用于难治性癫痫发作预测的电阻抗成像新方法研究

Surgical resection of the epileptogenic focus is the main treatment for refractory epilepsy. For patients in whom resection is not feasible, or in whom resection has failed, more effective treatment is needed. It is of great significance to relieve the patients' pain and cure the epileptic seizure if prediction can be made before seizure and intervention measures can be taken in advance to inhibit the upcoming epileptic seizures. In view of the shortcomings of the existing seizure prediction methods, based on the previous work which realizing seizure monitoring by electrical impedance imaging, this study proposes a new method for seizure prediction by electrical impedance imaging. It aims at two important scientific problems: the extraction of electrical impedance characteristic parameters during preictal period and real-time seizure prediction based on electrical impedance characteristic parameters. Real-time prediction of refractory epileptic seizures in rats was realized by designing an acquisition system that can collect electrical impedance signals and EEG signals simultaneously, extracting the characteristic parameters of electrical impedance during preictal period and establishing discriminant analysis method for seizure prediction. This study lays a foundation for further clinical application. Also, it is expected to open a new research field for electrical impedance imaging study, as well as to provide new ideas and methods for the refractory epilepsy prediction, which has important scientific and clinical value.

难治性癫痫的主要治疗方式是手术切除病灶，但对于手术也无效的患者，尚缺乏有效的治疗手段。若能在癫痫发作前进行预测，并提前采取干预措施抑制即将到来的癫痫发作，对于缓解患者病痛以及治愈癫痫发作具有重大意义。针对现有癫痫预测方法存在的不足，本项目在前期已实现癫痫发作期电阻抗监测的基础上，提出基于电阻抗成像的癫痫发作预测新方法。主要围绕基于电阻抗成像的癫痫发作前期判定及其电阻抗特征参数提取、以及实现基于癫痫发作前期电阻抗特征参数的判别分析预测方法两大关键问题，从设计能够同步采集电阻抗信号与脑电信号的采集系统、提取癫痫发作前期电阻抗特征参数、建立预测癫痫发作的判别分析法等关键环节入手，实现对大鼠难治性癫痫发作实时预测，为进一步实现临床应用奠定基础。本项目有望开辟电阻抗成像研究新领域，以及为难治性癫痫发作预测提供新思路和新方法，具有重要的科学和临床价值。

难治性癫痫的主要治疗方式是手术切除病灶，但对于手术也无效的患者，尚缺乏有效的治疗手段。若能在癫痫发作前进行预测，并提前采取干预措施抑制即将到来的癫痫发作，对于缓解患者病痛以及治愈癫痫发作具有重大意义。.针对现有癫痫预测方法存在的不足，本项目提出基于电阻抗断层成像技术（EIT）的癫痫发作预测新方法，利用EIT检测动物癫痫发作期以及发作间期的电阻抗特征变化，并提取癫痫发作前期的电阻抗特征参数进行癫痫发作预测研究。项目选用海人酸点燃法制备了急、慢性SD大鼠难治性癫痫模型，通过脑电癫痫特征波判断癫痫模型是否制备成功，以及记录特征波形、发作持续时间、发作期行为学等癫痫发作特征。研发了EIT-EEG同步采集系统，并基于硬件低通滤波器和软件梳状滤波器方法抑制了同步监测过程中EIT对EEG的干扰，实现了电阻抗与脑电的长时间同步监测癫痫发作，为准确提取与癫痫发作相关的电阻抗特征奠定了技术支撑。研究了癫痫发作期EIT感兴趣区域的电阻抗变化趋势，提取了癫痫发作期电阻抗先下降后上升的变化特征，并发现电阻抗下降时刻早于脑电癫痫发作开始时刻，证实了基于EIT预测癫痫发作具有可行性。实现了基于Higuchi和Hurst指数时间序列分型方法提取癫痫发作前期以及发作期的电阻抗特征参数，建立了基于主成分分析法和马氏距离的癫痫发作预测方法，突破了现有癫痫预测方法存在的不足，为难治性癫痫发作预测提供了新思路和新方法，同时为进一步实现临床应用奠定了基础。