Real time cortical connectivity imaging from EEG/MEG data

描述（由申请人提供）： 借助于人脑的非侵入性表面脑电图（EEC）或脑磁图（MEG）的神经反馈提供了用于控制脑-计算机接口（BCI）或用于诸如癫痫和注意缺陷多动障碍（ADHD）的障碍的潜在治疗的重要机制。它还允许研究人类大脑的功能方面，这在离线研究中是不可能观察到的，即那些只在反馈期间被激活的神经元回路。传统上，反馈通道是记录在头皮表面上的EEC电位或其光谱分量。由于体积导体效应，这些记录是潜在的皮层活动的模糊表示。这使得很难单独从通道信息评估特定皮层区域的反馈。此外，原始EEC的低信噪比（SNR）是BCI或反馈应用中的限制因素。一个逆模型，它结合了个别受试者的几何形状，允许直接成像的皮层区域，并可以提供特定的功能反馈，这些地区。使用逆方法结合连接性测量的应用，例如相干性、高阶谱分析（HOSA）、锁相或多变量自回归（MVAR）模型，将以两种方式增强神经反馈应用：由于噪声导致的区域之间的随机相互作用的概率将低于单通道测量，并且通过逆方法使用先验空间信息将利用所有可用的空间信息。通道，从而提高系统的SNR和特异性。非线性方法，如HOSA，具有对高斯白色噪声和线性串扰效应鲁棒的优点，这可能会混淆连接性测量。基于皮层同步测量的真实的时间系统可以用于训练受试者主动增加或减少所选皮层区域之间的同步，并且可以促进ADHD或癫痫的神经反馈治疗。此外，皮层同步检测可以直接应用于BCI。我们将开发一个EEG/MEG皮层成像系统用于神经反馈应用，它将真实的实时成像选定区域之间的皮层同步。我们将通过对健康受试者的反馈实验来验证该系统，其目标是使受试者能够使用所提出的成像系统来增加或减少选定区域中的皮质同步。

DESCRIPTION (provided by applicant): Neurofeedback by means of non-invasive surface Electroencephalography (EEC) or Magnetoencephalography (MEG) of the human brain provides an important mechanism for control of brain- computer-interfaces (BCI) or potential treatment for disorders such as epilepsy and attention deficit hyperactivity disorder (ADHD). It also allows studying functional aspects of the human brain, which are impossible to observe in off-line studies, i.e. those neuronal circuits, which only become activated during feedback. Traditionally, the feedback channels have been the EEC potentials recorded on the scalp surface or their spectral components. Due to volume conductor effects these recordings are blurred representations of the underlying cortical activity. This makes it difficult to assess feedback for specific cortical regions from the channel information alone. Also, the low signal to noise ratio (SNR) of the raw EEC is a limiting factor in BCI or feedback applications. An inverse model, which incorporates the individual subject geometry allows direct imaging of cortical regions and can provide specific functional feedback for these regions. Using an inverse method in combination with application of connectivity measures, such as coherence, higher order spectral analysis (HOSA), phase-locking or a multivariate autoregressive (MVAR) model will enhance neurofeedback applications in two ways: The probability of random interaction between regions due to noise will be lower than single channel measures and the use of a priori spatial information by an inverse method will utilize all avalable channels, thus increasing the SNR and specificity of the system. Non linear methods, such as HOSA, have the advantage of being robust against Gaussian white noise and linear crosstalk effects, which can confound connectivity measures. A real time system based on cortical synchrony measures can be used to train subjects to actively increase or decrease synchronization between selected cortical regions and can facilitate neurofeedback treatment of ADHD or epilepsy. Also, cortical synchrony detection can be directly applied to BCI. We will develop an EEG/MEG cortical imaging system for neurofeedback applications, which will image cortical synchrony between selected regions in real time. We will verify the system by feedback experiments on healthy subjects with the goal to enable subjects to increase or decrease cortical synchrony in selected regions using the proposed imaging system.