MRI basics: T1, T2 and T2* relaxation
Magnetic resonance imaging (MRI) uses the hydrogen nuclei in the body's water and fat. A strong magnet lines them up, a radio pulse knocks them over, and the scanner listens as they recover. Different tissues recover at different rates, described by the times T1 and T2, and that difference is what makes the picture.
The short version
- In the scanner's field, hydrogen nuclei (protons) produce a small net magnetization along the field.
- A radio-frequency (RF) pulse at the protons' resonance frequency tips that magnetization over; it then relaxes back.
- T1: how fast magnetization regrows along the field (longitudinal recovery). T2: how fast the signal in the sideways plane dies away (transverse decay).
- T2* is the observed, faster decay that also includes small unevenness in the magnetic field. T2* ≤ T2 ≤ T1.
- Choosing TR (repetition time) and TE (echo time) decides whether an image is T1-weighted (fat bright, fluid dark) or T2-weighted (fluid bright).
From spins to signal
The physics is the same as in NMR spectroscopy. Hydrogen-1 nuclei behave like tiny magnets. In a scanner field of, say, 1.5 or 3 tesla, slightly more of them point with the field than against it, giving the tissue a small net magnetization pointing along the field (the "longitudinal" direction).
The scanner then sends an RF pulse at the protons' resonance (Larmor) frequency, about 42.6 MHz per tesla: roughly 64 MHz at 1.5 T and 128 MHz at 3 T. A "90-degree" pulse tips the magnetization into the plane at right angles to the field (the "transverse" plane), where it precesses like a spinning top and induces a radio signal in the receiver coil. After the pulse, two things happen at once, at different speeds.
T1, T2 and T2*
| What it describes | Definition | Also called | |
|---|---|---|---|
| T1 | Regrowth of magnetization along the main field | Time to recover about 63% of the way back to equilibrium | Longitudinal or spin–lattice relaxation |
| T2 | Loss of signal in the transverse plane, as spins fall out of step with each other through their own interactions | Time for the transverse signal to fall to about 37% of its starting value | Transverse or spin–spin relaxation |
| T2* | The decay actually observed: T2 effects plus dephasing from small inhomogeneities in the magnetic field | Same idea as T2 but always shorter or equal | "T2-star", effective T2 |
Both processes are exponential. After one T1 the longitudinal magnetization is about 63% recovered; after three T1s, about 95%. Transverse decay is faster than longitudinal recovery, so in tissue T2 is always shorter than T1. In body tissues T1 values are typically hundreds to a few thousand milliseconds and T2 values tens to a couple of hundred milliseconds, depending on the tissue and the field strength (T1 gets longer at higher fields).
Why does T2* exist? No magnet is perfectly uniform, and tissue boundaries (for example between air and bone, or near iron in blood) distort the field slightly. Spins in slightly stronger field precess slightly faster, so they drift out of step even faster than T2 alone would cause. That extra dephasing can be undone, which leads to the two basic pulse-sequence families.
Spin echo versus gradient echo
- Spin echo: after the 90° pulse, a 180° pulse flips the spins so that fast ones are now behind and catch up. At the echo time they come back into step, cancelling the static field-inhomogeneity effects. The echo is weighted by true T2.
- Gradient echo: no 180° pulse; a magnetic field gradient is reversed to form the echo. This is faster, but static inhomogeneity is not refocused, so the signal reflects T2*. That is a feature when you want it: T2*-weighted images show bleeding and calcium deposits, and the BOLD signal used in functional MRI is a T2* effect.
TR, TE and image contrast
The scanner repeats the pulse many times to build an image. Two timing settings control contrast:
- TR (repetition time): the time between successive excitation pulses. A short TR does not leave long-T1 tissues time to recover, so it brings out T1 differences.
- TE (echo time): the time from excitation to reading the signal. A long TE lets short-T2 tissues fade first, so it brings out T2 differences.
| Weighting | TR | TE | Bright | Dark |
|---|---|---|---|---|
| T1-weighted | Short | Short | Fat; gadolinium contrast agent | Water and fluid (such as cerebrospinal fluid) |
| T2-weighted | Long | Long | Water and fluid; many areas of swelling and inflammation | Tissues with short T2; fat is intermediate to bright depending on the sequence |
| Proton density | Long | Short | Tissues with more hydrogen | Tissues with little hydrogen |
Fat looks bright on T1 because its magnetization recovers quickly (short T1). Fluid has a long T1, so it is dark on T1 images, but its long T2 keeps it bright on T2 images. Gadolinium-based contrast agents shorten T1 of nearby water, so tissues they reach light up on T1-weighted images.
Phantoms
In MRI, a phantom is not a ghost: it is a test object, often a container of fluid or gel with known relaxation properties and built-in shapes, that is scanned to calibrate the machine, check image quality and compare scanners over time.
Safety basics
- MRI uses no ionizing radiation.
- The magnet is always on. Iron-containing objects can become projectiles, so everyone and everything entering the room is screened.
- Some implants (certain pacemakers, cochlear implants, neurostimulators) can be unsafe; others are labeled as MR-conditional under specific conditions. The scanning team must check each one.
- Scanners are loud (some over 100 decibels), so hearing protection is used. Rapidly switching gradients can cause harmless twitching, and the RF energy slightly warms tissue, which scanners limit.
- Gadolinium agents are used only when needed in people with severe kidney disease, because of a rare serious condition called nephrogenic systemic fibrosis. MRI in pregnancy is usually avoided in the first trimester as a precaution unless needed.
Memory aid
"T1 is the tide coming back in; T2 is the choir falling out of tune." T1 is recovery along the field (the water level rising back up). T2 is a choir that starts singing in perfect unison and slowly drifts apart until the sound fades. T2* is the same choir in a room with bad acoustics: it sounds out of tune even sooner. A spin echo is the conductor who reverses the drift.
For contrast: "T1 = fat is the one; T2 = water is the view." And "short-short is T1, long-long is T2" (TR and TE).
A verse to remember it
Check yourself
What is the difference between T1 and T2?
T1 describes recovery of magnetization along the main field (longitudinal). T2 describes loss of signal in the transverse plane as spins dephase. T2 is always shorter than T1 in tissue.
Why is T2* shorter than T2?
T2* includes extra dephasing caused by small inhomogeneities in the magnetic field, as well as the spin–spin interactions that define T2.
An image shows cerebrospinal fluid as bright and was acquired with a long TR and long TE. What weighting is it?
T2-weighted.
What does a spin-echo sequence do that a gradient-echo sequence does not?
It uses a 180° refocusing pulse that reverses dephasing from static field inhomogeneity, so its echo reflects T2 rather than T2*.