Abstract: Background: Adolescence is an important period of brain maturation and development, and adolescents with autism spectrum disorder (ASD) have atypical brain activation and functional connectivity in prefrontal and temporal cortex, which may be related to the symptom severity. Objective: To examine the correlation between the development trajectories of the topological properties of the brain network and age, cognition and symptom severity in individuals with ASD via graph theory analysis. Design: Case series report. Methods: Children and adolescents with ASD who met the DSM-IV-TR/DSM-V/ICD-10/ICD-11 diagnostic criteria, and were confirmed by ADOS-G and ADI-R assessment were included. Inclusion criteria were as follows: age between 6 to 17 years old, first language being Mandarin Chinese, right-handed, and full-scale IQ (FSIQ)≥70. The 8-minute resting-state signals were collected by a continuous functional near-infared spectroscopy (fNIRS) device, measuring the concentration change of oxy-, deoxy- and total hemoglobin (HBO, HBR, HbT). The generalized linear model was applied to analyze the relationship between network properties and age. After the effect of age was controlled, partial correlation analyses were used to evaluate the brain-behavior relationships. Main outcome measures: The topological properties of the brain network, including global efficiency, local efficiency, betweenness centrality and degree. Results: A total of 44 children and adolescents with ASD (age 9.5±1.5 years, 40 males and 4 females) were enrolled. Regression analysis showed that the global efficiency of HBO data decreased with age (R2=0.233, P=0.026), while the local efficiency remained stable (R2=0.002, P>0.05). The betweenness centrality and degree of the right superior temporal gyrus and inferior frontal gyrus were decreased (channel 3, 4), while those of left middle and superior frontal gyrus were increased (channel 30, 40, 48). The topological properties of brain regions showed that significant age effects were related to the cognitive ability, while the core symptoms of ASD were related to the nodal metrics of brain regions that showed nonsignificant development. Conclusion: Results suggested a decreased functional integration and similar segregation development of brain network in ASD from childhood to adolescence. Current research also implied that the development of prefrontal network may be only related to the development of cognition ability of participants with ASD.

Key words: Autism spectrum disorder, Adolescence, Functional near-infrared spectroscopy, Resting state, Graph theory