To investigate, through density-matrix simulations, and phantom and in vivo experiments, the novel approach of Hadamard-encoded spectral editing.J-difference editing allows quantification of low-concentration metabolites, including N-acetyl aspartyl glutamate (NAAG) [1]. A disadvantage of this approach is that one experiment is required per edited molecule, limiting the number (and SNR) of edited experiments that can be incorporated into research studies. In this abstract, it is demonstrated that by using Hadamard-encoded combinations of editing pulse frequencies, it is possible to design a multiplexed experiment that simultaneously edits more than one metabolite (‘HERMES’). Compared to sequential acquisitions, SNR is improved since the full scan time is used for each metabolite. As an example of the method, a scheme for separately detecting NAA and NAAG through a single edited experiment is demonstrated.J-difference editing involves the acquisition of two experiments, which differ in their manipulation of a target spin system. Frequency-selective inversion pulses are applied to a target spin in one experiment (ON) to refocus the evolution of a target molecule’s J-coupling. In the other experiment (OFF), editing pulses are typically applied off-resonance, and the coupling is allowed to evolve. The difference spectrum then only contains those signals affected by the ON pulse. HERMES acquires all combinations of OFF and ON for multiple independent target molecules. For two targets, four experiments are needed: A (ON,ON); B (ON,OFF); C (OFF,ON); and D (OFF,OFF). The combination A+B-C-D gives the difference-edited spectrum for one target molecule, while A-B+C-D gives the other. By acquiring experiments simultaneously, Hadamard-editing of $$${N}$$$ species gives a theoretical $$$\sqrt{N}$$$ benefit in SNR.

Simulations

Density-matrix simulations were performed at a simulated B0=3T with sinc-Gaussian editing pulses, ‘GTST’ refocusing pulses (bandwidth 1300 Hz [2]), 2-kHz spectral width, 2048 points, 8-Hz exponential filter, TE 150 ms, fourfold zero-filling. HERMES NAA/NAAG editing was simulated as shown in Figure 1. Experiment A (ON,ON) was performed with 10-ms editing pulses, applied at 4.5 ppm to invert both NAA spins at 4.38 ppm and NAAG spins at 4.62 ppm. Experiments B (ON,OFF) and C (OFF,ON) were performed with more selective, 45-ms editing pulses to invert 4.38-ppm NAA and 4.62-ppm NAAG spins, respectively. To suppress residual NAA signal in the NAAG spectrum, Experiment D applied 45-ms editing pulses at 4.14 ppm, so C and D are symmetrical about NAA at 4.38 ppm. NAA and NAAG were simulated independently, and spectra were reconstructed to generate separate NAA and NAAG spectra (as in Figure 1); only peaks at ~2.6 ppm were plotted.

Phantom

HERMES NAA/NAAG experiments were performed on a 10 mM NAA phantom using a Philips Achieva 3T scanner. Scan parameters matched the simulations above, with TR 2.2 s, and a 20-Hz exponential filter. Spectra were reconstructed to generate NAA and ‘NAAG’ (residual NAA, since the phantom did not contain NAAG) spectra.

In vivo

HERMES NAA/NAAG experiments were performed in 8 healthy adults in a 5x3x3 cm³ voxel in the centrum semiovale using VAPOR water suppression, 384 transients, 5-Hz exponential filter, and prospective frequency drift correction based on interleaved water reference transients acquired every 24 dynamic scans. Individual transients were frequency-corrected based on the frequency of the NAA methyl peak using the ‘Gannet’ program [3], and reconstructed to give separate NAA and NAAG spectra. Spectra were fitted to a simulated model to extract a scaling factor indicative of the metabolites’ relative concentrations.

As shown in Figure 2, simulations show excellent separation of NAA and NAAG, with minimal residual NAA in the NAAG reconstruction. Phantom and in vivo data show excellent agreement with the simulations. Figure 3a shows that the lineshape and intensities of the in vivo NAA and NAAG signals are consistent between participants. Figure 3b shows the normalized fitted amplitudes (mean NAA = 1) having a ratio of NAA:NAAG concentration of 4.5:1.This abstract demonstrates that the simultaneous and separable detection of two metabolites using Hadamard Encoding and Reconstruction of Mega-Edited Spectroscopy, HERMES, is possible at 3T with minimal metabolite crosstalk in the reconstruction of two otherwise overlapping metabolites. There is excellent agreement between simulations and experiments in the NAA and NAAG reconstruction lineshapes - the differing lineshapes of the reconstructed NAA and NAAG spectra remain apparent in vivo. The acquired in vivo NAA:NAAG ratio of 4.5:1, is in good agreement with literature values [4]. These data demonstrate the capability of HERMES to separate NAA and NAAG within a single acquisition. HERMES can be applied to other overlapping edited species e.g. GABA/macromolecules, and, in principle, extended to more than two compounds by extending the Hadamard encoding matrix.