Zhu Min, Zhang Wenbo, Dickens Deanna, Ding Lei
School of Electrical and Computer Engineering, University of Oklahoma, Norman, OK 73019, USA.
Annu Int Conf IEEE Eng Med Biol Soc. 2012;2012:6744-7. doi: 10.1109/EMBC.2012.6347542.
The aim of the present study is to evaluate the capability of a recently proposed l(1)-norm based regularization method, named as variation-based sparse cortical current density (VB-SCCD) algorithm, in estimating location and spatial coverage of extensive brain sources. Its performance was compared to the conventional minimum norm estimate (MNE) using both simulations and clinical interictal spike MEG data from epilepsy patients. Four metrics were adopted to evaluate two regularization methods for EEG/MEG inverse problems from different aspects in simulation study. Both methods were further compared in reconstructing epileptic sources and validated using results from clinical diagnosis. Both simulation and experimental results suggest VB-SCCD has better performance in localization and estimation of source extents, as well as less spurious sources than MNE, which makes it a promising noninvasive tool to assist presurgical evaluation for surgical treatment in epilepsy patients.
本研究的目的是评估一种最近提出的基于l(1)范数的正则化方法，即基于变分的稀疏皮质电流密度（VB-SCCD）算法，在估计广泛脑源的位置和空间覆盖范围方面的能力。使用模拟和癫痫患者的临床发作间期棘波脑磁图（MEG）数据，将其性能与传统的最小范数估计（MNE）进行了比较。在模拟研究中，采用了四个指标从不同方面评估两种用于脑电/脑磁图逆问题的正则化方法。在重建癫痫源方面对两种方法进行了进一步比较，并使用临床诊断结果进行了验证。模拟和实验结果均表明，与MNE相比，VB-SCCD在源定位和范围估计方面具有更好的性能，并且伪源更少，这使其成为协助癫痫患者手术治疗术前评估的一种有前景的非侵入性工具。
