Magnetic Resonance Imaging (MRI) is a non-invasive medical technique used to generate detailed images of organs and soft tissues inside the human body. Unlike X-rays or Computed Tomography (CT) scans, MRI does not rely on ionizing radiation. Instead, it utilizes powerful magnetic fields and radio waves. The process begins by harnessing the natural properties of atoms within the body, allowing clinicians to visualize structures like the brain, muscles, and ligaments with clarity.
The Magnetic Foundation: Proton Alignment
The operation of an MRI machine depends on the physical properties of the hydrogen nucleus, which is a single proton. These protons are abundant in the body, primarily within water molecules and fat, and possess a property called spin. The spinning charge makes each proton behave like a tiny bar magnet, but normally, these magnets are oriented randomly, canceling out any net magnetic effect.
When a person enters the MRI scanner, they are exposed to a strong, static magnetic field, referred to as the $\text{B}_0$ field. This powerful field acts upon the randomly oriented protons. The external force causes the majority of the protons to align their axes of spin either parallel or anti-parallel to the direction of the $\text{B}_0$ field.
Slightly more protons align in the low-energy parallel state than the high-energy anti-parallel state, creating a small, measurable excess of protons aligned with the main magnetic field. This difference establishes a net magnetization vector pointing in the direction of the $\text{B}_0$ field. The aligned protons also begin to precess, or wobble, around the axis of the $\text{B}_0$ field, similar to how a spinning top wobbles.
The frequency at which these protons precess is known as the Larmor frequency, which is directly proportional to the strength of the $\text{B}_0$ field. This alignment and consistent precession frequency create the necessary conditions to introduce energy and extract a signal.
Generating the Signal: Radiofrequency Pulses and Resonance
With the net magnetization established, the machine introduces energy via a specialized radiofrequency (RF) coil. This coil emits a brief, focused burst of electromagnetic energy, the frequency of which is precisely matched to the Larmor frequency of the precessing protons. When the RF pulse hits the protons, they absorb this energy, a phenomenon called resonance.
The absorbed energy causes the net magnetization vector to be “tipped” away from the main $\text{B}_0$ field, moving into the transverse plane. This tipping angle is controlled by the RF pulse duration and intensity, often aiming for a 90-degree flip to maximize the signal. The energy causes the low-energy parallel protons to jump to the high-energy anti-parallel state, and forces the precessing protons to momentarily spin in phase with each other.
The synchronized precession of the protons in the transverse plane generates a detectable, oscillating magnetic field, which is the signal the MRI machine measures. The moment the RF pulse is turned off, the system begins to return to its original, low-energy equilibrium state. This return process is called relaxation, during which the protons release the absorbed energy as a faint radio signal.
The relaxation process is defined by two different time constants that provide tissue contrast. The $\text{T}1$ relaxation, or longitudinal relaxation, measures the time it takes for the net magnetization vector to realign with the $\text{B}_0$ field. The $\text{T}2$ relaxation, or transverse relaxation, measures the time it takes for the protons to lose their phase coherence and stop precessing in sync.
Different tissues, such as fat and water, have inherently distinct $\text{T}1$ and $\text{T}2$ relaxation times due to their unique molecular compositions and environments. It is the careful, time-dependent measurement of these specific signal decays that forms the basis of the final image.
Mapping the Body: Gradient Fields and Spatial Encoding
The radio signal generated by the relaxing protons, while rich in information about tissue type, is initially uniform and lacks spatial information. To create a recognizable picture, the MRI system must determine the precise location of every signal it receives. This localization is achieved using three sets of secondary coils known as gradient coils, which are arranged along the X, Y, and Z axes.
These gradient coils are engineered to produce temporary, localized magnetic fields that are much weaker than the main $\text{B}_0$ field. When activated, a gradient coil causes the total magnetic field strength to vary linearly across the patient’s body. By manipulating the total field strength, the machine can intentionally alter the Larmor frequency of the protons based on their physical location.
Localization begins with slice selection, where a gradient is applied along the axis perpendicular to the desired imaging plane. This linear change in field strength means that only protons within a specific, narrow band will have the exact Larmor frequency required to absorb the RF pulse. This ensures the signal is only generated from a defined slice of tissue.
To determine the location within that slice, the system uses two separate encoding methods. Frequency encoding applies a gradient across one axis of the slice, causing the signal from protons at different points along that line to have slightly different frequencies. Phase encoding applies a gradient across the perpendicular axis for a brief period, which causes a temporary shift in the phase of the protons’ precession based on their location.
The resulting signal is a complex mixture of frequencies and phases, each component corresponding to a specific location within the slice. By combining the information from the slice selection gradient, the frequency encoding gradient, and the phase encoding gradient, the machine effectively assigns three-dimensional coordinates to every part of the detected signal.
Reading the Data: Image Reconstruction and Tissue Contrast
The spatially encoded signals collected from the gradient process are raw frequency and phase data temporarily stored in a two-dimensional grid called K-space. This data represents the spatial frequencies of the image and is highly organized, but not visually intelligible. The data must undergo a complex mathematical transformation before it can be viewed.
The machine applies a computational technique known as the Fourier Transform to the raw data in K-space. This mathematical tool systematically decomposes the complex signal into its constituent frequencies and phases, translating the abstract frequency-domain data into a recognizable spatial-domain image. This conversion reveals the shape and structure of the internal anatomy.
The final visual appearance, or contrast, of the image is a direct result of the differences in $\text{T}1$ and $\text{T}2$ relaxation times measured earlier. The computer selectively weights the data to emphasize these differences, translating them into varying shades of gray. Tissues appear brighter or darker depending on the specific weighting applied, allowing for clear differentiation between structures.