Investigation of Face Processing in Autistic Spectrum Disorder (ASD) for the Development of Clinically Useful Biomarkers: An Electroencephalographic Approach

Background: Autistic Spectrum disorders (ASDs), are characterized by social deficits and restricted/repetitive behaviours. Underlying mechanisms are unclear, though they are thought to be caused by an interaction between genetic and environmental factors. Using imaging methods such as electroencephalography (EEG) to understand more about the neural mechanisms of ASDs may aid in the development of clinically useful biomarkers for earlier diagnosis as well as for drug treatment endpoints. 

Objectives: To conduct behavioural and electrophysiological endophenotyping to probe core social deficits of ASD to facilitate improved knowledge regarding underlying neural mechanisms of ASD. 

Methods: 5 participants with ASD and 10 typically developing (TD) participants have completed this research currently, data collection is ongoing. ASD participants have an unambiguous clinical diagnosis of ASD. EEG was recorded continuously using a high density 64 electrode array. Analysis was carried out offline. The paradigm consisted of schematic faces representing different emotions. There were three stimulus categories: Standard, Target and Deviant. The deviant varied between conditions A (Angry) and B (Happy). This paradigm was adapted from a study by Kreegipuu et al (2013). For this research particular focus was paid to the parietal region as this area is consistently associated with P300 generation.  

Results: A within group analysis comparing the difference between stimuli in the ASD group was carried out. All mean amplitudes between the time period of interest (250-550ms) were found to be significantly different from each other (p<0.001), a between group analysis compared mean P3 amplitudes, between the two groups, each stimulus response was compared. All three were found to be statistically significant from each other (p<0.001). A within group analysis of amplitude differences for each stimulus between the two conditions, exhibited that there was no significant differences found in mean P3 amplitudes in the ASD group (p<0.05). The control group conversely exhibited a significant difference in mean amplitudes in response to the deviant between conditions.

Conclusions: The aim here was to probe the core social deficits in ASD utilising EEG in TD individuals in comparison to individuals with ASDs. The significant differences in P3 amplitudes exhibited between the two groups suggest mechanistic differences in emotional processing between the two groups. Additionally, within group analysis of Condition A and B revealed a significant difference in amplitudes between the deviant stimuli in TD adults (p<0.001). There was no difference observed in the ASD group between the two conditions (p>0.05). Angry faces have been shown previously to elicit higher amplitudes than happy faces (Martens, Leuthold, & Schweinberger, 2010). There are several reasons this may have occurred, the ASD group have been more focused on details of the faces rather than the global representation, this more detail oriented observation has been exhibited in ASD previously. Secondly, ASD group may not be able to differentiate between the emotions on the schematic faces as easily as the TD group. Overall, significant differences in emotional face processing between the two groups were exhibited, further investigation into these differences could lead to the development of clinically useful biomarkers for earlier diagnosis.