Testing the role of learned regularities in visual working memory: The nature of chunking for continuous visual features

University Of California-San Diego, La Jolla CA

Investigators

Abstract

Working memory allows people to hold information in mind over a short period of time, and to work with this information. For example, while doing a math problem, working memory is used to hold in mind the numbers and manipulate them; and while driving, working memory is used to hold in mind the locations of other cars as you check whether it is safe to change lanes. Working memory has a severely limited capacity, such that only a few items can be accurately held in mind at any given time, and this limited capacity provides a strong constraint on people’s ability to perform many cognitive tasks. One well-known aspect of working memory capacity is that it is affected by learning: “chunking” frequently encountered information into compressed, meaningful units allows more efficient representations, and allows us to store more in working memory. For example, remembering a random string of 7 numbers is much more cognitively demanding than remembering our own phone number. The current research investigates the cognitive and neural basis of how learning increases the amount of information that we can process via chunking, and the extent to which the contents of “chunks” are actively stored in working memory vs. off-loaded into long-term memory after we have learned them. Importantly, the amount of information a person is able to hold in mind in working memory has been shown to be related to many other cognitive abilities (including intelligence), and disruptions of working memory are common in clinical disorders like attention deficit hyperactivity disorder and schizophrenia. Therefore, new knowledge regarding how this information is stored and used is critical for understanding the nature of important differences between individuals and how such differences may be impacted by learning. In addition, in many real-world situations it is vital to maintain detail about many objects or events at once, to make informed decisions in real time (e.g., changing lanes while driving). This research will inform how learning affects working memory capacity in such situations. In addition to this work on chunking, the proposal also includes the development of tools to allow scientists to measure memory performance and to do so in a wider variety of individuals (i.e., via internet-based experiments), as well as an outreach component, featuring the training and recruitment of students from underrepresented groups. The research uses behavioral and electrophysiological measures of working memory to test the role of item-specific, detailed information in chunk learning. The proposed studies are meant to distinguish between two potential theoretical explanations for how chunks are instantiated in working memory, one based on content-free pointers and the other based on hierarchical storage of objects and their features. A sequence of behavioral experiments asks how people’s representation of the visual features of objects is affected by learning, and in particular the extent to which learned regularities cause people to lose access to the details of visual objects in working memory. These behavioral experiments are followed by replications and extensions using electroencephalography (EEG), which will be analyzed to look at an event-related potential, the contralateral delay activity, which provides a measure of the amount of information held in working memory. Together, the behavioral and electrophysiological measures inform how learning and chunking affect our capacity to hold information in mind in working memory. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.