Poster A22, Thursday, August 16, 10:15 am – 12:00 pm, Room 2000AB

The left lateralization of print-specific negativity depends on visual attention in competitive stimuli.

Tomoki Uno1,2, Takashi Katakura3, Tetuko Kasai4;1Graduate School of Education, Hokkaido University, 2Japan Society for Promotion of Science Research Fellow, 3School of Education, Hokkaido University, 4Faculty of Education, Hokkaido University

INTRODUCTION: Fluent reading requires to implicitly decode the phonological information from logographic scripts. Previous studies using event-related potentials (ERPs) showed that such implicit grapheme-to-phoneme conversion was reflected by the print-specific N170, which was typically distributed over left occipito-temporal areas (Bentin et al., 1999; Maurer & McCandliss, 2007). Although early print processing may highly be automatized, recent studies have suggested that the leftward asymmetry of the N170 depends on visual attention based on observations in separate experiments (Okumura et al., 2014, 2015). The present study tested this notion, by directly manipulating visual attention for letter/symbol strings that were spatially overlapped with random-dot fields. We asked participants to attend either the strings or dots, so that the amount of attention available for letter strings were varied. METHODS: Twelve native Japanese speakers (6 females, 21-27 years) participated in this experiment. Hiragana words, nonwords and symbols strings (107 stimuli each) were spatially overlapped with the random-dot field (1° by 4.1°) and presented for 100 ms with interstimulus intervals of 300-600 ms (50 ms/step). In the attend-string condition, participants were required to detect navy letter/symbol strings (i.e., target) and to ignore navy-dot field (i.e., nontarget), while the targets and nontargets were reversed in the attend-dot condition. Each block contained 279 standard stimuli (both strings and dots were black), and 21 targets and 21 nontargets in both conditions. RESULTS: RTs were faster and Hit rates were higher in the attend-string task compared to the attend-dot task (ps<.01), and FA rates were higher in the attend-dot condition (p=.009). These behavioral results indicate that attending to strings was easier than attending to dots and that letter/symbol strings were more salient objects. In ERP results during 150-200 ms poststimulus, hiragana words induced a more negative deflection compared to the other stimuli (p=.006), which did not differ across the attention conditions (p=.26). Importantly, a left-lateralized negative enhancement for prints was observed only in the attend-string condition during 200-300 ms (p=.03). DISCUSSIONS: In the present study we observed a pattern of results similar to that in previous studies during C1/P1 and N170 latencies (e.g., Okumra et al., 2015; Proverbio & Adorni, 2009). However, interestingly, the letter-string/word effects were generally delayed, which may be due to additional time required for segregating letter strings from visual noise (i.e., dots). As the earliest effect, we found a word-specific negativity during 150-200 ms regardless of the attention conditions. Such insensitivity to attention suggests that word representations in early stages of the cortical hierarchy can be activated basically in an automatic fashion. Most importantly, attending to letters did elicit a left-lateralized negativity during 200-300 ms, which may be a correlate of the typical print-specific N170. Given that the left-lateralization of print-specific negativity reflects phonological mapping (Maurer & McCandliss, 2007), it is suggested that attentional resources are necessary for activating phonological information from prints.

Topic Area: Perception: Orthographic and Other Visual Processes