An MRI scanner uses a powerful magnet, radio waves, and the water in your body to produce detailed images of your organs, joints, brain, and other soft tissues. Unlike X-rays or CT scans, it involves no radiation. The process relies on the behavior of hydrogen atoms, which are abundant in every tissue because your body is roughly 60% water, and each water molecule contains two hydrogen atoms.
The Magnet and Your Body’s Hydrogen
Every hydrogen atom in your body has a single proton at its center, and that proton acts like a tiny bar magnet with its own magnetic field. Under normal conditions, these miniature magnets spin with their axes pointed in random directions, so they cancel each other out. When you lie inside an MRI scanner, you’re surrounded by an extremely strong magnetic field, typically 1.5 or 3 Tesla. For comparison, the Earth’s magnetic field is about 0.00005 Tesla, making a clinical MRI magnet roughly 30,000 to 60,000 times stronger.
That magnetic field forces the axes of your hydrogen protons to line up in the same direction, like compass needles swinging toward north. This collective alignment creates a measurable magnetic signal along the length of your body. The protons aren’t perfectly still, though. They wobble, or “precess,” at a specific frequency determined by the strength of the surrounding magnetic field. This wobble frequency is what makes the next step possible.
Radio Waves Knock Protons Out of Line
Once your protons are aligned, the scanner sends a brief pulse of radio waves tuned to exactly the frequency at which those protons are wobbling. Because the frequencies match, the protons absorb the energy. This is the “resonance” in magnetic resonance imaging. The pulse lasts only a few milliseconds, but it’s enough to knock the protons out of alignment with the main magnetic field, tipping them into a different plane.
When the radio pulse switches off, the protons begin snapping back to their aligned, low-energy state. As they do, they release the energy they absorbed, and that energy produces a faint electrical signal picked up by receiver coils built into the scanner. Different tissues release this energy at different rates, and that difference is what allows the scanner to tell muscle from cartilage, gray matter from white matter, or a healthy liver from a diseased one.
T1 and T2: Why Tissues Look Different
Protons return to equilibrium in two distinct ways, and each one gives the scanner different information. The first, called T1 relaxation, measures how quickly protons realign with the main magnetic field. Fat-rich tissues recover their alignment quickly, so they appear bright on T1-weighted images. Fluid-filled structures recover slowly and appear dark.
The second, T2 relaxation, tracks how quickly the protons fall out of sync with each other after the radio pulse ends. Water and other fluids stay in sync longer and appear bright on T2-weighted images, while dense tissues like tendons lose sync fast and appear dark. By choosing whether to emphasize T1 or T2 differences, radiologists can highlight specific types of tissue or detect abnormalities like swelling, tumors, or cartilage tears. This flexibility is one of the biggest advantages MRI has over other imaging methods.
Gradient Coils Build the Map
Aligning protons and listening to the signals they emit would only tell you that hydrogen exists somewhere inside your body. To turn those signals into a precise, cross-sectional image, the scanner uses three sets of gradient coils. Each set slightly varies the magnetic field along one of the three spatial axes: left to right, top to bottom, and head to toe. By layering these gradients, the scanner ensures that protons at every location in your body experience a slightly different magnetic field strength, which means they wobble at slightly different frequencies.
This frequency variation acts like an address system. When the scanner detects a signal at a particular frequency, it knows exactly where in the body that signal originated. The raw data collected this way is stored in a mathematical framework called k-space, where every data point contains partial information about the entire image. A mathematical process called a Fourier transform then converts that raw data into the sharp, cross-sectional images your doctor reviews. Every single point in the raw data matrix contributes to the complete final picture.
What MRI Sees Better Than CT
MRI’s greatest strength is soft tissue contrast. It excels at imaging the brain, spinal cord, joints, muscles, and internal organs. In the brain, MRI can detect swelling, nerve fiber injuries, and bleeding that CT scans miss entirely. In the spine, it reveals bone marrow changes and spinal cord damage invisible on CT. For joints like the shoulder and hip, MRI picks up cartilage tears, small cysts, and early stress reactions in bone that CT cannot show.
CT scans still have their place. They’re faster, less expensive, and better at imaging bone fractures and lung tissue. But when a doctor needs to evaluate ligaments, identify early inflammatory changes, or look for subtle brain abnormalities, MRI is the preferred tool.
Contrast Agents and Enhanced Images
Some MRI exams use a contrast agent injected into a vein to make certain structures stand out more clearly. The most common type is gadolinium-based. Gadolinium doesn’t show up directly in the image the way iodine-based CT contrast does. Instead, it works by altering how quickly nearby water molecules release their energy after a radio pulse. Tissue that absorbs the contrast agent appears brighter, making it easier to spot tumors, inflammation, or areas with increased blood flow. The chemical structure of the contrast agent is designed to attract water molecules to the gadolinium ion, amplifying this brightening effect.
Why MRI Machines Are So Loud
If you’ve ever had an MRI, you know the machine produces intense knocking, buzzing, and clanging sounds. Those noises come from the gradient coils. Each time the scanner rapidly switches the gradient magnetic fields on and off, the coils vibrate against their housing, producing acoustic noise. On a standard 3-Tesla scanner, sound levels routinely exceed 95 decibels and can peak above 105 decibels, roughly equivalent to standing near a jackhammer. Higher-powered 7-Tesla research scanners are even louder, reaching peaks above 120 decibels.
The FDA caps allowable sound levels inside an MRI at 140 decibels peak and 99 decibels average when hearing protection is worn. That’s why you’re always given earplugs or padded headphones before a scan. With that protection in place, exposure stays within safe limits set by occupational health standards, even for longer exams.
Who Can’t Have an MRI
Because the scanner uses an extremely powerful magnet, any ferromagnetic metal inside your body is a potential danger. Pacemakers, implantable defibrillators, and cardiac resynchronization devices pose serious risks because the magnetic field can interfere with their electronics or heat their leads. Metallic foreign bodies in the eye can shift and cause injury. Neurostimulators, certain cochlear implants, drug infusion pumps, cerebral aneurysm clips, and metallic fragments like shrapnel or bullet fragments are also considered unsafe.
Even items you might not think of, such as certain dental implants, tissue expanders with magnetic ports, metallic catheters, and body piercings, can be problematic. Before any MRI, you’ll be screened with a detailed questionnaire about your surgical and medical history. Some newer implants are specifically designed to be MRI-compatible, but the specific device model and the strength of the scanner both matter, so the screening process is thorough.
How Long a Scan Takes
A typical MRI exam takes anywhere from 15 to 45 minutes depending on the body part and the number of image sequences needed. You lie still on a motorized table that slides into the scanner’s cylindrical opening. Movement blurs the images, so staying motionless is important, especially for brain or spinal imaging where detail matters most.
Newer AI-powered reconstruction techniques are dramatically shortening scan times. By collecting less raw data and using deep learning algorithms to fill in the gaps, systems like NYU Langone’s fastMRI can complete exams up to four times faster than conventional scans. An exam that once took 30 minutes can potentially be finished in under 5 minutes, making MRI appointment times comparable to a CT scan. Faster scans also mean less time holding still, which is particularly helpful for children, elderly patients, and anyone with claustrophobia or chronic pain.