Developmental changes in gaze patterns in response to radial optic flow in toddlerhood and childhood

Abstract

A large field visual motion pattern (optic flow) with a radial pattern provides a compelling perception of self-motion; a radially expanding/contracting optic flow generates the perception of forward/backward locomotion. Moreover, the focus of a radial optic flow, particularly an expansive flow, is an important visual cue to perceive and control the heading direction during human locomotion. Previous research has shown that human gaze patterns have an "expansion bias": a tendency to be more attracted to the focus of expansive flow than to the focus of contractive flow. We investigated the development of the expansion bias in children (N = 240, 1-12 years) and adults (N = 20). Most children aged ≥ 5 years and adults showed a significant tendency to shift their gaze to the focus of an expansive flow, whereas the youngest group (1-year-old children) showed a significant but opposing tendency; their gaze was more attracted to the focus of contractive flow than to the focus of expansive flow. The relationship between the developmental change from the "contraction bias" in early toddlerhood to the expansion bias in the later developmental stages and possible factors (e.g., global visual motion processing abilities and locomotor experiences) are discussed.

Figure 1

(a) Schematic diagram of the motion paths of the area of interest (AOI) and the focus of the radial flow pattern. (b) Flowchart of the experimental procedure. These figures were adopted from Shirai & Imura (2016, Scientific Reports, https://doi.org/10.1038/srep34734), in accordance with a Creative Commons Attribution 4.0 International License. The apparatus, visual stimuli, and experimental procedures of the current study were identical to the method used by Shirai & Imura (2016).

Figure 2

Individual mean looking time under each experimental condition. The left and right panels summarize the results for the low- and high-speed conditions, respectively. In each panel, the vertical axis shows the mean time looking at the AOI, and the horizontal axis shows the age in years of the individual participants. Red dots and blue squares indicate individual mean looking times for the expansion and contraction flow patterns, respectively. Solid red and dotted blue lines represent logarithmic regression curves for the expansion and contraction flow conditions, respectively. Gray stripes indicate 95% confidence intervals. The regression curves and confidence intervals were obtained using the ggplot2 R package (ver. 3.3.5).

Figure 3

(a) Mean time looking at the AOI of the flow patterns. In each panel, the vertical axis shows the mean time looking at the AOI, and the horizontal axis shows the age group of the participants. Solid red lines with circles represent mean time looking at expansion flow patterns, and dotted lines with open squares represent mean time looking at contraction flow patterns. Left and right panels show the results under low-speed (5.8°/s) and high-speed (11.6°/s) conditions, respectively. Error bars indicate ± 1 SEM. (b) Mean time looking at the whole area (AOIW) of the flow patterns. In each panel, the vertical axis represents mean time looking at the AOIW, and the horizontal axis represents the age group of the participants. Solid red lines with circles represent mean time looking at expansion flow patterns, and blue dotted lines with open squares represent mean time looking at contraction flow patterns. Left and right panels show the results under low-speed (5.8°/s) and high-speed (11.6°/s) conditions, respectively. Error bars indicate ± 1 SEM. (c) Asymmetry index of looking time for the expansion and contraction flows in the AOIW analysis (vertical axis). The index for each participant was calculated as the total looking time for the expansion flow divided by the total looking time for both the expansion and contraction flows. Mean index values higher and lower than 0.5 indicate looking behaviors biased toward expansion and contraction, respectively. Error bars indicate ± 1 SEM.

Figure 4

Individual mean latency of the first gaze to AOI under each experimental condition. The left and right panels summarize the results for the low- and high-speed conditions, respectively. In each panel, the vertical axis shows the mean latency of the first gaze to AOI, and the horizontal axis shows the ages of the individual participants in years. Red circles and blue squares indicate the individual mean latency for the expansion and contraction flow patterns, respectively. Solid red and dotted blue lines represent logarithmic regression curves for the expansion and contraction flow conditions, respectively. Gray stripes indicate 95% confidence intervals. The regression curves and confidence intervals were obtained using the ggplot2 R package (ver. 3.3.5).

Figure 5

(a) Mean latency of the first gaze toward the AOI of the flow patterns. In each panel, the vertical axis shows the mean latency to the AOI, and the horizontal axis indicates the age group of the participants. Solid red lines with circles represent the mean latency for expansion flow patterns, and dotted blue lines with open squares represent the mean latency for contraction flows. Left and right panels indicate the results under low-speed (5.8°/s) and high-speed (11.6°/s) conditions, respectively. Error bars indicate ± 1 SEM. (b) Mean latency of the first gaze toward the AOI of the benchmark patterns. The vertical bar shows the mean latency to the AOI, and the horizontal bar indicates the age group of the participants. Error bars indicate ± 1 SEM.