25690
ERP Evidence of Semantic Processing in Children with ASD
Language outcomes vary in children with ASD, with up to 30% remaining minimally verbal (MV). Electroencephalography (EEG) yields information about an individual’s neural response to a stimulus in real time, and in the context of language impairment may elucidate neural mechanisms underlying language processing in ASD. Lexical-semantic processing can be quantified through a paradigm in which information is presented in a semantically congruent or incongruent way (Bentin 1985; Hagoort et al., 2008). This paradigm has been recommended for assessing receptive language in individuals with limited speech (Connolly et al., 1999). The EEG response after the incongruent presentation (N400) has been found to be reduced in both high-functioning (Dunn & Bates, 2005) and MV children with ASD (Cantiani et al., 2016). No study has examined variability in semantic processing within children with ASD and, in particular, whether EEG correlates of semantic processing relate to direct assessment of language.
Objectives:
We examine EEG correlates of lexical-semantic processing in a large, heterogeneous sample of children with ASD, compared with an age-matched typically developing (TD) group, and we investigated the relationship between lexical-semantic processing and receptive language. We hypothesized that children with ASD would show a reduced N400 response compared with TD children and that larger, faster N400 response would be associated with better receptive language.
Methods:
Verbal and MV children with ASD (ages 5-11; n=16 in each group), and an age-matched TD group (N=18) participated in a semantic congruence EEG paradigm, during which pictures were displayed followed by the expected (match) or unexpected (mismatch) spoken word. EEG data were collected using a 128-electrode Hydrocel Geodesic Sensor Net System (EGI), filtered at 0.3-30 Hz, and processed in NetStation. Participants with a minimum of 10 artifact-free segments per condition were included in analysis. Assessments included Differential Ability Scales (IQ), Peabody Picture Vocabulary Test (receptive vocabulary) and Autism Diagnostic Observation Schedule (diagnostic confirmation).
Results:
The mismatch condition elicited an N400 component in all groups (F=5.43, p=.024), with a shorter latency in the TD group (t=-3.15, p=.003). N400 latency was negatively correlated with receptive vocabulary (r=-.33, p=.03). A late negative component (LNC; 550-900ms) also was evident in the mismatch condition, with a group by condition by region interaction (F=2.5, p=.026). Post hoc analyses revealed that the LNC was present across frontal and central regions in the TD group (p-values .05-.002), in the mid-central region in the MV ASD group (t=2.5, p=.03), and approaching significance in the mid-frontal region in the V ASD group (t=1.99, p=.06).
Conclusions:
Findings indicate that both MV and verbal children with ASD demonstrate some EEG evidence of semantic processing, but their processing may be delayed and have reduced integration with mental representations. This supports the notion that at least some MV children show higher-level lexical processing, and performance on this task may be a meaningful stratification biomarker for understanding differential language trajectories. These findings differ from previous studies that did not find significant N400 response in children with ASD, likely due to our larger sample size and differences in the experimental paradigm.