PROJECT SUMMARY/ABSTRACT Episodic memories integrate the content of human experience in space and time and constitute the core of one's identity. Memory formation involves processing, and constructing interpretations of the incoming information in our daily lives and is one of the first functions compromised in neurodegenerative diseases such as Alzheimer's Disease. With population aging, we face a “Cognitive Tsunami” of millions of people with memory disorders. Thus, understanding neural mechanisms of memory, and finding interventions that enhance these processes is a critical endeavor with the potential to improve the lives of countless people world-wide. Although it is established that memory is critical for cognitive well-being, and sleep is critical for memory consolidation, the underlying mechanisms in the human brain are poorly understood. Research on memory and sleep so far has suffered from a gap between non-invasive cognitive research in humans and detailed electrophysiological research in animals. Moreover, most human studies are dominated by stimulus response methodologies where the presented stimuli constitute limited, discretized, aspects of memory. This approach, albeit well-controlled, is far from the rich narrative of episodes we experience. Thus, to mechanistically probe human memory consolidation, it is imperative to (a) employ methodologies that incorporate the continuous and multimodal nature of experience; (b) identify relevant neural activation patterns and how they are transformed and reactivated during sleep; (c) establish means to causally modulate memory processes during sleep. Building upon our exploratory U01 project, this proposal seeks a breakthrough in our understanding by going beyond the state-of-the-art, and via the application of integrative and multidisciplinary approaches. It capitalizes on a highly unique opportunity to record and modulate neuronal activity of a large number of single neurons and neuronal assemblies in the human brain. In parallel, it exploits the high dimensionality of the data as an asset through the use of cutting-edge Deep Learning (DL) algorithms, which have emerged as promising analysis tools. Specifically, the project will investigate the presence, and decoding, of distributed neural patterns associated with memory for different aspects of experience during wakefulness and identify their reactivation during sleep. The proposal aims to selectively modulate memory via application of novel closed-loop stimulation in sleep in concert with the DL model predictions. We anticipate that this project is poised to shed light on the relationship between sleep and memory, and its modulation from a novel perspective. Such an ambitious goal can only be achieved with unrivaled combination of experience, access to a clinical setting, and interdisciplinary collaborations such as those proposed in this project. By combining the opportunity to stimulate and record neural activity with the computation...