Abstract: Inference was a tool to establish coherence of the text and provided necessary insurance for correct comprehension. The study of inference in discourse comprehension using functional magnetic resonance imaging (fMRI) was a rather recent, but exciting development research area. The existing fMRI researches of inference in discourse comprehension nearly all based on narrative, studies which based on expository hadn’t been reported yet. Now, researchers had a clear consensus that inference was coded automatically on-line during narrative reading. Expository is another important literary form, inference in expository comprehension has been extensively investigated in behavioral paradigms, but it was somewhat surprising that there still great debated about if local coherence could be coded automatically during expository reading or not. And till now, the definitely and consistently explanation has not been coincided yet. Meanwhile, it was less known about the neural substrates of inference process in expository comprehension. In this study, brain activations of inference processing in familiar topics expository text comprehension were measured using fMRI, which aimed to investigate whether readers could drawn inference automatically during expository reading and pointed out the core brain area of inference processing in expository reading. Meanwhile, Mo had pointed out that when there existed reading focus in text, readers will make inference around the focus on their own initiative. So another aim of this test was to make sure whether focus-based reading will influence the inference process of expository reading. Participants were asked to read 80 passages which include five or six sentences that differed whether or not an inference needed to be drawn to understand them. Each sentence was present for 3 seconds. Half of the passages were focus-based, and half were not. The questions were about the text but not concerned with the inference. There whole test was divided into 5 runs and there were 16 trials in each run. Participants had to finish a probe task after reading each passage. There were two pre-experiments which were used to choose test materials and decide reading time for each sentence. The MRI signals were obtained when participants read the expository texts. The target sentence was the last sentence of the passage but not the question sentence. The scanner was a 3.0 T Siemens Magnetom Vision MRI scanner. The stimuli were presented with black characters on a gray background through the projector on the mirror. The functional MRI data from 22 participants were analyzed with SPM8. The general inference related activation appeared in frontal cortex (which included inferior frontal gyrus, superior frontal gyrus and middle frontal gyrus), inferior parietal gyrus and precuneus bilaterally in expository text reading. These areas were main areas which related to the inference process in expository comprehension. In the condition of focus-based reading, the cortical related activation appeared more concentrate on frontal gyrus bilaterally, especially the IFG. The present findings indicated that the basic brain areas about inference in expository reading were generally consistent with which researchers had found in narrative before. The focus-based reading helped readers to pay attention on inference processes, other processes which were not concerned with inference were restrained. This result supported the bi-processing theory of text comprehension.

Key words: expository, inference, focus-based reading, fMRI