Search Papers | Poster Sessions | All Posters
Poster B55 in Poster Session B - Thursday, August 8, 2024, 1:30 – 3:30 pm, Johnson Ice Rink
Dynamics of Multi-form Task Representations during Sequence Learning
Guochun Yang1 (), Jiefeng Jiang2; 1University of Iowa
While humans can rapidly learn new tasks, the underlying task representations are less known. We posited that a task, such as cooking, might be conceptualized in various formats—either as a sequence of steps (sequence-form), a collection of discrete tasks (task-form), or as interconnected subtasks linked by transitions (transition-form). To probe these ideas, we designed a delayed matching paradigm where participants were required to remember a stimulus composed of five distinct features and then select the matching option for a prompted feature after a brief interval. Five trials form a sequence, each having a fixed order of cued features. A good memory of sequence/transition can predict the upcoming task and enhance performance. We tested the dynamics of different representational forms by training participants (n = 37) with varying combinations of sequences at different stages. We developed a model with a hidden variable for each representational form. Model comparison results supported the presence of representations in different forms and characterized their dynamics in learning. In summary, our findings underscore the dynamic changes in task representation during learning.
Keywords: task representation learning computational modeling sequence memory