Triple-quantum filtered NMR imaging of sodium in the human brain

Abstract

In the past multiple-quantum filtered imaging of biexponential relaxation sodium-23 nuclei in the human brain has been limited by low signal to noise ratios; this thesis demonstrates that such imaging is feasible when using a modified gradient-selected triple-quantum filter at a repetition time which maximizes the signal to noise ratio. Nuclear magnetic resonance imaging of biexponential relaxation sodium-23 (23Na) nuclei in the human brain may be useful for detecting ischemia, cancer, and pathophysiology related to manic-depression. Conventional single-quantum NMR imaging of in vivo biexponential relaxation 23Na signals is complicated by the presence of single-exponential relaxation 23Na signals. Multiple-quantum filters may be used to selectively image biexponential relaxation 23Na signals since these filters suppress single-exponential relaxation 23Na signals. In this thesis, the typical repetition times (200--300 ms) used for in vivo multiple-quantum filtered 23Na experiments are shown to be approximately 5 times greater than the optimal repetition time which maximizes multiple-quantum filtered SNR. Calculations and experimental verification show that the gradient-selected triple-quantum (GS3Q) filtered SNR for 23Na in a 4% agarose gel increases by a factor of two as the repetition time decreases from 300 ms to 55 ms. The measured relaxation times of the 23Na in the 4% agarose gel were similar to in vivo 23Na relaxation times.