Poster E69, Monday, March 27, 2:30 – 4:30 pm, Pacific Concourse
Orthographic codes in the ventral visual system and the reading network revealed by complex grapheme manipulation
Florence Bouhali1,2, Zoé Bézagu1, Stanislas Dehaene3,4, Laurent Cohen1,5; 1Inserm U 1127, CNRS UMR 7225, Sorbonne Universités, UPMC Univ Paris 06 UMR S 1127, Institut du Cerveau et de la Moelle épinière, ICM, 2Université Paris Descartes, 3Collège de France, 4INSERM, CEA, NeuroSpin, U992, Gif-sur-Yvette, France, 5AP-HP Hôpital de la Pitié Salpêtrière
The meaning of a written word can be accessed through a direct lexical route, or through a phonological route that relies on grapheme-to-phoneme mapping. While the phonological route requires fine-grained information on the precise ordering of letters to isolate graphemes, the lexical route is likely to operate using parallel processing of letters to select candidate words efficiently from the lexicon (Grainger and Ziegler 2011). At the neural level, the Visual Word Form Area (VWFA) has been described as a heterogeneous region specialized in the recognition of letters and words, feeding both reading routes. How this region may implement both fine- and coarse-grained orthographic codes and feed the relevant information into the reading network remains unknown. Taking advantage of the high frequency of multiletter graphemes in French, we designed a reading fMRI experiment modulating the reliance on fine-grained orthographic encoding. Letter strings were split into subunits, corresponding either to graphemes (congruent condition) or non-graphemic letter groups (incongruent condition), made salient using both letter spacing and alternating font color. To further influence the reliance on each reading route, lexical status (word vs. pseudoword) and task (lexical decision vs. reading aloud) were manipulated. In agreement with the role of graphemes as a relevant sublexical unit, incongruently chunked words were processed with more difficulty, especially in conditions with highest phonological demands. When reading pseudowords aloud, the perturbation of grapheme encoding induced higher activations in Broca’s area and selected peaks of the VWF system, supporting the existence of two distinct codes in the VWFA.
Topic Area: LANGUAGE: Other