Project Summary The ability to remember information, whether after short or long delays, is a fundamental human ability. There is enormous research into understanding working memory and long-term memory in isolation, but also a longstanding debate about the interactions between working memory and long-term memory. This research proposal would address a critical knowledge gap by investigating and dissecting the gateway hypothesis, to characterize how working memory delay activity predicts what information is later remembered and to build brain- computer interfaces that leverage real-time insight into working memory delay activity to alter later memory. A better understanding how working memory and long-term memory interact would be beneficial for the numerous psychiatric disorders which are characterized by deficits in these systems. Recent empirical research has revealed links between working memory capacity and long-term memory, however multiple sub-processes underlie maintaining multiple items in working memory. For example, there are neural signatures that correspond to the number of items in working memory, and distinct neural signatures that correspond to the spatial locations of items in working memory. Either or both of those sub-processes, number and location, could predict long-term memory. I will use multivariate decoding in conjunction with time resolved neuroscience techniques, EEG (Aim 1) and intracranial EEG (Aim 2) to characterize these working memory sub-processes and their relationship to long-term memory. Then, as an independent investigator, I will build tools that can track delay activity in real time and adaptively design experiments contingent to the number and location of items of working memory. This will test a detailed and specific conceptualization of the relationship between working memory and long-term memory. The research and training goals of this research proposal will be furthered by an advising team of cognitive, systems, and clinical neuroscientists at the University of Chicago and UC Berkeley. This research proposal encompasses cognitive processes that are often siloed (working memory and long-term memory), complementary temporally resolved methods (EEG and iEEG), and computationally sophisticated multivariate analyses capable of sensitively decoding information in working memory. Finally, this research proposal will develop innovative real-time tools to track information held in mind and forecast future long-term memory performance. The short-term goal of this research proposal is to develop a composite model of how distinct moment-by-moment subprocesses of working memory delay activity predict long-term memory outcome. This will provide new insights into the relationship between working memory and long-term memory. The long-term goal for my research program is to comprehensively characterize the diverse factors that influence what we remember, in order to build tools that can enhance memory.