Recovery in reading:a treatment study of acquired deep dyslexia in Spanish

Background: Adult oral reading in consistent orthographies, like Spanish, is argued to proceed through sublexical spelling-to-sound conversion. Hence, it was surprising that Davies and Cuetos (2005) observed a Spanish-speaking aphasic patient, MJ, who showed some capacity to read words aloud-producing many semantic errors-alongside an inability to read nonwords. MJ's symptoms are congruent with deep dyslexia. We investigated whether her reading could be improved through therapy addressing those deep dyslexic symptoms (de Partz, 1986). Aims: We expected to see joint improvement in word and nonword reading if therapy aimed at facilitating improvement in nonword reading succeeded, due to the consistency of the orthography. We hypothesised that occurrence of semantic paralexias in reading should decrease as reading skills improved. Alternatively, if therapy had led to the establishment of a lexical-semantic reading strategy not available to the patient before injury, then we anticipated the possibility that semantic errors could actually increase in occurrence as more therapy was received over time. Methods & Procedures: Our study began 18 months after MJ's brain injury, and continued over 2 years. Training was given in completing spelling-sound mappings at the level of grapheme-phoneme correspondences, syllables and words, in phoneme awareness and the combination of phonemes, and in word-picture association. We tested oral reading with the same word and nonword stimuli at four time points through the period of our investigation. We analysed accuracy and error types. Outcomes & Results: By the end of our investigation, MJ presented near-normal word reading and greatly improved, though still impaired, nonword reading. Production of semantic errors disappeared. Thus, adding to some previous case studies (Friedman, 1996), our patient's symptoms evolved from those typical of deep dyslexia to those typical of phonological dyslexia. Conclusions: Our findings show that, even in a language with a transparent orthography, readers can be aided by therapy addressing both lexical and non-lexical reading. The implication is that premorbid reading in a transparent orthography may depend on lexical and non-lexical reading. This conclusion is supported by the superior improvement of word compared to nonword reading, and by the decline in production of semantic paralexias accompanying the improvement in reading performance