Amyloid Beta Secreted during Consolidation Prevents Memory Malleability

Memory allows organisms to predict future events based on their prior sampling of the world. Rather than faithfully encoding each detail of related episodes, the brain is thought to incrementally construct probabilistic estimates of environmental statistics that are re-evaluated each time relevant events are encountered [1]. When faced with evidence that does not adequately fit mnemonic predictions, a process called reconsolidation can alter relevant memories to better recapitulate ongoing experience [2]. Conversely, when an ongoing event matches well-established predictions, reactivated memories tend to remain stable [3, 4]. In part, the brain may confer selective mnemonic stability by shifting cell-intrinsic mechanisms of plasticity induction [5], which could serve to constrain maladaptive updating of reliably predictive representations during anomalous events. Based on evidence of decreased cognitive flexibility and restricted synaptic plasticity in later life [6], we hypothesized that some prevalent age-associated neurobiological changes might in fact contribute to mnemonic stability [7]. Specifically, we predicted that amyloid beta (Aβ)-a peptide that often accumulates in the brains of individuals expressing senescent dementia [8-10]-is required for memory stabilization. Indeed, we observe elevated soluble Aβ_x-42 concentrations in the amygdala shortly after young adult rats form reconsolidation-resistant auditory fear memories. Suppressing secretases required for Aβ production immediately after learning prevents mnemonic stabilization, rendering these memories vulnerable to disruption by post-reactivation amnestic treatments. Thus, the seemingly pathogenic Aβ₄₂ peptide may serve an adaptive physiological function during memory consolidation by engaging mechanisms that protect reliably predictive representations against subsequent modification.