SBIR Phase II: Information Theoretic Learning and Application to Fetal ECG

这个小企业创新第二阶段项目将开发信息理论方法，根据混合物的统计特性(盲源分离)将胎儿心电信号从母亲腹部的嘈杂电环境中分离出来。分离是使用最近引入的一种算法(美人鱼)完成的，该算法在计算和数据上都很高效。第一阶段的研究表明，美人鱼是对以往胎儿心电分离方法的显著改进。该项目将在一个非常紧凑的设备中开发全面的胎儿和产妇监测器技术，包括胎儿心率、胎儿心电和产妇宫电(EHG，测量宫缩信息)。该项目包括临床研究，旨在提供创建和验证NeuroDimension系统所需的信息，并说明其有效性。潜在市场包括基于医院的胎儿监测、家庭/医生办公室的胎儿监测和压力测试，以及用作研究工具。该监护仪不仅将比目前的监护仪更便宜，而且还将提供额外的信息，通过避免不必要的程序来显著改善患者护理并降低成本。

This Small Business Innovation Phase II Project will develop information theoretic methods to separate fetal electrocardiogram (FECG) signals from the noisy electrical environment of the maternal abdomen based on statistical properties of the mixtures (blind source separation). The separation is done using a recently introduced algorithm (Mermaid) that is computationally and data efficient. Phase I research showed that Mermaid is a marked improvement over prior methods of FECG separation. The project will develop the technology for a comprehensive fetal and maternal monitor including fetal heart rate, FECG, and maternal Electrohysterogram (EHG, which measures contraction information) in a very compact device. The project includes clinical studies designed to provide the information necessary to create and validate NeuroDimension's system and also to illustrate its effectiveness. Potential markets include hospital-based fetal monitoring, home/physician's office fetal monitoring and stress tests, and use as a research tool. The monitor not only will be less expensive than current monitors, but also will provide additional information that can dramatically improve patient care and reduce costs by avoiding unnecessary procedures.