A double dissociation between sensorimotor impairments and reading disability: A comparison of autistic and dyslexic children

Abstract

Does sensorimotor dysfunction underlie reading impairment? To investigate this question, a battery of literacy, phonology, auditory, visual, and motor tests were administered to age- and ability-matched groups of dyslexic, autistic, and control children. As in previous studies, only a subset of the dyslexic children had sensory and/or motor impairments, whilst some dyslexics were entirely spared, suggesting that sensorimotor impairments are not necessary to cause reading disability. A subset of autistic children was also found to have sensorimotor impairments; however, some of these children did not have reading problems, suggesting that sensorimotor impairments are not sufficient to cause reading disability. We conclude that sensorimotor and reading impairments are doubly dissociable. Sensorimotor impairments do not seem to be the cause of reading disability, but can be seen as nonspecific markers for developmental disorder.

Acknowledgments

We are grateful to Peter Hansen (Department of Physiology, Oxford University) for supplying the visual tasks. We would also like to thank the following: all the children who took part, and their parents, for their time, effort, and cooperation; All Saints Benhilton Primary School, the Autism Unit at Axton Chase School, Hillingdon Manor School, and the Marlborough Unit at Byron School for their patience and assistance; and Martin Turner for his help with recruiting participants. This study was funded by a Marie Curie fellowship of the European Community programme Quality of Life (QLGI-CT-1999–51305; F.R.), and by a Medical Research Council grant (G9617036; U.F.).

Notes

¹ Reliability analysis indicated that performance between the two trials of each task was inconsistent, more so for some tasks than others (Cronbach's alpha: /ba/–/da/ = .24, coat–goat = .33, FM 2 Hz = .65, speech formant discrimination = .35, nonspeech formant discrimination = .56). Such inconsistency may indicate additional task demands required to perform these tasks well; indeed, we believe that these tasks have a particularly high attentional load. We therefore took the best performance on a particular task as the most reliable measure of a child's auditory abilities, independent of other task demands.

² Reliability analysis indicated that performance between the two trials of each task was consistent; Cronbach's alpha: motion (ranked data as not normally distributed) = .74, form = .6. We therefore took the average performance to be the most reliable measure of a child's abilities.

³ Reliability analysis indicated that performance between the two trials of each task was highly consistent; Cronbach's alpha: bead threading (ranked data) = .84, finger and thumb = .87, heel-to-toe (ranked data) = .88, stork balance (ranked data) = .84. We therefore took the average performance to be the most reliable measure of a child's abilities.

⁴ It should be noted that the phonological measures used here may tap other abilities, particularly general processing speed in the rapid automated naming tests. We include these tests as they have been shown to be sensitive to the phonological impairment in dyslexia (see Frederickson et al., 1997). However, it is possible that some autistic children may therefore be defined as phonological outliers for nonphonological reasons.

⁵ Furthermore, of the 3 autistic and 2 control children who were outliers on the visual motion task but had good reading, only 1 of the autistic children was also an outlier on the control form task indicating that, for the majority, their sensorimotor impairments were specific rather than attributable to more general factors such as attention.