We tested the hypothesis that crowding effects are responsible for the reading slowness characteristic of developmental dyslexia. A total of twenty-nine Italian dyslexics and thirty-three age-matched controls participated in various parts of the study. In Experiment 1, we measured contrast thresholds for identifying letters and words as a function of stimulus duration. Thresholds were higher in dyslexics than controls for words (at a limited time exposure) but not for single letters. Adding noise to the stimuli produced comparable effects in dyslexics and controls. At the long time exposure thresholds were comparable in the two groups. In Experiment 2, we measured the spacing between a target letter and two flankers at a fixed level of performance as a function of eccentricity and size. With eccentricity, the critical spacing (CS) scaled in the control group with 0.62 proportionality (a value of b close to Bouma's law, 0.50) and with a greater proportionality (0.95) in the dyslexic group. CS was independent of size in both groups. In Experiment 3, we examined the critical print size (CPS), that is, the increase in reading rate up to a critical character size (S. T. Chung, J. S. Mansfield, & G. E. Legge, 1998). CPS of dyslexic children was greater than that of controls. Individual maximal reading speed was predicted by individual bs (from Experiment 2). The maximal reading rate achieved by dyslexics at CPS (and also for larger print sizes) was below the values observed in controls. We conclude that word analysis in dyslexics is slowed because of greater crowding effects, which limit letter identification in multi-letter arrays across the visual field. We propose that the peripheral reading of normal readers might constitute a model for dyslexic reading. The periphery model accounts for 60% of dyslexics' slowness. After compensating for crowding, the dyslexics' reading rate remains slower than that of proficient readers. This failure is discussed in terms of a developmental learning effect.

Vocal reaction times (RTs) in naming 3- to 8-letter words were measured in proficient and dyslexic readers (Study 1). In proficient readers, RTs were independent of word length up to 5-letter words, indicating parallel processing. In the 5- to 8-letter range, RTs increased linearly, indicating sequential processing. Reading experience was associated with both faster discrimination of individual elements and parallel processing of increasingly large word parts. In dyslexics, RTs increased linearly with increasing length indicating reliance on sequential decoding. Individual analysis indicated 2 profiles of RTs (Types A and B). In Study 2, the distinction between A and B dyslexics was not associated with the use of different reading procedures. However, a more marked speed deficit characterized Type B dyslexics.

The study aims to verify whether phonologic and rapid automatized naming (RAN) deficits are present and associated in Italian dyslexic children and whether they differentially affect dyslexics with and without a history of previous language delay (LD). According to the phonologic core deficit hypothesis, dyslexia may stem from impairment of the representation and manipulation of phonemes and may be closely associated with oral language deficits. However, deficits in tasks not requiring fine-grained phonologic representations, such as RAN, have also been described in dyslexic children. Thirty-seven children were selected on the basis of a reading deficit and were assigned to 2 groups according to whether or not they had a history of early LD as determined retrospectively by parental report. A battery of reading and writing, verbal working memory, metaphonologic, RAN, and visual search tests were administered. RAN deficits were shared by most dyslexics (with and without a history of LD), whereas phonologic deficits were mainly associated with a previous LD. This last condition did not result in a more profound impairment of reading and writing decoding skills. In a shallow orthography such as Italian, RAN, not phonologic deficits, may represent the main cognitive marker of developmental dyslexia.

This study examined the influence of rapid automatization naming (RAN) measures on various parameters of reading performance in children who were native speakers of a language with a shallow orthography (Italian). Participants included 281 children enrolled in first-to-sixth grade. They were given a Naming test, in which they had to name rapidly matrices of colors, objects, or digits, a Cancellation test, using the same stimulus materials, and an oral Articulation test. Performance on all tests improved steadily across ages tested. Performance on the Naming test, but not on the Cancellation and Articulation tests, predicted speed and accuracy in reading; none of these measures reliably predicted the reading comprehension measure. Data on a Blending test were also available for a subsample of first- and third-graders. Both RAN and phonological ability contributed independently to the prediction of reading ability (accuracy and speed) in these participants. The results extend observations on RAN to an orthographically shallow language (Italian) and suggest an element of continuity between languages with opaque and transparent orthographies.