It’s getting late and the party is noisy. Everybody around you is talking sat the same time and you have trouble keeping track of the conversation.
And then someone at the other end of the room whispers your name. All of a sudden you can hear quite clearly: someone said your name, and all the other conversations no longer matter.
We have all experienced this “cocktail party effect” and it is indicative of people’s attentional abilities: we can attend to our name with little effort and to the exclusion of other competing information.
It turns out that there is more to attention than just attention.
The Psychonomic journal Attention, Perception, & Psychophysics recently published a special issue on visual working memory that comprises 23 articles by leading experts in the field. Visual working memory refers to our ability to retain purely visual information—such as color patches or shapes—over brief periods of time.
Most people can remember around 4 color patches or similar visual stimuli. There is, however, considerable variation in the amount of information people can retain in their working memory. Some of us can remember 3 items, others might remember 4 or 5, or even 7 items.
Understanding those capacity limits appears important in light of the fact that working memory capacity (WMC) “…is strongly correlated with performance on a large range of cognitive tasks, scholastic achievements, and common cognitive failures”, as Jonathan Mall and colleagues put it in their contribution to the special issue.
Several theories have been put forward to explain variation in WMC. One influential view suggests that WMC varies as a function of people’s attentional abilities. Our ability to pay attention therefore determines how much we can remember. The paper by Mall and colleagues sheds new light on the relation between attention and working memory.
Let’s briefly return to the cocktail party effect. Turns out that people with low WMC are much more susceptible to hearing their own name while they focus on something else than people with high WMC—so high-WMC individuals can focus on the task at hand (a conversation) and filter out information that’s less relevant at the time.
Mall and colleagues followed up on this idea using a visual analog of a cocktail party. Their basic procedure is shown in the figure below:
People saw the memory display, which could contain either 2, 4, or 8 objects (the figure shows the condition with 8 objects; note that there is an equal number of circles and triangles). After a 3 second blank interval, a single test item (the “probe”) was presented, and people had to indicate via buttonpress whether the color of the probe was the same as or different (split 50/50 across trials) from the shape at the same location on the memory display.
So where is the cocktail party?
The cocktail party was introduced through another subtle manipulation: There were multiple “blocks” of trials, and during some blocks only circles (or triangles) were tested whereas on other blocks all items could be tested (there was a third condition which we ignore for simplicity). The former was called the “half-set” condition and the latter the “full-set.”
The half-set condition, in which one or the other shape was to be ignored, is a visual analog to a cocktail party: You are in a large room (lots of shapes) and you are trying to attend to a conversation (circles) while ignoring other people’s chatter (triangles).
The question of interest was whether people’s WMC—measured separately using standard tasks—would be associated with participants’ abilities to “filter out” the irrelevant stimuli in the half-set condition. Mall and colleagues measured people’s eye movements throughout, so they could tell whether people fixated on the irrelevant items during study in the half-set condition.
The results were intriguing: Mall and colleagues stated that “Contrary to the notion that WMC measures the ability to selectively focus attention by ignoring irrelevant information, we observed that individuals with low WMC spent less time looking at less relevant items in a visual memory display.” Although performance was better overall for high-WMC people than low-WMC people, there was no evidence that low-WMC people were unduly distracted by irrelevant items, which would have been reflected in them looking at those items more than their high-WMC counterparts.
What are the implications of this study?
Mall and colleagues suggest a few insights: For example, they believe that previous demonstrations that linked attentional capability to WMC may have exaggerated the degree of association because the two processes—attention and WMC—were measured within the same (or similar) experimental paradigm. In this study, by contrast, WMC was assessed very differently from the attentional manipulation involving visual working memory.
More generally, Mall and colleagues believe that “This research will help theorists determine appropriate boundary conditions to set upon theoretical relationships between attention and WMC, which is crucial to enabling prediction of how distraction and WMC are truly related.”