MRI Physics Explained: Magnetic Resonance and Spin Echo Sequences

The detected signal S(t) is the vector sum of all transverse spin magnetizations, mathematically resembling an inverse Fourier transform because phase factors depend linearly on spatial coordinates when gradients are applied. Frequency encoding uses gradients to encode spatial information by frequency—the Fourier transform yields 1D projections. Phase encoding applies perpendicular gradients briefly before readout, modifying spin phases for complete 2D reconstruction. K-space is a mathematical representation where each point corresponds to specific 2D sinusoidal patterns. Three rules govern discrete k-space sampling: spacing between frequencies is inversely proportional to image field of view; desired voxel size dictates Nyquist borders; and data point counts match between k-space and image grids. The mathematical connection is Kx = γ∫Gx(t)dt and Ky = γ∫Gy(t)dt.

This section covers the mathematical foundations of MRI image reconstruction. The raw data acquired by MRI is in the frequency domain (k-space or Fourier plane), not in the spatial domain. The Fourier transform, developed by Joseph Fourier in the 19th century, converts frequency domain data into spatial domain images, analogous to converting a musical score into audible music. The center of k-space contains low-frequency components that determine image contrast, while the periphery contains high-frequency components that provide image detail and resolution. MRI acquisitions typically fill k-space line by line over several minutes. To improve resolution, more lines must be acquired, which increases scan time. This fundamental trade-off between resolution and acquisition time is a key limitation of MRI compared to CT scanning.

The main differences between 2D and 3D Fourier Transform lie in data type and properties: 2D uses data for image formation and diffraction, while 3D uses data for light scattering and tomographic formation. 3D has all 2D properties plus an additional rotation property. K-space is a mathematical process in MRI that creates images from raw data through four steps: signal collection from excited tissue using gradient fields, mapping signals into K-space containing frequency components, data acquisition from K-space, and image reconstruction. The relationship between K-space data and image data is defined by Fourier Transform, making it fundamental to MRI image formation.

MRI image formation involves three sequential processes using gradients: slice selection applies a gradient while transmitting an RF pulse at a specific frequency, exciting only protons at that frequency; phase encoding applies a gradient during transverse precession, creating phase differences between protons at different positions; frequency encoding applies a perpendicular gradient to encode spatial information. Combined, these steps generate data stored in k-space—a mathematical representation where center lines contain intensity information and peripheral lines contain detail information. Fourier transformation of k-space reconstructs the final image. Motion during acquisition corrupts the entire image because each k-space line contributes to every pixel. This multi-dimensional encoding enables precise spatial localization, making MRI a true tomographic imaging technique.

In MRI, spatial encoding is achieved by applying magnetic field gradients that cause spins at different locations to precess at different frequencies, effectively winding them into Fourier basis functions; the resulting signal corresponds to Fourier coefficients in k-space, where each point represents how much of a specific oscillatory pattern exists in the object, and proper image reconstruction requires satisfying Nyquist criteria to ensure unique encoding of each spatial location, with non-Fourier encoding methods extending this principle by sampling weighted averages of k-space points rather than single points.

In MRI, image contrast is determined by manipulating the repetition time (TR) and echo time (TE) parameters of the pulse sequence: short TR and short TE create T1-weighted images where contrast reflects T1 relaxation differences (fat appears bright, CSF appears dark); long TR and long TE create T2-weighted images where contrast reflects T2 relaxation differences (CSF appears bright, muscle appears dark); and long TR with short TE creates proton density-weighted images where contrast reflects the actual number of free protons in tissues (fat and fluid appear bright).

In MRI, image contrast is determined by the interplay between T1 recovery time (TR) and T2 decay time (TE): T1-weighted images use short TR (100-1000ms) and short TE (10ms) to emphasize differences in tissue relaxation speeds; T2-weighted images use long TR and long TE to highlight differences in water content and molecular motion; proton density (PD)-weighted images use longish TR (1000-3000ms) and shortish TE (20ms) to correlate signal with proton concentration, with the key constraint that TR must always be longer than TE since the next RF pulse cannot be applied before signal acquisition.

MRI contrast is generated based on differences in T1, T2, and proton density among tissues. Cerebrospinal fluid has both long T1 and long T2 values. T2 values are typically about ten times shorter than T1 values. By selecting appropriate relaxation times, different tissues can be distinguished: tissues with shorter T1 appear whiter during T1-weighted imaging, while tissues with longer T2 appear whiter during T2-weighted imaging.

After the RF pulse is turned off, protons return to equilibrium through relaxation processes. T1 relaxation (spin-lattice relaxation) is the process by which longitudinal magnetization recovers to its equilibrium state as protons release energy to the surrounding molecular environment. T2 relaxation (spin-spin relaxation) is the process by which transverse magnetization decays as protons lose phase coherence with each other. Different tissues have different T1 and T2 relaxation times due to their molecular environments, creating contrast in MRI images. Fat has a short T1 relaxation time and appears hyperintense in T1-weighted images, while water has a long T1 relaxation time and appears hypointense.

MRI equipment consists of main magnet (creates primary field using superconducting magnets with zero electrical resistance at cryogenic temperatures), gradient coils (create linear field variations for spatial encoding), shim coils (improve field uniformity), RF coils (transmit/receive signals), and patient table. Superconducting magnets use liquid helium cooling but newer LTS (low-temperature superconductors) reduce maintenance costs. RF coils include transmit coils (send 90°/180° pulses), receive coils (detect emitted signals), and transmit-receive coils. Gradient coils along X, Y, Z axes enable precise spatial localization. Safety precautions include: claustrophobia assessment, pregnancy verification, contrast dye allergy check, and screening for metallic implants (pacemakers, stents, clips). During exams, patients wear ear protection for gradient coil sounds, remain still, and may hold breath for abdominal scans. Post-exam, patients drink 1L fluids to eliminate contrast dye and remain 30 minutes for observation.

MRI systems consist of three essential components: superconducting magnets generating B0 fields using niobium-titanium wire cooled by liquid helium; RF coils for spin excitation and signal reception; and gradient coils for spatial encoding. Field strengths range from clinical 1.5-3 Tesla to research systems exceeding 11.7 Tesla. Safety protocols mandate assuming magnets are always on, prohibiting metal objects, and confining fringe fields within the room using counter-winding coils. The 5 Gauss safety line must remain within the magnet room to prevent interference with implanted devices. Modern systems use receive-only coil arrays with 128 channels, replacing single-channel designs. Gradient amplifiers deliver up to 1 megawatt of power. Emerging high-temperature superconductors offer reduced weight and cost compared to traditional low-temperature systems requiring liquid helium.

An MRI scanner contains three nested components: (1) The main superconducting magnet creates a strong homogeneous B0 field (about 1-3 Tesla, 10,000-100,000 times Earth's magnetic field); (2) Gradient coils produce smaller linearly varying magnetic fields in x, y, and z directions to localize signals in space and create different imaging contrasts; (3) RF coils transmit radiofrequency pulses to excite spins and receive the resulting NMR signals. The main magnet operates continuously, while gradients and RF coils operate in precise timed sequences. MRI uses non-ionizing radiofrequency radiation, making it safe for repeated use including pediatric patients.

MRI systems consist of: the magnet (superconductor coil generating the main field), gradient coils (three orthogonal coils for spatial encoding), RF coils (for transmitting and receiving signals), cryogenic systems (maintaining superconductor temperature), and computer systems for image processing. The control system coordinates all components to execute imaging sequences, manage gradients, and process received signals. MRI magnets come in three types: permanent magnets (limited to 0.3 Tesla), resistive magnets (also around 0.3 Tesla), and superconducting magnets (1-3 Tesla, most common clinically). Superconducting magnets use niobium-titanium coils cooled to 4.2 Kelvin in liquid helium. Higher field strengths provide better image quality but require expensive helium refilling (15,000-20,000 dollars per load). Superconducting magnets have a quench valve that can release helium gas in emergencies to reduce the magnetic field to zero.

MRI (Magnetic Resonance Imaging) exploits hydrogen nuclei in water molecules (comprising ~70% of the human body) using strong magnetic fields and radio waves to generate detailed cross-sectional images without ionizing radiation. The system comprises five components: superconducting magnet generating the primary field, RF coil for signal transmission, gradient coils for spatial encoding, receiver coils for signal detection, and patient table. MRI produces characteristic knocking sounds from electromagnetic forces on gradient coils. Safety requires comprehensive metal screening: patients must be transported using approved methods (wheelchairs stopped at entrance, stretchers extended outside), and all external metal objects including watches, jewelry, and electronic devices must be removed before entering the scanning room.