LMS News

2005-07-31
Philip Warrick to host a poster session at the international joint conference on neural networks (IJCNN)

Montreal, Quebec, July 31, 2005 – Philip Warrick, a Senior Medical Research Engineer for LMS Medical Systems, a healthcare technology company and developer of the CALM™ system (Computer Assisted Labor Management), will be presenting a poster session at the upcoming International Joint Conference on Neural Networks Conference (IJCNN), July 31 – August 4, 2005 at The Hilton Montreal Bonaventure Hotel, Montreal, QC, Canada.

IJCNN 2005 is organized by the International Neural Network Society and the IEEE Computational Intelligence Society. This annual meeting is the largest international neural network conference in the world and will host a variety of special sessions and a series of post-conference workshops devoted to recent and important developments in neural networks.

Mr. Warrick’s presentation “Neural Network Based Detection of Fetal Heart Rate Patterns” will discuss a recent study on the development of automated detection of Fetal Heart Rate (FHR) patterns that can potentially improve intra-partum care by providing consistent and reliable measures that assist health-care professionals in their assessment of the state of the fetus. Although childbirth is a natural process and outcomes are generally good, approximately 1-7 in 1000 babies experience sufficient oxygen deprivation during labor to cause death or brain injury. Multiple reviews of such cases suggest that around 50% of these injuries are related to preventable medical errors, most often centering on incorrect analysis of the FHR recording. As a result, the goal of his work is to present a system that can provide a better assessment of FHR patterns in order to aid in reducing the incidence and severity of birth related brain damage.

Mr. Warrick will discuss how a software system that incorporates signal processing and neural networks as its engine can detect fetal heart rate patterns. The system applied in this study uses features derived from time-domain analysis and an extensive database of FHR tracings, marked by domain experts, to train neural-network subsystems to classify baselines from non-baselines or decelerations/accelerations from their negative examples. To date, most clinical studies relating FHR patterns to fetal outcomes have been severely limited as they use visual analysis on relatively few cases. Through this system, reliable automated detection now enables the analysis of a large number of cases and the application of modern probabilistic modeling to this clinical challenge.

Mr. Warrick is a Ph.D. student of the Department of Biomedical Engineering at McGill University. He also is a Senior Medical Research Engineer at LMS Medical Systems in Montreal where his work contributes to the development of cutting edge obstetrical tools for the evaluation of labor and fetal tolerance to labor in order to help teams improve outcomes and reduce risk. The CALM tools incorporate statistical processes to quantify normal and abnormal labor progression as well as digital signal processing and neural network applications for the identification of abnormal fetal heart rate patterns. The CALM Curve provides for consistent and objective assessment of labor progress at the bedside. CALM Patterns provides objective, real-time measurement and classification of fetal heart rate patterns. These tools address themes commonly found in adverse outcomes, namely failure to recognize and respond to abnormal tracings and prolonged labor.

For more information about IJCNN 2005 please see: http://faculty.uwb.edu/ijcnn05/