Brain Mechanisms for Language Processing in Adolescents With Autism Spectrum Disorder

About 30% of children with Autism Spectrum Disorder (ASD) fail to acquire spoken language. This group of children has been seriously neglected in research conducted over the past 2 decades. Little is known about them, in part because the field lacks the tools to assess them, and they often pose significant behavioral challenges that preclude their participation in research studies. Among the ~70% individuals with ASD who have spoken language skills, about 50% are language impaired; the remaining group have normal language scores on standardized tests. Thus, there is enormous heterogeneity in verbal abilities in ASD. To date, studies of the brain and cognitive mechanisms that underlie this heterogeneity remain quite limited.

At the cognitive level, current research suggests that language impairment in verbal children with ASD involves deficits in phonological working memory, as evident on tests of nonsense word repetition. The areas of language that are most affected in this group include complex syntax and morphology related to marking tense (Tager-Flusberg et al., 2005). There is also considerable evidence from MRI (magnetic resonance imaging) studies of both children and adults that there are differences in volume and asymmetry in language regions of the cortex, specifically Broca's and Wernicke's areas. Recent studies suggest that these differences do not necessarily track with degree of language impairment, though there are conflicting findings in the literature. Importantly, most of the participants in cognitive and neuroimaging studies of ASD have been adults and individuals who have relatively intact language. Far less is known about the processes (either cognitive or neural) that might be implicated in the minimally verbal group. This project is designed to address this issue using two different brain imaging methods and approaches.

1. fMRI and DTI: Converging lines of evidence support the view that ASDs are disorders of connectivity, in which abnormalities in white matter integrity and reduced coordination of activity across brain regions give rise to core features. Recent results from our group (Guenther; Manoach) have identified white matter anomalies in the speech network of high-functioning verbal adults with ASD, specifically in the pathway between the left supplementary motor area (SMA) and left ventral premotor cortex (vPMC), a pathway involved in the initiation of speech output according to the Directions into Velocities of Articulators (DIVA) model of speech production, a leading model of the neural computations underlying speech. These connectivity abnormalities are revealed in both DTI and functional connectivity analyses collected using fMRI. NOTE: IRB approval for the collection of fMRI/DTI data will be obtained from MGH, where this portion of the study will be carried out. Data analyses will be carried out at MGH and BU.
2. Electrophysiology: EEG/ERP It has long been known that some children with ASD fail to respond when spoken to, cover their ears in the presence of certain sounds, yet show highly acute hearing in other contexts. Little is known about the mechanisms that might underlie these unusual auditory processing profiles. Abnormal neural oscillations have been found to be an endophenotype for ASD in infants and adults. However, the picture of how oscillations differ from those of non-clinical subjects is unclear and depends on the age at testing. Characterizing the oscillations present in the brains of a range of adolescents will allow us to quantify the relationship between neural activity and severity of language impairment. The main hypothesis that will be tested in this project is that abnormal neural connectivity in minimally verbal children with ASD profoundly impedes the ability to perceptually organize auditory scenes into meaningful units or objects, which has a direct effect on language development in these individuals. Event-related brain potentials (ERPs) and frequency-based analysis of the electroencephalogram (EEG) will be collected during passive listening tasks, to evaluate auditory perception and the perceptual organization of tones and speech using the mismatch negativity (MMN) paradigm (the MMN reflects passive deviance detection processes indexing low-level acoustic processes in automatic sensory memory mechanisms, and it has been widely investigated in studies of auditory perceptual discrimination of tones and speech).

The investigators plan to collect these data from 3 groups of adolescents with ASD varying in language skills (minimally verbal, verbal-language impaired, verbal-language normal) and age and gender matched non-ASD typical adolescents. The participants will be diagnosed and tested using a range of standardized and non-standardized assessments including IQ/cognitive level; social abilities; ASD severity; communication skills (nonverbal). These assessments will be carried out by trained examiners in the clinical core housed at BU.

Because many of the participants present with highly challenging behaviors and are difficult to test, a number of innovative approaches will be taken to maximize participants' success in completing the behavioral assessments, and then to tolerate and provide artifact-free data in the electrophysiological and MRI portions of the study.

Overall, the investigators hypothesize that the groups tested will differ in their neural indices of auditory perception, neural indices of perceptual organization, and induced neural oscillations. Further, the investigators hypothesize that differences in the degree of language impairment across these groups will be correlated with differences in the degree of abnormality of the measured neural indices, but not in the form of the abnormality, suggesting that the root causes of the differences across the tested subject groups are quantitative rather than qualitative.