Functional MRI Based on Intermolecular Double-Quantum Coherences (iDQC) at 3 Tesla

Contrast generated by intermolecular double-quantum coherences (iDQC) is a novel approach to fMRI, which is qualitatively and quantitatively different from the standard blood oxygen level dependent (BOLD) contrast. Specifically, it was suggested that the sensitivity of iDQC to local susceptibility gradients can be adjusted on a mesoscopic scale (between 10µm and 1mm) externally by the experimenter. A drawback of iDQC experiments, however, is their inherently poor signal-to-noise ratio leading to substantial signal fluctuations in previous pilot studies. In the current work, parameters of modified CRAZED sequences were carefully optimized to achieve high signal stability in iDQC experiments at 3 T with the standard birdcage headcoil. Initial phantom experiments were utilized to verify that the detected signal was due to iDQC by recording its angular dependence. A four-step phase-cycling scheme was used to filter out specific multiple-quantum coherences. Contributions from unwanted coherence pathways at the center of the k-space were of the noise level. In in vivo experiments (series of 60 repetitions) in healthy volunteers a systematic variation of the repetition time (TR) yielded sufficient signal stability (as compared to ordinary EPI time series) if a relatively long TR 3 5s was used. A two-step phase cycling scheme was used in these experiments. A reason for the suboptimal signal stability at shorter TRs are contributions from additional stimulated echoes, which are not suppressed by the crusher gradients after image acquisition. fMRI studies were performed in 6 subjects with a simple visual paradigm (blocked design). In all subjects, both gradient-recalled echo (GRE) and spin-echo type iDQC sequences were recorded. Additional GRE or SE BOLD measurements were also performed. While no functional contrast was observed with SE-iDQC, functional maps were reproducibly obtained with GRE-iDQC (Fig. 1). These maps showed less activated pixels compared to routine BOLD experiments, however, the averaged functional signal change was >10% (GRE-iDQC) as compared to only 2% in the BOLD experiments. The optimized sequences provide a basis for future experimental investigations into the origin of the functional iDQC contrast.