Academic journal article Cognitive, Affective and Behavioral Neuroscience

A Masked Priming ERP Study of Letter Processing Using Single Letters and False Fonts

Academic journal article Cognitive, Affective and Behavioral Neuroscience

A Masked Priming ERP Study of Letter Processing Using Single Letters and False Fonts

Article excerpt

Previous event-related potential (ERP) research on letter processing has suggested that a P150 reflects lowlevel, featural processing, whereas a P260 reflects high-level, abstract letter processing. In order to investigate the specificity of these effects, ERPs were recorded in a masked priming paradigm using matching and nonmatching pairs of letters (e.g., g-g, g-j) and false fonts (e.g., ... ). If the P150 priming effect indexes featural processing, there should be no effect of condition on the P150, since the letters and false fonts shared visual features. If the P260 priming effect indexes the processing of abstract letter representations, it should be evident only in the letter condition. As was expected, the P150 priming effect was similar for letters and false fonts; however, the P260 priming effect was also similar between conditions. Thus, the P260 priming effect may not be sensitive to abstract letter processing per se, or such processing may be extremely abstract.

(ProQuest: ... denotes formula omitted.)

Although a number of studies have investigated word processing in the human brain (e.g., Cohen et al., 2000; McCandliss, Cohen, & Dehaene, 2003; Petersen, Fox, Snyder, & Raichle, 1990), few studies have explored single letters as the parts that make up those words. However, a handful of behavioral studies have investigated letter processing in terms of case (uppercase vs. lowercase; e.g., Boles & Eveland, 1983; Bowers, Vigliocco, & Haan, 1998; Haber & Cole, 1980) and identity (real letters vs. nonletter stimuli, such as symbols, numbers, or letter-like characters; e.g., Gibson, Gibson, Pick, & Osser, 1962; Lachmann & van Leeuwen, 2004). The results of such studies have been used to develop a theoretical model of single-letter processing (Jacobs & Grainger, 1991) that has been extended recently to include event-related potential (ERP) data outlining the time course of the various proposed stages of letter processing (Petit, Midgley, Holcomb, & Grainger, 2006). These ERP data suggest that a P150 priming effect reflects featural processing, whereas a P260 priming effect reflects abstract letter processing (Petit et al., 2006). Here, we used a masked priming ERP paradigm with real letters and letter-like false font stimuli to further investigate the nature of these priming effects. We hypothesized that if the P150 priming effect indexes featural processing, letters and false fonts created from letter features should elicit similar P150 priming effects. In contrast, if the P260 priming effect indexes abstract letter processing, only real letter stimuli should elicit this effect.

Models of Letter Processing and ERP Evidence

In the Jacobs and Grainger (1991) model of letter processing, an extension of an interactive activation model for word processing (McClelland & Rumelhart, 1981), activations occur at three levels: a feature level, a casespecific level, and an abstract (name) level. Others have suggested similar models of letter processing including visual featural, case-specific, and abstract levels (e.g., Arguin & Bub, 1995; Selfridge, 1959); with regard to the latter, some have proposed both font-free but casespecific abstract letter representations and font-free, casefree abstract letter representations (Brunsdon, Coltheart, & Nickels, 2006). Computational models have also been constructed that encode letters independently of case, in terms of abstract letter identities (Polk & Farah, 1997). However, these models do not speak to the neural time course of letter processing or the online automaticity of letter processing in fluent readers.

With real-time temporal sensitivity, the recording of ERPs is an excellent method for investigating the time course of letter processing. In a masked priming study, Petit et al. (2006) presented prime-target pairs of letters that were either matched or unmatched for both name and case. The authors reported that the ERP waveform elicited by letter reading included three different components indexing the early processing of letters. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.