MRIs are Cool AF
Having done my graduate school in the field of MRI, I think it is by far the coolest imaging modality out there. But before I dive into some reasons as to why, it’s necessary to have an idea of how an MRI works.
An Brief ELI5 to MRIs
At its core, MRI is looking at protons, or the H in H2O. Once inside an MRI, the protons in your body align with the strong magnetic field generated by the machine. The machine generates a radiofrequency pulse, which effectively knocks the protons out of alignment, and the rate at which they align back with the magnetic field determines how bright it is on an image.
There are these toys called “roly poly toys” or “okiagari” that act similar to the protons. Their low center of mass means that if you tip it over, it naturally stands back up again. The radiofrequency pulse effectively knocks down the toy. The toy naturally wants to stand back up again to align with the magnetic field. If the toy stands back up very quickly, it shows up bright on the image, and if not, it shows up darker.
With that out of the way, let’s get started!
1. It’s completely safe
An MRI uses a combination of radiofrequency (RF) waves and rapidly changing magnetic gradients to image parts of the body. RF waves are quite safe, and it’s reasonably evident by the fact that we’re constantly swimming in RF energy — it’s almost the same kind of radio waves that allow your regular AM/FM car radio receives a signal. And as far as we’re aware, magnets are perfectly safe around people as well. The biggest risk factor that is usually written down when conducting MRI studies on people is actually claustrophobia.
2. The FDA approves and governs MRIs
Surprisingly, MRIs are approved for medical use by the Food and Drug Administration, which also applies hard limits on how much energy an MRI can deposit into the body. The particular frequency range that MRIs use is in the MHz range — so right between radio waves and microwaves (yes, literally the ones that heat up your food). Because of this, the FDA does limit the amount of RF energy that can be deposited into the body — and that limit is that you cannot head up any part of the body more than 1 °C
3. The magnet is supercooled with liquid helium
An MRI is only able to create its extremely powerful magnetic field by the use of superconductors. The superconductor must be kept extremely cold (~7–8 Kelvin) in order to maintain its magnetic field, and this generally achieved using liquid helium. Previously, I used to say this was a concern since we’re actually slowly running out of helium because it’s being underpriced for the party balloon industry, but it seems after a little bit more research it seems that there should be plenty to go around for a while.
4. Only ~1 out of a million H’s actually give signal
There’s a large, green hardcover book that’s referred to as “The Bible of MRIs” by Haacke and I had to double check this one to make sure I remembered it correctly. Protons have only two spin states, and under the effect of a magnetic field can align with the field (lower energy state) or against the field (higher energy state). However, the energy difference between the two is miniscule, and interaction with thermal energy (residual heat) can easily flip a proton from the low energy state to the higher energy state. This means that out of a million protons, only about 1 out of a million are actually properly aligned; the rest are randomly split between the two energy states. And this single proton out of every million is what is providing the signal to create the image. Thankfully, even in a few grams of tissue, there are many Avogadro numbers of protons (10²³) and even 1 in a million (10⁶) adds up quite quickly, allowing us to get enough signal from the body.
5. Data is acquired in frequency space
This is probably the strangest but coolest concept to grasp. Unlike x-ray or ultrasound which have a fairly simple concept to understand, MRI is extremely unique in how data is actually acquired.
The ELI5 version of it is that the protons in your body are spinning (specifically gyrating, like a dreidel) because it’s aligned with the magnet. The magnet can control which parts of your body are spinning faster than others — it’s how it can tell the left side of the body from the right side — because the magnet forces one side to spin faster than the other. Therefore, the data that is being acquired is actually the speed of the protons, aka frequency space (in MRI land, this is called k-space). And by some miracle that is the fourier transform, we can convert that into an actual image.
6. Has clever ways to overcome movement
MRI faces the same problems as other imaging technologies, including cameras — that movement is a huge problem. Since MRIs can take a few seconds to get all the data it needs in an image, it runs into challenges when taking pictures of the heart.
There are primarily two different sources of movement when trying to image the heart — the first is fairly obvious, and that is that your heart is beating. We overcome movement from the heart by keeping track of movement via an EKG. The EKG has a “trigger” which detects when the heart is resting and only acquires data during those moments. Data acquisition is then split up over multiple heartbeats, ie the first fifth of the image is acquired during the first heartbeat, the second fifth of the image is acquired during the second heartbeat, etc.
The second source of movement is actually respiratory movement or movement from breathing. We can ignore respiratory movement by usually asking patients to hold their breath. It doesn’t always work as the people that need MRIs aren’t in the best of shape, and can have trouble holding their breath for even a few seconds. So instead of asking patients to hold their breath, we can also use the MRI to track their respiratory movement. This is achieved by imaging along the liver/lung, and then tracking where they meet.
And with that, until next time, where maybe I will write about why other modalities are not cool AF.
