【作者】 吴正华；

【导师】 尧德中；

【作者基本信息】 电子科技大学 ， 生物医学工程， 2008， 博士

【摘要】 稳态视觉诱发电位(Steady-state visual evoked potential,SSVEP)的产生机理到目前还没有统一的解释,主要有反馈网络理论、行波理论等。SSVEP的应用主要有两个方面,一是在脑机接口(Brain-Computer Interface,BCI)上的应用,另一个是在认知过程研究方面的应用。本文将就SSVEP的基本理论以及它的应用展开讨论,主要由以下一些内容组成。1.通过研究在频率一定(10Hz),而具有不同的占空比的闪光刺激下所产生的SSVEP的幅度,发现当占空比为0.4左右时,产生的SSVEP的幅度明显大于其它占空比时的幅度。2.通过用同样强度和频率的红、绿、蓝三种单色光分别刺激被试,结果显示不同单色光产生的SSVEP的分布基本一样,但强度不一样,蓝光对应的SSVEP最强,红光次之,绿光最弱。3.用位于不同频段的闪光单独和同时刺激被试,通过研究两种情况下所得的不同频率SSVEP的幅度和分布,发现当不同的SSVEP网络被同时激活时,彼此之间并没有明显的相互作用,这说明它们基本上是独立的。4.由于不同类型刺激器的发光原理不同,所产生的闪光的频谱结构就不一样。在同样刺激频率和亮度的情况下,所产生的SSVEP的强度将不一样。通过用位于不同频段的三种不同类型的闪光分别刺激被试,发现LED产生的SSVEP最强,而CRT和LCD所产生的SSVEP基波并无明显差别,但谐波差异很大。5.在常规的功率谱方法中,所获得的SSVEP是在一定长度时间段内的平均值,当选取的时间段足够长时(一般长于3—4秒),将会在一定程度上消除或削弱自发脑电的影响,获得较高的准确性和稳定性,但传输率比较低。在本工作中,通过对SSVEP进行小波分析,引入了一个新的参数---稳定系数,该系数在短的时间段内也具有较高的抗干扰性,这使得对目标识别的准确性提高,从而也提高了系统的传输率。6.为了比较同一个认知任务对不同频率SSVEP的影响是否相同,用位于不同频段的刺激分别诱发SSVEP,同时让被试执行同样的记忆任务。通过对两种情况的比较,发现同样的认知任务对不同频率SSVEP的影响基本是一样的,而对低频SSVEP的影响更为明显且稳定。7.通过研究注意与不注意两种情况下的SSVEP幅度,发现注意时诱发的SSVEP显著增强,而在低频时注意对SSVEP的作用比在高频时注意对SSVEP的作用更明显。该实验结果支持了注意的早期选择机制。8.首先让被试反复执行同一个记忆任务,通过传统的叠加平均方法得出该任务的ERP。再让被试在一个与任务无关的闪光刺激下执行同样的记忆任务,并用稳态检测地形图方法(Steady-State Probe Topography,SSPT)来提取SSVEP的变化过程。通过比较ERP和SSVEP的波形,发现ERP能对SSVEP进行幅度调制,在ERP和SSVEP都较强的区域,这种调制作用越明显。更多还原

【Abstract】 The genesis of SSVEP isn’t clear till now, and often explained by loop-loop feedback or wave theories. SSVEP is used mainly in two hands, one is in studying of cognitive task, the other is in Brain-Computer Interface (BCI). In this paper, we’ll discuss the fundamental theory about SSVEP and its application, and there include the contents below.1). From studying the SSVEP evoked by a pulse stimulus of 10 Hz with different duty-cycle, we found that the amplitude of SSVEP was the biggest one when the duty-cycle was about 0.4.2). From using the red, green or blue flicker to stimulate the eyes of subjects separately under the same luninance, we found that the distribution of SSVEP was similar under each situation, while the intensities of them were different, the SSVEP evoked by the blue flicker was the biggest one, that evoked by the red flicker was smaller than that evoked by the blue flicker, while that evoked by the green flicker was the smallest one.3). We use the flickers in different frequency band to evoke the SSVEP separately or simultaneously, from studying the SSVEP amplitude and distribution under these situations, it is found that there is no clear interaction between these networks when activated simultaneously, in other word, these networks are independent from each other.4). Because of the different lighting theory for different type stimulator, the spectum of the flickers are different too. Although the stimulating frequency and luminance are the same, the evoked SSVEPs are different from each other. From using different type stimulators in different frequency band, it is found that the SSVEP evoked by the LED flicker is the strongest one, while the fundamental frequency in SSVEP evoked by the CRT and LCD flicker is similar to each other, although the difference of the harmonics is significant.5). When using power spectrum (PS) method to extract SSVEP within a relative long time period, the influence by the potential mental activity can be eliminated to some extent, and a relative high accuracy but low transfer rate can be obtained. When using the stability coefficient (SC) method introduced here, which was obtained by Wavelet Analysis, the influence in a relative short time period by the noise can be eliminated greatly, which could further result in a relative high accuracy and transfer rate in a SSVEP-based BCI system.6). In order to compare the influence on the different frequency SSVEP by a same cognitive task, we use two frequencies in different band to evoke SSVEP, and make the subject execute the same memory task. From comparing the SSVEP under two situations, it is found that the influence on the different frequency SSVEP by the same task is similar to each other, while that on the low frequency is more clear and stable than that on the high frequency.7). From studying the SSVEP amplitudes with attention or without attention, it is found that the SSVEP amplitude is improved with attention, while this influence is clearer for the low frequency than for the high frequency. This result supports the theory of selection early about attention.8). The subject was asked to finish a same memory task repetitively, and the ERP of this task was obtained by the traditional average method. Then the subject was asked to execute the same memory task repetitively under the flicker stimulus, and a method named Steady-State Probe Topography (SSPT) was used to extract the SSVEP. From comparing the ERP and SSVEP, it is found that the ERP can modulate the amplitude of SSVEP, and this modulation is clear in the regions where two kinds of signals are both strong.更多还原