错误信息持续影响效应的神经基础

本研究通过分析任务态fMRI下相关脑区激活及功能连接的条件间差异以揭示CIEM神经基础并为心理模型更新及记忆提取失败假说提供更多证据。结果发现, 更正条件的推理分显著高于控制条件, 存在CIEM。编码阶段左颞中回在更正条件下的激活显著弱于控制条件, 提取阶段更正条件下额中回及前扣带回激活更弱、额中回与中央前回的功能连接更强。结果提示上述脑区可能参与了CIEM的形成, 并从神经层面提供了心理模型更新和记忆提取失败假说可能解释了CIEM形成的不同阶段的证据。

The neural basis of the continued influence effect of misinformation

Misinformation often continues to influence people’s beliefs and reasoning even after retracted—this persistence is termed the ‘continued influence effect of misinformation’ (CIEM). Both of the mental-model- updating hypothesis and the memory-retrieval-failure hypothesis attempt to explain this phenomenon. The neural substrates of CIEM are controversial, and results from different studies support different assumptions. The disputations may relate to neglection of potential contribution of inhibitory control in CIEM and of methodological differences between studies. This study aimed to investigate neural substrates and cognitive mechanism of CIEM using the functional magnetic resonance imaging (fMRI) from the view of inhibition control. Thirty-one participants (10 males) were recruited in this study. They were instructed to read brief, fictional news reports and answer three inference questions after reading while lying in a 3.0T Siemens Prisma MRI scanner. Each participant needed to read 40 reports (20 reports in their retraction versions and 20 in their control versions). Each fictional report contained six sentences and derived retraction and control versions based on whether the second sentence contained misinformation. Pseudorandom uniform temporal jitter was used for this fMRI design. Imaging data were preprocessed and processed using SPM, RESTplus and DPABI toolbox to obtain the functional activities of the ROIs and their functional connectivity. Multi-comparison tests were conducted for brain activities induced by target sentence 5 (encoding phase) and three reasoning sentences (retrieval phase) under different versions. The results showed: (1) retractions elicited less activity in the left middle temporal gyrus (BA21/22) than control during encoding phase; and retractions also elicited less activity in the left middle frontal gyrus (L_MFG, BA10) and right anterior cingulate cortex (R_ACC, BA32) than control during retrieval phase. Additionally, activation at the left dorsolateral prefrontal cortex (L_DLPFC, BA9) in the retraction condition was marginally significantly different from that in the control condition during retrieval phase. No significant activation difference was observed across conditions in others ROI. (2) A marginally significantly negative correlation was found between functional metrics (Beta) of the left MFG in retraction condition and individuals’ interference scores. (3) With regard to functional connectivity, we compared the connectivity between two seeds (left MFG and right ACC) and the rest of the brain in control versus retraction condition during retrieval phase. Results demonstrated that the right ACC showed decreased functional connectivity with the bilateral inferior occipital gyrus (IOG) under retraction condition when compared to control condition. The left MFG showed similar decreased connectivity with the bilateral IOG under retraction condition when compared to control, but increased functional connectivity with right precentral gyrus under retraction condition when compared to control. The results suggest that the CIEM be related to semantic encoding failure during information comprehension and inhibition failure of misinformation during information retrieval. The mental-model-updating hypothesis and the memory-retrieval-failure hypothesis can explain the different phase of CIEM. The findings provide more experimental evidence for neural basis of CIEM and refine corresponding theoretical accounts, and provide neurological clues for further exploration of ways to reduce the negative impact of CIEM in the future.