López-Peréz , P. J. , van der Meij, M. , Muñoz, S. & Barber, H. A.
Department of Cognitive Psychology, University of La Laguna, Spain
Parafoveal perception plays an important role in reading, but the exact extension and type of word processing in the parafovea is still a matter of discussion. In two ERP experiments we looked at two important variables that could constrain parafoveal word processing: the sentential context and the reading time. A modification of the standard RSVP procedure was used to study parafoveal perception during sentence reading in absence of eye movements. In experiment 1, volunteers read Spanish sentences presented word-by-word at fixation (SOA= 450 ms). Each of these words was flanked 2 degrees bilaterally by two different words: the next word in the sentence to its right, and the previous word in the sentence to its left. In some of the sentences, target words were replaced by a word that was semantically incongruous with the overall sentence context. This happened when the target word was presented at fixation (foveal manipulation), and/or in the right flanker of the previous word (parafoveal manipulation). Parafoveal and foveal manipulations resulted in qualitatively similar N400 effects, and these effects were observed for target words with high and low cloze probability. Although parafoveal N400 effects with high cloze probability words could be explained by the use of formal features (or just initial letters) to confirm contextually anticipated items, the same effect with low cloze probability words supports the claim that under these circumstances words identified in the parafovea were fully processed. In experiment 2, the same sentences were presented at a faster presentation rate, close to that of normal reading (4/s; SOA=250 ms). In this case, the parafavoeal N400 effect was observed only for words with high cloze probability. These results suggest that time can constrain parafoveal perception during sentence reading.