Back

MRI: the Bloch Equations, the FID and k-Space Imaging

An MRI (magnetic resonance imaging) physics playground. The Bloch equations are solved analytically in the rotating frame: after a 90-degree pulse the magnetization precesses while the transverse component decays with $T_2$ and the longitudinal component recovers to $M_0$ with $T_1$, tracing a spiral on the Bloch sphere and producing the free induction decay whose Fourier transform is a Lorentzian. A brain-like phantom is imaged with the spin-echo equation $S \sim \rho (1 - e^{-TR/T_1}) e^{-TE/T_2}$ or the spoiled gradient-echo Ernst-angle equation; the image is transformed to k-space and reconstructed by a 2D inverse Fourier transform, and discarding the outer k-space lines blurs it. Panel A is the Bloch sphere, Panel B the FID and spectrum, Panel C the image and its k-space. The magnetisation magnitude is conserved under pure precession, $T_1$ and $T_2$ relaxation follow the Bloch laws, the spin-echo signal reaches its expected limits at the Ernst angle, and discarding the outer k-space lines blurs the image.

Figure 1. The Bloch equations: a magnetization vector precesses and relaxes (T2 transverse decay, T1 longitudinal recovery), giving the free induction decay and its Lorentzian spectrum; a brain phantom is imaged by a spin-echo or gradient-echo sequence and reconstructed from k-space by the 2D inverse Fourier transform, with the T1, T2 and proton-density contrast set by TR and TE. Method: analytic Bloch solution, the SE/GRE signal equations and a radix-2 2D FFT; Canvas2D, deterministic.
weighting
sequence
phantom
k-space kept (%)100

WHAT TO TRY

  • Vary each control and watch the rail readouts respond.
  • Compare the diagnostic plot against the live scene.