Eye-Tracking and Upright and Inverted Face Recognition Performance in Adults with an Autism Spectrum Disorder

Background: This study examined the Face Inversion Effect (FIE) in individuals with an Autism Spectrum Disorder (ASD), and whether difference in attention to the mouth and eyes affects overall face recognition performance. The term FIE describes the effect on performance of inverting stimuli during a face recognition task. Inversion is thought to disrupt holistic face processing and forces reliance on inner face features. There is some evidence that individuals with an ASD may not be affected by inversion to the same degree as persons without an ASD (Hobson et al., 1988; McPartland et al., 2004). Some studies, however, indicate that older or higher functioning individuals with ASD are affected by inversion (Lahaie et al., 2006; Scherf et al., 2008). Interestingly, eye-tracking studies have found that individuals with an ASD spend less time fixating on the eyes and show increased attention to the mouth region (Corden et al., 2008; Pelphrey et al., 2002; Spezio et al., 2007). It is not known how this might impact inverted face recognition.   

Objectives: To examine the FIE in adults with an ASD. We predicted participants with an ASD would be less affected by face inversion than non-ASD participants. We used eye-tracking to examine between-group differences in attention to the eyes and mouth.

Methods: We compared upright and inverted face recognition performance for 26 participants with an ASD (M_age = 28.91, SD = 9.49 years) with that of 33 (M_age = 24.99, SD= 9.98 years) age and IQ matched non-ASD participants. Stimuli were 72 (36 targets) male faces cropped to reveal the inner face only. Upright and inverted images were presented on a Tobii eye-tracker which recorded eye-movements. A study phase was followed by a recognition test. Eye-movements, RT and recognition performance were examined.

Results: Face recognition was significantly better for upright compared to inverted faces, and non-ASD participants performed better than participants with an ASD. A non-significant interaction (F < 1) between diagnosis and inversion suggested that both groups were affected by inversion. Participants with an ASD did not, therefore, show an RT advantage for inverted faces compared to non-ASD participants. Effect size analysis of RT differences between inverted and upright faces did, however, hint at a larger advantage for upright compared to inverted faces for non-ASD participants (77 ms, d = .28) compared to participants with an ASD (6 ms, d =.01). Both groups directed more fixations to the eyes than the mouth, irrespective of image orientation, and no between-group differences were identified for eye-movement measures. Increased attention to the eyes was not associated with better recognition performance. 

Conclusions: Recognition and RT results provided evidence that adults with an ASD are affected by face inversion, although non-ASD participants showed a greater, albeit non-significant, RT advantage for upright over inverted faces. These results indicate that individuals with an ASD may not be advantaged in inverted face recognition compared to non-ASD persons when only inner face features are presented. No significant between-group differences in attention to the eyes and mouth were identified.