A Neural Network Model for Memory Decay of the Olfactory System in Alzheimer’s Disease

Abstract

The Olfactory System is receiving increasing attention in recent years as a potential biomarker for Alzheimer’s disease (AD). Early-stage AD is often associated with a decline in olfactory function, with studies suggesting that olfactory memory deficit may precede cognitive symptoms. We explore the intricate relationship between the Olfactory System and Alzheimer’s disease, examining both the neuroanatomical and physiological changes that occur in the olfactory pathways during the progression of AD. Memory, as a functional feature, was modeled using artificial neural networks, and it was related to the macro-parameters of the network. While these networks cannot completely capture the intricacies and functions of the human brain, they provide a clear understanding of how processes occur within it. Neural networks exhibited memory domains, defined by stable and unstable steady states; the former can be considered a prerequisite for memory storage and recall; the latter can be considered threshold values between stable steady states. Additionally, by introducing division of the neural population into subpopulations, the networks manifested multiple stable states, corresponding to multiple memory domains, which in a qualitative manner suggest hierarchically organized nonlinear complexity and multi-scale self-similarity of memory processes.