Time-frequency signal processing for seizure prediction in epileptic patients

目前近30万加拿大人患有癫痫，每年约有1.4万新病例被诊断出来。对许多患者来说，癫痫的发作是突然的、剧烈的，没有先兆迹象，可以通过药物或其他手段进行预防性干预。在精神健康的这一主要领域，只有当汇集各种科学、技术和医学专业的研究项目得到高度优先考虑时，才能增加对活跃的大脑功能(或功能障碍)的了解，就像在神经科学的其他领域，如衰老一样。本研究项目提出了一种用于预防癫痫发作的颅内脑电分析方法。其目标是能够准瞬时地从脑电信号中提取和量化尽可能多的信息，以检测癫痫发作的预警迹象。我们的方法不是试图绘制涉及复杂网络中相互连接的数十亿个神经元的基本过程，而是依赖于现代信号和信息处理工具，并实施借鉴复杂系统科学的方法和技术。使用真实数据进行评估，这一嵌入软件的方法可在互联网上获得，以便在研究界广泛传播，目的是提供一个能够快速而准确地分析的脑电信号处理平台，从而能够对癫痫发作进行早期预警。

Nearly 300,000 Canadians currently suffer from epilepsy and about 14,000 new cases are diagnosed each year. For many patients, the onset of seizures is sudden, violent and without forewarning signs to allow preventive intervention by either medicinal or other means. In this major area of mental health, increased understanding of active brain functioning (or malfunctioning) will be acquired only when research programs bringing together a variety of scientific, technological and medical specializations are given high priority, as is the case in other areas of neuroscience, such as aging. This research project proposes a methodology for intracranial EEG (electro encephalography) analysis aimed at seizure prevention. The goal is to be able to extract and quantify quasi-instantaneously as much information as possible from EEG signals to detect forewarning signs of seizures. Rather than attempting to map elementary processes involving billions of neurons interconnected in complex networks that defy any exhaustive analysis, our methodology relies on modern signal and information processing tools and implements approaches and techniques borrowed from the science of complex systems. Evaluated using real data, this software-embedded methodology, available on Internet for a large diffusion in the research community, aims to provide an EEG signal processing platform capable of quick and precise analysis allowing early warning of the onset of seizures.