Technological advances in psychological research have enabled large-scale studies of human behavior and streamlined pipelines for automatic processing of data. However, studies of infants and children have not fully reaped these benefits, because the behaviors of interest, such as gaze duration and direction, even when collected online, still have to be extracted from video through a laborious process of manual annotation. Recent advances in computer vision raise the possibility of automated annotation of video data. In this paper, we built on a system for automatic gaze annotation in human infants, iCatcher (Erel et al., 2022), by engineering improvements, and then training and testing the system (hereafter, iCatcher+) on two datasets with substantial video and participant variability (214 videos collected in lab and mobile testing centers, and 265 videos collected via webcams in homes; infants and children aged 4 months to 3.5 years). We found that when trained on each of these video datasets, iCatcher+ performed with near human-level accuracy on held out videos on distinguishing “LEFT” and “RIGHT” looking behavior, and “ON” versus “OFF” looking behavior, across both datasets. This high performance was achieved at the level of individual frames, experimental trials, and study videos, held across participant demographics (e.g., age, race/ethnicity) and video characteristics (e.g., resolution, luminance), and generalized to a third, entirely held-out dataset. We close by discussing next steps required to fully automate the lifecycle of online infant and child behavioral studies, representing a key step towards enabling rapid, high-powered developmental research.

Associations between children’s early language processing efficiency and later language, literacy, and non-verbal outcomes shed light on the extent to which early information processing skills support later learning across domains. Examining whether the strengths of associations are similar in typically developing and at risk populations provides an additional lens into the varying routes to learning that children take across development. We compared patterns of associations between early language processing efficiency (accuracy and reaction time) in the looking-while-listening (LWL) task and school-relevant skills in children born full-term (FT) and preterm (PT). Participants (n=94, 49 FT, 45 PT) were assessed in the LWL task at 18 months (corrected for degree of prematurity in PT group) and on standardized tests of expressive language, pre-literacy (print knowledge and phonological awareness), and non-verbal IQ at 4 ½ years. Early language processing efficiency was associated with later language and pre-literacy outcomes (r2 change ranged from 19.8 to 7.1, p < 0.01) to a similar extent in PT and FT children, controlling for age at test and SES, suggesting similar mechanisms of learning in these domains for PT and FT children. However, birth group moderated the association between reaction time and non-verbal IQ (r2 change 4.5, p < 0.05), such that an association was found in the PT but not the FT group. This finding suggests that information processing skills reflected in efficiency of real-time language processing may be recruited to support learning in a broader range of domains in the PT compared to the FT group.

Variation in skill at processing language in real time in toddlerhood, i.e., language processing efficiency, has been shown to be associated with variation in later language and non-verbal cognitive outcomes in full term (FT) and preterm (PT) children. However, no studies to date have assessed relations between early language processing efficiency and later pre-literacy skills and whether these relations operate directly on outcomes or indirectly, via relations to other skills, such as expressive language. In this study, participants (n=94, 49 FT, 45 PT) were assessed in the looking-while-listening (LWL) task at 18 months (corrected for degree of prematurity in PT group) and on standardized tests of pre-literacy skills (print knowledge and phonological awareness) and expressive language at 4 ½ years. Early language processing efficiency predicted pre-literacy outcomes (r2 change ranged from 7.8 to 19.5%, p < 0.01) in FT and PT children. These relations were significantly reduced when controlling for children’s expressive language abilities. However, processing speed and phonological awareness were also directly related. Parallel relations were observed in both birth groups. Thus, early real-time language processing efficiency in toddler years supports expressive language abilities at preschool age, which in turn supports developing pre-literacy skills in FT and PT children.

•Associations between early language processing efficiency in toddlerhood and later standardized test performance inform the extent to which information processing skills support learning across domains.•Comparing patterns of associations in children from different clinical groups (e.g., children born full term and preterm) further informs whether neurobiological risk alters developmental pathways.•Early language processing efficiency was associated with language and pre-literacy outcomes to a similar extent for preterm and full term children, suggesting similar underlying mechanisms. •Association between processing speed and non-verbal IQ differed by group; processing speed supports learning in a broader range of domains in preterm than term children.