Children's Semantic Knowledge Influences Visual Working Memory Capacity

• In Event: 2-160 - The More You Know: How Prior Knowledge Scaffolds Cognitive Development

Integrative Statement

A hallmark of visual working memory (VWM) is that it is capacity limited: we can only maintain a subset of the visual information that we encounter (Cowan, 2001). However, in contrast to earlier claims that VWM has a fixed capacity of only 3-4 items, recent work has demonstrated that the capacity of VWM is influenced by the type of visual information to be remembered. In particular, VWM capacity is greater for real-world objects than for abstract stimuli such as shapes or colored squares (Alvarez et al., 2004; Brady et al., 2016). This benefit is thought to stem from the presence of long-term semantic memory representations of real-world objects, which are thought to strengthen and help maintain representations of those objects within VWM. Here, we evaluate this hypothesis by exploring whether semantic knowledge about an object in fact makes an object easier to hold in VWM: Does the advantage for real-world objects hold only for objects that are familiar?

We first addressed this hypothesis in adults by testing VWM performance with arrays of familiar versus unfamiliar objects in a standard change detection task (Luck, 1997). This experiment, as well as three follow-up control experiments, confirmed that VWM capacity in adults is better for familiar objects than unfamiliar objects. In adulthood, our interactions with the world consist mainly of interactions with objects that are familiar. This experience is quite different for young children, who are constantly encountering and learning about new objects. If VWM capacity is also influenced by object knowledge in children, then children’s acquisition of semantic knowledge may contribute to previously-documented improvements in VWM capacity over development.

Thus, we next asked whether object familiarity also benefits VWM performance in older (aged 6-9 years, N=19) and younger (aged 4-5 years, N=25) children. Arrays of 3 or 4 pictures of objects were displayed for 2 or 3 seconds (in older and younger children, respectively), followed by a 1.5 second delay. Next, an image appeared in one of the locations and children had to respond whether it was the same or a different image than the one that had previously been in that location (Figure 1). Children in both age groups exhibited better memory for the familiar compared to the unfamiliar object pictures (ps<.001; Figure 2). These data suggest that knowledge about objects makes them easier to maintain in working memory, and that even preschoolers can use semantic knowledge to bolster their working memory. Interestingly, response strategies changed with age: whereas older children and adults were more likely to false alarm to unfamiliar compared to familiar items, false alarms in younger children were not affected by familiarity. In contrast, younger children had more misses for unfamiliar items compared to familiar items, which was not the case in older children and adults. These findings challenge traditional views of working memory capacity and its development by suggesting that semantic knowledge contributes to the processes supporting working memory.

Authors

• Ariel Starr, UC Berkeley
  Presenting Author

• Mahesh Srinivasan, UC Berkeley
  Non-Presenting Author

• Silvia A Bunge, UC Berkeley
  Non-Presenting Author