Physical Principles of MRI and Hydrogen Spin Mapping

Magnetic Resonance Imaging (MRI) relies on the principles of nuclear magnetic resonance to generate high-contrast images of the human body without ionizing radiation. This article provides an overview of the core physical mechanisms behind MRI, detailing how strong magnetic fields align hydrogen nuclei, how radiofrequency pulses disturb this alignment, and how magnetic gradients spatially encode the resulting signals to map tissue structures in clinical physics.

Nuclear Spin and Hydrogen Abundance

The foundation of clinical MRI is the hydrogen atom, which consists of a single proton. Because the human body is composed largely of water and fat, hydrogen is the most abundant element in biological tissue.

In quantum mechanics, the hydrogen proton possesses an intrinsic property called spin, which gives it angular momentum and turns the proton into a microscopic bar magnet with its own magnetic dipole moment. Outside of an MRI scanner, the orientations of these billions of proton spins are completely random, canceling each other out and resulting in zero net magnetization in the body.

Polarization and Larmor Precession

When a patient is placed inside the bore of an MRI scanner, they are exposed to a powerful, static external magnetic field, denoted as \(B_0\) (typically 1.5 or 3.0 Tesla in clinical settings). In the presence of \(B_0\), the magnetic moments of the hydrogen nuclei align either parallel (low-energy state) or anti-parallel (high-energy state) to the external field.

A slight excess of protons—just a few parts per million—align in the lower-energy parallel state. These unmatched protons combine to form a Net Magnetization Vector (NMV) aligned longitudinally with the \(B_0\) field.

Rather than remaining completely static, the protons wobble, or precess, around the axis of the magnetic field, much like a spinning top wobbles around the axis of gravity. The speed of this precession is governed by the Larmor equation:

\[\omega_0 = \gamma B_0\]

Here, \(\omega_0\) is the Larmor frequency (precessional frequency), \(B_0\) is the strength of the main magnetic field, and \(\gamma\) is the gyromagnetic ratio, a constant unique to each nuclear species. For hydrogen, \(\gamma\) is approximately 42.58 MHz per Tesla.

Radiofrequency Excitation and Resonance

The Net Magnetization Vector cannot be measured while it is aligned parallel to the massive \(B_0\) field. To detect a signal, the NMV must be tipped away from this longitudinal axis (Z-axis) into the transverse plane (XY-plane).

This is achieved through resonance. The scanner emits a secondary oscillating magnetic field in the form of a Radiofrequency (RF) pulse, denoted as \(B_1\). For the hydrogen protons to absorb energy from this pulse, the RF pulse must be transmitted at the exact Larmor frequency of the protons. When this resonance condition is met, two things occur simultaneously:

  1. Flip Angle: Protons absorb energy and transition to the higher-energy state, tipping the NMV away from the Z-axis and into the transverse XY-plane.
  2. Phase Coherence: The randomly precessing protons are forced to precess in sync with one another (in phase).

Signal Detection and Relaxation

Once the RF pulse is switched off, the coherent, precessing transverse magnetization acts like a rotating magnet. According to Faraday’s law of induction, this moving magnetic field induces an alternating electrical current in receiver coils placed around the patient. This induced current is the raw MR signal.

Immediately after the RF pulse ends, the hydrogen nuclei begin returning to their original equilibrium state through two independent relaxation processes, which provide the basis for tissue contrast:

  • T1 Relaxation (Spin-Lattice): Protons release their absorbed energy back into the surrounding molecular environment (the lattice). This causes the longitudinal magnetization along the Z-axis to recover exponentially.
  • T2 Relaxation (Spin-Spin): Protons interact with the tiny magnetic fields of neighboring protons, causing them to precess at slightly different speeds. They rapidly lose phase coherence, causing the transverse magnetization in the XY-plane to decay exponentially.

Because water, fat, muscle, and pathological tissues have different molecular structures, they exhibit different T1 and T2 relaxation times. By adjusting the timing of RF pulses and signal collection, physicists manipulate these relaxation differences to make specific tissues appear bright or dark.

Spatial Encoding with Magnetic Field Gradients

The receiver coil detects the total signal from all excited protons at once, providing no inherent information about where the signals originated. To map the hydrogen spins in three-dimensional space, MRI utilizes three orthogonal gradient coils (\(G_x\), \(G_y\), and \(G_z\)).

These coils superimpose weak, linear variations onto the main \(B_0\) magnetic field. Because the Larmor frequency is directly proportional to magnetic field strength, applying a gradient makes the precessional frequency of hydrogen protons dependent on their physical location within the scanner. Spatial mapping occurs in three steps:

  1. Slice Selection (\(G_z\)): A gradient is applied along the Z-axis while the RF pulse is transmitted. Because the magnetic field now varies from head to toe, only a narrow slice of tissue possesses the exact Larmor frequency matching the RF pulse. Only protons in this specific slice are excited.
  2. Phase Encoding (\(G_y\)): After excitation, a gradient is briefly turned on along the Y-axis. Protons at one end of the gradient temporarily precess faster than those at the other end. When the gradient is turned off, all protons return to the same frequency, but they now have permanent, position-dependent phase shifts along the Y-axis.
  3. Frequency Encoding (\(G_x\)): While the receiver coil is actively recording the signal, a gradient is applied along the X-axis. This causes protons across the X-axis to precess at different frequencies during readout.

K-Space and Image Reconstruction

By repeating this process across multiple excitations with varying phase-encoding gradient strengths, the MRI system collects a grid of complex raw data known as k-space. Every point in k-space contains spatial frequency and phase information from the entire imaged slice.

Finally, a mathematical operation called the Inverse Fast Fourier Transform (IFFT) is applied to the k-space data. The IFFT deconstructs the complex signal back into its individual frequency and phase components, directly translating them into X and Y spatial coordinates. The amplitude of each specific frequency and phase combination determines the brightness of the corresponding pixel, completing the spatial map of hydrogen nuclear spins.