Research
How we measure, what we measure
Our work spans two questions: what tools we use to quantify infant movement, and which outcomes we quantify with them.
All of our measurement happens through naturalistic home observation — we go to the infant's home rather than bringing families into a lab, so what we capture reflects how infants actually move in everyday life. Within that setting, we use three complementary techniques:
Methods
Video-based coding
Frame-by-frame behavioral coding of naturalistic home video — our core measurement approach for over a decade.
Sensor-based measurement
Wearable IMU sensors combined with a dynamic-threshold algorithm to detect and quantify movement automatically.
AI-assisted classification
Deep neural network models that classify infant posture directly from video, reducing manual coding burden at scale.