要旨
Introduction:
Episodic memory dysfunction is a clinical hallmark of Alzheimer's disease (AD) [1], and pathological change affecting the posterior cingulate and precuneus region (PCP) is considered to significantly contribute to memory decline [2]. Magnetic resonance spectroscopy (MRS) has been proven valid for non-invasive investigation of various brain metabolites in the PCP. In the current study, we used MR spectroscopic imaging (MRSI) at very high field strength of 7 Tesla to study metabolic correlates of episodic memory in grey matter and white matter sub-regions of the PCP in non-demented elderly subjects at risk for AD due to their high age.
Methods:
Thirty elderly subjects (mean age 70±6 years, 40 females, mean Mini Mental State Examination 29.4±4) received standardized cognitive testing and were administered MRI at 7 Tesla including a high resolution free induction decay (FID) MRSI (nominal voxel size 0.9 x 0.9 x 1.5 mm) and a T1-weighted MP-RAGE structural MRI (resolution 0.8 x 0.8 x 0.9 mm) for anatomical reference of the MRSI and definition of tissue classes. The PCP region of interest was anatomically defined using a standard reference atlas [3,4], and SPM12 was used for segmentation of the structural MRI into grey matter, white matter, and CSF [5]. Averaged MR spectra from grey matter and white matter voxels of each subject were analyzed using LC Model [6]. All MRSI metabolites were normalized to creatine (Cre). Episodic memory performance was assessed using the Verbal Learning Memory Test delayed recall (VLMT), and the study sample was dichotomized into high and low-performers of VLMT using a median split.
Results:
After quality control (Cramér Rao lower bounds < 0.2), metabolite-to-creatine ratios from the PCP grey matter and white matter were measurable for the neuronal biomarker N-acetylaspartate (NAA), the glial metabolite myo inositol (mI), total choline (tCho), glutamate-glutamine (Glx), and gamma-amino butyric acid (GABA). VLMT low-performers had significantly higher levels of NAA in the PCP grey matter compared with the VLMT high-performers (Wilcoxon ranksum test p=0.01). Similarly, VLMT low-performers had higher levels of mI (p=0.046), Glx (p=0.044), and GABA (p=0.035) in the PCP grey matter. No such effects were observed in the PCP white matter. There were inverse correlations of VLMT performance with mI (Spearman Rho=-0.42, p=0.02), Glx (Spearman Rho=-0.39, p=0.03), and GABA (Spearman Rho=-0.44, p=0.02).
Conclusions:
To our knowledge, this is the first study utilizing MRSI at very high field-strength of 7 Tesla for investigating metabolic correlates of memory dysfunction in an elderly, non-demented study population at risk for AD. Our data reveal grey matter-specific metabolic correlates of low episodic memory performance, as reflected by increased levels of NAA, mI, Glx and GABA. Moreover, our findings indicate that increased grey matter metabolism may serve as a compensatory mechanism for neuronal alterations [7], located in the memory processing systems of the elderly. Alternatively, increased grey matter metabolism in the PCP grey matter could reflect a more causal phenomenon of age-related memory dysfunction [8]. Further studies are warranted to evaluate the potential of altered metabolism in the PCP grey matter as an indicator of AD-risk and response-marker for preclinical intervention.
