Frequent, Scalable and Global Use of 'Intelligent Onesies' to Quantify Newborns' Spontaneous Movements in Natural Settings

Early identification of atypical neurological function in newborns is essential for the timely detection of cerebral palsy and other neurodevelopmental conditions such as Autistic Spectrum Disorder. Traditionally, assessing infants' spontaneous general movements provides a clinical standard for early detection. Such assessment requires specialized training and periodic in-person observation-conditions that can be difficult to fulfil for many families. Recent technological advances in wearable inertial sensors raise the exciting possibility of an automated, home-based protocol for detection of motor delays and their cascading consequences that could increase reach and reduce costs. In this project, we examined the feasibility of testing newborns' spontaneous movements in high frequency and in their natural setting. To that end, we created a wearable 'intelligent onesie' equipped with inertial measurement units. The data collections were administered by caregivers and were monitored remotely using secure tunnelling to local Raspberry-Pi devices. We detail the setup and methodological challenges encountered-from shipping the home recording kits to ensuring adequate sensor placement, data syncing, calibration, and retrieval-and how we overcame them. We demonstrate how to collect large-scale and global data that capture newborn spontaneous movements in home environments. We discuss implications for leveraging automated assessment for early intervention pathways and future directions for scalable infant neurodevelopmental screening.