High Performance Seizure Monitoring and Alert System

描述(由申请人提供)：患者经常因癫痫或急性神经损伤(如创伤、中风、感染和一些毒性和代谢障碍)导致的失控癫痫而住院治疗。然而，由于癫痫发作是间歇性的和不可预测的，因此患者的癫痫发作没有被发现的情况并不少见，这使得住院治疗癫痫的治疗变得复杂。这可能会导致不必要的住院时间延长，或者更糟糕的是，延误治疗和不可逆转的脑损伤。因此，迫切需要开发一套准确的床边癫痫监测与报警系统。该SBIR项目的总体目标是将一种准确、可靠、用户友好的基于EEG的癫痫监测和警报(EEGSMA)系统商业化，用于患者需要密切神经监测的临床环境中。预期的临床环境包括但不限于癫痫监测病房(EMU)、重症监护病房(ICU)、急诊科(ED)以及神经科和神经外科患者的普通护理病房。Optima神经科学的研究人员开发了一种自动化算法，通过分析头皮EEG信号的时空模式来准确检测癫痫发作。该算法被整合到我们的IdentEvent“癫痫检测软件中，该软件于2009年10月16日获得FDA批准。在该SBIR项目的一期工程中，我们进一步完成了SMA系统的床边硬件设计，并将IdentEvent进行了实时应用。SMA系统在模拟实时模式下测试成功，初步完成了临床可行性测试。在这个第二阶段的应用中，我们建议继续对SMA系统进行临床测试，并扩展EEGSMA系统的功能，以用于急性护理环境，例如ICU和急诊室。为了实现这一目标，我们不仅必须开发一个可靠、便携、用户友好、可以快速设置的脑电采集模块，而且我们还需要扩展IdentEvent，使其适用于儿童和ICU患者。因此，该应用的具体目标是：(1)在动车组环境下完成SMA模块的临床性能评估；(2)设计和测试EEG头盒的前端硬件和软件组件并将其与SMA系统集成；(3)完成集成EEGSMA系统的临床前测试和门诊中试研究；(4)在动车组中对集成的EEGSMA系统进行住院测试；(5)在动车组中测试检测算法(IdentEvent)；以及(6)进一步开发适用于ICU患者的癫痫检测算法。这种EEGSMA设备的成功商业化将通过允许检测间歇性和以前被误诊的事件来改善住院患者对癫痫发作的管理。 公共卫生相关性：尽管对关键心肺功能的自动监测是所有医院的标准护理，但目前对大脑功能的监测几乎完全依赖于床边临床观察。结果，每天都有大量的亚临床癫痫(只有微小的可观察到的变化)得不到诊断。该项目的主要目标是建立和测试一个非常需要的自动化系统的原型，以提醒未受过神经学培训的工作人员癫痫发作活动的存在。总体目标是改善癫痫患者的诊断和治疗，特别是在社区医院，那里可能没有经过脑电培训的神经科医生。

DESCRIPTION (provided by applicant): Patients are frequently hospitalized for management of uncontrolled seizures due to epilepsy or acute neurological insults such as trauma, stroke, infections, and a number of toxic and metabolic disorders. However, inpatient management of seizures is complicated by the fact that they occur intermittently and unpredictably, and thus it is not infrequent that patients' seizures go unrecognized. This can result in unnecessarily prolonged hospital stays, or worse, delay of treatment and irreversible brain injury. Therefore, there is a great need to develop an accurate bed-side seizure monitroing and alert (SMA) system. The overall goal of this SBIR project is to commercialize an accurate, reliable, and user-friendly EEG-based seizure monitoring and alert (EEGSMA) system for use in clinical settings where patients require close neurological monitoring. Intended clinical settings include but are not limited to, epilepsy monitoring units (EMUs), intensive care units (ICUs), emergency departments (EDs), and general care units for neurology and neurosurgical patients. Researchers at Optima Neuroscience have developed an automated algorithm to accurately detect seizures by analyzing the spatiotemporal patterns of scalp EEG signals. The algorithm was incorporated in our IdentEvent" seizure detection software, which received FDA approval on October 16, 2009. During the Phase I of this SBIR project, we have further completed bed-side hardware design of the SMA system and transformed IdentEvent for real-time application. The SMA system was successfully tested in simulation real-time mode, and initial clinical feasibility testing has been completed. In this Phase II application, we propose to continue the clinical testing for the SMA system as well as expand the functions of the EEGSMA system for use in acute care environments, e.g., ICUs and EDs. To accomplish this, we not only have to develop a reliable, portable, and user-friendly EEG acquisition module that can be set up quickly, but also we need to expand IdentEvent for use in children and for ICU patients. Therefore, the specific aims of this application are: (1) to complete the clinical performance evaluation of the SMA module in an EMU setting, (2) to design and test front end hardware and software components of the EEG head-box and integrate them with the SMA system, (3) to complete pre-clinical testing and pilot outpatient study of the integrated EEGSMA system, (4) to conduct inpatient testing of the integrated EEGSMA system in EMUs, (5) to test the detection algorithm (IdentEvent) on pediatric patients (age 3 ~ 17) in EMUs, and (6) to further develop a seizure detection algorithm for ICU patients. Successful commercialization of this EEGSMA device will improve inpatient management of seizures by allowing for detection of intermittent and previously misdiagnosed events. PUBLIC HEALTH RELEVANCE: Although automated monitoring for critical heart and lung function is the standard of care in all hospitals, monitoring the function of the brain currently relies almost exclusively upon bedside clinical observations. As a result, a large number of subclinical seizures (only subtle observable changes) go undiagnosed every day. The primary goal of this project is to build and test a prototype for a greatly needed automated system to alert staff untrained in neurology to the presence of seizure activities. The overall goal is to improve the diagnosis and treatment of patients suffering from seizure disorders, particularly in community hospitals where EEG trained neurologists may not be available.