Cognitive function and brain structure correlates in individuals with subjective cognitive decline: Voxel‐based morphometry results from a lifestyle intervention to prevent Alzheimer’s disease (PENSA study)

Abstract Background Interventions based on the modification of several lifestyle factors have provided promising results to maintain cognitive function in late/middle‐aged individuals at higher risk of developing Alzheimer’s disease (AD). Framed under the WW‐FINGERS network, the PENSA clinical trial is focused on the prevention of cognitive decline in individuals with subjective cognitive decline who are also APOEε4 carriers. It consists of a 12‐month multimodal lifestyle intervention supplemented with EGCG, a natural flavonoid from green tea. In this study, we looked for the brain structure correlates of cognitive performance for PENSA participants at baseline. Methods One hundred thirty‐one (N = 131) participants of the PENSA study at baseline [ Table 1 ] with valid T1w, T2w and FLAIR MRI scans were included. They underwent a neuropsychological assessment battery consisting of the modified Alzheimer Disease Cooperative Study ‐ Preclinical Alzheimer Cognitive Composite (ADCS‐PACC‐Plus‐exe) that includes MMSE, Free and Cued Selective Reminding Test (FCSRT), Delayed Recall Logical Memory (DRLM), Five Digits Test, Digit Symbol Substitution Test, and Stroop. Additional cognitive measurements were obtained from semantic verbal fluency (SVF), Boston Naming Test, digit span subtest, visual puzzle subtest (VPT) and the Montreal Cognitive Assessment. FLAIR images were used to segment white matter hyperintensities and perform a lesion filling procedure in T1w and T2w images. These two later modalities were then entered into a multimodal segmentation in the SPM12 voxel‐based morphometry pipeline. Associations between test scores and grey matter volumes (GMv) were sought (p<0.001) including age, sex, education, and total intracranial volume as confounders. Results A positive association was found (p uncorr <0.001; k>200) between GMv and ADCS‐PACC‐Plus‐exe composite [ Figure 1 ] in the bilateral hippocampus, amygdala, entorhinal and parahippocampal gyri. Additionally, positive correlations (p uncorr <0.001; k>200) were found between GMv and FCSRT, DRLM, VPT and SVF tests albeit in less consistent regional patterns. No associations were found between GMv and the remaining tests. Conclusion In a sample at high risk of developing AD, the ADCS‐PACC‐Plus‐exe composite is associated with GMv in brain areas known to be affected in AD. These results suggest that the ADCS‐PACC‐Plus‐exe composite might be sensitive to detect subtle cognitive changes in cognitively unimpaired at‐risk individuals.