Distinct Neural Systems For Semantic Competition And Retrieval Processing In Chinese Word Recognition

Purpose and Significance of Research: Semantic competition was a relatively low level process which happened in the early stage of word recognition.A cue word would automatically evoke multiple candidates which compete with each other.In contrast,semantic retrieval was a relatively high level process which happened in a later period.Among multiple competing word candidates,only one target word was selected at the end,and all other distractors decay gradually.Here we aim to shed light on this issue by manipulating explicit variables of competition and retrieval processes in a finegrained parametric modulation design using a semantic similarity judgment task.Method: There were three task types: semantic similarity judgment of concrete words,semantic similarity judgment of abstract words and perceptual similarity judgment of non-linguistic visual patterns.Each of the three task types was made up of a set of words or monochrome visual pattern triplets that varied respectively in semantic or perceptual similarity between the cue and target and the cue and distractor.Within each event-related f MRI run,participants made one type of judgment.We manipulated the measurements of competition and retrieval as two modulators.Increasing competition measurement referred to the higher competition demand between target and distractor,while increasing retrieval measurement referred to the lower difficulty of retrieving the target.Results: Increasing competition demands generated greater activation in bilateral middle frontal gyri(MFG),superior medial frontal gyri(SMFG),and left left inferior parietal lobule/angular gyrus(IPL/AG),while the retrieval process activated leftlateralized inferior frontal regions,SMFG and IPL/AG.These two networks might interact with each other through bilateral SMFG and left IPL/AG.The competition network was only activated for concrete words and increasing SF produced greater activation in bilateral fusiform,indicating that this neural network is sensitive in detecting competition among concrete SF.Conclusion: Low level language processing was supported by the bilateral frontotemporo-parietal network,while high level language processing relied on the left fronto-temporo-parietal network and there were interactions between these two processes.High level language processing would activate left neural network,however,whether the phenomenon of left-lateralization was caused by linguistic properties remained unclear.