Abstract
Fixation durations in infancy research are considered to be indicative of attention and information processing. Previous studies showed how fixation durations can be affected by (a) individual differences between short lookers and long lookers, (b) stimulus characteristics (e.g., static vs. dynamic), and (c) developmental changes in saccadic inhibitory control. However, the interaction between individual differences in saccade timing and top-down modulation of fixation durations by scene semantics is not known. In this study a group of infants aged 6 months were presented with (1) a set of customized naturalistic videos whereby three people perform several baby-friendly actions, (2) a second set of abstract non-social videos created from the first set, and (3) static complex images while fixation location and durations were measured. Further, all the infants performed a gap-overlap task in order to measure inhibitory control. Results from this study revealed stable individual differences in fixations durations as well as systematic changes across viewing conditions. Interestingly, fixation durations during the naturalistic videos were significantly shorter than in the abstract non-social videos, suggesting an adaptation of their looking behavior to the semantic content of the stimuli. Moreover, all infants showed very short fixations during the presentation of the static images. Corroborating previous research, the gap and overlap disengagement latencies correlated with individual mean fixation durations, evidencing the influence of inhibitory control on gaze allocation. These findings form the basis for building the first computational model of fixation durations in early infancy, which will be grounded on the CRISP adult model (Nuthmann et al, 2010; Psych.Rev). The infant model will aim to capture developmental changes in fixation durations enabling valuable insights into oculomotor control in typically developing children.
Meeting abstract presented at VSS 2013