Special Coverage on the 2020 Invited Talk

The Role of Meaning in Visual Working Memory Capacity (143)

Timothy Brady (University of California, San Diego) and Viola Störmer (University of California, San Diego & Dartmouth College)

Summary by Laura Mickes, Digital Content Editor

This recap is part of a special series of session summaries from the Psychonomic Society's 61st Annual Meeting. To read the rest of the series, click here.

Tim Brady, one of the Psychonomic Society’s 2019 Early Career Award Recipients, discusses some of his research with Viola Störmer on visual working memory.

Brady starts his talk equating the impoverished visual displays to understanding visual working memory to mouse models, as shown in the screenshot below. His question is, “How good are simple stimuli as a ‘model system’?

In all the experiments discussed, participants memorize six objects or colors. Comparing color stimuli to objects, memory accuracy is higher for objects. Why? Brady hypothesizes that objects give rise to greater working memory capacity.

He asks and answers questions that people have in the past asked him about this work. For each question, he provides empirical support for his hypothesis.

1. Is comparing performance across stimuli fair?

Answer: It can be. Discrimination accuracy increases with foil dissimilarity (see the screenshot below). To compare colors with objects, they used convolutional neural nets to compute similarities for pairs of objects. When the comparisons are fair e (e.g., color foils and objects foils are selected in the same manner), discrimination accuracy is higher for objects.

2. Do encoding limits work the same for both colors and objects?

Answer: No. Objects can be processed deeper than color patches. One test is to compare performance on simultaneously vs. sequentially presented color patches and objects. The result is that discrimination accuracy is higher for objects, but there’s also an interaction.

Discriminability accuracy is higher for object when presented sequentially and higher for colors when presented simultaneously. Brady proposed that simple features support parallel, global attention, but objects may require more individual processing. 

3. Do the objects have more dimensions, and that improves memory?

Answer: No. That should make memory performance worse for the objects.  

4. Do meaningful objects allow for more effective use of the same capacity?

Answer: No. Brady gave a couple of lines of evidence to refute this possibility, including EEG results (shown in the screenshot below) that show more activity for stimuli that participants perceive as meaningful than less meaningful.