Rethinking the function of brain regions for reading Chinese characters in a meta-analysis of fMRI studies

作者:Zhao, Rong; Fan, Rong; Liu, Mengxing; Wang, Xiaojuan*; Yang, Jianfeng*
来源:Journal of Neurolinguistics, 2017, 44: 120-133.
DOI:10.1016/j.jneuroling.2017.04.001

摘要

Neuroimaging studies have proposed specific brain regions for reading Chinese characters. For example, compared to alphabetic languages, Chinese reading has revealed more activation of the left Middle Frontal Gyrus (MFG) and null or deactivation at left posterior Superior Temporal Gyrus (pSTG). However, the model of specific regions for Chinese was contested by recent findings showing similar activities across languages under the same task demand. Thus, whether the engagement of the Chinese-specific regions was due to stimulus selectivity or to task demand was called into question. The current meta-analysis, using Activation Likelihood Estimation (ALE), summarized brain regions in previous NM studies from two types of contrast, stimulus contrasts (SC), in which more regions were activated for Chinese characters than for other visual stimuli (e.g., Korean words, faces) or for other types of characters (e.g., real > artificial characters), and task contrasts (TC), in which more regions were activated for one task than for another (e.g., lexical decision > symbol detection) in processing Chinese characters. ALE activation maps revealed a shared network for the two types of contrast, including the right Inferior Frontal Gyrus (IFG, BA47) and four clusters in the left hemisphere, the IFG (BA45/46), Superior Parietal Lobule, Superior Frontal Gyrus, and posterior Middle Temporal Gyrus (pMTG). The SC > TC comparison demonstrated the greater involvement of the left Angular Gyrus and left Fusiform Gyrus. The TC > SC comparison, in contrast, revealed the engagement of left hemisphere regions in the IFG, MFG, pSTG, pMTG, and Medial Frontal Gyrus. These results suggest that the previously identified Chinese-specific areas (e.g., MFG, pSTG) might be task dependent rather than stimulus selective. This work shed light on the language general neural network of visual word reading.