Neonatal seizures are often the first, and, in some situations, the only clinical sign of central nervous system dysfunction in the newborn. Despite the importance of early detection and characterization of neonatal seizures with regard to diagnosis and management of underlying neurological problems, most neonatal intensive care units and nurseries have limited resources for seizure monitoring, detection and characterization. The research outlined in this proposal is the first attempt ever to utilize recent advances in video and computer technology toward the development of automated video processing and analysis procedures that can facilitate the characterization and recognition of neonatal seizures. These procedures will rely on quantitative information regarding the behavioral characteristics of neonatal seizures, which will be extracted from videotaped neonatal seizures in the form of temporal motion strength and motor activity signals. The proposed research is expected to produce novel computational tools for extracting quantitative information from image sequences, which may be utilized to support diagnosis and extend human analysis. The automated video processing and analysis procedures developed in this project will be evaluated and tested on an existing library of videotaped clinical events, which include neonatal seizures and normal or abnormal infant behaviors not due to seizures. The long-term goal of this research is the integration of the proposed video processing and analysis procedures into the development of a stand-alone automated seizure detection and characterization system that could be used as a supplement in the neonatal intensive care unit to: 1) provide 24-hour a day noninvasive monitoring of infants at risk for seizures, and 2) facilitate the analysis and characterization of videotaped neonatal seizures by physicians during retrospective review.
