摘要目的 通过记录恒河猴颞叶癫痫模型癫痫发作前的丘脑前核(ANT)电信号变化,运用优化后的GOFA算法,实现对癫痫临床发作的检测.方法 以10只健康成年雄性恒河猴作为实验对象,通过立体定向手段放置ANT深部电极,采集正常ANT深部电信号作为基线信号.制备恒河猴颞叶癫痫动物模型,采集、分析双侧ANT电信号,应用改进后的GOFA算法对采集到的ANT电信号进行数学分析,从准确性、特异性、敏感性和检测延迟时间等方面对检测方法的可靠性进行验证.结果 运用改进后的GOFA算法对37段、共计452.4 min的ANT局部电信号进行分析,当t=22,D =0.84 s时,平均检测延迟时间55 s,敏感性为93.2％.结论 通过改进的GOFA算法对猴颞叶癫痫模型ANT脑电信号分析,可对癫痫进行预测.根据对癫痫预测的准确率和假阳性率要求不同,对预测算法中的时间阈值T和空间阈值D进行调整,可以实现癫痫预测的个体化方案,为今后RNS的个体化治疗方案提供基础.

abstractsObjective To record the change of anterior thalamic nucleus (ANT) electrical signals before the rhesus monkeys model of temporal lobe epilepsy seizures,use the optimized GOFA algorithm for the prediction of epileptic seizures,and lay the theory foundation for feedback electrical stimulation treatment of intractable epilepsy.Methods ANT deep electrodes were placed in 10 healthy adult male rhesus monkeys by stereotactic method.Normal ANT deep electrical signals were collected as a baseline signal.The rhesus monkeys temporal lobe epilepsy animal models were prepared.The bilateral ANT electrical signals were studied.The mathematical analysis of ANT electrical signals by improved GOFA algorithm was collected.The accuracy,specificity,sensitivity and detection delay time and so on were used to validate the reliability of the detection methods.Results The local electrical signals of ANT of 37 periods and 452.4 minutes were analyzed by the improved GOFA algorithm.According to the different characteristics of every period of EEG signals,different results by different space threshold T and time threshold D were collected.The experiment results showed that when t =22,D =0.84 s,the average detection delay time was 55 seconds,the sensitivity was 93.2％.Conclusions Epilepsy could be predicted through the improved GOFA algorithm.The epileptic prediction of individualized programs could be implemented,according to the different requirements for accuracy and false positive rate of epileptic prediction.The foundation for feedback electrical stimulation personalized therapy for the treatment of refractory epilepsy could be founded.