Wyss Center Fellow, Sleep-Wake-Epilepsy Center, Center for Experimental Neurology, NeuroTec, Department of Neurology, Inselspital Bern University Hospital, University of Bern, Bern, Switzerland.
Department of Neurology and Weill Institute for Neurosciences, University of California, University of California, San Francisco, California, USA.
Epilepsia. 2023 Dec;64 Suppl 4:S99-S113. doi: 10.1111/epi.17406. Epub 2022 Sep 22.
Epilepsy is characterized by spontaneous seizures that recur at unexpected times. Nonetheless, using years-long electroencephalographic (EEG) recordings, we previously found that patient-reported seizures consistently occur when interictal epileptiform activity (IEA) cyclically builds up over days. This multidien (multiday) interictal-ictal relationship, which is shared across patients, may bear phasic information for forecasting seizures, even if individual patterns of seizure timing are unknown. To test this rigorously in a large retrospective dataset, we pretrained algorithms on data recorded from a group of patients, and forecasted seizures in other, previously unseen patients.
We used retrospective long-term data from participants (N = 159) in the RNS System clinical trials, including intracranial EEG recordings (icEEG), and from two participants in the UNEEG Medical clinical trial of a subscalp EEG system (sqEEG). Based on IEA detections, we extracted instantaneous multidien phases and trained generalized linear models (GLMs) and recurrent neural networks (RNNs) to forecast the probability of seizure occurrence at a 24-h horizon.
With GLMs and RNNs, seizures could be forecasted above chance in 79% and 81% of previously unseen subjects with a median discrimination of area under the curve (AUC) = .70 and .69 and median Brier skill score (BSS) = .07 and .08. In direct comparison, individualized models had similar median performance (AUC = .67, BSS = .08), but for fewer subjects (60%). Moreover, calibration of pretrained models could be maintained to accommodate different seizure rates across subjects.
Our findings suggest that seizure forecasting based on multidien cycles of IEA can generalize across patients, and may drastically reduce the amount of data needed to issue forecasts for individuals who recently started collecting chronic EEG data. In addition, we show that this generalization is independent of the method used to record seizures (patient-reported vs. electrographic) or IEA (icEEG vs. sqEEG).
癫痫的特征是在不可预测的时间自发发作。尽管如此，我们之前使用多年的脑电图（EEG）记录发现，当间歇性癫痫样活动（IEA）在数天内周期性增加时，患者报告的发作始终会发生。这种多日（多天）的发作间-发作关系在患者之间共享，即使个体发作时间模式未知，也可能为发作预测提供阶段性信息。为了在大型回顾性数据集上严格检验这一点，我们在一组患者的数据上对算法进行了预训练，并在其他以前未见过的患者中预测了发作。
我们使用 RNS 系统临床试验中参与者（N=159）的回顾性长期数据，包括颅内 EEG 记录（icEEG），以及 UNEEG 医疗临床试验中皮下 EEG 系统（sqEEG）中两名参与者的数据。基于 IEA 检测，我们提取瞬时多维相位，并训练广义线性模型（GLM）和递归神经网络（RNN）来预测 24 小时内发作的概率。
使用 GLM 和 RNN，在以前未见过的 79%和 81%的受试者中，发作可以被预测，中位数区分曲线下面积（AUC）=0.70 和 0.69，中位数 Brier 技能评分（BSS）=0.07 和 0.08。直接比较时，个体化模型的表现中位数相似（AUC=0.67，BSS=0.08），但适用的受试者较少（60%）。此外，还可以保持预训练模型的校准，以适应不同的受试者的发作率。
我们的发现表明，基于 IEA 多日周期的发作预测可以在患者之间推广，并且可以大大减少为最近开始收集慢性 EEG 数据的个体发出预测所需的数据量。此外，我们还表明这种推广与用于记录发作的方法（患者报告的 vs. 电描记法）或 IEA（icEEG vs. sqEEG）无关。
