Multiple TR approach for direct detection of fast oscillating magnetic fields

Abstract

Various non-invasive brain imaging tools have been developed to directly detect the neuronal activity. However, electroencephalography (EEG) has limitation in spatial resolution and magnetoencephalography (MEG) has poor source localization problem. Functional MRI (fMRI) has been utilized for brain neural activity mapping due to higher spatial resolution, however, fMRI using blood oxygenation level dependent (BOLD) measures secondary hemodynamic response to infer neural activity. Many researchers have studied to diverse MRI techniques for directly detecting the neuronal activity such as stimulus-induced rotary saturation (SIRS) and spin-locked oscillatory excitation (SLOE). Although most of the techniques demonstrated significant detectability in phantom experiment, the possibility of detecting brain neuronal oscillation in vivo is still controversial over a decade.

In this study, we suggested a novel method, multiple TR approach, which extended from previous Two-TR approach. This approach could directly detect weak and fast oscillating magnetic fields without any signal synchronization through phantom experiments. By multi-phase acquisition and Fourier analysis, we demonstrated that the mean SNR at the oscillation frequency on the multiplied frequency spectra was remarkably enhanced with the higher number of TRs under almost the same scan time by the principle of amplifying oscillation frequency component while suppressing other noises. We also showed the targeting frequency on absolute frequency spectrum without a priori oscillation frequency information from multiple TR dataset with different time-to-repeat (TR) and could detect the weak and fast oscillating magnetic field (~1nT) through spatial mapping.