Abstract
Postnatal development of the visual system depends on retinal image quality and correspondence. These are defined by an infant's ability to accommodate and align their eyes to targets in a dynamic 3D environment. Numerous factors impact development of these coupled responses including immaturities in sensory sensitivity to blur and disparity, the primary cues to accommodation and vergence. Infants' accommodation responses in the absence of disparity cues (condition 1), and vergence responses in the absence of blur feedback (condition 2) were determined, to understand the relative reliability of the two motor responses. Methods: Accommodation and vergence responses were recorded simultaneously, at 25Hz, using the MCS PowerRefractor (photorefraction and Purkinje image eye tracking). 95 infants recruited and tested between 3 & 5 months of age, were tested again between 7 & 9 months of age. They viewed a 7cm square animated cartoon moving repeatedly on a motorized track between 80cm (1.25D) and 33cm (3D). In condition 1, the right eye was occluded with a near-IR filter eliminating disparity cues (while recording from both eyes). In condition 2, the cartoon screen was covered with a spatially low-pass filter, with a 2D Difference of Gaussian (DOG) printed on it to remove blur feedback. Results: Raw responses were filtered to remove outliers, based on manufacturer's recommendations and physiological plausibility. Correlations between responses and the stimulus profile were then calculated. At both ages, the vergence response in the absence of blur feedback was more highly correlated with the stimulus than accommodation responses in the absence of disparity cues (mean correlation: 3-5m = 0.27 (SD+/- 0.21) vs 0.02 (+/-0.23), p<0.001, 7-9m = 0.15 (SD+/-0.23) vs 0.01 (+/-0.24), p<0.001). Discussion: Although quite variable, vergence responses in the absence of blur feedback to accommodation were driven more reliably than accommodation in the absence of disparity, suggesting increased senstivity or reduced tolerance to error in vergence.
Meeting abstract presented at VSS 2014