even do this in real-time meaning we can use this method for brain-machine interface. Also what we do in the lab I’m in now.
The best part about doing this non-invasively is that we can do testing with humans. We don’t have to use animal models or are restricted to humans who already have an ECoG implant in the area of the brain we are interested in studying. In fact, I’ve come up with a whole new technique for doing recording (and I’m new to all of this, so it can be done) and I can jump right into human testing because it doesn’t involve surgery!
EEG recording is painless and completely safe. It takes forever to set up (upwards of an hour in some cases) and there is a conductive gel that will get put into your hair under each of the sensors, which means washing your head when you’re done. However, if we compare that to having a surgery, I think most people would feel more comfortable with needing to wash their hair. It’s so safe we demo it, we recently had a demo in a local children’s museum! The kids love it, they get to see their brain activity in real time on large monitors and we get to teach them how we do neuroscience research.
There you have it, the intro to how we can record signals from your brain! Stick around and I’ll talk about my latest experiment (in a few months, after we publish). Unfortunately my super secret technique won’t be revealed until the end of the year at the earliest, but hey it gives you something to look forward to and I can talk vaguely about the progress towards being able to unveil it (assuming it pans out of course).
Until next time, don’t stop learning!