Evidence for Language Phenotypes in Children with ASD Based on Varied, Longitudinal Assessment Measures

Saturday, May 14, 2016: 11:30 AM-1:30 PM
Hall A (Baltimore Convention Center)
L. Naigles1, A. M. Mastergeorge2, R. T. Johnson3, C. W. Nordahl4 and D. G. Amaral5, (1)University of Connecticut, Storrs, CT, (2)University of Arizona, Tucson, AZ, (3)UC Davis M.I.N.D. Institute, Sacramento, CA, (4)University of California at Davis, Sacramento, CA, (5)UC Davis The M.I.N.D. Institute, Sacramento, CA
Background:  Language assessment is challenging in young children with ASD, yet is crucially important for designing interventions and predicting outcomes.   Standard scores situate children with ASD in relation to their TD peers, yet specifics of which aspects of language are known vs. unknown are lacking.  Moreover, language-brain comparisons may suffer because children of comparable standard scores may be at different stages of language development; specifics of language growth also become hard to document.  We hypothesize that scrutiny of raw scores from standardized assessments might be revealing concerning children with ASD’s actual levels of language knowledge, plus relationships among their early language, brain development, and language growth.  

Objectives:   1) To examine how raw Mullen (MSEL) scores might distinguish language knowledge in a large group of 3-year-olds with ASD, and 2) to investigate relationships among these scores, acquired MRI data and subsequent language levels.

Methods: 239 families of children with ASD participated in the APP study of autism phenotypes at three years old (Mage = 35.06 months; MMullenVR-T=28.88, SD=12.97); 98 of these provided additional data at 5 years (Mage=68.63 months; MDAS NV SS-score=81.39, SD=27.51).   Three-year-olds were administered the MSEL and CDI; five-year-olds were administered the ADOS, DAS Verbal-Comprehension, EOWPVT and PPVT.  104 of the 3-year-olds underwent structural MRI imaging, enabling DTI analyses.  


Based on the MSEL receptive and expressive language, the 3-year-olds were divided into three language subgroups based on raw scores:  Nonverbal,  WordLearner; Words&Grammar:  Subgroups’ language scores differed (p<.05); however,  Nonverbal and WordLearner subgroups differed only marginally on MSEL VR (p<.10) and WordLearner and Words&Grammar groups did not differ on ADOS scores. Fractional anisotrophy (FA) in the DTI analyses varied in several tracts according to these subgroups; that is children in the  Nonverbal group manifested the lowest FAs in the left inferior longitudinal fasciculus (ILF). 

The 5-year-olds were divided into four subgroups based on standard scores: Low Verbal , High Verbal, Low Normal; and Possible SpecificLanguage Impairment .  Comparison of the children at T1 vs. T2 revealed stability: Words&Grammar children at T1 predominantly scored as HighVerbal at T2.  However, 28 children who were Nonverbal at T1 scored in one of the three higher subgroups at T2.  

Conclusions: 3-year-olds’ raw MSEL language scores revealed subgroups that implicated graded vocabulary development.  The DTI findings supported this interpretation, as children at lower vocabulary levels manifested less mature white matter tracts, particularly tracts connecting visual and auditory cortex. These children may exhibit impaired word-object mappings.  Many of the 5-year-olds remained at similar verbal levels relative to their chronological age. However, a significant minority demonstrated marked improvement.  Ongoing analyses consider the children’s actual speech, comparing lexical with grammatical functioning and potential neural correlates thereof.  Understanding language profiles of young children with ASD provides important information about stability and variability in language symptomatology and furthers our understanding of emergent language phenotypes.