Target Audience: Clinicians and Physicists who work for neurodegenerative diseases

Background: Proton magnetic resonance spectroscopy (MRS) is a non-invasive method that repeatedly and directly measures the levels of metabolites in the brain. Functional MRS (fMRS) is a technique that measures the metabolic response of the brain during neuronal stimulations using MRI, which is referred to as dynamic MRS (1). To the best of our knowledge, there has been only one study to date that has used fMRS to examine metabolite changes using a cognitive task as the activation paradigm (2).

Purpose: To investigate the metabolite changes in subjects with Alzheimer’s disease (AD), mild cognitive impairment (MCI), and cognitively normal (CN) elderly during a memory task using dynamic MRS at a 3T MRI system.

We included 95 subjects who included: 23 young normal control (YC) (age range: 23 to 49), 24 cognitively normal (CN) elderly (age range: 51 to 81), 24 aMCI (age range: 55 to 82), and 24 mild and probable AD (age range: 56 to 89). All subjects were assessed by the full Seoul Neuropsychological Screening Battery (SNSB) examination, except YC. The stimulation paradigm for the face-name (FN) association task had three conditions, which were fixation (i.e., baseline), novel, and repeat. The session was divided into 8 blocks. The blocks of “fixation (F)-novel (N)-fixation (F)–repeat (R)” were altered twice, but we used different face-name pairs for every block. The fMRS data were measured during the face-name association task from the rectangular VOI, which were carefully placed at the precuneus and posterior cingulate, extending inferiorly into the retrosplenial cortex. fMRS data were obtained with the PRESS single-voxel-spectroscopy (SVS) sequence which had 36 dynamic fMRS data per run (i.e. 36 dynamic scans). Acquired in vivo 1H fMRS data were analyzed using the Linear Combination Model software (LCModel, Version 6.3–0L, Oakville, Ontario, Canada). Statistical analyses were performed to evaluate differences of the metabolite values between the same stimulation conditions, among the three different conditions, and among the four subject groups.

Comparisons within the same condition: There were no significant differences between different blocks in the same condition for the three fixation blocks (DF=2, F<2.97, p>0.068), the two novel blocks (p>0.060) and the two repeat blocks (p>0.5207).

Comparisons among the three conditions: Glx (DF=2, F=3.28, p=0.044) was statistically significantly different among the three conditions. Glx in the novel condition was significantly lower than that in the repeat condition (Bonferroni corrected p=0.0124).

Comparisons among the four subject groups: The result of the one-way ANCOVA test of tNAA among the four subject groups showed that in the fixation condition (DF=3, F=7.673, p<0.001) the tNAA value was significantly lower in the AD group than that in the YC group (Bonferroni corrected p=0.0116), as well as in the CN group (Bonferroni corrected p=0.0001).

Overall, Glx was statistically significant different among the three conditions (i.e., fixation, novel, and repeat). The Glx in the novel condition was significantly lower than that in the repeat condition with a quadratic trend in the order of fixation, novel, and repeat conditions. In the four subject groups, Glx in the novel condition was reduced 1.6%, compared with that in the repeat condition. With an improving sensitivity of the technique, the dynamic measurement of Glx may be used as a biomarker that more closely reflects functional brain impairment in early AD stages, and may also be suitable to demonstrate functional recovery after treatment.Functional MRS data were acquired during the face-name associate task at the 3T MRI system in AD and MCI subjects. Glx was altered during the stimulation conditions, which can be used in neuronal dysfunction for a patient with dementia. The tNAA and tCr were reduced in the patients with AD and tNAA; moreover, they can be used in differentiating subject groups.