Parallel developmental changes in children's production and recognition of line drawings of visual concepts

Abstract

To what extent do visual concepts of dogs, cars, and clocks change across childhood? We hypothesized that as children progressively learn which features best distinguish visual concepts from one another, they also improve their ability to connect this knowledge with external representations. To examine this possibility, we investigated developmental changes in children's ability to produce and recognize drawings of common object categories. First, we recruited children aged 2-10 years to produce drawings of 48 categories via a free-standing kiosk in a children's museum, and we measured how recognizable these >37K drawings were using a deep convolutional neural network model of object recognition. Second, we recruited other children across the same age range to identify the drawn category in a subset of these drawings via "guessing games" at the same kiosk.We found consistent developmental gains in both children's ability to include diagnostic visual features in their drawings and in children's ability to use these features when recognizing other children's drawings. Our results suggest that children's ability to connect internal and external representations of visual concepts improves gradually across childhood and imply that developmental trajectories of visual concept learning may be more protracted than previously thought.

Developmental changes in drawing production under different memory demands in a U.S. and Chinese sample

“…In doing so, we make three key contributions to our understanding of the development of children's drawings. First, we replicate prior findings (Long et al, 2021) that the recognizability of children's drawings increases steadily throughout this age range (4-9 years).…”

“…A common view is that these changes are driven primarily by children's increasing ability to plan and control their motor movements (Freeman, 1987;Rehrig & Stromswold, 2018) Prior work has provided evidence that changes in children's drawings partly reflect changes in children's mental representations of visual concepts. In a large observational dataset, older children produced drawings of object categories that were more diagnostic of the categories they were trying to depict (Long, Fan, Chai, & Frank, 2021;Long, Fan, & Frank, 2018). This result held even when accounting for differences in basic shape tracing abilities and the amount of effort children expended on individual drawings.…”

“…Finally, we assessed the degree to which visuomotor development accounts for observed developmental changes in drawing recognizability. While it is uncontroversial that visuomotor control can constrain how and what we can draw, little work has directly related measures of visuomotor control to measures of drawing recognizability (Long et al, 2021). To do so, we measured each child's visuomotor control via a shape tracing task and related these measurements of tracing accuracy to the recognizability of the drawings that each child produced.…”

A tutorial on capturing mental representations through drawing and crowd-sourced scoring

“…This method can serve as a tool for understanding schemas or concepts. Drawing has been used to investigate the emergence of schemas in children (Freeman & Janikoun, 1972;Long et al, 2021), the influence of culture on schemas (Axia et al, 1998), and to question whether children even can use drawings to represent schemas (Kosslyn et al, 1977). Drawings have also revealed how experts differentially organize information through chunking (e.g., investigating electronic technicians' memories for circuit diagrams: Egan & Schwartz, 1979).…”

“…Some DNNs also create heatmaps across the image, such as MemNet, which creates a heatmap of what image information most contributes to its predicted memorability score. There has also been burgeoning work in relating DNN-based representations of drawings to those of the corresponding image (Fan et al, 2018) and using DNNs to score drawing recognizability (Long et al, 2021).…”