Early Social Perception in Infants at Risk for ASD Using Functional Near-Infrared Spectroscopy (fNIRS)

Background: Converging evidence is implying that individuals with autism spectrum disorder (ASD) show atypical brain response to social stimuli at an early stage. However, the neural processing during early infancy in ASD has not been fully characterized. Studies of infants who are at substantially higher risk to develop ASD than the general population (HR), have the potential to clarify early functional manifestations of the brain for individuals with ASD.

Objectives: In this study, our aim was to find out whether cortical response to socially stimuli in HR infants differs from low risk (LR) controls by using fNIRS.

Methods: Until now, twenty 5-months-old infants participated in the fNIRS study; 4 of them were excluded because they failed to complete the whole study or made too much head movements.Finally, sixteen 5-months-old infants (n=8 for HR, n=8 for LR) were included for further analysis. Infants in the HR group had an older sibling with an ASD diagnosis or feeding problem or were premature babies. All subjects passively listened to a set of human vocal sounds (e.g., coughing and laughing) and familiar environmental sounds (e.g., rattles and car moving). fNIRS measurements were made using the NIRScout (NIRX, Germany/USA) device.

Results: The cortical activation by human sounds (i.e., the social task) relative to environmental sounds (i.e., the non-social task) was first measured by paired-sample t-tests within each group, primarily using the changes of oxy-hemoglobin (HbO). For the LR group, we found a greater hemodynamic response to the social task (compared to the non-social task) in the bilateral temporal cortex, mainly located in the left posterior temporal region (channel 32: t (7) = - 3.26, p < 0.05, d = 1.36; channel 34: t (7) = - 2.67, p < 0.05,d = 1.06), and the right temporal region (channel 53: t (7) = - 3.32, p < 0.05, d = 1.37). By contrast, no channels showed significant activation for the social task compared to the non-social task in the high-risk group (p > 0.05 for all tests) after Bonferroni correction. To further investigate whether patterns of activation were different in the two groups, we carried out a 2*2 (groups: HR, LR; tasks: social, non-social) mixed ANOVA analysis in the left regions (Ch.32 and Ch.34)and right regions (Ch.53) separately. We observed a significant interaction effect between groups and tasks in the right hemisphere (F (1,15) = 5.04, p = 0.04, ηp² = 0.14) but not in the left hemisphere. For the social task, HR infants showed lower activation (t (7) = - 2.58, p = 0.02, d = 0.98) in the right hemisphere than LR infants. For the non-social task, no group differences were found.

Conclusions: Our initial results demonstrate that HR infants show different neural activation to social stimuli in the temporal regions compared to LR infants. This suggests that social development may already be impaired in the first year of life for (some of the) infants in the HR group. Data collection is still ongoing. At the meeting we expect to present results of at least 50 infants.