Safi-Harb M, Salam M Tariqus, Mirabbasi S, Nguyen D K, Sawan M
Polystim Neurotechnologies Laboratory, École Polytechnique de Montréal, Montréal, Canada.
Annu Int Conf IEEE Eng Med Biol Soc. 2011;2011:5847-50. doi: 10.1109/IEMBS.2011.6091446.
In this paper, we present a new seizure detection algorithm and the associated CMOS circuitry implementation. The proposed low-power seizure detector is a good candidate for an implantable epilepsy prosthesis. The device is designed for patient-specific seizure detection with a one variable parameter. The parameter value is extracted from a single seizure that is subsequently excluded from the validation phase. A two-path system is also proposed to minimize the detection delay. The algorithm is first validated using MATLAB® tools and then implemented and validated using circuits designed in a standard 0.18-μm CMOS process with a total power dissipation of 7.08 μW. A total of 13 seizures from two drug-resistant epileptic patients are assessed using the proposed algorithm and resulted in 100% sensitivity and a mean detection delay of 9.7 s after electrical onset.
在本文中，我们提出了一种新的癫痫发作检测算法及其相关的CMOS电路实现。所提出的低功耗癫痫发作检测器是植入式癫痫假体的理想选择。该设备专为针对特定患者的癫痫发作检测而设计，具有一个可变参数。该参数值从单个癫痫发作中提取，随后在验证阶段被排除。还提出了一种双路径系统以最小化检测延迟。该算法首先使用MATLAB®工具进行验证，然后使用采用标准0.18μm CMOS工艺设计的电路进行实现和验证，总功耗为7.08μW。使用所提出的算法对两名耐药性癫痫患者的总共13次癫痫发作进行了评估，结果显示灵敏度为100%，电发作后平均检测延迟为9.7秒。
