Automated Seizure Detection for Home Seizure Monitoring with Epilog Sensors

摘要 Aura是一个使用Epilog™的数字健康平台，Epilog ™是一种微型、无线、可穿戴的EEG传感器， 发际线以下的头皮，可以记录临床和亚临床发作。在初步诊断为癫痫后， 癫痫病学家将使用有关患者癫痫发作的已知信息来指导最佳头皮位置放置 Epilog EEG传感器（A）。EEG数据连续传输（B）到个人智能手机上的Aura应用程序（C） 使用安全的蓝牙TM，通过WiFi（D）与Aura云平台（E）通信。Epilog EEG是 分析癫痫发作，并与癫痫病学家（F）共享每日数字癫痫发作日记，并将其推回 Aura app（G）. Epilog每天充电，可重复使用一年。Epilog的设计非常谨慎， 在日常生活的各个方面持续使用。 数据是一个10秒的片段开始的局灶性癫痫与运动障碍和完整的意识（ILAE 图1A 1）从放置在左前额上的Epitel的单通道Epilog传感器记录。患者入院 视频脑电图监测作为标准治疗这次癫痫发作由三位癫痫病学家独立证实。 在第一阶段，基于机器学习的自动化癫痫发作检测算法将被设计为首先在 先兆云在Epilog EEG中检测癫痫发作，包括一个人可能没有意识到的癫痫发作 有（>50%的癫痫发作），如在睡觉时。Aura将运行这些专门为 Epilog的单通道EEG提供每日数字化癫痫日记。 在第二阶段，Aura系统将作为一种基于EEG的自动化系统进入FDA批准的临床验证试验。 家庭癫痫检测和警报系统。在第二阶段的早期，Aura将作为一种医疗器械商业化。 设备支持服务的商业模式。癫痫患者的自付费用平均为 380元/年有了长期的、可重复的脑电图，癫痫学家现在将有一个更精确的、定量的 记录癫痫发作次数，使他们能够更迅速、更成功地调整患者治疗， 生活质量Aura将使癫痫患者恢复正常生活。Aura为你提供了确定性 在整个第二阶段，生理、心理、行为和环境因素 将在Aura应用程序中结合起来，从300名患者那里收集27,000天的多模态数据， 前所未有的数据集的特征已知的沉淀癫痫发作。 这些数据将用于第三阶段，以创建一个强大的，可穿戴的癫痫发作预测系统， 智能结合多模态癫痫发作诱发因素，创造一个小时的癫痫发作概率。 Aura将深刻地颠覆癫痫的管理方式，并改善癫痫患者的生活质量。

ABSTRACT Aura is a digital health platform that uses Epilog™, a miniature, wireless, wearable EEG sensor worn on the scalp below hairline that can record clinical and subclinical seizures. After an initial diagnosis of epilepsy, an epileptologist will use known information about patients’ seizures to guide the best scalp location to place the Epilog EEG sensor (A). EEG data is continuously transferred (B) to the Aura app on a person’s smartphone (C) using secure BluetoothTM where it communicates over WiFi (D) to the Aura cloud platform (E). Epilog EEG is analyzed for seizures and a daily digital seizure diary is shared with epileptologists (F) and pushed back to the Aura app (G). Epilog is recharged daily, and reusable for a year. Epilog is designed to be discreet, allowing for continuous use in all facets of daily life. Data are a 10 s snippet of the beginning of a focal seizure with motor impairment and intact awareness (ILAE 1A1) recorded from Epitel’s single-channel Epilog sensor placed on the left forehead. The patient was admitted for video-EEG monitoring as standard-of-care. This seizure was verified independently by three epileptologists. In Phase I, automated, machine learning-based seizure detection algorithms will be designed to first work in the Aura cloud to detect seizures in Epilog EEG, including seizures a person may not consciously know they are having (>50% of all seizures), such as while sleeping. Aura will run these algorithms developed exclusively for Epilog’s single-channel of EEG to provide a daily digital seizure diary. In Phase II, the Aura system will enter clinical validation trials for FDA clearance as an EEG-based automated home seizure detection and alerting system. Early in Phase II Aura will be commercialized as a medical device-enabled-service business model. Out-of-pocket costs for a person living with epilepsy is an average of $380/year. Armed with long-term, reproducible EEG, epileptologists will now have a more precise, quantitative record of seizure counts, enabling them to adapt patient treatment more rapidly and successfully to improve quality of life. Aura will give people living with epilepsy their lives back. Aura provides certainty where you are and when you need it. Throughout Phase II, physiological, psychological, behavioral, and environmental factors will be combined in the Aura app to collect 27,000 days of multi-modal data from 300 patients to create an unprecedented dataset of features known to precipitate seizures. These data will be used in Phase III to create a robust, wearable seizure forecasting system using artificial intelligence that combines multi-modal seizure precipitating factors, creating an hourly seizure probability. Aura will profoundly disrupt how epilepsy is managed and improve the quality of life of people living with epilepsy.