Parafoveal previews and lexical frequency in natural reading: Evidence from eye movements and fixation-related potentials.

Abstract

Participants’ eye movements and electroencephalogram (EEG) signal were recorded as they read sentences displayed according to the gaze-contingent boundary paradigm. Two target words in each sentence were manipulated for lexical frequency (high vs. low frequency) and parafoveal preview of each target word (identical vs. string of random letters vs. string of Xs). Eye movement data revealed visual parafoveal-on-foveal (PoF) effects, as well as foveal visual and orthographic preview effects and word frequency effects. Fixation-related potentials (FRPs) showed visual and orthographic PoF effects as well as foveal visual and orthographic preview effects. Our results replicated the early preview positivity effect ( Dimigen, Kliegl, & Sommer, 2012 ) in the X-string preview condition, and revealed different neural correlates associated with a preview comprised of a string of random letters relative to a string of Xs. The former effects seem likely to reflect difficulty associated with the integration of parafoveal and foveal information, as well as feature overlap, while the latter reflect inhibition, and potentially disruption, to processing underlying reading. Interestingly, and consistent with Kretzschmar, Schlesewsky, and Staub (2015) , no frequency effect was reflected in the FRP measures. The findings provide insight into the neural correlates of parafoveal processing and written word recognition in reading and demonstrate the value of utilizing ecologically valid paradigms to study well established phenomena that occur as text is read naturally.

Eye Movements and Fixation-Related Potentials in Reading: A Review

“…Fairly consistent results have been observed between EM and FRP measures for PoF effects of parafoveal preview [106,109,120], preview effects of predictability [26,108,113], frequency [113,117,120], and type of preview [110,114,117,120], foveal effects of text type [111], inter-letter spacing [118], repetition [107,110,112], word predictability [26,104,108,113], syntactic and semantic violations [116,119] and foveal load [114]. However, inconsistencies in EM and FRP results have been observed for PoF effects of semantic relatedness [106,[108][109][110]115], preview effects of semantic relatedness [110,115], foveal effects of semantic relatedness [108,110,115] and word frequency [113,120]. It remains the case, though, that co-registration investigations of aspects of reading are in their formative stages (with some effects being investigated in single studies only) and a greater body of experimental data is a necessity before firm conclusions may be formed as to the kinds of experimental manipulations that regularly and consistently produce FRP effects of specific kinds.…”

“…The second neural correlate of identity preview benefit is observed between 360-400 ms, when valid previews elicit more positive amplitudes than invalid previews over central sites of the scalp [102,110,115], and between 300-500 ms, when valid previews elicit more negative amplitudes than invalid previews over occipital areas of the scalp [121]. This late preview effect on the N400 component was not observed in Degno et al [120], indicating that this might be related to naturalness of the reading task or to baseline choices. The eye movement results also mapped onto the FRP data.…”

“…First, studies have consistently observed two neural correlates of identity preview benefit. The first neural correlate is such that between 200-280 [110,114], between 140-200 and 200-300 [117], and between 120-300 [120,121], N1 amplitudes are more positive for identical previews compared to invalid previews, largely over occipito-parietal and temporal areas of the scalp. The only study which failed to find such an effect [115] involved a methodological difference such that linked mastoids were used as an offline reference (see [39] for discussion).…”

“…Indeed, both responses appear to be associated with the uptake of visual information [61], and to have a common neural generator in the visual cortex [127]. The P1 component seems to be modulated by the nature of parafoveal preview [120,121], as well as the visual form of the text [111].…”

Jalapeno or jalapeño: Do diacritics in consonant letters modulate visual similarity effects during word recognition?

“…First, although most effects obtained with the masked priming paradigm have been extended to a sentence reading using, for instance, Rayner's (1975) boundary, it is important to examine in future research the role of diacritics in a standard reading scenario (e.g., see Johnson, Perea, & Rayner, 2007, for evidence of transposed-letter similarity effects during reading; see also Marcet & Perea, 2018b, for evidence of visual similarity effects during reading). In a ideal scenario, this could be combined with a co-registration of the fixation-related potentials (see Degno et al, 2019). Second, Spanish orthography only contains one consonant letter with a diacritical sign (i.e.,ñ), and this may limit the generality of the findings.…”