System and method of performing magnetic resonance spectroscopic imaging

A magnetic resonance localized spectroscopy and imaging system, comprising:

• a magnetic resonance imaging scanner;
• and
• a data processing system configured to communicate with said magnetic resonance imaging scanner to receive magnetic resonance spectroscopy signals of an object,
• wherein said data processing system is configured to:
• receive a magnetic resonance image of said object,
• display said magnetic resonance image to permit identification of a plurality C of compartments that include all magnetic resonance spectroscopy signal sources in said object and at least one compartment of interest,
• segment in at least one spatial dimension said magnetic resonance image of said object into said C compartments,
• receive magnetic resonance spectroscopy signals from said object corresponding to said magnetic resonance image by applying a plurality M′ of phase encodings in at least one spatial dimension, wherein M′≥C, and wherein said plurality M′ of phase encodings are further chosen from a central portion of a k-space representing said object to optimize at least one of a signal-to-noise ratio or a spatial selection of the at least one compartment of interest,
• calculate a spatially localized magnetic resonance chemical shift spectrum from the at least one compartment of interest using a linear algebraic method, and
• provide a spatially localized magnetic resonance spectrum substantially equal to a spatial average of the magnetic resonance chemical shift spectra from the at least one compartment of interest,
• provide an improved signal-to-noise ratio in at least one compartment of interest, as compared to a conventional chemical shift imaging process in which signals from all volume elements that contribute to the compartment are co-added, wherein said improvement is substantially proportionate to a square root of a ratio of a volume of the at least one compartment of interest to the volume resolution of said chemical shift imaging process.