The possibility and extent of thought without language have been subject to much controversy. Insight from nonor minimally verbal humans can inform this debate empirically. Since most such individuals are on the autism spectrum, of which they make up a sizable 25-30%, an important connection between language and autism transpires. Here we propose a model which makes sense of this link and explains why the non-verbal human mind, as present evidence suggests, represents a fundamentally different cognitive phenotype. This model views the relevant part of the autism spectrum as reflecting the breakdown of a cognitive phenotype of which language is an inherent element and which will manifest principled limitations in the latter's absence.

Background & aims Individuals with non- or minimally verbal autism (nvASD) are primarily characterized by a severe speech production deficit, with speech limited to no or only a few words by school age. Significant unclarity remains over variability in language profiles across the lifespan, the nature of the language impairment seen, and (dis-) associations between linguistic and nonverbal cognitive measures. Methods To address these questions, we recruited both a school-age and an adult group with nvASD (total N = 49) and investigated relations between expressive and receptive language, and between these and nonverbal intelligence quotient (NVIQ) and sense-making capacities (the ComFor test). Results Results revealed limited variation across this sample in receptive language, which in turn predicted expressive language levels. Importantly, an upward trend in verbal mental age (VMA) across increasing chronological age was seen in the youngsters (only). A radical dissociation between NVIQ and both expressive and receptive language transpired as well, and a subset of individuals with normal NVIQ were comparable in terms of any other cognitive aspect. Sense-making reached symbolic levels in 62.2% of the sample and loaded on both verbal and nonverbal factors. Conclusions These patterns inform theories of nvASD by revealing an impairment that is not conceptualizable as one of expressive language only, sharply limits learning opportunities across the lifespan, and cannot be compensated for by nonverbal cognition. Implications These findings stress the need to seize developmental opportunities that may disappear when youngsters turn into adults, via therapies that specifically target language as a central cognitive system comprising both production and comprehension.

Language capacities in autism spectrum disorders (ASD) range from normal scores on standardized language tests to absence of functional language in a substantial minority of 30% of individuals with ASD. Due to practical difficulties of scanning at this severe end of the spectrum, insights from MRI are scarce. Here we used manual deterministic tractography to investigate, for the first time, the integrity of the core white matter tracts defining the language connectivity network in non-verbal ASD (nvASD): the three segments of the arcuate (AF), the inferior fronto-occipital (IFOF), the inferior longitudinal (ILF) and the uncinate (UF) fasciculi, and the frontal aslant tract (FAT). A multiple case series of nine individuals with nvASD were compared to matched individuals with verbal ASD (vASD) and typical development (TD). Bonferroni-corrected repeated measure ANOVAs were performed separately for each tract—Hemisphere (2:Left/Right) × Group (3:TD/vASD/nvASD). Main results revealed (i) a main effect of group consisting in a reduction in fractional anisotropy (FA) in the IFOF in nvASD relative to TD; (ii) a main effect of group revealing lower values of radial diffusivity (RD) in the long segment of the AF in nvASD compared to vASD group; and (iii) a reduced volume in the left hemisphere of the UF when compared to the right, in the vASD group only. These results do not replicate volumetric differences of the dorsal language route previously observed in nvASD, and instead point to a disruption of the ventral language pathway, in line with semantic deficits observed behaviourally in this group.

No abstract

Autism spectrum disorder (ASD) involves a large variety of language capacities, from normal scores on standardized language tests to absence of functional language in a substantial minority of 30% of individuals with ASD. Due to practical difficulties of scanning at this severe end of the autism spectrum, insights from MRI are scarce. Here we used manual deterministic tractography to investigate, for the first time, the integrity of the core white matter tracts defining the language connectivity network in nonverbal ASD (nvASD): the arcuate (AF), inferior fronto-occipital (IFOF), inferior longitudinal (ILF) and uncinate (UF) fasciculi, and the frontal aslant tract (FAT). A multiple case series of nine individuals with nvASD were compared to normative benchmarks from matched individuals with verbal ASD (vASD) and typical development (TD). Bonferroni-corrected repeated measure ANOVAs were performed separately for each tract –Hemisphere (2:Left/Right) x Group (3:TD/vASD/nvASD). Results revealed (i) a main effect of group consisting in a reduction in fractional anisotropy (FA) in the IFOF in nvASD relative to TD; (ii) a significant interaction of hemisphere and group in the UF, which showed reduced volume in the left hemisphere when compared to the right, in the vASD group only; and (iii) main effects of hemisphere in both the AF (left-lateralized in volume) and the ILF (left-lateralized in FA). These results do not replicate deficits of the dorsal language route previously observed in nvASD, and instead point to a disruption of the ventral language pathway, which is in line with semantic deficits observed behaviourally in this group.