Parallel Imaging

Parallel imaging is an MRI acceleration technique that exploits the spatially varying sensitivity profiles of multi-element receiver coil arrays to reduce the number of k-space samples required for image reconstruction. Rather than relying solely on gradient encoding for spatial localization, parallel imaging treats each coil as an independent sensor with a known spatial response, effectively multiplying the encoding capacity of the system by the number of coils. The reconstruction problem becomes a joint estimation across all coil channels — a distributed sensing architecture that connects MRI to the broader theory of sensor fusion and multi-channel signal acquisition. The canonical algorithms, SENSE and GRAPPA, represent different trade-offs between reconstruction speed and noise amplification, but both embody the same principle: locality in sensor response can substitute for density in spatial sampling.