Christian Gerlach¹, Linnea Brønnum Hansen², Mette Elmose Andersen³; ¹University of Southern Denmark, ²University of Southern Denmark, ³University of Southern Denmark

Introduction: Does less efficient face processing in autism reflect reduced holistic processing? To examine this, we compared the magnitude of the face inversion effect (FIE) for 41 autistic adults (AA) and 41 neurotypical adults (NT). Methods: The FIE, which is the difference in performance with upright and inverted faces, is often taken to reflect face-specific processing, with upright faces being subjected to holistic processing and inverted faces being processed in a more parts-based manner. The FIE measurement was based on the Cambridge Face Perception Test (CFPT). Results: The AA-group performed worse than the NT-group with both upright (p < .001, d = 1.04) and inverted faces (p < .05, d = .45), and also exhibited a reduced FIE (p < .01, d = -.58). Comparing AAs who performed 2SD below controls on the CFPTUpright (AAPositive, 27%) with those who did not (AANegative), we find that the AAPositive-group performed worse than the AANegative-group with both upright (p < .001, d = 3.07) and inverted (p < .01, d = 1.4) faces, and also exhibited a reduced FIE (p < .05, d = -.67). In comparison, while the AANegative-group did differ from the control group on the CFPTUpright (p < .01, d = .75) it did not on the CFPTInverted or in the magnitude of the FIE (all p’s > .1 and all 95%CI’s crossed 0). Conclusion: Like in previous studies we find that many AAs perform abnormally on the CFPTUpright (27%). Interestingly, while the AA-group does exhibit a smaller FIE than the NT-group, the AA-group is also impaired on the CFPTInverted. To the degree that the FIE is a specific index of holistic processing, holistic processing is more affected in AAs than parts-based processing. Importantly, however, parts-based processing (inverted faces) is also affected.

Acknowledgements: This project was funded by the Independent Research Fund Denmark (0132-00122B)