Modeling longitudinal trajectories of fixation durations in infants

Eye movements are the earliest fine-motor skill to develop in infancy and their progression to maturity provides a window into the development of cognitive functions over the first few years of life. However, the mechanisms underlying oculomotor control in scene perception, their development over the first few months, and the factors responsible for the large variations in fixation duration distributions observed in infants remain unexplored. We report a novel longitudinal study combining precise empirical developmental methods with computational modeling in order to investigate the development of the underlying mechanisms modulating fixation durations. Eye-movements were monitored in a group of infants at 14, 20 and 26 weeks while they looked at dynamic and static complex stimuli. In each assessment infants were presented with: (1) a set of naturalistic videos in which three people performed baby-friendly actions, (2) a set of abstract videos created from the first set, and (3) static complex images. Empirical findings indicated greater top-down eye-movement control from 14 to 20 weeks, and highlighted the increased influence of cognitive processing on fixation duration distributions as infants grew older. Data were then modeled in a series of simulations using an adaptation of CRISP, a model accounting for variations in fixation durations in adult scene viewing. Results confirmed the CRISP theoretical framework’s applicability to infant data, and highlighted the influence of both visual/cognitive processing and the developmental state of the visuo-motor system on fixation durations during the first few months of life. Model simulations showed how the extent to which each of these factors affected saccadic control varied across visits, suggesting a greater influence of visual/cognitive processing as infants developed. Furthermore, results suggested that already at 14 weeks infants are able to program saccades in two stages (an initial labile stage subject to cancellation and a subsequent non-labile stage that can no longer be cancelled). The relative length of the labile stage increased as compared to the non-labile stage especially from 14 to 20 weeks, indicating a greater ability to cancel saccade programs as infants grew older. The present work confirms that infant fixation durations reflect on-line perceptual and cognitive activity in a similar way to adults, but that the individual developmental state of the visuo-motor system still affects this relationship even at 26 weeks. Furthermore, the computer simulations allowed us to analyze the effects of these two factors on fixation durations providing greater insights into the development of oculomotor and attentional control.