The Influence of Semantic Richness on Novel Word Learning in Children with Autism Spectrum Disorder

Background:  The development of words (Howlin, 2003) is delayed in children with autism spectrum disorder (ASD). These children often continue to lag behind their peers in their vocabulary growth into their school years, and almost 20% remain functionally nonverbal (Lord et al., 2004). When children with typical language development (TLD) first map a novel word form with meaning, they benefit from learning words in rich semantic contexts (McGregor et al., 2007). Further, their speech motor stability increases when novel word forms are taught with visual referents (Heisler, Goffman, & Younger, 2010). Motor deficits (Bhat et al., 2012; Landa et al., 2013; Liu & Breslin, 2013) and vocal stereotypies (Rapp et al., 2013) are well documented in children with ASD. However, little is known about how their articulatory systems change across the timecourse of word learning, especially as a function of semantic knowledge.

Objectives:  The inclusion of semantic cues enhances semantic, phonological, and articulatory performance on word learning tasks in children with TLD (Gladfelter & Goffman, 2013; Heisler et al., 2010; Storkel & Adlof, 2009). The goal of our study was to investigate whether and how comprehension and production of words and phonetic accuracy and articulatory variability are influenced by semantic cues in children with ASD.

Methods:  Twelve children with ASD and 12 children with TLD (matched for expressive vocabulary level, all typical nonverbal IQ) were taught six novel words across three experimental sessions. The words were divided into three semantic learning conditions: no semantic cues, sparse semantic cues (visual referent only), and rich semantic cues (a story including multiple attributes of the word). Traditional measures of learning included: comprehension and production probes and a definition task. Phonetic transcriptions of imitated productions during the pre- and post-tests were analyzed to compare phonetic accuracy before and after learning the novel words. Articulatory movement recordings were collected from the upper and lower lip and jaw and analyses of speech motor variability (e.g., Gladfelter & Goffman, 2013) were conducted both in the pre- and the post-tests. Critical comparisons were whether semantic cues facilitated word learning in children with ASD and TLD.

Results:  Both children with ASD and their TLD peers improved their performance on comprehension, production, and definition tasks as semantic richness increased. Children with ASD were even more responsive than their peers to the highly informative cues in the rich story condition, and this benefit was observed from the onset of the first session and maintained across sessions. Articulatory measures were especially sensitive to the depth of semantic context, with the children with ASD only improving in movement stability for the words taught with the rich semantic cues included in the story condition. Children with TLD also benefitted from semantic cues, but responded similarly in the sparse (visual referent) and the rich (story) contexts.

Conclusions:  Measures of knowledge, sound accuracy, and articulatory movement stability all converge to reveal that children with ASD, even more than their TLD peers, benefit from exposure to rich semantic contexts during their earliest acquisition of new words.