Oxytocin impacts top-down and bottom-up social perception in adolescents with ASD: a MEG study of neural connectivity

Abstract

Background: In the last decade, accumulative evidence has shown that oxytocin can modulate social perception in typically developed individuals and individuals diagnosed with autism. While several studies show that oxytocin (OT) modulates neural activation in social-related neural regions, the mechanism that underlies OT effects in ASD is not fully known yet. Despite evidence from animal studies on connections between the oxytocinergic system and excitation/inhibition neural balance, the influence of OT on oscillatory responses among individuals with ASD has been rarely examined. To bridge these gaps in knowledge, we investigated the effects of OT on both social and non-social stimuli while focusing on its specific influence on the neural connectivity between three socially related neural regions-the left and right fusiform and the medial frontal cortex.

Methods: Twenty-five adolescents with ASD participated in a wall-established social task during a randomized, double-blind placebo-controlled MEG and OT administration study. Our main task was a social-related task that required the identification of social and non-social-related pictures. We hypothesized that OT would modulate the oscillatory connectivity between three pre-selected regions of interest to be more adaptive to social processing. Specifically, we focused on alpha and gamma bands which are known to play an important role in face processing and top-down/bottom-up balance.

Results: Compared to placebo, OT reduced the connectivity between the medial frontal cortex and the fusiform in the low gamma more for social stimuli than for non-social ones, a reduction that was correlated with individuals' performance in the task. Additionally, for both social and non-social stimuli, OT increased the connectivity in the alpha and beta bands.

Limitations: Sample size was determined based on sample sizes previously reported in MEG in clinical populations, especially OT administration studies in combination with neuroimaging in ASD. We were limited in our capability to recruit for such a study, and as such, the sample size was not based on a priori power analysis. Additionally, we limited our analyses to specific neural bands and regions. To validate the current results, future studies may be needed to explore other parameters using whole-brain approaches in larger samples.

Conclusion: These results suggest that OT influenced social perception by modifying the communication between frontal and posterior regions, an attenuation that potentially impacts both social and non-social early perception. We also show that OT influences differ between top-down and bottom-up processes, depending on the social context. Overall, by showing that OT influences both social-related perception and overall attention during early processing stages, we add new information to the existing understanding of the impact of OT on neural processing in ASD. Furthermore, by highlighting the influence of OT on early perception, we provide new directions for treatments for difficulties in early attentional phases in this population. Trial registration Registered on October 27, 2021-Retrospectively registered, https://ichgcp.net/clinical-trials-registry/NCT05096676 (details on clinical registration can be found in www.

Clinicaltrial: gov , unique identifier: NCT05096676 ).

Keywords: Autism; Connectivity; Face perception; Gamma; MEG; Oxytocin.

Conflict of interest statement

The authors declare that they have no competing interests and no competing personal or financial interests that could influence the study in this paper.

© 2022. The Author(s).

Figures

Fig. 1

Behavioral paradigm. The task consisted of eight blocks—half contained social stimuli (pictures of eyes during emotional expression, see example above the line) and the other half contained non-social stimuli (pictures of vehicles, see example below the line). Trials began with a fixation cross (500 ms). Next, an image appeared for approximately one second, followed by a single word. Participants were asked to decide whether the word described the image. The three dots at the end of each row mark the continuity of the block

Fig. 2

Violin plots of the connectivity between posterior and frontal social-related regions in alpha, beta, and gamma frequencies. A Selected regions of interest. Three neural regions were chosen prior to the experiment: left fusiform, right fusiform, and medial frontal cortex. All areas were marked based on AAL atlas’ locations. Black dots represent the selected voxels in each region. For connectivity analysis, we used one voxel from each region. Power analysis was calculated from all the voxels. B–E The neural coherence score between ROIs, in each frequency band: B alpha (8–13 Hz); C beta (14–25 Hz); D low gamma (30–60 Hz); and E high gamma (60–100 Hz). OT and PL sessions are represented as separate lines. Asterisks represent a significant result (p < .05) after FDR correction. Dots represent individual data. Black horizontal lines represent the mean. Vertical lines represent mean values ±1 SE

Fig. 3

Brain–behavior correlation. A significant negative correlation was observed between low-gamma-band coherence and participants’ performance. The X-axis represents normalized coherence—the connectivity between the fusiform and mPFC in the social trials divided by the connectivity in the non-social trials. Y-axis represents normalized behavioral performance—accuracy rates in social trials divided by the rates in the non-social trials