1Harvard Medical School, Boston, Massachusetts.
2Department of Neurosurgery, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts.
J Neurosurg. 2024 May 24;141(4):1115-1123. doi: 10.3171/2024.3.JNS2441. Print 2024 Oct 1.
Interictal epileptiform discharges (IEDs) are intermittent high-amplitude electrical signals that occur between seizures. They have been shown to propagate through the brain as traveling waves when recorded with epicortical grid-type electrodes and small penetrating microelectrode arrays. However, little work has been done to translate experimental IED analyses to more clinically relevant platforms such as stereoelectroencephalography (SEEG). In this pilot study, the authors aimed to define a computational method to identify and characterize IEDs recorded from clinical SEEG electrodes and leverage the directionality of IED traveling waves to localize the seizure onset zone (SOZ).
Continuous SEEG recordings from 15 patients with medically refractory epilepsy were collected, and IEDs were detected by identifying overlapping peaks of a minimum prominence. IED pathways of propagation were defined and compared to the SOZ location determined by a clinical neurologist based on the ictal recordings. For further analysis of the IED pathways of propagation, IED detections were divided into triplets, defined as a set of 3 consecutive contacts within the same IED detection. Univariate and multivariate linear regression models were employed to associate IED characteristics with colocalization to the SOZ.
A median (range) of 22.6 (4.4-183.9) IEDs were detected per hour from 15 patients over a mean of 23.2 hours of recording. Depending on the definition of the SOZ, a median (range) of 20.8% (0.0%-54.5%) to 62.1% (19.2%-99.4%) of IEDs per patient traversed the SOZ. IEDs passing through the SOZ followed discrete pathways that had little overlap with those of the IEDs passing outside the SOZ. Contact triplets that occurred more than once were significantly more likely to be detected in an IED passing through the SOZ (p < 0.001). Per our multivariate model, patients with a greater proportion of IED traveling waves had a significantly greater proportion of IEDs that localized to the SOZ (β = 0.64, 95% CI 0.01-1.27, p = 0.045).
By using computational methods, IEDs can be meaningfully detected from clinical-grade SEEG recordings of patients with epilepsy. In some patients, a high proportion of IEDs are traveling waves according to multiple metrics that colocalize to the SOZ, offering hope that IED detection, with further refinement, could serve as an alternative method for SOZ localization.
发作间期癫痫样放电（IEDs）是发作间期间歇性高振幅电信号。当使用皮质网格型电极和小穿透微电极阵列记录时，已证明它们以游走波的形式在大脑中传播。然而，很少有工作将实验性 IED 分析转化为更具临床相关性的平台，如立体脑电图（SEEG）。在这项初步研究中，作者旨在定义一种计算方法来识别和描述从临床 SEEG 电极记录到的 IED，并利用 IED 游走波的方向性来定位癫痫发作起始区（SOZ）。
连续采集 15 例药物难治性癫痫患者的 SEEG 记录，通过识别最小突出度的重叠峰来检测 IED。定义 IED 传播途径，并将其与临床神经病学家根据发作记录确定的 SOZ 位置进行比较。为了进一步分析 IED 传播途径，将 IED 检测分为三组，定义为同一 IED 检测中连续 3 个接触点的集合。采用单变量和多变量线性回归模型将 IED 特征与 SOZ 的共定位相关联。
15 例患者平均记录 23.2 小时，每小时检测到中位数（范围）为 22.6（4.4-183.9）个 IED。根据 SOZ 的定义，中位数（范围）为 20.8%（0.0%-54.5%）至 62.1%（19.2%-99.4%）的患者每个 IED 穿过 SOZ。穿过 SOZ 的 IED 遵循离散途径，与穿过 SOZ 之外的 IED 途径几乎没有重叠。在同一患者中多次出现的接触组对发生在穿过 SOZ 的 IED 中的检测更为重要（p < 0.001）。根据我们的多变量模型，具有较大比例 IED 游走波的患者具有明显更大比例的 IED 定位于 SOZ（β=0.64，95%CI 0.01-1.27，p=0.045）。
通过使用计算方法，可以从癫痫患者的临床级 SEEG 记录中对 IED 进行有意义的检测。根据与 SOZ 共定位的多个指标，一些患者的 IED 中存在较大比例的游走波，这为进一步改进 IED 检测可能作为 SOZ 定位的替代方法提供了希望。
