Neural Response to Eye Gaze Differentiates ASD Diagnostic Status Among Adults Meeting ADOS-2 Criteria

Background: Individuals with autism spectrum disorders (ASD) process social information differently than individuals with typical development (TD), as indicated by a slower latency of face-specific processing components (McPartland et al., 2004). However, there has been little investigation of the relationship between neural processing of social stimuli in individuals who have social communication challenges but do not receive an ASD diagnosis from a clinician compared to individuals who receive an ASD diagnosis.

Objectives: To examine the relationship between EEG response to faces, social communication symptoms, and autism diagnosis in a clinically heterogeneous population.

Methods: The study included (a) adults who met criteria for autism or autism spectrum on the ADOS-2 Module 4 and received an ASD diagnosis (n=20; male=15); (b) adults who met criteria on the ADOS-2 but did not receive an ASD diagnosis (n=10; male=8; DX: n=9 schizophrenia spectrum, n=1 generalized anxiety disorder); and (c) adults who did not meet criteria on the ADOS-2 and did not receive an ASD diagnosis (n=40; male=26). Groups did not differ on age or IQ, and the two groups that met criteria on the ADOS did not differ in their communication and social interaction totals (ps>0.05). Each EEG trial began with a crosshair in the upper, middle, or lower portion of the screen directing gaze to specific regions of a subsequently appearing stimulus (eye/nose/mouth region of a face or upper/middle/lower region of a house). The N170 event-related potential, a component reflecting early face processing, was segmented to the appearance of the face or house.

Results: There was a significant main effect of diagnostic group on N170 latency (F(2, 67)=8.16, p<0.01). Adults who met criteria on the ADOS and received an ASD diagnosis had significantly longer N170 latencies to all stimuli than adults who met criteria but did not receive a diagnosis (p<0.01) and adults who did not meet criteria (p<0.01). Adults who met criteria but did not receive an ASD diagnosis did not have significantly different N170 latencies from adults who did not meet criteria on the ADOS (p=0.72). There was a significant main effect of condition on N170 latency (F(3, 201)=5.30, p<0.01) and amplitude (F(3, 201)=54.18, p<0.01), with faster N170 components to eyes than mouths, noses, or houses, and more negative N170 components to facial features than houses. Among adults who met criteria on the ADOS, those who did and did not receive an ASD diagnosis differed significantly in their N170 latency to eyes (p=0.01), noses (p=0.03), and houses (p=0.03), but not mouths (p=0.11), with only the latency difference to eyes remaining significant after post-hoc correction.

Conclusions: The presence of differences in N170 latency between groups that scored similarly on the ADOS but had different diagnostic outcomes underscores the utility of N170 latency as a biomarker to differentiate individuals with ASD from individuals without an ASD diagnosis. This finding further highlights the importance of considering brain-based measures of social functioning, as they provide practicable and complementary information to commonly used clinical instruments such as the ADOS.