Abstract
Understanding facial expressions is crucial for human social interaction as they convey emotions and intentions. Autistic adults often show marked differences (compared to neurotypically developed adults) in interpreting these cues, impacting communication and empathy. To aid autistic individuals, it’s essential to comprehend the neural basis behind these differences. However, the diverse nature of behavioral reports in autism impedes efficient study design. Current models for interpreting facial emotion judgments overlook individual image-level sensory representations, pivotal in understanding differences between neurotypical and autistic adults. In a recent study by Kar (2022), behavioral disparities between these groups were more evident at the image level rather than through broad categorical descriptors like “happiness” or “fear.” To investigate this further, we established an image-level framework using 360 diverse facial expression images from the Montreal Set of Facial Displays of Emotion (MSFDE). Through binary emotion discrimination tasks, we observed subtle yet significant differences in image-level behavioral error patterns between neurotypical and autistic individuals. Addressing the challenge of heterogeneity, we pinpointed shared variances in our developed image-level metrics, serving as a critical behavioral benchmark. To delve into the neural underpinnings, we conducted extensive neural recordings in the inferior temporal (IT) cortex of rhesus macaque monkeys. Initial findings align with previous predictions (Kar 2022), indicating stronger correlations between macaque IT-based decodes of facial emotion responses and neurotypical behavior compared to autistic behaviour. Our study aims to create an innovative framework merging non-human primate neural investigations with the autistic behavioral phenotype. By focusing on shared variances in image-level behavioral metrics, we aim to identify more sensitive neurobehavioral markers