1Department of Neurological Surgery and.
2Vanderbilt University Institute of Imaging Science, Vanderbilt University Medical Center.
J Neurosurg. 2022 Apr 1;137(6):1571-1581. doi: 10.3171/2022.1.JNS212031. Print 2022 Dec 1.
The most common surgically treatable epilepsy syndrome is mesial temporal lobe epilepsy (mTLE). Preoperative noninvasive lateralization of mTLE is challenging in part due to rapid contralateral seizure spread. Abnormal connections in both the mesial temporal lobe and resting-state networks have been described in mTLE, but it is unclear if connectivity between these networks may aid in lateralization.
In 52 patients with left mTLE (LmTLE) or right mTLE (RmTLE) and 52 matched control subjects, the authors acquired 20 minutes of resting-state functional MRI (fMRI) and evaluated functional connectivity of bilateral hippocampi and amygdalae with selected resting-state networks. They used Pearson correlation, network-based statistic, and dynamic causal modeling. Also, to evaluate the clinical utility of a resting-state connectivity model in lateralizing unilateral presurgical mTLE patients, they used receiver operating characteristic curve analysis.
RmTLE patients demonstrated decreased nondirected connectivity between the right hippocampus and default mode network compared with LmTLE patients and control subjects. Network-based statistic analysis revealed that the network with most decreased connectivity that distinguished LmTLE from RmTLE patients included the right hippocampus and amygdala, right lateral orbitofrontal cortices, and bilateral inferior parietal lobules, precuneus, and medial orbitofrontal cortices. Dynamic causal modeling analysis revealed that cross-hemispheric connectivity between hippocampi and amygdalae was predominantly inward toward the epileptogenic side. A regression model incorporating these connectivity patterns was used to accurately lateralize mTLE patients with an area under the receiver operating characteristic curve of 0.87.
Evaluating fMRI connectivity between mesial temporal structures and default mode network may aid in mTLE lateralization, reduce need for intracranial monitoring, and guide surgical planning.
最常见的可手术治疗的癫痫综合征是内侧颞叶癫痫（MTLE）。由于对侧癫痫快速传播，MTLE 的术前非侵入性定位具有挑战性。在 MTLE 中已经描述了内侧颞叶和静息态网络中的异常连接，但尚不清楚这些网络之间的连接是否有助于定位。
作者在 52 例左侧 MTLE（LmTLE）或右侧 MTLE（RmTLE）患者和 52 例匹配的对照受试者中获得了 20 分钟的静息态功能磁共振成像（fMRI），并评估了双侧海马和杏仁核与选定的静息态网络的功能连接。他们使用 Pearson 相关、基于网络的统计和动态因果建模。此外，为了评估静息态连接模型在单侧术前 MTLE 患者定位中的临床效用，他们使用了接收者操作特征曲线分析。
与 LmTLE 患者和对照组相比，RmTLE 患者右侧海马与默认模式网络之间的无定向连接减少。基于网络的统计分析显示，区分 LmTLE 和 RmTLE 患者的最具降低连接性的网络包括右侧海马和杏仁核、右侧外侧眶额皮质和双侧下顶叶、楔前叶和内侧眶额皮质。动态因果建模分析显示，海马和杏仁核之间的跨半球连接主要向内朝向致痫侧。一个包含这些连接模式的回归模型被用来准确地定位 MTLE 患者，其接收者操作特征曲线下面积为 0.87。
评估内侧颞叶结构与默认模式网络之间的 fMRI 连接可能有助于 MTLE 的定位，减少对颅内监测的需求，并指导手术计划。
