Motivation: Wireless coil arrays are a major long term goal for MRI engineering. However, data rates and spectral bandwidth are more limited than in cabled arrays. Goal(s): We wish to assess feasible amounts by which k-space data can be compressed prior to transmission. Approach: We use prior data sets and explore a range of lossless compression algorithms, including lossess JPEG algorithms, Fibonacci and Huffman variants. These are compared against an estimated Shannon entropy lower bound. Results: Because MRI k-space data has very few large integers, 3x compressions are typical. Even the most basic lossless algorithms come within about 1 bit of the Shannon limit. Impact: MRI k-space data appears compressible by 3x before data transmission. This means band-limited wireless coil arrays (WiFi or ultra-wideband) could support image 3x higher throughput, or operate with significantly slower link rates. This would be crucial for robust performance.
Zhang et al. (2025) studied this question.