Our behaviour emerges as the result of many systems interacting at different scales, from low level biology to high level social interaction. Is it possible to create naturalistic explanatory models which can integrate these factors? This paper describes the general approach and design of a framework to create autonomous expressive embodied models of behaviour based on affective and cognitive neuroscience theories. The goal of our research is to integrate many different current theories and models to create a large functioning sketch of several fundamental aspects of human behaviour including face to face interaction, to explore how it may emerge from interaction of low level and high level systems in a top-down bottom up approach. A key feature of the approach is (given the constraints of the medium), to create as naturalistic models as possible in order to elicit and respond to the appropriate behaviours from the user, involving both sensing and synthesis of visual and auditory stimuli. The models are intended to be as autonomous as possible, so interactions are unscripted and co-created. We illustrate how these features come together in a psychobiological simulation of an infant Baby X currently under development, which aims to combine models of the facial motor system and models of basic neural systems involved in early interactive behaviour and learning.

Infants demonstrate remarkable growth in their ability to learn words across the first two years of their lives. To help infants achieve this goal, adults use a number of cues and strategies. We know much about the cues infants use to learn words but a comprehensive evidence-based model of the process during early interaction does not exist. The present research takes a critical first step to filling this gap.

In an ongoing longitudinal study using multiple high-speed cameras and a novel experimental paradigm we are characterizing the evolving nature of learning interactions in unprecedented detail. The sample consists of 10-month-old infants’ interactions with their own mum (n=8 dyads) and an unfamiliar mum (n=5 dyads). Mothers were given a novel toy called a blicket and were asked to teach infants the toy’s name.

Initial results of our detailed frame-by-frame coding are already revealing behavioural differences, namely, infants spend more time looking at their own mum’s face than the unfamiliar mum’s face. Infants also spend more time touching the toy with their own compared to an unfamiliar mum. This is the first evidence that there are subtle differences in 10-month-old’s behaviour with their own mum compared to another mum in a word-learning context.

Future work will examine mothers’ gaze, touch, speech, and identify mother/infant behaviours relating to enhanced learning. This data provides the critical foundation to develop 3D models of the interactions, offering novel insights into the complex interplay of mother and infant behaviour that shape early language learning.

Inhibitory control and sustained attention are critical skills for everyday living, however we have a poor understanding of how these executive functions develop in childhood. This longitudinal study tested 40 6-year-olds, 36 8-year-olds and 44 10-year-olds on the Random version of the Sustained Attention to Response Task (SART), three times over 12 months. There were initially no age differences in inhibitory control, as measured by commission errors, but the 10-11 group improved over time and made fewer commission errors than the 6-7 group at Time Three. All children showed improvement in top-down executive control of attention: they all showed a decrease over time in fast frequency response time variability (RTV). The 10-11 group performed the task with lower fast frequency RTV than the 6-7 group at all times, and lower fast frequency RTV than the 8-9 group at Time Three. The 8-9 group performed with lower fast frequency RTV than the 6-7 group at Times One and Three. The 10-11 group showed improvement in bottom-up arousal processes: only this group performed with decreased slow frequency RTV over time. The 10-11 group performed with lower slow frequency RTV than the 6-7 group at all times, whilst this occurred for the 8-9 group at Time Three only. Results suggest the ability to inhibit a prepotent response improves between 10 and 11 years, whilst sustained attention components show differing developmental trajectories between 6 and 11 years. This may reflect differing rates of maturation for neural networks involved in inhibitory control and sustained attention.

The aetiology of depression can be attributed to both environmental and genetic factors. Studies on adolescents and adults have frequently linked genes influencing the regulation of monoamines to the disorder. However, there is limited research investigating the genetic determinants of depression in children and the moderating role of environmental factors during gestational development. The current study investigated whether single nucleotide polymorphisms (SNPs) from monoaminergic genes interacted with measures of early life stress to influence depressive symptoms in children from the Auckland Birthweight Collaborative cohort. Indicators of early life stress included antenatal maternal stress and small birth weight for gestational age (SGA). Depressive symptoms were measured using the Centre for Epidemiological Studies Depression Scale for Children (CES-DC) and DNA samples were collected when the cohort was 11 years. Results revealed that SGA status and two SNPs from the dopamine transporter gene DAT1 had interactive effects on children’s depressive symptoms. Specifically, depressive symptoms were greater in children born SGA who are T homozygous for the rs1042098 SNP and C homozygous for the rs3863145 SNP. These findings suggest that poor intrauterine environments leading to a small birth weight for gestational age also seem to exacerbate the effects of certain DAT1 SNPs on depressive symptoms in children.