Effects of menopausal estrogen loss on the functional brain activity underlying associative memory

Abstract

Background Ovarian removal via bilateral salpingo‐oophorectomy (BSO) prior to spontaneous menopause (SM) is related to increased Alzheimer’s disease (AD) risk (Rocca et al., 2007). Associative learning deficits are considered the earliest AD symptoms, heralding preclinical AD (Fowler et al., 2002). Performance and brain activation during a face‐name associative memory task differ based on reproductive stage and are linked to fluctuating levels of 17β‐estradiol (E2; Rentz et al., 2017). We hypothesized that BSO would affect memory and functional brain activity during associative encoding. Method Middle‐age women underwent functional magnetic resonance imaging (fMRI) while completing a face‐name associative memory task (Sperling et al., 2003). Recognition performance and brain activation during face‐name pair encoding were assessed in women with BSO taking E2‐based hormone therapy (BSO+E2; n=10; mean age=46), women with BSO taking no hormone therapy (BSO; n=12; mean age=49), age‐matched women with intact ovaries (AMC; n=14; mean age=44), and older women in spontaneous menopause (SM; n=15; mean age=56). Result No group differences in face‐name pair recognition accuracy were found. Multivariate partial least squares analyses (McIntosh & Lobaugh, 2004) revealed significant differences in brain‐behaviour correlations between BSO and SM groups. Accuracy in the SM group correlated positively with activation of the hippocampus, medial temporal, parietal, and frontal lobes, while accuracy in the BSO group correlated negatively with activation of these regions (see Figure). Region‐of‐interest (ROI) analyses revealed that functional activity in the right superior frontal lobe correlated positively with E2 levels in the BSO+E2 group (r=0.83, p=0.01), and negatively with E2 levels in the BSO group (r=‐0.66, p=0.03). Conclusion Activation of distinct brain regions underlying associative memory depends on E2 and age. The BSO group, who experienced menopause approximately 10 years earlier than the SM group, showed significantly different patterns of brain activation compared to the SM group, ultimately to achieve similar recognition accuracy. Importantly, there were no significant differences in performance, indicating that brain changes may precede associative memory changes, and that E2 depletion could play an important role in brain activity underlying women’s associative memory.

Cardiometabolic health across menopausal years is linked to white matter hyperintensities up to a decade later

“…Participants with missing values (“ Not a Number (NaN) ,” “ prefer not to answer ,” “ do not know ”), were excluded (missing datapoints 520 for demographic factors, 2,974 for WMH volume, 8,448 for cardiometabolic risk factors, 4,562 for menopausal status, and 101 for hysterectomy/bilateral oophorectomy, with 10% of all participants having missing values for more than 2 variables). 1,015 participants who had undergone a hysterectomy and/or bilateral oophorectomy were excluded in order to focus the study on variation in natural menopause, as surgical menopause may involve independent risks for cardiometabolic diseases ( 4 , 44 ), as well as brain ageing and dementia ( 45 – 47 ). 17 participants were excluded due to implausible menopause status data or age at menopause outliers (see Section 2.2 for details).…”