Towards automatic assessment of atypical early motor development?

Abstract

Atypical motor development is an early indicator for several neurodevelopmental conditions, including cerebral palsy and Rett Syndrome, prompting early diagnosis and intervention. While not currently part of the diagnostic criteria for other conditions like Autism Spectrum Disorder, the frequent retrospective diagnosis of motor impairments alongside these conditions highlights the necessity of a deeper understanding of the relations between motor and cognitive development. Traditional clinical assessments, while considered the gold standard, rely on movement characteristics discernible to the trained eye of professionals. The emergence of automated technologies, including computer vision and wearable sensors, promises more objective and scalable detections. However, these methods are not without challenges, including concerns over data quality, generalizability, interpretability, and ethics. By reviewing recent advances, we highlight the potential and the challenges of integrating automated detections into research and clinical practice. While we agree that these technologies can revolutionize pediatric care, we believe their use must be tempered with caution and supported by clinical expertise to ensure effective outcomes.