Visual working memory is a fundamental cognitive resource, critical for visuospatial reasoning, guidance of eye movements, and the stability of perception over time. Much of the previous research on visual working memory has focused on its capacity limitations, with a general consensus that visual working memory can store detailed visual information from at most 3 or 4 individuated objects at any one time. A separate body of research has focused on a parallel capacity for storing lower-resolution representations of average features (ensemble representations) derived from collections of objects - but not much work has explored connections between the two, and they are often treated as separate topics in vision research. By contrast, the proposed research evaluates two competing hypotheses about how these two modes of visual representation relate to one another. The Flexible Representation model predicts that the visual system can flexibly alternate between individuation and ensemble representation, choosing, as it were, the type of representation most appropriate at any given time. The Encapsulated Representation model holds that these modes of representation are automatic and inflexible, with both occurring in parallel, all of the time. Te proposed studies evaluate these two models, relying on behavioral and neural signatures of memory precision to determine whether individuation is reduced (and ensemble representation enhanced) under conditions that should (under the Flexible Representation model) favor storing ensembles rather than individuals. These studies will use EEG/ERP to measure a purported neural signature of individuation - the contralateral delay activity (CDA), and will help reveal whether this signal truly reflects individuated object-like representations, or more flexible, hierarchically structured representations. Critically, the proposed studies are designed to yield a definitive result - confirming either the Flexible model or the Encapsulated model - and will represent a major advance in the visual memory literature in either case. These studies may also shed light on previously-documented visual processing deficits in autism spectrum disorder (ASD). Specifically, prior research has shown that ASD individuals have difficulty with global or holistic perception, but seem to be better than typically-developing individuals at tasks that involve narrow focus or precise representations of parts. If our studies confirm the Flexible Representation model, this would suggest a more fundamental impairment underlying these effects - ASD individuals may have impaired flexibility, individuating even when they should be forming ensemble representations. Finally, this research will provide the applicant with vital training in both visual memory and ERP research. While the applicant has previously studied visual attention and perception, he has not extensively studied visual working memory, nor has he previously done EEG/ERP research. Thus, both aspects of the project represent critical training that will broaden the scope of the applicant's research skillset.

Public Health Relevance

The proposed studies will establish whether human visual working memory can alternate flexibly between two modes of representation: either storing high-precision memories of single objects or low-precision summaries of groups of objects when most appropriate. These findings will inform several important debates about the fundamental nature (and neural basis) of visual memory, but also have important implications for research into autism and visual comparison ability.

National Institute of Health (NIH)
National Eye Institute (NEI)
Postdoctoral Individual National Research Service Award (F32)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1)
Program Officer
Agarwal, Neeraj
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Northwestern University at Chicago
Schools of Arts and Sciences
United States
Zip Code