DIVA(Directions Into of Articulators)模型是一种为了生成单词、音节或者音素而控制模拟声道运动的自适应神经网络模型,其依赖的输入信号是从人体大脑中采集到的脑电信号。针对汉语神经分析系统研究中非侵入式脑机接口采集到的脑电数据存在的分辨率低、干扰大的问题,文中提出一种基于DIVA模型对脑电信号进行约束处理的方法。首先利用独立分量分析方法剔除原始信号中的噪声,提取有效事件相关电位(Event-Related Potentials,ERP)成分;然后以模拟生成的功能性磁共振成像(functional Magnetic Resonance Imaging,fMRI)数据的激活点的空间信息作为限制条件,对提取出的ERP成分进行精确定位。通过对实验数据进行分析处理并模拟受试者的激活脑区,验证了所提方法的正确性和有效性。
【英文摘要】
DIVA( Directions Into Velocities of Articulators) is an adaptive neural network model which is used to control the movement of analog vocal tract for generating the words,syllables or phonemes. The input signal that DIVA model depends on is the EEG( ElectroEncephalo Gram) signal collected from the human brain. However,EEG data collected through non- invasive brain- computer interface shows lowresolution and large interference in the research of Chinese Neuralynx System. A constraint processing method for th...