Language Experience and the Arcuate Fasciculus in Autism

Background: Delayed echolalia, altered prosody and speech onset delay are among the main language characteristics of autism spectrum disorders (ASD). At the cerebral level, a very common observation in autism is an atypical frontotemporal functional connectivity. The arcuate fasciculus (AF), is the main white matter tract that connects the posterior temporal region of the brain involved in the comprehension of language (Wernicke's area) with the inferior frontal gyrus (Broca’s region), which is involved in language production. The structural properties of the AF have been linked with word acquisition, vocabulary growth (Su et al. 2018) and other linguistic abilities (Salvan et al. 2017). Moreover, lesions damaging the fibers of the AF disconnecting the temporal region from the inferior frontal region result in conduction aphasia, an impairment in the repetition of speech, phonological paraphasias, and impaired speech monitoring and learning.

Objectives: Considering that ASD individuals with speech onset delays (SOD) contrast with those without speech onset delays (No-SOD) in terms of brain activation during speech-like processing (Samson et al, 2015), the aim of this study was to use diffusion imaging tractography protocols to reconstruct the AF and investigate its properties in high-functioning autistic individuals who ranged in age of onset of first words and in language ability.

Methods: Thirty-eight autistic and 28 typically developing individuals, all right-handed and matched for age (14-35 years) and intellectual functioning were scanned using a T1 structural and a diffusion-weighted sequence. Reconstruction of the AF was achieved using a novel deterministic ROI-based approach with FSL, Diffusion toolkit and Trackvis software.

Results: Overall, the autistic group had a significantly smaller left AF volume than the non-autistic group. Within the autistic group, there was a significant correlation between the age of acquisition of speech and the volume of the left AF. Stratifying the autistic group into those with SOD (n=15) or No-SOD (n=23) individuals revealed that the smaller AF volume result was driven by the No-SOD group. Within the SOD group only, greater Fractional Anisotropy (FA) in the AF was also correlated to higher ADOS and ADI Communication scores.

Conclusions: The structural properties of the AF in autism may be associated with communication symptom severity and speech acquisition experience, even in adulthood. Our results are consistent with Moseley et al. (2016), who also found AF reductions in a group of Asperger individuals without SOD compared to typicals. No-SOD autistic individuals, despite a typical age of speech onset, present as adults a less developed AF relative to matched controls. By contrast, in autistics individuals with SOD, their delayed speech is associated with a more developed left AF, that does not differ from the one of typical individuals. Since all the autistic participants under study ended up developing proficient speech as adults, at time of testing, this suggests that SOD and No-SOD autistic people use different neuronal mechanisms for developing speech.