Aging is a risk factor for cognitive decline and transition to dementia, including Alzheimer’s disease (AD). One well-known age-related change involves altered energy efficiency mediated in part by changes in mitochondrial function. Our goal was to learn the role of mitochondrial mechanisms in cognition in a sample of older adults. To accomplish this, we used phosphorous (³¹P) magnetic resonance spectroscopy (MRS), a unique, non-invasive, and powerful method for examining in vivo mitochondrial function via high energy phosphates and phospholipid metabolism ratios. Given that declines in executive function and recent memory are present in aging, we hypothesized that ³¹P MRS ratios within a focal brain region would map onto cognitive behaviors referable to that brain region. Thus, we predicted better executive function or recent memory performance would be related to greater frontal or temporal ³¹P MRS ratios, respectively. Alternatively, it is possible that this relationship does not exist or may only emerge when neural systems are stressed.

We leveraged baseline neuroimaging data from an ongoing multisite study at the University of Florida and University of Arizona. Participants were cognitively healthy older adults with memory complaints and a first-degree family history of AD [N = 70; mean [M] age [years] = 70.9, standard deviation [SD] = 5.1; M education [years] = 16.2, SD = 2.2; M MoCA = 26.5, SD = 2.4; 61.4% female; 91.5% non-Latinx white]. All participants completed cognitive tasks sensitive to frontal-executive function (Dot Counting, N-Back) and mesial temporal-hippocampal function (Mnemonic Similarity Task [MST], Logical Memory [LM] Stories). To achieve optimal sensitivity, we used a single voxel method to examine ³¹P MRS ratios (bilateral prefrontal and left temporal). Mitochondrial function and energy metabolism were estimated by computing 5 ratios for each voxel: summed adenosine triphosphate to total pooled phosphorous (ATP/TP; momentary energy), ATP to inorganic phosphate (ATP/Pi; energy consumption), phosphocreatine to ATP (PCr/ATP; energy reserve), phosphocreatine to inorganic phosphate (PCr/Pi; oxidative phosphorylation), and phosphomonoesters to phosphodiesters (PME/PDE; cellular membrane turnover rate). All ratios were corrected for voxel size and cerebrospinal fluid fraction.

Separate analyses of variance revealed that lower MST scores were related to higher temporal momentary energy (ATP/TP; 𝛽 = -.26, p = .04) as well as lower temporal energy reserves (PCr/ATP; 𝛽 = .25, p = .04). No other significant relationships emerged.

In contrast to predictions, executive function tasks were unrelated to frontal ³¹P MRS ratios, which may reflect methodological issues (test selection, range restriction), assessment impurity, or some combination. In terms of recent memory, a computerized (MST), but not a traditional (LM), task was differentially related to two out of five temporal ³¹P MRS ratios (ATP/TP, PCr/ATP). In line with the primary function of PCr as an energy replenisher, these latter findings suggest energy stores may compensate for an altered momentary energy state and have a supportive role in recent memory performance in older adults. Future work is warranted to determine whether observed findings reflect meaningful relationships or the influence of unaccounted factors (task complexity, computer versus paper-pencil administration).