Word form access dyslexia: Understanding the basis of visual reading errors

Abstract

We report the case of a patient V.E.M. who exhibits the infrequently described syndrome of visual dyslexia. V.E.M. read many single words accurately and rapidly; however, when words were read inaccurately, the majority of responses were alternate real words bearing a strong visual relationship to the target items. A series of observational and experimental investigations of her reading skills revealed that response accuracy was strongly affected by word frequency and mildly affected by concreteness but not influenced by word length, orthographic neighbourhood size, or semantic priming. It is argued that V.E.M.'s visual dyslexia results from a deficit of the visual word form system. More specifically, we propose that this deficit has an “access” quality, with visual word form representations remaining relatively intact despite an impairment in the processes by which these representations are activated. Taking the results of this study together with previous reports of dyslexic patients who make a high proportion of visual paralexic errors, we also suggest that visual dyslexia represents a multicomponent dyslexic syndrome for which phonological impairment and not semantic impairment may be a necessary condition.

We are very grateful to Professor Martin Rossor for allowing us to study a patient under his care. We are also grateful to Professor Max Coltheart for providing us with the orthographic neighbourhood ratings for our test stimuli and Dr John Stevens for his opinion on the neuroimaging data. Finally, we are indebted to V.E.M. for her patience and willingness to help us with this work.

Notes

¹ The difference between visual input logogens and visual word forms can be considered here purely in terms of the activation required to access semantics: Logogens can only activate semantic units once a threshold has been reached, whereas Shallice and Warrington (1975) suggested simply that the most active graphemic unit would elicit semantic activation.

² Since the original report of an “access” deficit (patient A.R.; Warrington & Shallice, 1979), access deficits have been classified into refractory and nonrefractory subtypes (Crutch & Warrington, 2001), the former being indicated not only by inconsistency but also by a sensitivity to temporal factors (i.e., the length of the response–stimulus interval; RSI). It is unclear whether access deficit observed in the original case report (A.R.) had a refractory quality as the influence of temporal factors was not explored at that time. The access disorders described in the current patient are held to be of a nonrefractory type, as indicated by a relative insensitivity to temporal factors such as RSI duration.

³ Orthographic neighbours are all those words that can be formed by changing one letter in the target word.

⁴ The influence of such semantic information upon S.C.'s word-reading accuracy was not directly tested.

⁵ Shallice and Warrington (1980) used the term “central” deep dyslexia to distinguish the dyslexic syndrome arising from damage direct to semantic representations from the “input” and “output” forms or the disorder. Here we use the term “core” rather than “central” to refer to storage damage or access deficits of word form representations in order to avoid confusion with the broader distinction between central and peripheral dyslexic syndromes.