Presenter(s): Min Zhang
Faculty Mentor(s): Ulrich Mayr & Melissa Moss
Poster 111
Session: Social Sciences & Humanities
In order to execute complex sequences, such as playing a piece of music, people group sequential elements (e.g., individual notes) into “chunks”. The dominant theory assumes that chunks are merely independent bundles of elements (Lashley, 1951), though little is known about the organization of chunked information. Applying the efficient coding principle, chunks may be coded in a relational manner, based on structural similarities, to allow people to smoothly execute complex tasks (Botvinick et al., 2015). In two experiments, we assessed how performance is affected by whether chunks contain similar, abstract patterns of elements (“matching” chunks, e.g., ABA pattern of elements in both chunks). Participants completed a task in which they needed to remember and execute sequences of rules. The rule sequences contained two 3-element chunks with various patterns of rule elements. Some sequences contained matching chunks, while others contained non-matching chunks. In Experiment 1, participants executed sequences in which both chunks contained the same types of rules arranged in matching or non-matching patterns. In Experiment 2, the two chunks contained different sets of rules, thus requiring a greater degree of abstraction across chunks. Results showed better performance for sequences containing matching chunks. The effect of pattern similarity on sequential performance was much weaker in Experiment 2 than in Experiment 1. Generally, these findings indicate that our cognitive system makes use of abstract patterns to efficiently code sequential information.