Combined fMRI-MRS acquires simultaneous glutamate and BOLD-fMRI signals in the human brain

Abstract

Combined fMRI-MRS is a novel method to non-invasively investigate functional activation in the human brain using simultaneous acquisition of hemodynamic and neurochemical measures. The aim of the current study was to quantify neural activity using combined fMRI-MRS at 7T. BOLD-fMRI and semi-LASER localization MRS data were acquired from the visual cortex of 13 participants during short blocks (64s) of flickering checkerboards. We demonstrate a correlation between glutamate and BOLD-fMRI time courses (R=0.381, p=0.031). In addition, we show increases in BOLD-fMRI (1.43±0.17%) and glutamate concentrations (0.15±0.05 I.U., ~2%) during visual stimulation. In contrast, we observed no change in glutamate concentrations in resting state MRS data during sham stimulation periods. Spectral line width changes generated by the BOLD-response were corrected using line broadening. In summary, our results establish the feasibility of concurrent measurements of BOLD-fMRI and neurochemicals using a novel combined fMRI-MRS sequence. Our findings strengthen the link between glutamate and functional activity in the human brain by demonstrating a significant correlation of BOLD-fMRI and glutamate over time, and by showing ~2% glutamate increases during 64s of visual stimulation. Our tool may become useful for studies characterizing functional dynamics between neurochemicals and hemodynamics in health and disease.

Keywords: BOLD fMRI; Functional Spectroscopy; Glutamate; Neurochemistry; Visual cortex.

Graphical abstract

Fig. 1

(a) Diagram shows the MR sequence consisting of a 3D BOLD echo planar imaging (3D-EPI) and semi-LASER MR-spectroscopy sequence in the same TR, with diagram below illustrating the 2×2×2 cm MRS voxel in the occipital lobe (blue square) and the EPI slice coverage (red outline), overlaid on a high resolution anatomical image. (b) Experimental design showing stimulus conditions, consisting of a baseline (black screen, 64 s) and a flashing checkerboard (64 s). Each participant took part in a single functional MRS visual stimulation experiment, consisting of four cycles of baseline and stimulation blocks. Subjects performed a fixation task throughout the 8 min 24 s experiment.

Fig. 2

(a) Red heat map shows thresholded group activation generated by the comparison of stimulation vs. baseline (z-stat >2.3, cluster-corrected at p<0.05) overlaid on sagittal (left) and horizontal (right) standard brain sections. Blue colour map shows region where >50% of the group MRS-voxel overlapped (N=13). Combined MRS data during stimulation (b) and baseline (c) conditions. Each baseline and stimulation spectra are composed of the sum of 13 subjects×4 repetitions×14 spectra/block=728 total spectra/condition. (d) Difference spectrum, created by subtracting baseline from stimulation with LCModel fit. (e) Residual of LCModel fit of the difference spectrum. (f) LCModel fit of glutamate on the difference spectrum. Pale vertical gray lines indicate the singlet positions Glutamate and singlet of N-acetylaspartate (NAA) at 2.01 ppm. Spectra in b-c are plotted to the same y-axis scale; d-f have been plotted on an expanded vertical scale.

Fig. 3

(a) Glutamate response over time, as a percentage of average baseline glutamate concentration. Data represent the responses to a single 512 s stimulation experiment collected per subject, averaged across subjects (N=13, ±s.e.m). Pale red line and filled circles indicates transient response in the first baseline period. BOLD-fMRI response is shown in blue (N=13, ±s.e.m). R- and p-values in bold font reflect Pearson's Correlation using data from the full time course. Italicized R- and p-values are Pearson's Correlation using data after the first baseline period. The baseline periods are indicated as white and stimulus periods as gray shaded background. (b) Bar graph shows ΔGlu (Stim-Base). (c) Bar graph shows resting state ΔGlu analyzed as if stimulation had been delivered (‘Sham’). During the resting state scan, participants kept their eyes shut and no stimulation was delivered (N=13, ±s.e.m). P-value indicates result of a one-sample T-test. NS=not significant, *=p <0.05.