• Citer

Does bigram frequency genuinely impact visual word recognition? An ERP study

par Sauval, Korinne ;Chetail, Fabienne
Référence 20th conference of the European Society for Cognitive Psychology (ESCoP) (September 3-6th: Potsdam)
Publication Publié, 2017

Poster de conférence

• Accès en ligne
• Détails
• Contenu
• Statistiques

Résumé :

Repeatedly exposing readers to print makes them sensitive to the fact that some letters or letter groups occur more frequently than other ones. Among the different types of orthographic regularities, the role of bigram frequency has been the most hotly debated in the field. At an empirical level, some studies reported detrimental effects of bigram frequency in the lexical decision task, with words including high-frequency bigrams being processed less efficiently than words with low-frequency bigrams. The most developed proposal to explain this detrimental effect is that low-frequency bigrams would help to reduce the set of lexical candidates, since they are the most diagnostic to identify a word. At the sublexical level, however, bigram and letter frequency would facilitate processing. Hence, depending on the task, the early facilitative and the later detrimental effects could cancel each other, leading to null effects. Another proposal is that bigram frequency has no influence in visual word recognition per se and the few effects reported in the past would be due to confounds with word frequency or orthographic neighborhood. Given this mixed pattern of results and the alternative explanations, the aim of the present study was to test the impact of bigram frequency using EEG, since this method enables one to observe joint facilitative and detrimental effects during the timing of word recognition. Two categories of words contrasted on summed bigram frequency were selected, while carefully controlling for potential confounded variables. We recorded neural activity through 64 channels while participants performed a semantic detection task. Overall, waveforms showed some evidence of a facilitative effect around 150 ms and of a detrimental effect from 250 ms after the word onset. These results support the hypothesis that bigram frequency has an influence during word processing. Results will be discussed in the interactive activation framework of visual word recognition.