Time Series Analysis of Physical/Mental process in Human Brain

这个项目开始的目的是客观/统计/时间序列分析人类大脑的精神和身体活动之间的关系。主要结果是关于脑电图数据。对癫痫患者内侧颞叶皮层和基底颞叶皮层的脑电图记录进行分析，发现存在约15赫兹（范围从10到20赫兹）的规律β -1振荡。脑磁图数据：在感知水平上，电偶极子的估计位置与侵入性电生理方法确定的位置吻合良好。在认知水平上，海马海马在P300m时间范围内活动显著。在内隐运动成像任务中，脑磁图揭示了脑皮层活动的时间过程，这是PET研究无法做到的。w.r.t.t建模：建立了一个描述脑时间序列数据的非线性模型——交互频带模型；并开发了基于信息标准方差估计的“模型置信集”程序。这些结果为进一步研究大脑的生理/心理过程奠定了良好的基础。

This project started aiming at objective/statistical/time-series analysis of relation between mental and physical activity of human brain. Main results are with respect to EEG data. Analysis of th4e ECoG recorded form the medial and basal temporal cortices of human epileptic subjects revealed the presence of regular beta-1 oscillations of approximately 15 Hz (ranging from 10 to 20Hz).W.R.T.MEG data : At perception level, the estimated locations of electric dipoles demonstrated good coincidence with those determined by invasive electrophysiological methods. At cognition level, significant activities in the dippocampus were found in the P300m time range. In an implicit motor imaging task, MEG revealed time course of ortical activities, something which had not benn possible in PET studies.W.R.T.modelling : A nonlinear model, Interactive frequency Band Model, for the description of brain time-series data was developed ; and a program to produce 'Confidence Set of model' based on the estimate of variance of Infromation Criteria was developed.Those results make a good foundation of future study of the Physical/Physiiological/Mental Process of the Brain.