INDEPENDENT COMPONENT ANALYSIS OF SINGLE-TRIAL EVENT-RELATED POTENTIALS
INDEPENDENT COMPONENT ANALYSIS OF SINGLE-TRIAL EVENT-RELATED POTENTIALS
复制标题
DOI：
--
复制
发表时间：
1999
期刊：
影响因子：
--
通讯作者：
Tzyy-Ping Jungl;S. Make;M. Westerfield;J. Townsend
中科院分区：
文献类型：
--
作者：
Tzyy-Ping Jungl;S. Make;M. Westerfield;J. Townsend
事件相关电位（ERP）实验的单次试验由脑电图（EEG）记录的大脑活动时间锁定在实验事件。这些通常在一组相似或相同的事件中平均，以增加相对于非锁相脑电图活动和非脑伪影的信噪比，而不考虑反应活动在时间过程和头皮分布上可能在不同试验中有很大差异的事实。因此，平均可能不适合研究涉及短暂和间歇性受试者认知状态的神经元脑动力学。另一方面，对单个ERP时代的分析虽然是理想的，但由于与眨眼、眼球运动和肌肉噪声相关的显著脑电信号伪影、大量非锁相的背景脑电信号活动以及每次试验中ERP波形的潜伏期和幅度的广泛变化而造成混淆。本研究引入了一种新的可视化工具，即“ERP图像”，用于研究自发性脑电图或脑磁图记录中事件诱发反应的潜伏期和振幅的变异性。其次，我们将一种新的线性分解工具——独立分量分析（ICA） [I]应用于多通道单次脑电记录，得出空间滤波器，将单次脑电时代分解为由不同或重叠的脑或脑外网络产生的时间独立和空间固定的分量之和。我们通过海军研究办公室和霍华德休斯医学研究所的部分资助，证明了建议的分析和可视化工具在单试验ERP分析中的作用。本文仅代表作者个人观点，不代表美国海军部、国防部或美国国防部的官方政策或立场。政府。一个正常受试者和一个自闭症受试者的样本数据集分析。
Single-trials in event-related potential (ERP) experiments consists of electroencephalographic (EEG) recordings of brain activity time-locked to experimental events. These are usually averaged across a set of similar or identical events to increase their signal/noise ratio relative to non-phase locked EEG activity and non-brain artifacts, regardless of the fact that response activity may vary widely across trials in time course and scalp distribution. Averaging thus may not be suitable for investigating neuron brain dynamics involving transitory and intermittent subject cognitive states. Analysis of single ERP epochs, on the other hand, while ideal, suffers from confusions caused by significant EEG artifacts associated with blinks, eye-movements, and muscle noise, by large non-phase locked background EEG activities, and by the wide variability in latencies and amplitudes of ERP waveforms from trial to trial. This study introduces a new visualization tool, the 'ERP image', for investigating variability in latencies and amplitudes of event-evoked responses in spontaneous EEG or MEG records. Second, we apply a new linear decomposition tool, Independent Component Analysis (ICA) [I], to multichannel single-trial EEG records to derive spatial filters that decompose single-trial EEG epochs into a sum of temporally independent and spatially fixed components arising from distinct or overlapping brain or extra-brain networks. We demonstrate the power of the proposed analysis and visualization tools for single-trial ERP analysis through This report was supported in part by grants from the Office of Naval Research and Howard Hughes Medical Institute. The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Navy, Department of Defense, or the U S . Government. analysis of sample data sets from one normal and one autistic subject.