Enhancing associative learning in rats with a computationally designed training protocol

Abstract

AbstractAssociative learning requires the activation of protein kinases with distinct temporal dynamics. Learning protocols with computationally designed intertrial intervals (ITIs) that maximize the interaction between fast-activated protein kinase A (PKA) and slow- activated extracellular signal-regulated kinase (ERK) enhance nonassociative learning inAplysia. Here, we tested whether an optimal learning protocol with irregular ITIs, predicted computationally to increase the overlap between PKA and ERK signaling in rat hippocampus, would enhance associative learning in mammals. We simulated ∼1000 training protocols with irregular ITIs to identify this optimal protocol, predicted to induce stronger associative learning than standard protocols with fixed ITIs. With auditory fear conditioning, we showed that male adult rats exposed to the optimal conditioning protocol exhibited stronger fear memory retrieval and impaired fear memory extinction, compared to rats that received either massed or spaced conditioning protocols with fixed ITIs. With fear extinction, we likewise observed that fear conditioned rats exposed to the optimal extinction protocol showed improved extinction of contextual fear memory, compared to rats that received standard extinction protocols. Immunohistochemical staining confirmed that the optimal conditioning protocol increased phosphorylated cAMP responsive element binding (pCREB) protein levels in the dentate gyrus (DG) of the dorsal hippocampus. These findings demonstrate the capacity of a behavioral intervention driven by a computational model of memory-related signaling pathways to enhance associative learning in mammals, and may provide greater insight into strategies to improve cognition in humans.Significance StatementTo test if associative learning in mammals can be enhanced by predetermined training protocols, we used a computational model of the dynamics of essential biochemical cascades for learning and memory to predict an optimal training protocol in rats. Using auditory fear conditioning and fear extinction tests, we demonstrated that the optimal protocol enhanced the acquisition and the extinction of fear memories. Immunostaining confirmed that the optimal protocol induced greater levels of pCREB in rat hippocampus. Our results suggest that a non-invasive behavioral intervention guided by computational modeling can be used to enhance associative learning in mammals, with potential relevance for facilitating memory formation as well as enhancing extinction-based therapies in humans.