Randomisation

The random allocation sequence was generated in two blocks (n = 10 each) using randomly generated numbers from the uniform distribution. These were generated using Stata 13 [63], and constraining the intervention/control ratio to be 1:1.

The main researcher conducting the trial received the allocation in a sealed opaque envelope when a new participant is due to commence testing. The researcher has been instructed to open the envelope only when the child finishes the baseline assessment.

The sequence generation was produced by the first author and project PI. Participants are enrolled by the Experimenter 1 (training/control implementation). Experimenter 1 does not administer the baseline and post-test outcome measures, but during these, he attends to the cameras recording task administration. In doing so, he controls the camera remotely through a laptop while sitting outside the testing room.

To ensure blinding, Experimenter 2 does not contribute to the administration of the training/control procedures, nor is he involved in coding or data analyses concerning training and control procedures. We inform parents that we will not tell them in which group their child is allocated. Parents are thus intended to remain blind to group allocation. Parents, however, keep their infant on their lap while the infant takes part in the training or control procedure, which raises the possibility that parents may identify whether the games are part of the training (i.e. they are responding in real time to the infant’s gaze) or the control procedure (i.e. games are not interactive). Based on experience from previous studies, we estimate it is unlikely parents will recognise the infant’s group allocation. Furthermore, while keeping the infant on their lap parents cannot monitor the infants’ gaze direction (i.e. they may not have key information to detect whether the games are interactive or not). However, at the end of the study, we ask parents to indicate whether they thought they had recognised to which study arm the child had been allocated, and in which study arm they thought their child was.