Abstract
Adults can decide rapidly if a string of letters is a word or not. However, the exact time course of this discrimination is still an open question. Here we sought to track the time course of this discrimination and to determine how orthographic information—letter position and letter identity—is computed during reading. We used a go/no-go lexical decision task while recording event-related potentials (ERPs). Subjects were presented with single words (go trials) and pseudowords (no-go trials), which varied in orthographic conformation, presenting either a double consonant frequently doubled (i.e., “ss”) or never doubled (i.e., “zz”) (identity factor); and a position of the double consonant was which either legal or illegal (position factor), in a 2 × 2 factorial design. Words and pseudowords clearly differed as early as 230 msec. At this latency, ERP waveforms were modulated both by the identity and by the position of letters: The fronto-central no-go N2 was the smallest in amplitude and peaked the earliest to pseudowords presenting both an illegal double-letter position and an identity never encountered. At this stage, the two factors showed additive effects, suggesting an independent coding. The factors of identity and position of double letters interacted much later in the process, at the P3 level, around 300–400 msec on frontal and central sites, in line with the lexical decision data obtained in the behavioral study. Overall, these results show that the speed of lexical decision may depend on orthographic information coded independently by the identity and position of letters in a word.