Individual differences in how infants change behaviours from spontaneous to instrumental.

Young infants can change their behaviour and learn through interactions with novel environments. This ability has been demonstrated through group-averaged analyses. However, it remains unclear whether averaged behavioural changes accurately capture the diverse changes occurring at the individual level. To address this, we measured limb movement alterations in 185 infants aged 2 to 3 months before and after their arm was tethered to an overhead mobile and analysed individual differences in addition to conventional group analyses. While the group-averaged data showed a gradual increase in arm movements, individual learning curves rarely exhibited such simple gradual increases and instead displayed more complex patterns. To disentangle the complex movement patterns, we applied time-series clustering and dynamical systems modelling to our large-scale dataset. As a result, the infants were divided into distinct clusters with significant differences in spontaneous movements before learning, rather than after. A dynamical systems model further demonstrated that only differences in spontaneous movements could explain the diversity of overall behavioural changes. These findings indicate that the varying degrees of behavioural change reflect infants' unique learning processes rather than their learning capabilities. Furthermore, learning, as a process that reduces individual difference, suggest that infants harness their unique spontaneous movements to acquire instrumental behaviours.