Contribution and chronometry of left ventral occipito-temporal cortex and posterior middle temporal gyrus in reading: evidence from transcranial magnetic stimulation

Pattamadilok, C. 1 , Bulnes, L. C. 1 , Leonard, E. 1 , Devlin, J. T. 3 , Morais, J. 1 & Kolinsky, R. 1, 2

1 Unité de Recherche en Neurosciences Cognitives. Université Libre de Bruxelles. Brussels, Belgium
2 Fonds de la Recherche Scientifique-FNRS. Brussels, Belgium
3 Cognitive, Perceptual and Brain Sciences; Institute of Cognitive Neuroscience, University College London. London, UK

The left ventral occipito-temporal cortex (vOTC) and posterior middle temporal gyrus (pMTG) are among the most stable and crucial regions to reading, although presumably at different processing levels. The first area has been shown to play a causal role in written word processing. The second area is strongly involved in semantic processing, although its causal role in reading has yet to be demonstrated. Transcranial magnetic stimulation (TMS) was used to investigate the contribution of these two areas as well as the time course of their activation during regular word, irregular word and pseudoword reading. Paired-pulse TMS delivered at 3 time periods post-stimulus onset (-20/20ms, 40/80ms, 100/140ms) showed that both areas are causally involved in regular and irregular word but not in pseudoword reading. Interestingly, TMS interrupted the function of the left vOTC as early as 40/80ms post-stimulus onset while its effect on the left pMTG became significant only in the later, 100/140ms, time window. Finally, while the TMS effect on the left vOTC was more reliable on regular than irregular words, the opposite tendency was found when the stimulation was applied over the left pMTG. This latter observation suggests a stronger involvement of the left pMTG when the pronunciation of written words cannot be correctly retrieved by using the grapheme-phoneme conversion rules. Together, the present study provides evidence that complements the existing brain imaging data and suggests that the time course of the activation of these areas may be earlier than predicted by ERP literature. The implications of our finding on reading models will be discussed.