Abstract
Girls with fragile X syndrome (FXS) often manifest significant symptoms of avoidance, anxiety, and arousal, particularly in the context of social interaction. However, little is currently known about the associations among neurobiological, biobehavioral such as eye gaze pattern, and social-cognitive dysfunction in real-world settings. In this study, we sought to characterize brain network properties and eye gaze patterns in girls with FXS during natural social interaction. Participants included 42 girls with FXS and 31 age- and verbal IQ-matched girls (control). Portable functional near-infrared spectroscopy (fNIRS) and an eye gaze tracker were used to investigate brain network alterations and eye gaze patterns associated with social-cognitive dysfunction in girls with FXS during a structured face-to-face conversation. Compared to controls, girls with FXS showed significantly increased inter-regional functional connectivity and greater excitability within the prefrontal cortex (PFC), frontal eye field (FEF) and superior temporal gyrus (STG) during the conversation. Girls with FXS showed significantly less eye contact with their conversational partner and more unregulated eye gaze behavior compared to the control group. We also demonstrated that a machine learning approach based on multimodal data, including brain network properties and eye gaze patterns, was predictive of multiple domains of social-cognitive behaviors in girls with FXS. Our findings expand current knowledge of neural mechanisms and eye gaze behaviors underlying naturalistic social interaction in girls with FXS. These results could be further evaluated and developed as intermediate phenotypic endpoints for treatment trial evaluation in girls with FXS.