Neurometabolic timecourse of healthy aging

Abstract

Purpose: The neurometabolic timecourse of healthy aging is not well-established, in part due to diversity of quantification methodology. In this study, a large structured cross-sectional cohort of male and female subjects throughout adulthood was recruited to investigate neurometabolic changes as a function of age, using consensus-recommended magnetic resonance spectroscopy quantification methods. Methods: 102 healthy volunteers, with approximately equal numbers of male and female participants in each decade of age from the 20s, 30s, 40s, 50s, and 60s, were recruited with IRB approval. MR spectroscopic data were acquired on a 3T MRI scanner. Metabolite spectra were acquired using PRESS localization (TE = 30 ms; 96 transients) in the centrum semiovale (CSO) and posterior cingulate cortex (PCC). Water-suppressed spectra were modeled using the Osprey algorithm, employing a basis set of 18 simulated metabolite basis functions and a cohort-mean measured macromolecular spectrum. Pearson correlations were conducted to assess relationships between metabolite concentrations and age for each voxel; paired t-tests were run to determine whether metabolite concentrations differed between the PCC and CSO. Results: Two datasets were excluded (1 ethanol; 1 unacceptably large lipid signal). Statistically significant age-by-metabolite correlations were seen for tCr (R2=0.36; p<0.001), tCho (R2=0.11; p<0.001), sI (R2=0.11; p=0.004), and mI (R2=0.10; p<0.001) in the CSO, and tCr (R2=0.15; p<0.001), tCho (R2=0.11; p<0.001), and GABA (R2=0.11; p=0.003) in the PCC. No significant correlations were seen between tNAA, NAA, GSH, Glx or Glu and age in either region (all p>0.25). Levels of sI were significantly higher in the PCC in female subjects (p<0.001) than in male subjects. There was a significant positive correlation between linewidth and age. Conclusion: The results indicated age correlations for tCho, tCr, sI, and mI in CSO and for tCr, tCho and GABA in PCC, while no age-related changes were found for NAA, tNAA, GSH, Glu or Glx. Our results provide a normative foundation for future work investigating the neurometabolic time course of healthy aging using MRS.

Cohort-mean measured macromolecules lead to more robust linear-combination modeling than parameterized and subject-specific ones

“…contamination). Finally, since this study focuses on MM modeling strategies, we did not include a detailed interpretation and discussion of associations between metabolite levels, age and sex, which can be found elsewhere 10 .…”

“…The data in this study are of exceptionally high spectral quality (high SNR, narrow linewidth, and no lipid contamination) and our observations and conclusions may be substantially differed for spectra of lower spectral quality (low SNR, broader linewidth, and lipid contamination). Finally, since this study focuses on MM modeling strategies, we did not include a detailed interpretation and discussion of associations between metabolite levels, age and sex, which can be found elsewhere 10 .…”

“…However, it enforces fixed ratios between the MM signals and is unable to capture any subject-specific variation of the MM signals. This approach has previously been employed in the same cohort 10 and closely corresponds to the MRS data analysis consensus recommendations 4,5 .…”

“…Interestingly, it has not been investigated whether including subject-specific measured MMs (mMM) in the LCM has advantages over cohort-mean mMMs, possibly because prohibitively long acquisition times have disincentivized such a study, and current MRS analysis software does not provide easily accessible workflows to do so. A recent large-scale lifespan cohort 9,10 now provides sufficient publicly available data to compare the two strategies. This dataset further allows investigation of other important issues of MM modeling in short-TE spectra to be investigated: Parameterized vs. measured MM: Only a few low-n studies have systematically compared metabolite estimation between parameterized MM (pMM) and mMM modeling, or studied updated MM parameterizations 11–13 . Other studies focused on individual differences between measured MM spectra, but did not investigate the impact on the metabolite estimation 9,14 . LCM algorithms apply modeling parameters to the basis functions, including a single Gaussian linebroadening parameter to account for B 0 inhomogeneities in the MRS volume 7,15,16 .…”

“…We sought to substantiate expert-consensus recommendations on MM handling 4,5 and provide readily usable workflows for all strategies in our analysis pipeline ‘Osprey’ 15 . To this end, different MM modeling strategies were implemented and applied to a large lifespan dataset from healthy volunteers 9,10 . Macromolecular modeling strategies include cohort-mean and subject-specific measured macromolecules, the most commonly used Gaussian parametrizations, and the use of a single or two separate Gaussian linebroadening parameters.…”