Electrophysiological Endophenotypes in Autism: A Family Study

Background: Attempts to categorize genetic/behavioral components of ASD reveal diverse causal mechanisms, leading to diagnostic heterogeneity. One promising avenue of research is the identification of endophenotypes of ASD. The error-related negativity (ERN), a scalp-recorded event-related potential that represents performance monitoring processes, may qualify as an endophenotype of ASD. However, previous research is mixed, with some studies pointing to a reduced-amplitude ERN in ASD relative to neurotypical controls and other studies revealing minimal or no group differences.

Objectives: We examined the ERN as an endophenotype of ASD by determining whether ERN amplitude and behavioral (error rates, reaction times [RT]) differences were present in first-degree relatives of ASD probands. We compared electrophysiological and behavioral data in male ASD probands and each proband’s father, mother, and one non-ASD male sibling relative to control youth and their parents. We hypothesized that youth with ASD would display reduced-amplitude ERN and impaired behavioral modification. We also hypothesized that higher presence of ASD characteristics in first-degree relatives would be associated with reduced EEG and behavioral activation. 

Methods: We collected data from 19 ASD probands, 53 first-degree relatives, and 75 control youth and parents during a modified Eriksen flanker task. One-way ANOVAs were used to examine group differences in ASD symptoms. Group (ASD, ASD Sibling, Control Youth, ASD Mothers, ASD Fathers, Control Mothers, Control Fathers) x Accuracy (Error, Correct) repeated-measures ANOVAs were used to examine group differences.   

Results: Levels of autism symptomology significantly differed between youth groups. ASD probands displayed significantly higher levels of symptomology relative to controls and siblings (ts < 9.14, ps > 0.001) but siblings and control youth did not significantly differ on measures of autism symptomology (ts < 0.77, ps > .44). Fathers of ASD probands displayed significantly greater autism symptomology relative to control fathers (t = 2.33, p = 0.03) but group differences were not present among mothers (|ts| < 1.30, ps > .20). Repeated measures ANOVAs on ERN amplitude revealed a significant main effect of Accuracy (F = 161.96, p < 0.001) but no significant main effects or interactions with Group (|Fs| < 1.82, ps > .10). Significant main effects of Accuracy were observed for error rates (F = 430.91, p < 0.001) and RTs (F = 739.99, p < 0.001). For error rates, the interaction with Group was not significant (F = .47, p = 0.83) but the main effect of Group was significant (F = 7.74, p < 0.001). For RTs, the Accuracy x Group interaction (F = 9.59, p < 0.001) and main effect of Group were significant (F = 1.22, p = 0.30). However, there were no significant differences between youth or parent groups for error rates or RTs (|Fs| < 1.07, ps > .35).

Conclusions: Although groups differed on measures of autism symptomology, no group differences were present for electrophysiological or behavioral data. These findings suggest that autism may not be consistently associated with reduced performance monitoring and that the ERN may not qualify as an endophenotype for ASD. Given the heterogeneity of previous studies, replication is warranted.