The relationship between reflex eye realignment and the percept of single vision in young children

Abstract

Effective binocular vision is dependent on both motor and perceptual function. Young children undergo development of both components while interacting with their dynamic three-dimensional environment. When this development fails, eye misalignment and double vision may result. We compared the range of image disparities over which young children display reflex motor realignment of their eyes with the range over which they report a single versus double percept. In response to step changes in the disparity of a 2.2° wide stimulus, 5-year-olds generated an adult-like reflex vergence velocity tuning function peaking at 2° of disparity, with a mean latency of 210 ms. On average, they reported double vision for stimulus disparities of 3° and larger, compared to 1° in adult reports. Three-year-olds also generated reflex vergence tuning functions peaking at approximately 2° of disparity, but their percepts could not be assessed. These data suggest that, by age 5, reflex eye realignment responses and percepts driven by these brief stimuli are tightly coordinated in space and time to permit robust binocular function around the point of fixation. Importantly, the plastic neural processes maintaining this tight coordination during growth control the stability of visual information driving learning during childhood.

Figure 1

(A) Mean proportion of trials for which participants reported seeing two (rather than one) characters as a function of disparity, for adults (n = 10) and 5-year-olds (n = 10). Error bars reflect standard error across participants. (B) The median and distribution of participants’ 63% diplopia thresholds, collapsed across sign, for each age group. Symbols to the right represent individual participant estimates. C: Mean proportion of stimuli reported as ‘saw two’ by the group of 5-year-olds who took part in the interleaved binocular control condition, collapsed across disparity sign. Error bars reflect standard error across participants.

Figure 2

Representative averaged vergence responses collected from one participant in each age group, for stimuli stepped to − 2° disparity (negative disparity indicates a stimulus for convergence). The responses represent the mean vergence position across all trials completed at the relevant disparity (20 trials for the adult and 5-year-old; 10 trials for the 3-year-old). Parameters estimated from the bilinear function are labeled in the middle trace, where C is the stable vergence value for the first zero-disparity period, T_L is the latency to the vergence response, and V is the slope representing the summary velocity of the open-loop vergence response.

Figure 3

The latency between the step in stimulus disparity and initiation of the vergence response (T_L). (A) The distributions of individual participant latencies, with median values, as a function of disparity step size for each age group (n = 10 per group). (B) Mean latency as a function of disparity step size for each group. Error bars reflect standard error. In all plots, the value plotted at a disparity of 0 reflects the mean latency across all trials, since no vergence movement is elicited from these stimuli.

Figure 4

The velocity estimates from the averaged vergence responses collected from each participant in each age group (open circles; n = 10 per group). The curves fit to these data are LOESS functions. The horizontal green (lower) and blue (upper) lines indicate the peaks for the convergent and divergent responses, respectively, with the 95% confidence interval for these peak estimates (shaded region). Triangles indicate the stimulus disparity at these peak estimates.

Figure 5

Top: Demonstration of a dichoptically presented disparity used in the diplopia threshold estimation task. The left eye image is slightly closer to the left side of the reference frame and the right eye image is slightly closer to the right side of the frame. This small disparity should result in the cartoon face appearing nearer than the reference frame if convergence is used to fuse the images. In this case, an observer should report seeing one face; if they are experiencing diplopia, they should report seeing two faces. Bottom: Example of a binocular stimulus used in the control condition. This stimulus has the same separation between the faces as the top row, but both faces are presented to both eyes. This stimulus should always lead to a report of seeing two characters. Note: the rectangular boxes used for illustration here were not included in the actual stimulus. The stimuli used in the experiment were green images of Pokemon characters, which have been replaced here with cartoon faces for copyright reasons.

Figure 6

The fitting approach used to estimate the latency and velocity of vergence responses to disparity stimuli. C is a constant reflecting the stable vergence position during the initial latency period, T_L specifies one latency interval before the vergence position begins to change, and V is the slope of the line fit to the dynamic vergence response.