Poster #198 - The fractal structure of gaze patterns for young children watching naturalistic social stimuli

• In Event: 2-055 - Poster Session 06
  In Poster Session: PS 06 Section - Social Cognition

Integrative Statement

Background
The manner in which infants and young children visually explore their surroundings is one of the earliest ways they learn about the world. Guided by both bottom-up stimulus properties and top-down motivations, the processes by which children deploy their gaze are thought to be self-organizing and increasingly complex over time.
We propose that clarifying the fractal properties of children’s gaze patterns will provide meaningful insights into the emergence of these complex systems. Specifically, complex, non-linear processes often carry organizational structures known as fractals—repeated, self-similar patterns across a variety of scales (Mandelbrot, 1977). This mathematical approach has been used to characterize the temporal and spatial structure of adults’ gaze behavior (e.g., Stephen & Anastas, 2010; Aks et al., 2002; Marlow et al., 2015). However, the feasibility of its use with children, the dynamic micro-structure of children’s gaze behavior, and its relationship with stimulus properties and developmental change, are unknown.

Objective
The current study aims to fill these gaps by examining the fractal micro-structure of 5- to 55-month-old children’s gaze patterns (N=23) while watching videos of dynamic, naturalistic social stimuli.

Method
Eye-tracking data were collected as participants watched eight 20-second movies. Four movies included 3 females performing child-friendly actions. The other 4 were identical except that the social information was pixelated. Each movie was divided into 3-4 segments based on movement sequences. Time-series, comprised of amplitude changes between consecutive gaze coordinates (Fig.1a), were used for analyses.
We then calculated the fractal complexity of each movie segment for each child (N=23, 13 females, 428 segments), using detrended fluctuation analysis (DFA; Ihlen et al., 2015). DFA produces a parameter called the Hurst exponent (H), an index of the overall fractal structure of the time-series, on a continuum from white noise (i.e., more random; ~0.5); to pink noise (i.e., optimal flexibility; ~0.7-1.0); to brown noise (i.e., more rigid; ~1.5).

Results
The H values for children’s gaze behavior were normally distributed (M=0.86, SD=0.12; Fig.1b). Linear mixed effects models indicated that, on average, H was 0.76 for 5-month-olds, suggesting a gaze structure in the optimally flexible, pink noise range. Preliminary results suggest that H increased linearly with age by 0.0025 per month (p=0.022), but remained within the pink noise range for all ages. There was a main effect of stimulus type such that, on average, H was 0.034 lower (0.28 SDs, p<0.001; Fig.2) for pixelated social content compared to coherent social content.

Conclusions
Results suggest that, similar to findings in adults, infant gaze patterns exhibit a fractal micro-structure. Moreover, this fractal structure is partially determined by stimulus properties; children’s gaze patterns have higher H values for stimuli with coherent social content, suggesting children exhibit comparatively more random gaze patterns when the social salience of the stimuli is reduced. Preliminary results also suggest that H values show developmental change, as children become more sophisticated explorers of their visual worlds. Future work will probe these associations further and examine the extent to which nonlinear fractal methods can shed light on the development of other complex socio-cognitive processes.

Authors

• Isabella Stallworthy, Institute of Child Development, University of Minnesota
  Presenting Author

• Robin Sifre, Institute of Child Development, University of Minnesota
  Non-Presenting Author

• Carolyn Lasch, Institute of Child Development, University of Minnesota
  Non-Presenting Author

• Jed Elison, Institute of Child Development, University of Minnesota
  Non-Presenting Author

• Tim Smith, Department of Psychological Sciences, Birkbeck University of London
  Non-Presenting Author

• Dan Berry, University of Minnesota
  Non-Presenting Author