Introduction

While functional MRI (fMRI) is used to localize regions of activation, functional MRS (fMRS) provides metabolic response to functional activation. fMRS, using short echo-time (TE) non-edited ¹H-MRS protocols, has detected robust lactate increase in sensory-induced activations^1-3. Given that short TE non-edited ¹H-MRS is susceptible to functional hyperemic effects as well as contamination from signals attributed to lipids and macromolecules, we posited if long TE J-edited ¹H-MRS⁴ can reliably detect lactate changes.

Methods

Ten healthy volunteers (right-handed, 9 male, 1 female, age 36.5±3.8) participated in the experiment that consisted of one fMRI run and 2-4 fMRS runs spanned over 2-4 days (in total 28 runs, 2.8±0.2 runs per subject). The fMRI run consisted of five 97s regulation blocks that were interleaved with five 97s baseline blocks (16.2min). The fMRS run consisted of three 5min finger-tapping blocks interleaved with three 5min fixation blocks (30min). Subjects were asked to perform the visually-cued finger tapping at a rate of 3Hz. Flashing numbers from 1 to 4 for each of four alternated fingers were indicated. Left motor cortex (MC) was identified in the fMRI data using statistical parametric mapping (SPM12). Single voxel for J-edited acquisition protocol was placed around the identified area. The experiments were performed at the MRRC (Yale University) on a 4T Bruker spectrometer using 16 channel transmit-receive head coil and single-shot T2*-weighted FLASH for fMRI (74 scans, TR = 13.135 s, TE = 30 ms, voxel size = 4×4×5 mm), and J-difference editing for fMRS (150 paired spectra, TR = 3330 ms, TE = 144 ms, voxel size = 22×28×22 mm). Prior to fMRS acquisitions, we acquired B0 field map and water spectrum, adjusted basic frequency and shimming globally and locally, and optimized RF power. The acquired spectra were corrected for a basic frequency drift, aligned, phase-corrected, apodized (gaussian 2Hz, exponential 2Hz) and averaged to 30 pairs. On a group-level, individual spectra were centered and aligned to the group average reference NAA peak. The residual BOLD linewidth narrowing was estimated using line-shape differences in NAA peak (0.03±0.02 Hz) between the fixation and tapping conditions and then nulled using exponential linewidth adjustment. The same centering, alignment, and BOLD corrections were applied to the lactate spectra. Lactate and BHB integrals were estimated using LCM quantification. The integrals of the LC modelled lactate and BHB peaks, and noise integral (10.5±0.5 ppm), were normalized to the corresponding NAA integrals (2.01±0.15 ppm).

Results

Because β-hydroxybutyrate (BHB) at 1.19ppm has the same J evolution profiles as lactate at 1.32ppm, our J-edited spectra captured both BHB and lactate reliably in all acquisitions. We found significantly higher normalized lactate integrals during tapping than in fixation conditions (Fig. 1A, 1.32±0.15 ppm, t = 3.2, p < 0.01), and significantly higher lactate integrals estimated based on the LCM quantification of the same spectra (Fig. 1B, t = 2.4, p < .05), and no difference between the corresponding NAA and noise integrals (ps > 0.5). BHB levels did not change with stimulation (t = 1.3, p < 0.22). Narrow linewidth for lactate (9.7±0.3 Hz) and water (5.1±1.7 Hz) peaks were estimated using LC modeling, as wel­­l as relatively low CRLB for lactate (3.7±0.2) and BHB (14.1±1.3) quantifications ensured the high quality of the acquired data and the reliability of the results. The stimulation-induced lactate changes in our J-edited spectra correspond to 11.2±4.6% from basal levels, which is in good agreement with prior non-edited spectra^1-3.

Discussion & Conclusion

Our J-edited fMRS results at 4T showed the physiological modulation of the MC lactate level estimates during standard finger-tapping experiment. In summary, these results confirm that previously detected lactate changes are probably devoid of functional hyperemic effects and are not significantly contaminated by spurious lipids/macromolecules signals.

Acknowledgements

This study was supported by the Swiss National Science Foundation (P300PB_161083) and the National Institute of Health, USA (R01 NS-100106, R01 MH-067528, R21 MH-110862, P30 NS-052519).