Phosphate Brain Energy Metabolism and Cognition in Alzheimer’s Disease: A Spectroscopy Study Using Whole-Brain Volume-Coil ³¹Phosphorus Magnetic Resonance Spectroscopy at 7Tesla

Introduction: Mitochondrial dysfunction is a neurometabolic hallmark signaling abnormal brain energy metabolism (BEM) targeted as a potential early marker of Alzheimer’s disease (AD). Advanced imaging technologies, such as ³¹phosphorus magnetic resonance spectroscopy (³¹P MRS) at ultra-high-field (UHF) magnetic strength 7T, provide sensitive phosphate-BEM (p-BEM) data with precision. The study’s first goal was to develop a methodology to measure phosphate energy and membrane metabolites simultaneously across the whole-brain using volume-coil ³¹P MRS at 7T in three groups-cognitively normal (CN), amnestic mild cognitive impairment (aMCI), and AD. The second aim investigated whether p-BEM markers in the four brain regions-frontal, temporal, parietal, and occipital were significantly different across the three groups. The final goal examined correspondence between the p-BEM markers and cognition in the three groups. Methods: Forty-one participants (CN = 15, aMCI = 15, AD = 11) were enrolled and completed cognitive assessment and scan. The cognitive domains included executive function (EF), memory, attention, visuospatial skills, and language. The p-BEM markers were measured using energy reserve index (PCr/t-ATP), energy consumption index (intracellular_Pi/t-ATP), metabolic state indicator (intracellular_Pi/PCr), and regulatory co-factors [magnesium (Mg²⁺) and intracellular pH]. Results: Thirteen metabolites were measured simultaneously from the whole brain for all three group with high spectral resolution at UHF. In the aMCI group, a lower p-BEM was observed compared to CN group based on two markers, i.e., energy reserve (p = 0.009) and energy consumption (p = 0.05) indices; whereas in AD a significant increase was found in metabolic stress indicator (p = 0.007) and lower Mg²⁺ (p = 0.004) in the temporal lobes compared to aMCI using ANOVA between group analytical approach. Finally, using a linear mixed model, a significant positive correlation was found between Mg²⁺ and cognitive performance of memory (p = 0.013), EF (p = 0.023), and attention (p = 0.0003) in CN but not in aMCI or AD. Conclusion: To our knowledge, this is the first study to show that it is possible to measure p-BEM in vivo with precision at UHF across the three groups. Moreover, the findings suggest that p-BEM may be compromised in aMCI even before an AD diagnosis, which in future studies should explore to examine whether this energy crisis contributes to some of the earliest neuropathophysiologic changes in AD.