Assessing the robustness of mental rotation in infancy: the importance of task and context

• In Event: Infants’ learning about object properties and categories in diverse environments

Abstract

Mental rotation, or the ability to recognize different views of an object and predict how it should look from a new perspective, is an important spatial ability that develops during infancy (Constantinescu et al., 2018) and predicts later mathematical aptitude (Lauer & Lourenco, 2016). However, we cannot directly probe infants’ mental rotation. Instead, we infer infants’ mental rotation from their discrimination of different rotations of the same object versus views of a distinct object, a mirror image (Moore & Johnson, 2020; Lauer et al., 2015). These results have been inconsistent; some studies suggest that mental rotation is evident in young infants (Moore et al., 2011; Quinn & Liben, 2008; Moore & Johnson, 2011), whereas others show that even in older infants’ mental rotation emerges with specific motor experience (Schwarzer et al., 2013; Möhring et al., 2013). Such discrepancies suggest that that performance on these tasks may be influenced by minor variations in task or stimuli. One possibility is that mental rotation is fragile and can only be observed in extremely sensitive contexts. Alternatively, mental rotation may be robust in infancy, but our tasks may not be effective at isolating it. Our goal was to establish whether one assessment of infant mental rotation, the change-detection task, is sufficiently robust to replicate.
We tested infants’ mental rotation in an exact (Experiment 1) and conceptual (Experiment 2) replication of the change detection task used by Lauer et al. (2015). In this task, infants were presented two simultaneous streams in which a Tetris-shaped object repeatedly appeared and disappeared, rotating from one view to another. In the non-changing stream, the item is the same on every rotation, and in the mirror stream, the object is a mirror image on every third appearance. Lauer et al. (2015) found that overall infants’ mirror preference score (i.e., the proportion of looking at the mirror stream divided by looking at both streams) was significantly different from chance (.50), indicating that they could mentally rotate the object, with boys showing a stronger preference.
We tested 59 infants (M = 296 days, SD = 53, 28 boys) in Experiment 1. Infants did not exhibit mirror preference scores (M = .49) that differed significantly from chance, t(58) = 0.12, p = .90, d = .02. Thus, infants showed no evidence of mental rotation, and we observed no gender differences. We then tested 32 infants (M = 290 days, SD = 32; 18 boys) in a follow-up experiment (Experiment 2; identical stimuli but shorter-duration trials and multiple monitors) and found that these infants also did not significantly prefer the mirror stream (M = .51), t(31) = 0.88, p = 0.38, d = .16, nor was there any difference between boys and girls. Our results suggest that mental rotation in this task is not robust. Our ongoing work leverages an online testing platform to test infants’ mental rotation. Comparison of lab and online assessment will provide further insight into how infants’ abilities differ when tested in well controlled lab settings versus noisy home environments.

Authors

• Aaron Gerald Beckner, University of South Carolina
  Presenting Author

• Annika Voss, University of California - Davis
  Non-Presenting Author

• Laura Sheridan, Cornell University
  Non-Presenting Author

• Marianella Casasola, Cornell University
  Non-Presenting Author

• Lisa M Oakes, University of California - Davis
  Non-Presenting Author