Understanding associative vs. abstract pictorial relations: An ERP study

NEUROPSYCHOLOGIA

卷:133

Article

Zucker, Leemor¹  Mudrik, Liad^1,2 

[1] Tel Aviv Univ, Sagol Sch Neurosci, POB 39040, IL-69978 Tel Aviv, Israel

[2] Tel Aviv Univ, Sch Psychol Sci, POB 39040, IL-69978 Tel Aviv, Israel

关键词: N400;    Late negativity;    Relational processing;    Semantic integration;    Expectations mechanisms;    Object processing;   

DOI  :  10.1016/j.neuropsychologia.2019.107127

来源: Elsevier

【 摘 要 】

One of the most remarkable human abilities is extracting relations between objects, words or ideas - a process that underlies perception, learning and reasoning. Yet, perhaps due to its complexity, surprisingly little is known about the neural basis of this fundamental ability. Here, we examined EEG waveforms evoked by different types of relations, conveyed by pairs of images. Subjects were presented with the pairs, that were either associatively related, abstractly related or unrelated, and judged if they were related or not. Evidence for a gradual modulation of the amplitude of the N400 and late negativity was found, such that unrelated pairs elicited the most negative amplitude, followed by abstractly-related pairs and lastly associatively-related ones. However, this was confined to first encounter with the pairs, and a different, more dichotomous pattern was observed when the pairs were viewed for the second time. Then, no difference was found between associatively and abstractly related pairs, while both differed from unrelated pairs. Notably, when the pairs were sequentially presented, this pattern was found mostly in right electrodes, while it appeared both in left and right sites during simultaneous presentation of the pairs. This suggests that while two different mechanisms may be involved in generating predictions about an upcoming related/unrelated stimulus, online processing of associative and abstract semantic relations might be mediated by a single mechanism. Our results further support claims that the N400 component indexes multiple cognitive processes that overlap in time, yet not necessarily in neural generators.

【 授权许可】

Free