Ju Suyeon, Horien Corey, Shen Xilin, Abuwarda Hamid, Trainer Anne, Constable R Todd, Fredericks Carolyn A
Department of Neurology, Yale School of Medicine, New Haven, CT, United States.
Interdepartmental Neuroscience Program, Yale School of Medicine, New Haven, CT, United States.
Front Dement. 2023 Mar 20;2:1126016. doi: 10.3389/frdem.2023.1126016. eCollection 2023.
Alzheimer's disease (AD) takes a more aggressive course in women than men, with higher prevalence and faster progression. Amnestic AD specifically targets the default mode network (DMN), which subserves short-term memory; past research shows relative hyperconnectivity in the posterior DMN in aging women. Higher reliance on this network during memory tasks may contribute to women's elevated AD risk. Here, we applied connectome-based predictive modeling (CPM), a robust linear machine-learning approach, to the Lifespan Human Connectome Project-Aging (HCP-A) dataset ( = 579). We sought to characterize sex-based predictors of memory performance in aging, with particular attention to the DMN. Models were evaluated using cross-validation both across the whole group and for each sex separately. Whole-group models predicted short-term memory performance with accuracies ranging from ρ = 0.21-0.45. The best-performing models were derived from an associative memory task-based scan. Sex-specific models revealed significant differences in connectome-based predictors for men and women. DMN activity contributed more to predicted memory scores in women, while within- and between- visual network activity contributed more to predicted memory scores in men. While men showed more segregation of visual networks, women showed more segregation of the DMN. We demonstrate that women and men recruit different circuitry when performing memory tasks, with women relying more on intra-DMN activity and men relying more on visual circuitry. These findings are consistent with the hypothesis that women draw more heavily upon the DMN for recollective memory, potentially contributing to women's elevated risk of AD.
阿尔茨海默病（AD）在女性中的发病过程比男性更具侵袭性，患病率更高且进展更快。遗忘型AD特别针对默认模式网络（DMN），该网络负责短期记忆；过去的研究表明，老年女性的后扣带回默认模式网络存在相对高连接性。在记忆任务中对该网络的更高依赖可能导致女性患AD的风险升高。在此，我们将基于连接组的预测建模（CPM）（一种强大的线性机器学习方法）应用于生命周期人类连接组计划-衰老（HCP-A）数据集（n = 579）。我们试图确定衰老过程中基于性别的记忆表现预测因素，尤其关注默认模式网络。使用交叉验证在整个组以及分别针对每种性别对模型进行评估。全组模型预测短期记忆表现的准确率范围为ρ = 0.21 - 0.45。表现最佳的模型来自基于联想记忆任务的扫描。性别特异性模型揭示了男性和女性在基于连接组的预测因素方面存在显著差异。默认模式网络活动对女性预测记忆分数的贡献更大，而视觉网络内和视觉网络间的活动对男性预测记忆分数的贡献更大。虽然男性的视觉网络分离度更高，但女性的默认模式网络分离度更高。我们证明，男性和女性在执行记忆任务时使用不同的神经回路，女性更多地依赖默认模式网络内的活动，而男性更多地依赖视觉回路。这些发现与以下假设一致，即女性在回忆记忆时更多地依赖默认模式网络，这可能导致女性患AD的风险升高。
