Modelling the reading of isolated words
Abstract: I will talk about further joint work carried out within the split-fovea approach to visual word recognition. In this research we have used connectionist architectures to simulate the processing going on in the two halves of the brain, based on the fact that the peripheral anatomy of the visual pathways shows that a fixated word has its two parts initially contralaterally projected to the two cerebral hemispheres. I will talk about the principles of such modelling, and in particular the representation of letter position in words, and then present simulations of novel data from normal reading experiments. In particular, I will show that some of the hemispheric differences in semantic processing may be emergent effects due to the information structure of words in the English lexicon.