Spectral dependent signal to noise ratio and edge enhancement : its applications to NMR microscopy and ultrafast imaging

Abstract

The self diffusion of nuclear spins has been suggested edge enhancement in images especially on a microscopic scale. According to previously published works, theory suggests that edge enhancement is caused by the motional narrowing due to the boundaries and spin self-diffusion during the data acquisition period. However, it is found that the edge enhancement depends greatly on the data acquisition mode; therefore, the images obtained are different depending on the pulse sequence employed. For example, excessive attenuation of DC components due to diffusion can result in edge enhancement in the spin echo signal. However, in the case of FID like signal, DC components are preserved while positive high frequency parts are attenuated, thereby resolution is degraded. The new phenomenon observed has been termed selective spectral suppression since the observed edge enhancement results from the selective attenuation of certain frequency components in the nuclear signals due to diffusion dependent signal attenuation for a given pulse sequence. The selective spectral suppression can be applied to signal attenuation during data acquisition if signal attenuation is very large. For example, in EPI, read-out time is relatively very long and it severely suffers from T2 decay during data-acquisition which can make the edge enhancement or the image-resolution degradation according to selective spectral suppression. The paper shows that this selective spectral suppression can be applied to any pulse sequences with long read-out time. In EPI, we can choose the unattenuated frequency components by changing coding method. SNR will not be decrease, but resolution will be degradation by T2 signal decay if DC is unattenuated. However, if high frequency components are unattenuated by T2 effect, SNR will be severely decreased and resolution will be degraded by this poor SNR. The paper have also simulated and revealed the SNR and resolution variation by changing coding method in various ...