MRI scanners are loud because their internal coils vibrate rapidly inside a powerful magnetic field, and those vibrations resonate through the enclosed bore like a drum. Peak noise levels can reach 120 to 130 decibels during certain scanning sequences, which puts them in the same ballpark as a rock concert or a jet engine at close range. The noise is not a design flaw or a sign that something is broken; it is a physical consequence of how the machine creates images, and the louder, faster sequences often produce the most diagnostically useful pictures.
What Creates the Noise
An MRI scanner uses three kinds of magnetic fields. The main field is the constant, powerful one (measured in Tesla) that aligns hydrogen atoms in your body. The radiofrequency pulses are the brief signals that nudge those atoms to emit detectable signals. Neither of those is particularly noisy. The culprit is the third component: the gradient coils.
Gradient coils are loops of wire built into the scanner’s bore. During a scan, electrical current is switched rapidly through these coils to create small, precisely shaped variations in the magnetic field, which is how the machine figures out where in the body each signal is coming from. The problem is that a wire carrying current inside a strong magnetic field experiences a physical push, known as a Lorentz force. Every time the current switches direction, the coil flexes in the opposite direction. This happens thousands of times per second during a scan, and the resulting vibration radiates outward as sound.1PubMed Central. Overview of Methods for Noise and Heat Reduction in MRI Gradient Coils
The vibrating gradient coil acts like a loudspeaker cone inside a tube. The scanner’s bore amplifies the sound and directs it toward the patient lying inside. Because the current-switching patterns change depending on the type of image being acquired, different scan sequences produce different pitches and rhythms. Some sound like rapid hammering; others produce a sustained, almost musical drone. The common thread is that they are all startlingly loud for something happening inches from your head.
How Loud It Actually Gets
Noise levels vary by machine, field strength, and the specific imaging sequence. At a 4-Tesla research scanner running echo-planar imaging (EPI), one of the fastest and loudest sequence types, sound pressure levels were measured between 120 and 130 decibels inside the bore.2PubMed. Acoustic noise characteristics of a 4 Telsa MRI scanner For context, sustained exposure above 85 decibels is considered hazardous to hearing by occupational safety standards, and 130 decibels is near the threshold of pain.
Most clinical scanners run at 1.5 or 3 Tesla, which are quieter than that worst case but still far from comfortable. Measurements comparing 3-T and 7-T scanners found that the time-averaged sound level across sequences was about 91 decibels at 3 T and roughly 106 decibels at 7 T, with maximum peaks exceeding 105 decibels for the majority of scans at both field strengths.3PubMed Central. Acoustic Noise Levels in High-field Magnetic Resonance Imaging Scanners Greater magnetic field strength increases the force that vibrates the gradient coils, so stronger magnets mean louder scans.3PubMed Central. Acoustic Noise Levels in High-field Magnetic Resonance Imaging Scanners
Not every sequence is equally punishing. A slow, conventional spin-echo scan can be noticeably quieter than a fast EPI sequence, and the pitch profile differs too. EPI sequences pack their gradient switching into tight bursts that produce sharp, high-frequency noise, while other sequences spread the switching over longer intervals and generate a lower-pitched hum. If a technologist tells you “this next one will be louder,” they are not guessing; different scan types genuinely produce different sound levels.
Can MRI Noise Damage Your Hearing?
The short answer is that a single scan with proper ear protection does not cause lasting hearing loss, but the noise is intense enough to produce a temporary shift in how well you hear. A study of healthy volunteers who underwent a roughly 50-minute 3-T brain MRI while wearing hearing protection found an average temporary threshold shift of about 5 decibels immediately afterward. That shift was well below levels associated with nerve damage, and by 25 days later, hearing had returned to baseline with no significant difference from the pre-scan measurement.4PubMed. Temporary Hearing Threshold Shift in Healthy Volunteers with Hearing Protection Caused by Acoustic Noise Exposure during 3-T Multisequence MR Neuroimaging
A separate study looking at patients who had 1.5-T head and neck MRIs found a similar pattern: measurable threshold shifts at certain frequencies immediately after scanning, but no significant difference between pre-scan and 24-hour post-scan hearing.5PubMed Central. Hearing Thresholds Changes after MRI 1.5T of Head and Neck The consistent finding across research is that these temporary shifts resolve and do not translate into permanent damage when hearing protection is worn. The concern, though, is what happens if protection is inadequate or absent, or if someone undergoes frequent scans over a career. MRI technologists and research volunteers who spend hours near running scanners have a different risk profile than a patient getting a single scan.
Why Hearing Protection Is Not Always Enough
Every MRI facility provides hearing protection, typically foam earplugs, over-ear headphones, or both. The effectiveness of these devices varies widely depending on the product, the fit, and the frequencies involved. In the frequency range where MRI noise is most intense (roughly 1 to 1.4 kHz), earplugs alone reduce perceived sound by about 25 to 28 decibels, earmuffs alone by about 30 to 37 decibels, and the two together by about 39 to 41 decibels. Adding a specialized acoustic helmet on top of earmuffs and earplugs pushed attenuation to 55 to 63 decibels in that same frequency band.6PubMed Central. Isolating the auditory system from acoustic noise during functional magnetic resonance imaging: examination of noise conduction through the ear canal, head, and body
Those numbers look reassuring until you consider the gap between laboratory ratings and real-world use. A comprehensive review of MRI acoustic outputs and hearing protection performance concluded that the loudest MRI equipment requires the best available passive protection to meet patient safety guidelines, and only when that protection is fitted correctly. When government derating formulas are applied, which estimate how much protection people actually achieve in practice rather than under ideal test conditions, the numbers become less comfortable. The various metrics do not consistently confirm that even the best devices provide sufficient protection during the loudest scans.7PubMed. A Review of MRI Acoustic Noise Outputs and Hearing Protection Device Performance
This gap between rated and actual attenuation is a known issue in industrial hearing protection generally, but it matters more in MRI because you cannot simply step away from the noise. Once you are in the bore, you are committed for the duration of that sequence. A poorly seated earplug that lets in 10 extra decibels above its rated attenuation could mean the difference between a safe exposure and one that exceeds recommended limits.
Engineering the Noise Away at the Source
Since the fundamental problem is a vibrating coil inside a tube, engineers have attacked the noise from several angles. One approach seals the gradient coil assembly inside a vacuum chamber, because sound cannot travel through a vacuum. A demonstration system that combined vacuum enclosure with vibrationally isolated gradient mounting and low-eddy-current components reduced acoustic noise in the patient bore by about 20 decibels compared to a standard scanner, bringing levels below 85 decibels for several typically noisy pulse sequences.8PubMed. Making MRI quieter That is a substantial improvement: 20 decibels corresponds roughly to a fourfold reduction in perceived loudness.
The same research group confirmed that the key strategies are blocking both airborne vibration (through the vacuum seal) and structure-borne vibration (by mechanically decoupling the gradient coil from the rest of the scanner), along with reducing eddy currents in surrounding metal components that also vibrate and add to the noise.9PubMed. Quiet MRI with novel acoustic noise reduction Some modern commercial scanners have adopted versions of these ideas, though the most aggressive sound-reduction designs add cost and mechanical complexity.
Another strategy works on the software side rather than the hardware. Because the noise frequency depends on how fast the gradients switch, you can design pulse sequences that avoid exciting the gradient coil’s most resonant frequencies. “Soft” gradient pulses that limit their energy to frequencies below about 200 Hz, where the coil’s acoustic response is weak, can cut noise substantially.10PubMed. “Silent” MRI with soft gradient pulses The trade-off is speed: gentler gradient ramps take longer, which means longer scan times. Some approaches compensate by using spiral readout trajectories that collect data more efficiently, partially offsetting the time penalty of quieter gradients.11PubMed. Quiet imaging with interleaved spiral read-out Several MRI manufacturers now offer “silent” or “quiet” scan options as standard features, though these typically work best for certain anatomical targets and sequence types.
Neonates and the Noise Problem
The noise question becomes especially pointed for newborns. Premature infants and neonates who need brain or body imaging have immature auditory systems, including a weak or absent middle-ear muscle reflex that normally provides some protection against sudden loud sounds. Standard adult earplugs do not fit tiny ear canals, and sedation, while sometimes used to keep babies still, does not protect the inner ear from acoustic damage.
A purpose-built neonatal MRI system installed in a neonatal intensive care unit was measured to be on average about 14 decibels quieter than a conventional scanner across sequences, and roughly 30 decibels quieter in the low-frequency vibrotactile range below 200 Hz. With appropriate hearing protection, high-quality exams were performed without exposing infants to noise above 65 decibels, and the lower noise levels also increased the likelihood of completing the scan without sedation.12PubMed Central. Characterization of acoustic noise in a neonatal intensive care unit MRI system Beyond hearing, the noise reduction also mattered because loud sound can trigger autonomic instability in fragile newborns, including swings in heart rate and blood pressure.
A study evaluating whether 3-T brain MRI causes measurable auditory effects in neonates found that a single 40-minute scan, with equivalent sound pressure levels measured at 104 to 114 decibels, did not produce significant disruption of auditory function as measured by standard neonatal hearing tests, provided adequate hearing protection was in place.13PubMed. Auditory Effects of Acoustic Noise From 3-T Brain MRI in Neonates With Hearing Protection That finding is reassuring, but researchers have acknowledged that the risk of noise-induced hearing loss in neonates is real enough that specialized hearing protection devices designed for the smallest patients continue to be developed and tested.14Journal of Radiology Nursing. Magnetic Resonance Imaging Evaluation of a Hearing Protector Device Designed for Neonatal Patients
How Noise Interferes with Brain Imaging Research
MRI noise is not just an annoyance or a safety issue; it creates a genuine scientific problem for researchers studying how the brain processes sound. Functional MRI (fMRI) detects brain activity by tracking changes in blood flow, and the auditory cortex lights up whenever it receives sound input. The scanner’s own noise is a constant, intense sound stimulus that activates auditory brain regions whether the researcher wants it to or not. This inflated baseline makes it harder to detect the brain’s response to the specific sounds being studied.15PubMed Central. Methodological challenges and solutions in auditory functional magnetic resonance imaging
The interference is frequency-specific. When scanning noise has its peak energy at a particular frequency, the brain’s measurable response to experimental tones near that frequency drops. Change the scan protocol so its peak noise shifts to a different frequency, and the suppressed response shifts too, following the scanner noise like a shadow.16PubMed Central. The effect of MR scanner noise on auditory cortex activity using fMRI This means that maps of auditory cortex organization can be distorted by the very machine measuring them. Researchers have developed workarounds, including “sparse” imaging protocols that acquire brain images in brief bursts with silent gaps in between, giving the auditory cortex time to settle before the next sound stimulus is played. But these techniques slow data collection and require careful experimental design.
Animals in the Scanner
Veterinary MRI is a growing field, with scanners used to diagnose everything from cruciate ligament tears in dogs to neurological disease in horses. Animals experience the same acoustic environment as human patients, but with some added vulnerabilities. Many species have more sensitive hearing than humans, especially at high frequencies, and animals cannot be instructed to lie still or told that the noise is harmless. A review of noise exposure during animal MRI found that many scans expose animals to noise levels and durations that would exceed occupational exposure limits set for humans, and recommended that hearing protection be used for animals as a matter of course.17Journal of Magnetic Resonance Imaging. MRI acoustic noise can harm experimental and companion animals
A study focused specifically on dogs found that MRI noise produced a significant reduction in frequency-specific cochlear function, though it was unclear whether the effect was reversible or permanent. The authors recommended that all dogs undergoing MRI should receive ear protection as a routine precaution.18PubMed. The effect of magnetic resonance imaging noise on cochlear function in dogs In practice, veterinary clinics vary in how consistently they provide hearing protection for animal patients, partly because suitable devices for animal ear anatomies are less standardized than those for humans.
Noise Beyond the Bore
The noise problem extends beyond the patient inside the scanner. MRI suites have to be designed so that the sound does not bleed into adjacent rooms. Vibrations from the gradient coils can travel through the floor and walls as structure-borne sound, reaching spaces that might be quite distant from the scanner room itself. Facility design for MRI installations involves both airborne sound containment and vibration isolation of the floor slab, because manufacturer data indicate that sound emissions across a broad frequency range can exceed acceptable noise levels for neighboring occupied rooms without proper treatment.19Building Acoustics. Structural Floor Vibration and Sound Isolation Design for a Magnetic Resonance Imaging System If you have ever wondered why MRI suites tend to be tucked away in hospital basements with thick walls and heavy doors, this is a big part of the reason.
The control room, where the technologist sits during your scan, has its own acoustic challenges. Even with a sealed window and solid construction, gradient vibrations conducted through the building structure can make the environment fatiguing for staff who spend entire shifts running scans. Some newer installations use floating floor designs or elastomeric isolation pads under the magnet to decouple it from the building, treating the scanner like a piece of heavy industrial equipment rather than a standard medical device.
Why Scanners Are Not Getting Quieter as Fast as You Might Expect
Given all the engineering solutions that exist, it is reasonable to wonder why MRI scans are still so loud. The answer is that loudness and diagnostic power are linked in ways that resist easy separation. Faster, stronger gradient switching produces sharper images with better spatial resolution and shorter scan times. Hospitals want scans done quickly so they can see more patients, and clinicians want the best possible image quality to catch small lesions or subtle pathology. Quieter sequences that use gentler gradient ramps tend to sacrifice speed, resolution, or both.
The trend toward higher field strengths compounds the issue. Research and specialized clinical scanners at 7 Tesla are becoming more common, and these are substantially louder than the 1.5-T and 3-T machines that make up most of the installed base. Ultra-high-field systems open up new diagnostic capabilities, particularly in neuroimaging and musculoskeletal imaging, but they bring noise levels that challenge even the best passive hearing protection. Meanwhile, the push for faster acquisition techniques like simultaneous multi-slice imaging and compressed sensing involves more aggressive gradient activity, not less. The industry is essentially running in two directions at once: engineering quieter gradient hardware while simultaneously demanding more from that hardware to improve image quality and speed.
Implanted medical devices add another wrinkle. Patients with active auditory implants such as cochlear implants or bone-conduction hearing devices face additional concerns beyond simple noise exposure. The gradient field’s rapid switching can induce forces and vibrations in the implant’s metallic components, and the static magnetic field exerts torque on any magnetic parts.20ScienceDirect / Journal of Otology. Review Safety of active auditory implants in magnetic resonance imaging For these patients, the acoustic environment inside the scanner is only one piece of a more complex safety evaluation that includes heating, displacement forces, and device malfunction. MRI compatibility labeling for implants has become an entire subspecialty of safety engineering, with detailed condition-specific guidelines for which devices can safely enter which scanners.