MRbildning

MRbildning is the process of generating magnetic resonance images from raw data in MRI. It encompasses the acquisition, reconstruction, and presentation of anatomical and functional information.

The physical principles: A strong static magnetic field aligns protons; RF pulses tip spins; signals re-emerge during relaxation. Spatial encoding uses gradient fields to vary precession frequency and phase, enabling sampling of k-space. The image is determined by tissue properties: proton density, T1, T2, T2*, and diffusion.

Data acquisition and reconstruction: Data collected in k-space; two-dimensional Fourier transform reconstructs spatial image. Modern methods include parallel imaging (SENSE, GRAPPA) and compressed sensing to improve speed. Multi-coil arrays increase SNR and enable acceleration.

Image contrast and sequences: Weighted images based on T1, T2, proton density; sequences like spin-echo, gradient-echo, diffusion-weighted imaging, perfusion; contrast agents increase information content.

Quality, limitations, artifacts: Signal-to-noise ratio, resolution trade-offs; motion, susceptibility, Gibbs, aliasing artifacts; artifact mitigation.

Applications and significance: Central to neuroimaging, musculoskeletal, cardiovascular, oncology; MRbildning is critical for diagnosis, planning, and follow-up.