Effect of phase encoding steps on 1D chemical shift imaging of lactate during brain activation

Lactate is a unique indicator of brain activation and is detectable in vivo by proton magnetic resonance spectroscopy. Previous brain activation studies have been confined to single-voxel localization of lactate. To extend this work to 1D chemical shift imaging, computer simulation, test-object and human studies were conducted to examine tradeoffs among the number of phase encoding steps, signal-to-noise ratio (SNR) and resolution. An iterative algorithm was developed to reduce truncation artifacts arising from a limited number of phase encoding steps. The results indicate that the resolution and SNR attained with 8 phase encoding steps and 16 averages per step after applying the truncation reduction algorithm are approximately equal to those attained with 32 encoding steps and 4 averages per step. Thus, 32 steps are preferred since contamination is minimized with increasing steps. A human study with 32 phase encodings showed a factor of two increase in lactate in the right auditory cortex during left-ear tonal stimulation