Facial Processing in Low-Functioning Individuals with Autism: An N170 Event-Related Potential Study

Background: Abnormalities in the neural processing of faces have been previously documented in high-functioning individuals with autism (HFA).  In EEG studies of normal controls (NC), the N170 event-related potential (ERP) shows a larger amplitude for faces relative to objects, and a delayed latency for inverted faces relative to upright faces. Previous work with HFAs has also demonstrated increased N170 amplitude for faces relative to objects; however, the delay in N170 latency for inverted faces relative to upright faces was not observed.  The processing of faces in low-functioning individuals (LFA) with autism might also be altered but has previously not been explored. Abnormal development of the fusiform gyrus in autism may underlie patterns of weak central coherence, specifically the tendency towards feature-based rather than configural face processing. This processing style means that individuals with autism have less experience processing faces holistically, which may be reflected in atypical N170 effects. Atypical face processing could contribute to abnormal social development in ASD; therefore this is an important aspect of cognition to study in this population. To our knowledge, this study is the first to examine facial processing in LFAs using ERPs.

Objectives: The current study characterizes and contrasts facial processing in LFAs, HFAs, and NCs. Specifically, this study evaluates the response of the N170 ERP component to faces vs. inverted faces vs. non-facial stimuli (houses) in order to compare differences between the three groups. 

Methods: Participants were LFAs, HFAs, and NCs. Stimuli consisted of gray-scale digital images of upright faces, inverted faces, houses, and filler items (furniture), which served as targets. While EEG data was recorded, images were presented for 500ms each, with an inter-stimulus interval of 500-1000ms. To maintain attention, participants were instructed to respond only to the targets; participants unable to make a behavioral response were asked to passively watch.

Results: For both NCs and HFAs, we found increased N170 amplitude in response to upright faces and inverted faces compared to houses. For both groups, we found a greater negative amplitude and a delay in N170 latency for inverted faces relative to upright faces. For LFAs, we found no N170 amplitude difference between the three conditions in the typical time window.  However, in the later window of 200-250 ms, we observed a greater negativity to upright faces, and a greater positivity to inverted faces, relative to houses. 

Conclusions: Overall, we replicated previous finding in normal controls, showing a typical processing delay for inverted faces relative to upright faces. For high-functioning individuals with autism, we also found a delay in N170 latency for inverted faces relative to upright faces, contradicting previous findings showing no such delay for this group. For the low-functioning individuals with autism, the lack of amplitude difference between conditions in the canonical N170 time window may suggest that they are not differentially processing the images.  Later differences observed in the 200-250ms time window may be indicative of delayed N170 effects; however, patterns across the conditions in this time window still differ from those observed for NCs and HFAs.