基于ICA的颞叶癫痫缺省模式网络的研究

STUDY ON THE "DEFAULT MODE NETWORK" IN PATIENTS WITH TEMPORAL LOBE EPILEPSY BASED ON ICA

很多fMRI研究表明部分癫痫患者缺省模式网络存在中断现象,但均采用广义线性模型的假设驱动方法。作者尝试运用独立成分分析(independent component analysis,ICA)分离出15例单侧颞叶癫痫(temporal lobe epilepsy,TLE)患者和17例正常对照的缺省模式网络,并采用拟合度值(goodness-of-fit scores)分析对感兴趣成分进行挑选,将其结果进行组内分析和组间分析。结果表明颞叶癫痫患者的缺省模式网络大部分区域功能连接度下降,以前额叶和同侧颞上回为著,这可能是由于颞叶癫痫患者的大脑功能内源性组织发生破坏所致。拟合度值下降表明缺省模式网络激活区域为单侧TLE患者提供了一个灵敏的生物信号特征。

Many functional magnetic resonance imaging （fMRI） studies have revealed the deactivation phenomenon of default mode network in the patients with epilepsy; however, they were on the basis of general linear model （GLM）. In this paper, independent component analysis （ICA） was used to isolate the default mode network of 15 patients with temporal lobe epilepsy （TLE） and 17 healthy controls, and a goodness-of-fit analysis applied at the individual subject level to choose the interesting component. Intra-group analysis and inter-group analysis were performed. The results indicated the TLE patients showed decreased connectivity in most of the default model regions with the predominance in the prefrontal lobes and ipsilateral superior temporal gyrus. It suggested that the intrinsic organization of the brain function might be disrupted in TLE. In addition, the decrease of fitness-of-good scores suggests that activity in the default mode network may ultimately prove a sensitive biomarker for TLE.