ERPs to Words In Toddlers with ASD Predict Behavioral Measures at 6 Years of Age

Background: Language/communication deficits and social impairment are key components of Autism Spectrum Disorder (ASD). Event related brain potentials (ERPs) have been shown to be a sensitive measure of differences in speech processing abilities of 3 year olds with typical development (TD) and ASD (Kuhl, et al., 2005). ERP measures of word processing have also been related to differences in language abilities in TD toddlers (Mills, et al., 1997).

Objectives: Our goals were to use ERP measures of word processing in a longitudinal study to (1) investigate differences in patterns of ERP response to words in TD children and children with ASD; and (2) study the relationship between ERP measures and later outcomes in children with ASD (cognitive, language and social domains). 

Methods: Participants were children with ASD and age-matched TD controls. ERPs were collected at 18-30 months (T1) from all participants. Behavioral measures of cognitive, language, and social function were collected at 18-30 months (T1) and again at 6 years (T4) in the children diagnosed with ASD. ERPs were recorded using 22 channel electrocaps with standard 10/20 arrangement. Stimuli consisted of 3 word types: words known to the child, unknown words and backwards words (known words reversed). Parents of participants signed a University IRB approved consent form prior to study participation.

Results: Significant differences in patterns of ERP latency, amplitude and timing were observed at T1 between TD children and children with ASD. In general TD children show a more focused response to known and unknown words, with a differential left temporal response to known and unknown words. Children with ASD show a broadly distributed response across electrode sites for known and unknown words. High functioning children ASD (HF) exhibited a more localized response to words, similar to that of TD controls. Low functioning children with ASD (LF) had more diffuse patterns of response. In addition, T1 ERP measures from children with ASD were significantly correlated with measures of language and cognitive development collected four years later at T4: mean amplitudes (200-500 msc) at left parietal region for known words at 18- 30 months predicted scores on the Differential Abilities Scale, Peabody Picture Vocabulary Test and Vineland Adaptive Behaviors Scale taken 4 years later. 

Conclusions: In general, HF children with ASD exhibited more localized ERP response, similar to TD controls. LF children with ASD had more diffuse patterns of response. Significant predictive correlations were obtained between ERPs to known words and various measures of language and cognition/adaptation obtained four years after ERP testing. Further research using brain measures of speech processing in children with ASD and TD controls is important to theory building and to understanding the nature of the relationship between the linguistic and social deficits which are characteristic of autism, and may have implications for diagnosis and treatment in children with developmental disabilities that involve language.