What Do Young Infants Do During Eye-Tracking Experiments? IP-BET - A Coding Scheme for Quantifying Spontaneous Infant and Parent Behaviour.

Eye-tracking measurement of looking is the fundamental method in infancy research. Over the last few decades it has provided many of the most significant discoveries in developmental psychology. Infants engage in looking tasks and use their bodies for learning differently from adults, yet, the breadth of their behavioural repertoire and the constraints that the testing situation places on them remain under-explored. Young infants are tested in close physical proximity to their parent, interact during the experiment and rely on the parent to stay engaged in the task. Infants may also engage a different set of skills (e.g. when self-regulating) to perform the very same looking tasks in comparison with adult participants. We devised a coding scheme to systematically analyse task-relevant (attention to the screen) and extraneous behaviours [body movement, self-touch, non-nutritive sucking (NNS), affect] that infants exhibit during an eye-tracking session. We also measured parental behaviours (attention to the screen or the child), including dyadic interactions with the infant (talking, physical contact). We outline the rationale for the scheme and present descriptive data on the behaviour of a large group of typical 5- and 6-month-olds ( = 94) during three standard eye-tracking tasks in two seating arrangements. The majority of infants showed very high and consistent within-group attention to the screen, while there were large individual differences in the amount of limb and body movement and the use of self-regulatory behaviours (NNS, self-touch, object manipulation). Very few sex differences were found. Parents spent most time attending to the screen, but engaged in some forms of dyadic interaction, despite being given standard task instructions that minimise parental interference. Our results demonstrate the variability in infants' extraneous behaviours during standard eye-tracking despite comparable duration of attention to the screen. They show that spontaneous interactions with the parent should be more systematically considered as an integral part of the measurement of infant looking. We discuss the utility of our scheme to better understand the dynamics of looking and task performance in infant looking paradigms: those involving eye-tracking and those measuring looking duration.