Room: Exhibit Hall | Forum 6
Purpose: GABA is a critical inhibitory neurotransmitter in human brain, but it is difficult to be measured in 1H-MRS. At present, a J-coupled difference editing method (MEGA-PRESS) is widely used to analyze GABA peaks (3.0ppm) in MRS. Extensive research have suggested that although MEGA-PRESS is able to study the change of GABA signal, it suffers from a number of limitations such as contamination of MM signal, long acquisition time and availability in clinical scanners.
Methods: Unlike physical editing technology such as MEGA-PRESS, this paper proposed to use wavelet transform and prior knowledge to quantitatively study GABA with normal PRESS sequence. Optimum coefficient threshold was selected in the wavelet transform to distinguish signals between Cr and GABA, based on prior knowledge in MRS.
Results: Our proposed method was first validated in analytical simulation and the difference between the input GABA peak and reconstructed GABA peak is less than 3%. Second, the results of 16 human MRS data indicate that the correlation between the proposed method and MEGA-PRESS method is up to 0.932 (P<0.01), and P>0.05 (t=-1.475) in the paired T test. Finally, the proposed method was applied in animal studies with 5 healthy rhesus monkeys and 5 Coriaria lactone-induced status epilepticus monkeys, the calculated GABA levels (GABA/Cr) is of 1.81Â±0.75 and 0.64Â±0.21 for the health group and the pathological group, respectively.
Conclusion: Compared to MEGA-PRESS, the proposed method is able to achieve comparable performance while offering several distinct benefits and great potential to be used in clinical settings.