There is a brief flash on the screen—just 1/10th of a second. A few color patches are flashed, and 2 seconds later another array appears that stays on the screen until you respond. Your task is to decide whether any one of the patches of color in the second array differs from the colors briefly flashed 2 seconds earlier.

When visual working memory is tested in this manner, people can remember in the vicinity of 4 color patches.

But why only 4 items? What determines that capacity limit? How come one’s memory doesn’t seem to lose precision gradually, but that items suddenly disappear from memory without prior warning?

Jeremy Wolfe of Harvard University recently edited a special issue of the Psychonomic Society’s journal Attention, Perception, & Psychophysics that explored those kinds of research questions. Jeremy thought that “Visual working memory has been a hot topic for the past few years so it seemed timely to put together a Special Issue on that topic.”

The special issue comprises 23 articles by leading experts in the field of visual working memory. No wonder Jeremy was “… pleased with the quantity and quality of the response to our call for papers.”

How does one summarize 23 articles in a single post? Jeremy directed me to an article by Jordan Suchow and colleagues, because “it does a really nice job of framing the issues in thinking about visual working memory.”

Let’s begin with the basics. Suppose we can only remember exactly 4 different visual stimuli, such as the color patches that were briefly flashed on the screen. How do we explain that capacity limit?

Conventionally, there are at least two views on this issue: According to a “slot model” our visual memory system contains a number of discrete slots into which visual representations can be inserted. So if we have 4 slots, then we can hold 4 representations in memory at a given time—one color patch in each discrete slot. An opposing view holds that there are no discrete slots but that there is a supply of something called a “resource” that is shared among representations in support of their retention. One can think of that “resource” as a kind of cognitive fuel that keeps things alight in our minds—and because the supply is limited, as more information is presented, each representation necessarily becomes dimmer.

Reduced to their essence, Suchow and colleagues suggest that the “slot models propose that only a handful of representations can be stored in working memory… In contrast, the core idea of resource models is that memory stores a limited amount of information, in unspecified units.”

There has been much research dedicated to differentiating between those two models. Although the situation is quite nuanced, one crucial prediction of the resource model has found considerable support: The resource model predicts a trade-off between the number of items that can be stored and the amount of information in each item. To illustrate, you may retain 4 color patches but only 1 or 2 Chinese characters or hieroglyphs.

The novelty of the approach by Suchow and colleagues is their suggestion that dichotomizing between slot and resource models is inadvisable, and that what we should do instead, it to focus on the underlying issues. As they put it, “a dichotomous framing muddies the space of possible theories and, in doing so, leads to the confounding of what are distinct questions about the nature of working memory.”

And indeed, the trade-off findings that supported a resource view over the slot model were promptly followed by modifications to the slot model—for example, the idea that the same object may be stored in multiple slots—which has empowered the slot model to handle data previously thought to be quite challenging. However, as so often happens in psychology, it has also made the models less testable and less parsimonious.

So what are those underlying issues that Suchow and colleagues suggest as an alternative to the dichotomization of theories of visual working memory?

The authors suggest that there are 7±2 core questions in visual memory research. (By the way, the magical number 7±2 is one of the most famous numbers in cognitive psychology, but that’s another story.).

Here are the 4 questions that I found to be most fascinating:

• Is there an upper bound on the number of items about which information can be stored? Is it impossible for observers to store information about more than a handful of objects such that nothing is retained about additional items
• What is the “quantization” of the mental commodity used for memory storage? Is it discretely divided or continuously divisible?
• What is the relationship between quantity (the number of items about which information is stored) and quality (the amount of information stored about each of them)?
• How do we use a memory representation to make a response? This is a non-trivial question because turning a memory into an overt response requires modeling of a large variety of cognitive processes.

For answers, turn to the special issue.

There are many reasons why researchers are fascinated by questions relating to visual working memory. Understanding working memory also tells us something about attention; as for example revealed by the role of eye movements in retention. Understanding working memory also tells us something about “higher-level” cognition, including reasoning abilities and intelligence; the capacity of working memory (at least for verbal material) is highly related to one’s IQ.

In fact, knowledge of working memory capacity explains fully half of the variation in IQ.