Robust Space-Time Signal Processing for MEG/EEG

描述（由申请人提供）：该计划的目标是开发信号处理方法，该方法将扩大MEG/EEG功能脑成像作为诊断脑功能障碍的临床工具以及作为研究脑发育和功能的科学工具的作用。该研究计划介绍了一种新的方法，其中测量数据和底层皮层表面表示使用多分辨率基函数扩展。通过开发利用基函数扩展的信号处理、源定位和成像算法来获得改进的性能。提出的研究的具体目标是：1）开发基函数扩展，共同代表传感器和皮层表面上的大脑活动的时空分布。我们提出的多分辨率方法采用了一个简约的表示在不同的时间和空间尺度的大脑活动，一个关键的属性，有效的解决方案的各个方面的反问题。2)开发鲁棒的信号处理算法，利用基函数表示来执行多个源的检测、定位和监测，以及皮层表面上的活动的成像。所提出的算法将具有降低的复杂性和改善的性能相对于现有的方法。3)使用非参数统计分析和计算机模拟来表征根据目标1和2开发的算法的性能。4)获得证明根据目标1和2开发的算法有效性的人体试验数据。

DESCRIPTION (provided by applicant): The goal of this program is development of signal processing methods that will expand the role MEG/EEG functional brain imaging as a clinical tool for diagnosis of brain dysfunction and as a scientific tool for the study of brain development and function. The research program introduces a new approach in which both the measured data and underlying cortical surface are represented using multiresolution basis function expansions. Improved performance is obtained by developing signal processing, source localization, and imaging algorithms that exploit the basis function expansion. The specific aims of the proposed studies are: 1) To develop basis function expansions that jointly represent the spatio-temporal distribution of brain activity at the sensors and on the cortical surface. Our proposed multiresolution approach employs a parsimonious representation for brain activity at different temporal and spatial scales, a critical property for effective solution of various aspects of the inverse problem. 2) To develop robust signal processing algorithms that exploit the basis function representation to perform detection, localization and monitoring of multiple sources, and imaging of the activity on the cortical surface. The proposed algorithms will have reduced complexity and improved performance relative to existing methods. 3) To characterize the performance of the algorithms developed under Aims 1 and 2 using nonparametric statistical analysis and computer simulation. 4) To obtain human subject pilot data demonstrating the effectiveness of the algorithms developed under Aims 1 and 2.