ERP correlates of lexical analysis: N280 re£ects processing complexity rather than category or frequency e¡ects Angele Brunelliere, Michel Hoen and Peter F. Dominey CA Sequential Cognition and Language Group, Institut de Sciences Cognitives, CNRS UMR 5015, Blvd Pinel, 69675 BRON Cedex, France CA Corresponding Author: dominey@isc.cnrs.fr Received 31May 2005; accepted 4 July 2005 In the context of language processing, the N280 is an anterior ne- gative event-relatedpotential pro¢le associated with the lexical ca- tegorization of grammatical function words versus content words. Subsequent studies suggested that this e¡ect was related to word statistics including length and frequency in the lexicon.The current research tests the hypothesis that the N280 e¡ect is related to an index of grammatical complexity. We recorded event-related po- tentials during a sentence reading task. Comparing content versus function words revealed the classic N280. Within function words, we compared the relative pronouns ‘qui’ and ‘que’ (which are iden- tical for length and frequency) that in French indicate a subsequent simple (subject^ subject) and complex (subject^object) relative clause, respectively. A left anterior N280 e¡ect was observed only for ‘que’, supporting our hypothesis that the N280 re£ects grammatical complexity that can be confounded with lexical cate- gory and statistical properties. NeuroReport 16:1435^1438  c 2005 Lippincott Williams & Wilkins. Key words: Content words; Event-related potential (ERP); Function words; Language; Lexical categorization; N280; Syntactic complexity INTRODUCTION Behavioral studies [1] have shown that open class words or content words (e.g. nouns, verbs, adjectives, adverbs, etc.) and closed class words or function words (e.g. prepositions, relative pronouns, determiners, etc.) implicate distinct neu- robiological processes. This process of ‘lexical categorization’ was first associated with a negative shift observed for the reading of function versus content words [2] and classified as a variant of the contingent negative variation [3,4]. Subsequent studies often reported this left frontal negativity peaking around 280 ms as the N280, which distinguished between closed class versus open class words [5–7]. Others argued, however, that the N280 reflected word length and word frequency, which are confounded with lexical categories, as function words are in general shorter and more frequent than content words [8–10]. Interestingly, ter Keurs et al. [11] observed an N280 for closed and not open class words that was not present for patients with Broca’s aphasia who can perceive but not grammatically integrate function words [12,13], again suggesting a link between the N280 and some aspect of grammatical processing. The current study attempts to aid in clarifying the functional significance of the N280. Part of the problem is that lexical category, word length and word frequency are often confounded. Our hypothesis is that the N280 reflects an index of the grammatical complexity of the upcoming processing that is triggered by the lexical item. To test this hypothesis, we should compare event-related potentials (ERPs) for words that indicate different levels of gramma- tical complexity, but otherwise have the same length, frequency and lexical category. French provides such an opportunity, by comparing responses to the relative pro- nouns ‘qui’ and ‘que’ in the context as in: (1) Le chien qui chasse le chat enterre la balle. (The dog that chases the cat buries the ball.) (2) Le chien que le chat chasse enterre la balle. (The dog that the cat chases buries the ball.) Sentence (1) is a ‘subject–subject’ relative sentence in which the principal noun phrase (the dog) is the subject of both the main and the relative phrases. In contrast, in sentence (2), the principal noun phrase is the subject of the main phrase, and the object of the relative phrase, so it plays a double thematic role. This ‘subject–object’ syntactic construction was demonstrated to be of greater processing complexity [14,15], likely owing in part to increased processing costs [16], and the recruitment of specialized processes. Indeed, ERP [17] and brain imaging [18] studies have indicated that processing of the syntactically complex portions of such sentences is associated with distinctive patterns of cortical activation. The interesting cross-linguistic point here is that while in English the same word ‘that’ is used in both cases, in French, these two functional roles are expressed by two different function words ‘qui’ and ‘que’ that have the same length and same frequency, 7121 per million for ‘qui’ and 9208 per million for ‘que’ [19]. Thus, by comparing the possible N280 effects associated with these two relative pronouns, we can COGNITIVE NEUROSCIENCE AND NEUROPSYCHOLOGY NEUROREPORT 0959-4965  c Lippincott Williams & Wilkins Vol 16 No 13 8 September 2005 1435 Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.