中文摘要：

失匹配负波（MMN）是否受注意力的调控,一直是MMN研究领域的争论问题。目前鲜有实验范式能够很好地控制受试者注意力在注意通路与非注意通路之间的分配,并缺乏一个量化尺度来反映受试者注意力资源分配的变化。扩散模型是一种认知过程模型,主要通过研究完成任务过程中的反应时间和准确率等行为数据来揭示其潜在的神经加工机理。因此,建立一种新的跨通路延迟反应实验范式,以更好地描述受试者注意力的分配;然后结合扩散模型参数拟合方法,量化受试者注意力资源分配情况,并明确MMN的潜伏期和峰值与注意力之间的关系。采集18名受试者在3组不同图像对比度实验条件下的脑电信号,研究对比9个导联的标准刺激与偏差刺激ERP波形。实验结果表明,扩散模型拟合参数可以解释实验中不同设定条件对受试者注意力分配的影响,从而定量地确定受试者的注意力。同时表明,MMN的峰潜伏期随实验条件变化显著,与对应的扩散模型参数有显著的相关性：MMN峰潜伏期与扩散模型拟合参数边界间隔、漂移率和非决策时间之间的相关系数分别为0.63、0.58和0.63。结果证实,扩散模型参数可以作为MMN测试中受试者的注意力资源分配的指标,且MMN的峰潜伏期与受试者注意力呈正相关,因此可以认为MMN具有半自动加工的性质。

英文摘要：

Whether MMN is regulated by attention has been a debate in the MMN research field. There are few experimental paradigms which could control the attention allocation between the attended and unattended channels. Moreover, there is no index to reflect the subtle fluctuation of attention resource of the subjects. The diffusion model is one kind of cognitive process models, to reveal the potential neural processing mechanism based on the behavior data such as reaction time and accuracy of the task. Therefore, this paper developed a new cross-modal and delayed response experimental paradigm to better control subjects＇ attention resources. Then the parameters fitted by the diffusion model were used to establish the relationship between the MMN and the attention. Experimental results showed that the parameters fitted by the diffusion model using behavior data could explain the underlying decision processes under different conditions on the subjects＇ attention. As a result, the method could quantitatively determine the subjects＇ attention. Meanwhile, we also found that the variation of the MMN peak latency was affected by variations in task difficulty and had a significant correlation with attention-related parameters of diffusion model： the correlation coefficient r between the MMN peak latency and b, v and tER were 0. 63, 0. 63 and 0.58 respectively. This proves that the parameters of the diffusion model can he used as an index of the subjects＇ attention resource allocation, and MMN peak latency is positively correlated with intensity of attention. Therefore, we consider that the MMN reflects a semi-automatic process.