NCS-FO: Using computational cognitive neuroscience to predict and optimize memory

点击翻译按钮获取中文摘要

The last decade has seen an explosion of research concerning the neural processes underlying memory formation and learning. As the basic research in this field becomes more mature, exciting possibilities for application of this knowledge have begun to emerge. This proposal aims to capitalize on these findings by developing assistive learning technologies that may revolutionize the way we teach and train people. Researchers at New York University will develop automated "adaptive teaching" technologies that guide learners through material in an individualized way. The goal is to increase retention or mastery of materials by tailoring instruction to individual learners. The novel contribution of this project is to combine insights from cognitive neuroscience and machine learning in the design and operation of such technologies. Successful development of this synergistic research would be a transformative application of neuroscience to our daily lives and may lead to new commercial technologies. The research will also provide a post-doctoral training opportunity for the next generation of scientists working at the intersection of neuroscience and computer science. The award is from the Integrated Strategies for Understanding Neural and Cognitive Systems program, with funding from the EHR Core Research (ECR) program, which supports fundamental research that advances the research literature on STEM learning, and from the Behavioral and Cognitive Sciences division in the SBE directorate. The specific scientific goal is to explore novel applications of neuroscience methods (particularly fMRI) to improve how people learn. Neuroscience research has identified robust neural correlates of successful memory formation (e.g., activity in brain areas such as the medial temporal lobe and the hippocampus). The goal of this project is to use these variables to help predict the information needs of learners in an adaptive way. The project design involves scanning an individual's brain during the learning phase of a task. The research team then will identify which materials would benefit from additional study by combining computational models of the time course of learning and forgetting with theories mapping neural activation to successful memory formation. A computer algorithm then selects new materials for learners to re-study in a subsequent session. The goal is to show that generating a training sequence from a computer-based "neurofeedback" algorithm can enhance long-term memory retention more than when learners choose for themselves which items to re-study. This is a high-risk but potentially large reward project that merges basic science findings from neuroscience and cognitive science in ways that may transform the way we educate people.