Pseudohomophones and word recognition

Orthography-phonology consistency in English: Theory- and data-driven measures and their impact on auditory vs. visual word recognition

Research on orthographic consistency in English words has selectively identified different sub-syllabic units in isolation (grapheme, onset, vowel, coda, rime), yet there is no comprehensive assessment of how these measures affect word identification when taken together. To study which aspects of consistency are more psychologically relevant, we investigated their independent and composite effects on human reading behavior using large-scale databases. Study 1 found effects on adults' naming responses of both feedforward consistency (orthography to phonology) and feedback consistency (phonology to orthography). Study 2 found feedback but no feedforward consistency effects on visual and auditory lexical decision tasks, with the best predictor being a composite measure of consistency across grapheme, rime, OVC, and word-initial letter-phoneme. In Study 3, we explicitly modeled the reading process with forward and backward flow in a bidirectionally connected neural network. The model captured latent dimensions of quasi-regular mapping that explain additional variance in human reading and spelling behavior, compared to the established measures. Together, the results suggest interactive activation between phonological and orthographic word representations. They also validate the role of computational analyses of language to better understand how print maps to sound, and what properties of natural language affect reading complexity.

The supramarginal and angular gyri underlie orthographic competence in Spanish language

Orthographic competence allows automatic word recognition and reading fluency. To elucidate how the orthographic competence in Spanish-speaking adults might affect the neurofunctional mechanisms of visual word recognition, 32 young adults equally divided in two groups (HSS: High Spelling Skills, and LSS: Low Spelling Skills) were evaluated using fMRI methods, while they performed an orthographic recognition task involving pseudohomophones. HSS achieved significantly more correct responses and lower reaction times than LSS. Interestingly, LSS showed greater activation in the left angular and supramarginal regions with increased bilateral activation pattern in the inferior frontal gyrus, and the anterior temporal and posterior parietal regions. In contrast, HSS showed a more left-lateralized pattern over these regions along with higher activation of the anterior cingulated gyrus for misspelled words. Results suggest that the differences found in cortical activation patterns might be explained by the higher degree of specialization for word recognition in HSS, a group of participants that due to their greater orthographic skills require less engagement of processing resources to succeed in the task.

Modulation of brain regions involved in word recognition by homophonous stimuli: an fMRI study

We used rapid event-related fMRI to explore factors modulating the activation of orthographic and phonological representations of print during a visual lexical decision task. Stimuli included homophonous word and nonword stimuli (MAID, BRANE), which have been shown behaviorally to produce longer response times due to phonological mediation effects. We also manipulated participants' reliance on orthography by varying the extent to which nonword foils were orthographically typical (wordlike context) or atypical (non-wordlike context) of real words. Key findings showed that reading low-frequency homophones in the wordlike context produced activation in regions associated with phonological processing (i.e., opercular region of the left inferior frontal gyrus [IFG; BA 44]), the integration of orthography and phonology (i.e., the inferior parietal lobule (IPL), and lexicosemantic processing (i.e., left middle temporal gyrus, [MTG]). Pseudohomophones in the wordlike context produced greater activity relative to other nonword trials in regions engaged during both phonological processing (i.e., left IFG/precentral gyrus; BA 6/9]), and semantic processing (triangular region of the left IFG; BA 47). Homophonous effects in the non-wordlike context were primarily isolated to medial extrastriate regions, hypothesized to be involved in low level visual processing and not reading-related processing per se. These findings demonstrate that the degree to which phonological and orthographic representations of print are activated depends not only on homophony, but also on the word-likeness of nonword stimuli. Implications for models of visual word recognition are discussed.

Is there an effect of print exposure on the word frequency effect and the neighborhood size effect?

Orthographic and phonological processing skills have been shown to vary as a function of reader skill (Stanovich & West, Reading Research Quarterly, 24, 402-433, 1989; Unsworth & Pexman, Quarterly Journal of Experimental Psychology, 56A, 63-81, 2003). One variable known to contribute to differences between readers of higher and lower skill is amount of print exposure: higher skilled readers read more often than lower skilled readers, and their increased print exposure is associated with faster responding to words and nonwords in lexical decision tasks. The present experiments examined the effect of print exposure on the word frequency effect and neighborhood size effect. We conclude that the different outcomes reported in previous studies (Chateau & Jared, Memory and Cognition, 28, 143-153, 2000; Lewellen, Goldinger, Pisoni, & Greene, Journal of Experimental Psychology: General, 122, 316-330, 1993) were due to the type of nonwords used in the lexical decision task (regular nonwords versus pseudohomophones). Our results are explained in terms of differences in the reliance on orthographic and phonological information between readers of higher and lower print exposure.

Spelling consistency affects reading in young Dutch readers with and without dyslexia

Lexical-decision studies with experienced English and French readers have shown that visual-word identification is not only affected by pronunciation inconsistency of a word (i.e., multiple ways to pronounce a spelling body), but also by spelling inconsistency (i.e., multiple ways to spell a pronunciation rime). The aim of this study was to compare the reading behavior of young Dutch readers with dyslexia to the behavior of readers without dyslexia. All students participated in a lexical-decision task in which we presented pronunciation-consistent words and pseudowords. Half of the pronunciation-consistent stimuli were spelling consistent and the other half were spelling inconsistent. All three reader groups, that is, students with dyslexia, age-match students, and reading-match students, read spelling-consistent words faster than spelling-inconsistent words. Overall reading speed of students with dyslexia was similar to that of reading-match students, and was substantially slower than that of age-match students. The results suggest that reading in students with or without dyslexia is similarly affected by spelling inconsistency. Subtle qualitative differences emerged, however, with respect to pseudoword identification. The conclusion was that the findings were best interpreted in terms of a recurrent-feedback model.

An fMRI investigation of strategies for word recognition

A common procedure used to study visual word recognition is the lexical decision task (LDT). Behavioral studies have demonstrated that overall performance in this task is modulated by the type of foils presented. There are divergent claims about the impact of different types of foils on overall processing strategies in the LDT: some researchers claim that pseudohomophone foils (e.g., BRANE) encourage participants to selectively de-emphasize phonological processing, while other researchers claim that pseudohomophone foils encourage participants to engage in more extensive processing of all types (orthographic, phonological, and semantic). To evaluate these claims, we conducted a functional magnetic resonance imaging study. Participants (N = 18) completed three lexical decision tasks, each with a different foil type: consonant strings (e.g., BVRNT), pseudowords (e.g., BLINT), and pseudohomophones (e.g., BRANE). We presented homophones (e.g., MAID) and nonhomophones (e.g., MESS) on word trials in order to be able to calculate the homophone effect as a marker of phonological processing for word stimuli in each foil condition. Comparison of behavioral results in the different foil conditions showed that reaction times were longest, error rates were highest, and homophone effects were largest in the LDT with pseudohomophones. Imaging results showed the greatest magnitude of activity in several regions, including the inferior frontal cortex, during the LDT with pseudohomophone foils. A comparison to inferior frontal activity produced during an additional task (rhyme judgment) supported the conclusion that LDT with pseudohomophone foils is a difficult task in which readers engage in an overall response strategy involving extensive processing of phonological information.

Semantic and phonological influences on the processing of words and pseudohomophones

In two experiments, we investigated the relationship between semantics and phonology in the lexical decision task. In the first experiment, lexical decisions to words with large semantic neighborhoods were faster than those to words with sparse semantic neighborhoods. Conversely, this effect of semantic neighborhood was reversed for pseudohomophones (e.g., nale). That is, pseudohomophones based on words with large semantic neighborhoods took longer to reject than did those based on words with sparse semantic neighborhoods. In the second experiment, we found the magnitude of the semantic neighborhood effect for words to be a function of nonword foil type. Taken together, these results indicate that semantic neighborhood size affects processing of both words and pseudohomophones, and that the effect of semantic neighborhood size for words is more pronounced when pseudohomophone foils are employed. These effects are discussed in terms of a model in which the orthographic, phonological, and semantic systems are fully interactive.