Semantic loss without surface dyslexia

Abstract
Recent models of reading, including some computational models, suggest that error-free performance in reading aloud may be reliant, at least in part, on support from the semantic system. Such models predict that, following semantic impairment, a pattern of acquired dyslexia known as ‘surface dyslexia’ will be an inevitable consequence. We present data from a patient with dementia of Alzhelmer type who shows a severe semantic impairment but near error-free performance on reading aloud real words (including low-frequency, irregular, abstract words) and non-words. The data are discussed in light of various models of reading and in relation to previously reported cases.