Developmental differences in spatial transformation processes: A pupillometry study

• In Event: 3-029 - Process-based approaches to understanding spatial development: the role of attention, language, and gesture

Integrative Statement

The ability to generate and transform mental representations of objects is an integral component of spatial intelligence that facilitates problem-solving across myriad contexts (e.g., Hegarty & Waller, 2005; Uttal & Cohen, 2012). In adulthood, this ability is often assessed via mental rotation tasks that require visualizing two objects rotating into alignment (Shepard & Metzler, 1971). In adulthood, the angular disparity between the two to-be-rotated objects linearly relates to manifold outcome measures, including reaction time, accuracy, and neural activity (Heil, 2002; Shepard & Cooper, 1986; Zacks, 2008), indicating that mental rotation represents an analogous process to the perception of physical rotation.

Relative to adults, evidence of analog mental rotation processes in young children is less clear. Preschool-aged children exhibit impoverished mental rotation abilities when tested via explicit measures, yet infants perform above chance on implicit looking-time tasks designed to assess mental rotation skills (cf., Frick, Möhring, & Newcombe, 2014). These paradoxical results introduce the possibility that preschool-aged children would engage in mental rotation if presented with implicit measures. To address this possibility, we developed a novel implicit measure to assess mental rotation processes in a passive-viewing context, allowing for the direct comparison of young children’s and adults’ performance. The measure (the spatial-oddball task) was modelled after the visual-oddball paradigm to index pupillary responses to changes in the orientation of two-dimensional stimuli (see Figure 1). Pupillary responses exhibit differential sensitivity to variations in task parameters; thus, if mental rotation processes are recruited during the spatial-oddball task, we would expect pupillary responses to linearly relate to the angular disparity between stimuli.

We administered the spatial-oddball task to 56 undergraduates (34 female) and 25 four-year-olds (15 female; data collection is ongoing) using a Tobii T120 eye-tracker. The task consisted of three conditions that differed in the angular disparity between adaptation and test stimuli (Figure 1). Participants also completed an explicit mental rotation task designed for their respective age groups (Levine et al., 1999; Peters et al., 1995). A mixed-model ANOVA revealed a linear relation between angular disparity and pupillary response across conditions (p=.022, η²p=.09; Figure 2), as would be expected if mental rotation processes were recruited. Importantly, the linear relation between pupillary response and angular disparity was greater among individuals with higher explicit mental rotation performance (p<.01, η²p=.11), suggesting the spatial-oddball task and the explicit mental rotation measure recruited similar underlying cognitive processes. Critically, age also moderated the effect of condition (p=.037, η²p=.07): whereas adults exhibited the expected linear relation between pupillary responses and angular disparity, children did not (Figure 2).

The present study is the first to utilize a passive-viewing task to directly compare mental rotation processes in young children and adults. Validating our novel adaptation of the oddball paradigm, results suggested that the spatial-oddball task elicited mental rotation processes in adulthood. However, in line with previous studies evaluating children’s explicit mental rotation performance, findings indicated that mental rotation processes are not manifest during the preschool years and thus emerge later in development.
 

Authors

• Jillian E Lauer, New York University
  Presenting Author

• Patricia Bauer, Emory University
  Non-Presenting Author