Do Toddlers at Familial Risk for ASD Differ in Their Electrophysiological Responses to Known and Unknown Words?

Saturday, May 13, 2017: 12:00 PM-1:40 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
K. H. Finch1, A. Seery2, H. Tager-Flusberg1 and C. A. Nelson3, (1)Psychological and Brain Sciences, Boston University, Boston, MA, (2)New York University School of Medicine, New York, NY, (3)Boston Children's Hospital, Boston, MA

Language is often impaired in individuals with autism spectrum disorder (ASD). Additionally, subtle atypicalities in language processing have been found in unaffected individuals at familial risk for ASD. In typically developing toddlers, neural responses, measured by event-related potentials (ERPs), vary depending on word familiarity and language ability (Mills et al., 2005). Kuhl and colleagues (2013) found atypical ERPs to words in toddlers with ASD. However, they did not control for language, so it is difficult to interpret whether these ERP differences were driven by ASD or the children’s language abilities.


Our study investigates electrophysiological responses to words in 24-month-old toddlers with and without ASD, including those at familial risk, while controlling for language abilities.



59 monolingual, English-speaking 24-month-olds were divided into three groups: low-risk control (LRC; N=34), high-risk for ASD (HRA; older sibling with ASD) without ASD (HRA-, N=21), and HRA children with ASD (ASD; N=4). 


ERPs were recorded while children passively listened to a stream of words playing from bilateral loudspeakers. Forty nouns were presented up to three times across two categories: ‘known’ words, confirmed through parent report, and ‘unknown’ words.


Analysis focused on the mean amplitude of the N200-500 (negative component from 200-500ms post-stimulus onset) as its distribution varies depending on word familiarity (Mills et al., 2005). To analyze group differences, we used a mixed-model ANOVA with two conditions (known, unknown), two regions (parietal, occipital), and two hemispheres (left, right) as repeated measures and group (LRC, HRA-, ASD) as a between-subjects factor. We controlled for language abilities in our analyses using the verbal developmental quotient (VDQ) from the Mullen Scales of Early Learning (Table 1). We also investigated relationships between the N200-500 and language abilities and ADOS severity scores.


The results of the ANOVA found that there was a main effect of condition (F(1,55)=4.34, p=0.042) with a more negative amplitude for known words than for unknown words. There was also a significant group by condition by hemisphere interaction (F(2,55)=3.86, p=0.027). This was driven by a significant difference between groups in the left hemisphere in response to the known words (p=0.037). However, follow up on this revealed LRC, HRA-, and ASD groups were not significantly different from each other (all p>0.10; Figure 1). There were no other significant effects.

There was a significant relationship between the N200-500 in the right hemisphere sites and VDQ (rs=0.328, p=0.011), such that toddlers with a less negative response in the right hemisphere had better language abilities. There were no significant relationships with ADOS severity scores.


When controlling for language abilities, toddlers with ASD and unaffected familial members of ASD showed similar lateralization patterns in response to known words as their typically developing peers. Moreover, there was a significant relationship between the lateralization pattern of the N200-500 and language abilities but not with ASD severity. Thus, any differences in ERPs in response to language within these groups may be a result of varying language abilities as opposed to an ASD diagnosis or familial risk of ASD.