Functional and time-course changes in single word production from childhood to adulthood Marina Laganaro ⁎, Hélène Tzieropoulos, Ulrich H. Frauenfelder, Pascal Zesiger Faculty of Psychology, University of Geneva, Switzerland abstract article info Article history: Accepted 11 February 2015 Available online 19 February 2015 Keywords: Development Language Picture naming ERP Picture naming tasks are widely used both in children and adults to investigate language production for research and for assessment purposes. The main theoretical models of single word production based on the investigation of picture naming in adults provide a detailed account of the principal mental operations involved in the trans- formation of an abstract concept into articulated speech and their temporal dynamics. These models and in particular their time-course do not apply directly to children who display much longer production latencies than adults. Here we investigate the functional processes and the temporal dynamics of word encoding in school-age children and adults. ERPs were analysed from picture onset to the onset of articulation in 32 children and 32 adults performing the same overt picture naming task. Waveform analyses were not informative since differences appeared throughout the entire period, due to an early shift of waveform morphology and to larger amplitudes in children. However, when the sequences of periods of topographic stability were considered, differ- ent patterns of electric fields at scalp only appeared in approximately the first third of the analysed period, cor- responding to the P1–N1 complex. From about 200 ms in adults and from 300 ms in children to articulation onset similar patterns of global topography were observed across groups but with a different time distribution. These results indicate qualitative changes in an early time-window, likely corresponding to pre-linguistic processes, and only quantitative changes in later time-windows, suggesting similar mental operations underlying lexical processes between age-school children and adults, with temporal dynamic changes during development. © 2015 Elsevier Inc. All rights reserved. Introduction Both adults and children can produce a familiar word corresponding to the concept they wish to communicate fluently and effortlessly. Such a transformation of a pre-linguistic concept into articulated speech sounds requires a series of cognitive and neurophysiological processes which have been investigated widely with picture naming tasks in psy- cholinguistic (Levelt et al., 1999; Dell, 1986) and neuroimaging studies (Indefrey and Levelt, 2004; Price et al., 2005) on (young) adult speakers. Although the sequential or parallel nature of the processes underlying word production is still under debate, there is a general agreement on the necessity and existence of a minimal set of mental operations involved in such tasks. When speakers produce a word corresponding to a depicted object, they first analyse the picture visually and recognize it before beginning linguistic planning. Word encoding then entails lexical–semantic processes, that is, the lexical selection or the retrieval of the word – the lemma – in the mental lexicon and lexical–phonological processes or the encoding of the phonological form – lexeme – , which constitutes the input to the preparation of motor plans (phonetic encoding) for articulation. The time-course of these processes has been estimated in a meta-analysis of behavioural and event-related po- tential (ERP) studies by Indefrey and Levelt (2004), and then updated on the basis of more recent evidence (Indefrey, 2011). In picture nam- ing tasks, visual and conceptual processes are estimated to take place during the first 190–200 ms after picture presentation, followed by lexical–semantic processes (lemma retrieval) until about 275 ms. Word form (lexeme) retrieval and phonological encoding processes are thought to be engaged until 400–500 ms followed by phonetic encoding and motor execution. These estimates of the timing of the different cognitive processes have been made for adult word production. They provide a general framework for studying the dynamics of word production in adults, although mean production latencies vary considerably across experi- ments, raising the question of rescaling the duration of processing stages. Indefrey (2011) suggested that linear rescaling in the case of faster or longer response latencies is probably not the correct approach as, depending on the experimental conditions, specific processes may take less or more time. Laganaro et al. (2012) suggested that variation in object naming speed in a group of young adults is attributable to a single underlying electrophysiological process starting around 200 ms that lasts longer in slower participants. The question of rescaling the du- rations of processing stages is all the more relevant when populations NeuroImage 111 (2015) 204–214 ⁎ Corresponding author at: FPSE, University of Geneva, 40, Bd Pont d'Arve, CH-1211 Geneva 4, Switzerland. E-mail address: marina.laganaro@unige.ch (M. Laganaro). http://dx.doi.org/10.1016/j.neuroimage.2015.02.027 1053-8119/© 2015 Elsevier Inc. All rights reserved. Contents lists available at ScienceDirect NeuroImage journal homepage: www.elsevier.com/locate/ynimg