← MRI subtopic / Sim: k-space Visualizer

k-space Visualizer

k-space is the spatial-frequency representation of an MR image. Its center encodes overall contrast (low frequencies); its periphery encodes edges and fine detail (high frequencies). Toggle which region contributes.

k-space (sampled region)

Reconstructed image

Center only — blurred, low-contrast shape visible, no fine detail. The overall "what is this tissue" information lives here.

Periphery only — edges and fine structure, but no overall contrast. Used in partial-Fourier acceleration and understanding Gibbs ringing.

Why this matters

In a spin-echo or gradient-echo acquisition, k-space is filled line-by-line as phase encoding steps through. Any motion or inconsistency during the acquisition of center lines produces large contrast artifacts (ghosts, shading); inconsistency during peripheral lines produces edge artifacts. This is why view-ordering strategies (low-to-high vs centric) matter. Partial-Fourier and half-scan methods exploit the Hermitian symmetry of k-space to skip half of it — at the cost of SNR.