Poster #2 - Bridging the “toddler gap” in our understanding of visual working memory development

• In Event: 1-115 - Poster Session 03
  In Poster Session: PS 03 Section - Attention, Learning, Memory

Integrative Statement

Visual working memory (visual WM) is allows us to combine pieces of visual input to briefly maintain representations of the world. Successful visual WM requires encoding and maintaining visual information, binding visual features to their specific locations in space, and updating visual WM as new information comes in. While these processes have been studied in infancy (Kibbe, 2015), childhood (Cowan, 2016), and adulthood (Brady et al., 2011), little is known about the developmental trajectory of visual WM between infancy and childhood. To our knowledge, no studies have examined visual WM in toddlerhood. To address this gap, we created a visual WM task designed to examine multiple components of visual WM (encoding, binding, and updating) in two-year-old children.

In our task (Figure 1), we showed children a box containing multiple covered cups. On each trial, we hid either two or three different colored beads one at a time, each in their own cup. We then probed toddlers’ recall for the bead in one of the locations by showing them a colored card and asking them to find the bead that matched the color on the card. Thus, the task engaged multiple visual WM processes: to succeed, toddlers were required to encode representations of the hidden beads, to maintain those representations while updating the contents of visual WM as subsequent beads were hidden, and to bind those representations to specific spatial locations. Toddlers completed 10 total trials (4 Set Size 2, 6 Set Size 3), with each location (first-hidden, last-hidden, etc.) probed twice.

Eighteen toddlers (M=35.6 months, SD=4.25, 10 girls) participated. Results are summarized in Figure 2. We first computed toddlers’ mean proportion correct for each Set Size. Toddlers chose the correct location at rates significantly above chance for both Set Size 2 and Set Size 3 (ps<0.05). However, toddlers’ success varied as a function of which location was probed. For both Set Sizes, toddlers were above chance when probed on the last-hidden object, and performance on this object was better than for the other hidden object(s), consistent with previous results in both infants (Kibbe & Leslie, 2013) and children (Berry et al., 2018). At both Set Sizes, toddlers also were significantly above chance when probed first-hidden object. However, at Set Size 3, toddlers were below chance when probed on the middle-hidden object.

These results suggest that, when tasked with tracking the locations of objects hidden sequentially, toddlers can reliably recall two of those objects, binding specific colors to specific locations. Further, toddlers can maintain bindings (e.g. between the first-hidden object and its location) while updating the contents of visual WM with new information (e.g. of the second-hidden object). Finally, as set size increases, toddlers allocate limited WM to recalling the first- and last-hidden objects. Toddlers’ tendency to avoid the middle-hidden location may reflect their uncertainty about the contents of this location. These data contribute new insights into the development of visual WM, providing data on a previously unstudied population.

Authors

• Jessica R. Beal, Boston University
  Presenting Author

• Melissa M. Kibbe, Boston University
  Non-Presenting Author