Slide Slam S13 Sandbox Series
Semantic processing during the natural reading
Yali Pan1, Steven Frisson1, Kara Federmeier2, Ole Jensen1; 1University of Birmingham, 2University of Illinois at Urbana-Champaign
Introduction: Through reading, we crack meaning out of the rich text. Event-Related Brain Potential (ERP) literature regards N400 as an index for semantic processing (Kutas and Federmeier, 2010). Kutas & Hillyard (1980) first characterized N400 as a negative potential between 200 and 600ms, peaking around 400ms, which was higher for semantic incongruent than congruent sentences. However, the classic N400 paradigm is not natural reading and parafoveal information is not available. Words from a sentence are presented one by one on the screen centre at a pace of around one word per second. Although N400 inspires the conceptualization of how meaning processing might unfold, its time course is strikingly slow regard the fast natural reading. So, how the meaning processing unfolds in the natural reading where parafoveal information is available? Methods: Participants will read 180 one-line sentences silently while eye movements and brain activities are recorded by an eye-tracker and MEG simultaneously. Every sentence is plausible and contains an unpredicted critical target word. Half of the target words make sentences semantic congruent, the other half incongruent. In our previous study (Pan, Frisson, and Jensen, 2021), Rapid Invisible Frequency Tagging (RIFT) has been shown to be a powerful tool to capture parafoveal processing during natural reading. Here we apply RIFT to flicker the critical target word by adding a patch underlie it. The luminance of this patch changes from black to white as a sinusoid at 60Hz, which is invisible on a grey background and will not interfere with reading. Tagging responses will be estimated as the coherence at 60Hz between tagging signal and brain activities. Coherence during the fixations of pre-target words reflects the previewing of flickering target words in the parafovea. We will also calculate the N400 difference between semantic congruent and incongruent conditions aligning with fixation onset to the target words. Results: We expect stronger 60Hz coherence for the fixations of pre-target words when followed by semantic incongruent compared with congruent target words. Besides, a classic semantic violation effect is expected for the fixations of target words: stronger N400 amplitude for semantic incongruent compared with congruent target words. We also expect that the pre-target coherence difference correlates positively with reading speed (estimated from congruent sentences only), but negatively with the N400 difference of the foveal target words. Conclusion: Semantic information can be extracted even before making saccades to that word. This semantic parafoveal processing facilitates reading performance. Therefore, rapid meaning processing unfolds to both foveal and parafoveal words in a distributed way during the natural reading.