认知控制在工作记忆表征引导注意中的作用:来自眼动的证据

What do eye movements reveal about the role of cognitive control in attention guidance from working memory representation

研究同时采用首次注视点百分率与行为反应时指标,探讨了认知控制在无关工作记忆表征引导视觉注意选择过程中的作用。实验1发现:当视觉搜索容易时,首次注视点百分率和反应时指标都出现了注意引导效应;而当视觉搜索困难时,在反映早期注意定向的首次注视点百分率指标上,搜索快的被试组的注意引导效应显著大于搜索慢的被试组,而在行为反应时指标上,搜索快的被试组表现出注意引导效应,但搜索慢的被试组则表现出相反的注意拒绝效应。实验2发现当采用更为突显的靶子刺激时,首次注视点百分率指标上表现出的注意引导效应并未受到视觉搜索靶子显著性的影响,甚至还表现出一定的优先性。这些结果表明认知控制在工作记忆表征引导视觉注意的过程中所起的作用是动态变化的,视觉搜索早期注意定向阶段的注意引导效应比较稳健,认知控制只能减弱其效应的量值,而在随后的视觉搜索过程中,当视觉搜索比较缓慢时,认知控制可能会将注意引导效应反转为相反的注意拒绝效应。

The biased competition model of attention proposes that stimuli in visual scene compete for attentional selection with the ＇winner＇ gaining control of both perceptual and response systems. The activation of relevant object features in working memory （WM） biases neural activity in specific brain regions that encode those particular features so that the object could win the competition and get the priority of attention. In decades, researchers have found that irrelevant representations in working memory （WM） could guide visual attention bias to the distractors that sharing the features of the WM representations in visual search task. However, the issue whether the cognitive control could modulate the process of attentional guidance was controversial yet. Here, we adopted the eye movement tracking technique to examine （1） whether the cognitive control made the participants intentionally inhibit or reject the bias of attention towards the WM matched distractor and （2） when did cognitive control play the role in the processing of attentional guidance. The participants in the present study were asked to search for a target among three distractors while holding a colored shape online in WM. During the visual search task, the item in WM either reappeared as one of the distractors in half of trials or did not reappear in other half of trials. Most importantly, at the beginning of the experiment, all the participants were indefinitely informed that the WM item could never reappear as target in visual search task and they could adopt the ＇rejecting＇ or ＇inhibiting＇ strategy to bias attention away towards such distractor so as to facilitate the efficiency of visual search. Both the eye movement deployment and the response time （RT） were recorded while the participants performing the visual search task. The results of experiment 1 showed that, when using the easy visual search task, an obviously attentional guidance effect was observed for both the first fixation proportion and RTs, and such effect was not affected by the visual search speed. However, when using the difficult visual search task, the results of the first fixation proportion showed an significant attentional guidance effect, and the magnitudes were significantly greater for the participants who searched fast （fast group） than for those who searched slowly （slow group）. In contrast, the results of RTs failed to obtain any above attentional guidance effect, and the further analysis indicated that thisnull effect was due to the opposite attentioanl bias pattern represented for two groups of visual search speed, i.e., attentional guidance pattern for fast group and attentional rejection pattern for slow group. The results of experiment 2 found the attentional guidance remain robust at the early stage when the more prominent visual search target was adopted. Most importantly, the attentional guidance was stable even when the WM matched distractor directly competed with the prominent search target. In conclusion, the results showed that the attentional guidance from WM representation was dynamically changed along with the process of visual search and the cognitive control could modulate the attentional guidance effect only when the visual search was so slowly that the time for cognitive control working was sufficient. The modulation of cognitive control could weaken the magnitudes of attentional guidance effect at the early stage of visual search task and reverse the attentional bias pattern from guidance to rejection in the following processes of visual search task.