The relationship between depressive symptoms and cognitive function in Alzheimer's disease: The mediating effect of amygdala functional connectivity and radiomic features

Yang Du,Jie Yu,Manhua Liu,Qi Qiu,Yuan Fang,Lu Zhao,Wenjing Wei,Jinghua Wang,Xiang Lin,Feng Yan,Xia Li
DOI: https://doi.org/10.1016/j.jad.2023.02.129
2023-06-01
Abstract:Background: Depressive symptoms are common in Alzheimer's disease (AD) and are associated with cognitive function. Amygdala functional connectivity (FC) and radiomic features related to depression and cognition. However, studies have yet to explore the neural mechanisms underlying these associations. Methods: We enrolled eighty-two AD patients with depressive symptoms (ADD) and 85 healthy controls (HCs) in this study. We compared amygdala FC using the seed-based approach between ADD patients and HCs. The least absolute shrinkage and selection operator (LASSO) was used to select amygdala radiomic features. A support vector machine (SVM) model was constructed based on the identified radiomic features to distinguish ADD from HCs. We used mediation analyses to explore the mediating effects of amygdala radiomic features and amygdala FC on cognition. Results: We found that ADD patients showed decreased amygdala FC with posterior cingulate cortex, middle frontal gyrus (MFG), and parahippocampal gyrus involved in the default mode network compared to HCs. The area under the receiver operating characteristic curve (AUC) of the amygdala radiomic model was 0.95 for ADD patients and HCs. Notably, the mediation model demonstrated that amygdala FC with the MFG and amygdala-based radiomic features mediated the relationship between depressive symptoms and cognitive function in AD. Limitations: This study is a cross-sectional study and lacks longitudinal data. Conclusion: Our findings may not only expand existing biological knowledge of the relationship between cognition and depressive symptoms in AD from the perspective of brain function and structure but also may ultimately provide potential targets for personalized treatment strategies.
What problem does this paper attempt to address?