Osamu Katayama ^1,2,3* Yaakov Stern ⁴ Christian Habeck ⁴ Annabell Coors ⁴ Sangyoon Lee ¹ Kenji Harada ¹ Keitaro Makino ¹ Kouki Tomida ¹ Masanori Morikawa ¹ Ryo Yamaguchi ¹ Chiharu Nishijima ¹ Yuka Misu ¹ Kazuya Fujii ¹ Takayuki Kodama ³ Hiroyuki Shimada ¹

• ¹ National Center for Geriatrics and Gerontology (NCGG), Ōbu, Japan
• ² Japan Society for the Promotion of Science (JSPS), Tokyo, Japan
• ³ Kyoto Tachibana University, Kyoto, Kyōto, Japan
• ⁴ Columbia University Vagelos College of Physicians and Surgeons, New York, United States

Background and Objectives: Cognitive reserve (CR) is a property of the brain that allows for better–than–expected cognitive performance relative to the degree of brain change over the course of life. However, neurophysiological markers of CR remain under-investigated. Electroencephalography (EEG) features may function as suitable neurophysiological markers of CR. To assess this, we investigated whether the dorsal attention network (DAN) and ventral attention network (VAN) activities, as measured during resting–state EEG, moderate the relationship between hippocampal volume and episodic memory. Methods: Participants were recruited as part of the National Center for Geriatrics and Gerontology–Study of Geriatric Syndromes. Hippocampal volume was determined using magnetic MRI, and episodic memory was measured using word lists. After testing the effect of hippocampal volume on memory performance using multiple regression analysis, we evaluated the interactions between hippocampal volume and DAN and VAN network activities. We further used the Johnson–Neyman technique to quantify the moderating effects of DAN and VAN network activities on the relationship between hippocampal volume and word list memory, as well as to identify specific ranges of DAN and VAN network activity with significant hippocampal–memory association. Results: A total of 449 participants were included in this study. Our analysis revealed significant moderation of DAN with a slope of β = -0.00012 (95% CI: -0.00024; -0.00001, p = 0.040), and VAN with a slope of β = 0.00014 (95% CI: 0.00001; 0.00026, p = 0.031). Further, we found that a larger hippocampal volume was associated with improved memory performance, and that this association became stronger as the DAN activity decreased until a limit of DAN activity of 944.9, after which the hippocampal volume was no longer significantly related to word-list memory performance. For the VAN, we found that a higher hippocampal volume was more strongly associated with better memory performance when VAN activity was higher. However, when VAN activity extended beyond -914.6, the hippocampal volume was no longer significantly associated with word-list memory. Discussion: Our results suggest that attentional networks help to maintain memory performance in the face of age-related structural decline, meeting the criteria for the neural implementation of cognitive reserve.