Had my first MRI yesterday. I’d heard all sorts of horror stories about them, but while I’m not fond of tight spaces I closed my eyes and concentrated on the music piped through the headphones. Actually dozed off at least twice. Tech said she wished everyone was that calm and relaxed. Woke up, though, when the noise started.
Damn, those things are LOUD!
Yes, yes, I understand that it has to do with the magnetic fields reversing. But why? Does something inside the machine actually bang back and forth in there? What’s causing all that noise?
At a guess, it’s from eddy-current forces. Whenever you have a magnetic field changing quickly, it’s going to induce forces on anything conductive in the vicinity (note: This is different from the forces on specifically ferromagnetic metals like iron, and affects all conductors). MRI machines are of course made with little or no iron, but you can’t get away from using all metals entirely. The effect on the housing and other metal parts is much the same as if you put an abrupt force on them through other techniques, like hitting them with a hammer.
MRI machines uses metal coils to generate their magnetic field. It’s your basic really simple electromagnet. Run current through a coil and it creates a magnetic field. Power transformers and motors all take advantage of this basic principle, as do many other things.
When an MRI is functioning, the current in those coils goes from nothing to a very high level very quickly. If the coils were infinitely strong, there wouldn’t be a noise, but those very high current levels and very high magnetic forces cause the coils to move and deform slightly, resulting in the noise that you hear.
Power transformers hum faintly for basically the same reason, though in transformers you also have metal laminate plates moving slightly which contributes to the noise. Power transformers don’t pop like an MRI because the current through a transformer alternates back and forth 60 times a second (50 in some parts of the world) and the current level changes much more gradually, resulting in a hum instead of a pop.
I’ve mentioned coin-shrinking before, a mad-scientist stunt in which gigantic electrical currents (with gigantic ramp rates) are used to create gigantic magnetic fields (probably a lot more than the MRI machine) that literally crush coins to some percentage of their original diameter. That link provides a detailed description of the physics involved. Along the way, they point out that the electrical coil used to generate the magnetic field is destroyed in the process. The coil doesn’t melt: it is violently torn apart by tensile stresses related to the magnetic field - point being that strong magnetic fields do in fact produce substantial forces in the coils that generate them, and are no doubt responsible for the vibration/noise that happens during an MRI.
See magnetostriction for an explanation of this phenomenon. This explains why transformers hum even when they’re mounted in plastic cases.
And it’s intrinsic to the way MRI works that the field changes repeatedly, that’s why the noise continues. If it just powered up to a constant magnetic field, you’d just get one bang.
That is pretty much what’s happening. They are not “loose” in the normal sense, but they are being subjected to a huge force each time the field changes, and they are not inifinitely rigid.
Just to clarify - the critical words in the answer are “gradient coils”
In order to make an image the magnetic field needs to vary across the subject. The precise resonant frequency varies with the magnetic field strength. This is the key to how MRI works. In the simplest form you can vary the magnetic field strength so thua only one small volume (voxel) in the entire subject is at a desired frequency - pulse the subject with RF at just that frequency and you get a response from just that voxel. .The first ever MRI images were made this way. It isn’t viable for clinical purposes. Luckily with an FFT you can retrieve energy across a range of frequencies and thus if you impose a carefully constructed magnetic field gradient across the patient you can get information from a more than one voxel. The gradient is very shallow compared to the overall field strength - so whilst the main magnet is of insane power, and may use a superconducting magnet, the gradient coils are just ordinary wire wound around a former, and are driven by a modest power source. (For some reason the first ever research MRI machines all used Crown audio power amplifiers - I guess they were readily available and worked.) In order to get an image is useful time you need to switch the field gradients quickly - the image construction needs a large number of different gradients in order to cover the volume of the subject and resolve signal from each voxel. In addition the precise signal retrieved can vary, and the overall schedule of gradient and activation RF vary (it has become very sophisticated, back in the day you had simple things like proton density, T1, T2 times - relaxation times for the protons, and that was about it.) But you need to be varying the gradients pretty much as fast as you can and hitting the subject with pules of RF to energise the protons - listening to the result and moving on. So the gradients are switched quickly. Which makes the entire gradient coil react against the main magnetic field with a bang, and bang each time the gradient changes.
OK, let me see if I can construct an analogy that makes sense.
Imagine you are holding an aluminum can. By exerting a force on it by closing your fist sightly you get a sound flexing the metal. When you release your grip the metal flexes again and you get another noise.
In an MRI, the magnetic field(s) are likewise exerting forces on coils, which flex, and like the aluminum can, make a noise when the flex. When the force stops the coils flex back, making another noise.
But in an MRI the forces and the coils are enormously larger, so the noise is likewise enormously louder. And the force/release cycle is fast, so you get **BANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzng
**
*WHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbingWHAMbing
* BANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzngBANGzng
They put ginormous ear-muff type things on me, like what the guys on the pavement at the airport or on aircraft carrier wear. Judging by how they cut the ambient room noise I’d guess at least 30 dB of protection, maybe more. I’m sure you could wear those foam ear plugs under them, too, if you wanted to do so.
It’s the level of noise where you actually feel it vibrating your sternum and other large bones.
The obvious question is why not make them quieter since it’s unpleasant for the patient. Efforts are underway but most of the steps to reduce acoustic noise degrades signal-to-noise ratio or increases manufacturing cost. It’s not like a noisy mechanism you can just mount on rubber bushings. Some of the techniques involve placing the gradient coils within a vacuum then attempting to decouple the noise that leaks through the support posts.
Personally if I needed an MRI I’d rather have the highest fidelity image possible, regardless of acoustic noise or acquisition time. It’s usually just a few minutes and if needed I’ll wear both ear plugs and ear muffs.
However I accept it bothers some more than others. Even people who thought they had no trace of claustrophobia can be disturbed by the combination of tightly confined space and sustained loud noises.
OTOH MRI and CT were so revolutionary it’s common in medical circles to speak of medical history in two different eras: the pre-CT/MR era and the post-CT/MR era. When discussing the scientific literature on, say, certain brain tumors, they will often qualify the articles saying “well that data is interesting but you must consider it was before the CT/MR era”.
If it’s any consolation, in 1977 the first human MRI took 5 hours. That’s not a whole body scan – that was just a single image slice.
It may also help to keep in mind that before the CT/MR era, “exploratory surgery” was often necessary to examine a condition. The noise and confinement may be unpleasant but it’s better than “going under the knife”.
A lot of noise in machinery is often an indication of inefficiency, which would be another reason to make them quieter. But this case may not be one of inefficiency as much as physics, which don’t always cooperate with what humans want (that pesky lightspeed limit, for example!)
^ Yeah, this.
Actually, it’s NOT just a few minutes you’re lying in the machine, but it’s only a few minutes out of, say, 30-90 minutes (depends on what they’re doing) of the whole process. I’ll happily wear hearing protection for that time period if it results in better imaging.
Yes. Although I was pretty relaxed throughout most of it I can see where this could hit the panic button for some people. It is a VERY tight space you’re in, in a very big scary-looking machine that you don’t understand. The noise initially did bother me, but once my lizard-brain figured out what the normal sounds of the machinery were they stopped bothering me (other than being loud, but with the hearing protection not unbearably so) and eventually became somewhat reassuring from the “oh, it’s operating normally” manner. But then, I’ve worked around loud machinery and have learned that when the noise is regular and normal all is well. It’s when it stops having that pattern you need to worry. Not everyone has that experience or reaction to loud machinery. We are a bit hardwired to find sudden loud noises alarming and an MRI machine is full of sudden loud noises. And on top of that you’re in a narrow hole, which can make you feel trapped. Yep, I can see animal-panic happening in some people. The tech did say she’s had people freak out in the machine before.
They do give you a panic button so if you do feel a freak out about to happen you can let them know and they’ll pull you out and get you calmed down. I’m pretty sure they wind up sedating some people to get the scan done, too.
For me, having to lie so still for a prolonged period was probably the most uncomfortable part of the test, and frankly, that was a minor annoyance.
Yeah, my mom had exploratory surgery back in the pre-CT/MRI era. At the time they had nothing better, but it took her a couple months to really recover from it. No thanks. I’ll take an hour in an MRI machine over that.
Mind you, anything to be done to improve the technology and make it more comfortable I fully support, but I’m glad I live in an era where this is an option.
Ear protection. Pretty standard. (I’ve heard of using noise-canceling headphones, but not seen them much in practice- I think they’d be quite effective, assuming you could get electronics to work properly in MR environment.)
Mechanical mitigation. Mostly stuff like more sound-proofing between coils and patient. Sometimes making the resin holding the coils together stronger. A scanner came out in which the gradient coils were encased in vacuum. This was apparently quite effective, but (probably) makes for a smaller bore, which seems to bother patients more than the noise.
Advanced pulse sequence design. The noise occurs mostly because the gradient fields are “slammed on and off”, basically square waves (as much as possible). Newer pulse sequence designs (probably way more computationally intense to process into images) permit quieter or even silent imaging. They basically soften the edge of the gradient pulses, rounding the shoulders of the square waves. There is a loss of signal-to-noise ratio (about 10%), which is a drawback, because SNR is what it’s all about in MRI.
Most modern scanners have a significantly wider bore than they used to.
Many people do get sedation. Most don’t need any. Many do well with mild (oral) sedation. Some adults still do need to be “put out”, usually handled by anesthesia. Most children’s hospitals routinely use anesthesia.
I guess the people who most hate the loud noise while trapped in an enclosed space are veterans with PTSD.
