Developing a Methodology for Quantifying Infant Interaction with a Robotic Toy

Preterm infants are at increased risk for developmental delays and disability such as cerebral palsy. When these delays are detected early, movement training can significantly improve patients' quality of life. Instrumented toys have the potential to solve the pressing need for easily accessible, inexpensive diagnostic tools for infant development. While many studies have focused on developing toys for sensor-based assessment, the body of work dedicated to developing quantitative measurements from this data is minimal and has several limitations. Using a consistent and novel video coding methodology, a case study of $\text{1 8}$ preterm and full-term infants' videotaped interactions with the Play and Neuro Development Assessment (PANDA) Gym, an affordable, mechatronic, and sensor-based play environment, were analyzed. The coding methodology was able to quantify infant interactions with the PANDA toy and differences between preterm and full-term infants, but future analysis of involuntary interactions and a larger sample size are needed to determine whether results are consistent with expected developmental milestones.