Relationship Between Handedness and Language Lateralization In Autism

Background: Most neuroimaging studies investigating language in autism group-match participants with autism and typically developing controls on qualitative handedness of right or left. Evidence suggests, however, that the relationship between handedness and language lateralization may be quite different in autism versus in typical development. Lack of similar relationship between handedness and language lateralization in autism and control groups could affect the results of studies that simply control for handedness.

Objectives: The primary objective of the study was to examine the relationship between language lateralization and handedness in autism in comparison to typical development.

Methods: Forty-six males with autism and thirty-six typically developing control males, ages 9 to 42 years, were presented with an auditory language task during fMRI. The task required participants to think of a word in response to recorded, spoken phrases, such as “the water that falls from the sky” or “the funny man at the circus”. Signal during rest periods was subtracted from that occurring in response to the sentence stimuli. A language lateralization index (LI) was calculated for Broca’s and Wernicke’s areas for each participant using the formula LI = (Left – Right) / (Left + Right), where Left and Right represent the total number of active voxels within the respective hemisphere. Handedness index scores, which range between 100 (completely right-handed) and -100 (completely left-handed) were obtained using the Edinburgh Handedness Inventory. Data from only right-handed individuals was analyzed; the number of left-handed individuals was too few to be examined.

Results: A smaller proportion of the right-handed participants with autism had leftward language lateralization index scores (75.6%) compared to typically developing controls (94.4%; p = 0.0115). The difference in the autism group appeared to be primarily driven by less left-lateralization in Broca’s area and to a lesser extent by Wernicke’s area. When the interaction between lateralization in Broca’s and Wernicke’s areas was examined, there was less left-lateralization concordance in the autism group; more autism participants had discordant lateralization (e.g, L>R Broca's with R>L Wernicke's) and rightward lateralization in both Broca’s and Wernicke’s areas.

Conclusions: The differing relationship between language lateralization and handedness in autism and typical development has implications for genetic and neurobiological studies of autism. The different relationship could also affect the results of neuroimaging studies that attempt to control for group differences in language lateralization by using handedness as a surrogate. The effects of other covariates on the outcomes of interest in these studies could be masked without accounting for this previously unknown relationship.