Mapping the Neural Taxonomy of Mental Objects in Moment-to-Moment Cognition

Abstract

We mentally represent all kinds of objects across a variety of tasks and source modalities (i.e., mental objects). One intriguing theory is that in our moment-to-moment processing in working memory, objects are represented by content-free, reassignable pointers (or indexicals). Do all mental objects share the same set of pointers in working memory? If not, where should we draw the lines between different kinds of pointers? This dissertation’s goal is two-fold. One goal is theoretical: I propose a novel research program precisely centering around the above questions – aiming at unraveling the neural taxonomy of mental objects, by testing the extent to which the neural markers for pointers generalize across different paradigms, task goals, source modalities, and beyond. I review this research program in Chapter 1. The other goal is empirical: I aim to provide a set of initial results within this research program. These investigations make up Chapters 2 to 6, where I offer initial evidence suggesting a shared set of pointers regardless of input modalities (vision and language), and a shared set of pointers representing both objects and relations between them. I also offer novel MEG markers for pointers in visual working memory tasks that would facilitate future research on this topic: a right posterior sensor-level response that gets reflected in responses in inferior parietal regions. The last but one chapter (Chapter 7) marks methodological development important for this endeavor in a new variety of MEG (OPM-MEG), laying the foundation for future discoveries of relevant neural markers and better characterization of existing neural markers with this novel technology. Chapter 8 concludes, summarizing this dissertation and proposing future steps of research.