Anamarie Casper, “A Novel Paradigm for Measuring Acquisition and Transfer of Knowledge Schemas”
Mentor: Neal Morton, Psychology, Letters & Science (College of)
Poster #59
Generalizing knowledge across situations is thought to depend on schemas that represent structure that is shared among different situations. For example, one’s knowledge of airports may be stored in a general-purpose schema that represents common features learned from experience with multiple different airports. However, the mechanisms underlying the learning and organization of these schemas still remain unclear. The recently proposed Tolman-Eichenbaum Machine (TEM) model proposes that schemas are represented by cognitive maps that contain information about general structure of situations and help apply this knowledge to new situations. According to the TEM model, a process of path integration, which involves making predictions about the outcomes of actions within a situation, is critical for learning. We developed a novel behavioral paradigm to examine the role of actions, or “operators”, and their outcomes in recognizing similar situations. The paradigm is similar to navigating blocks within a city; each block corner has a landmark, and participants are “moved” between these corners through north, south, east, and west movements. Their task is to predict which landmark will appear next, based on what they have previously learned. Participants are shown novel non-words in place of the directions; for example, a participant might see “tane” instead of “north”. Thus, this task requires participants to learn how to navigate a completely unfamiliar situation. After learning about one situation, participants navigate a new situation using the same operators, allowing us to assess how well they have learned a generalizable schema. Based on the TEM model, we predict that participants who learn one movement at a time will be less able to apply their knowledge in a new situation, compared to participants who are forced to make multi-step predictions that encourage cognitive map formation. Findings from this paradigm will clarify the cognitive mechanisms involved in schema learning.