Abstract：
OBJECTIVE: Temporal lobe epilepsy (TLE) patients often exhibit varying degrees of cognitive impairments. This study aims to predict cognitive performance in TLE patients by applying a connectome-based predictive model (CPM) to whole-brain resting-state functional connectivity (RSFC) data.
METHODS: A CPM was established and leave-one-out cross-validation was employed to decode the cognitive performance of patients with TLE based on the whole-brain RSFC.
RESULTS: Our findings indicate that cognitive performance in TLE can be predicted through the internal and network connections of the parietal lobe, limbic lobe, and cerebellum systems. These systems play crucial roles in cognitive control, emotion processing, and social perception and communication, respectively. In the subgroup analysis, CPM successfully predicted TLE patients with and without focal to bilateral tonic-clonic seizures (FBCTS). Additionally, significant differences were noted between the two TLE patient groups and the normal control group.
CONCLUSIONS: This data-driven approach provides evidence for the potential of predicting brain features based on the inherent resting-state brain network organization. Our study offers an initial step towards an individualized prediction of cognitive performance in TLE patients, which may be beneficial for diagnosis, prognosis, and treatment planning.
METHODS: A CPM was established and leave-one-out cross-validation was employed to decode the cognitive performance of patients with TLE based on the whole-brain RSFC.
RESULTS: Our findings indicate that cognitive performance in TLE can be predicted through the internal and network connections of the parietal lobe, limbic lobe, and cerebellum systems. These systems play crucial roles in cognitive control, emotion processing, and social perception and communication, respectively. In the subgroup analysis, CPM successfully predicted TLE patients with and without focal to bilateral tonic-clonic seizures (FBCTS). Additionally, significant differences were noted between the two TLE patient groups and the normal control group.
CONCLUSIONS: This data-driven approach provides evidence for the potential of predicting brain features based on the inherent resting-state brain network organization. Our study offers an initial step towards an individualized prediction of cognitive performance in TLE patients, which may be beneficial for diagnosis, prognosis, and treatment planning.
摘要：
目的：颞叶癫痫（TLE）患者常表现出不同程度的认知障碍。这项研究旨在通过将基于连接体的预测模型（CPM）应用于全脑静息状态功能连接（RRFC）数据来预测TLE患者的认知表现。
方法：建立CPM，并基于全脑RSFC采用留一法交叉验证来解码TLE患者的认知表现。
结果：我们的发现表明，TLE的认知表现可以通过顶叶的内部和网络连接来预测，边缘叶,和小脑系统。这些系统在认知控制中起着至关重要的作用，情绪处理，社会认知和沟通，分别。在亚组分析中，CPM成功预测了有或没有双侧强直阵挛性癫痫发作（FBCTS）的TLE患者。此外,两个TLE患者组与正常对照组之间存在显著差异.
结论：这种数据驱动的方法为基于固有的静息状态大脑网络组织预测大脑特征的潜力提供了证据。我们的研究为TLE患者的认知表现提供了个性化预测的第一步，这可能对诊断有益，预后,和治疗计划。
方法：建立CPM，并基于全脑RSFC采用留一法交叉验证来解码TLE患者的认知表现。
结果：我们的发现表明，TLE的认知表现可以通过顶叶的内部和网络连接来预测，边缘叶,和小脑系统。这些系统在认知控制中起着至关重要的作用，情绪处理，社会认知和沟通，分别。在亚组分析中，CPM成功预测了有或没有双侧强直阵挛性癫痫发作（FBCTS）的TLE患者。此外,两个TLE患者组与正常对照组之间存在显著差异.
结论：这种数据驱动的方法为基于固有的静息状态大脑网络组织预测大脑特征的潜力提供了证据。我们的研究为TLE患者的认知表现提供了个性化预测的第一步，这可能对诊断有益，预后,和治疗计划。
