Absence of higher order association based mnemonic percepts in Alzheimer’s disease

Abstract

AbstractBackgroundEveryday learning involves higher order associations (HOA) across new and old experiences when a response is encoded without reinforcement. However, the neural and behavioral mechanisms of HOA‐based learning during remote memory retrieval are unknown.MethodWe implement novel behavior to test the neural and behavioral mechanisms of HOA during remote memory retrieval. We train the animals in context A and sequentially test these animals for remote retrieval in same (A), similar(B), or novel (C) context, while invoking HOA‐based learning to investigate the neural correlates of HOA using hippocampal lesions, chemo genetics, and 2‐photon in vivo imaging.ResultCohorts of AD mice show overall differences in freezing vs. controls (1‐w ANOVA, P>F = 1.48E‐11, n = 5). Subgroups of AD (n = 12) and control (n = 14) mice during remote retrieval in BAC show overall difference (1‐w ANOVA, P>F = 4.59E‐6), with AD mice showing similar freezing in B and A (t = ‐0.70). Freezing in C is different for control mice in ABC vs. BAC order (P>F = 1.84E‐8, n = 5,14) whereas, interestingly, AD mice show similar increase in C freezing on Day3 of remote testing for ABC vs. BAC testing orders (P>F = n.s, n = 5, 12). AD (n = 3) and WT (n = 4) mice were trained with tone and tested after 70 days. Freezing before and during the tone was significantly different for WT mice (1‐t Paired‐sample‐t‐test, P<t = 0.015). Discrimination‐index indicates better discrimination ability for WT as compared to AD mice (1‐t two‐sample‐t‐test, P<t = 0.026). We further evaluate the neural correlates of such behavior in WT mice by timed‐hippocampus lesions, chemogenetic inhibition of PVcre interneurons in dorsal CA1, and longitudinal 2p‐in vivo imaging.ConclusionWe find that PVcre neurons are necessary for HOA during remote learning in WT mice and there is a loss of specificity in HOA‐based new learning during remote memory retrieval with AD mice. Our novel findings establish a basis for early behavioral and neural markers for predicting AD diagnosis to assist with early medical interventions.