The Effects of Oxyotcin on Socially Rewarded Learning in Autism Spectrum Disorder

Background: The neuropeptide oxytocin plays an important role in social cognition and behavior and has emerged as a promising candidate for targeting social impairment in ASD. While several studies have shown beneficial effects of oxytocin on core social deficits in ASD only a few studies have examined the underlying neural mechanisms. The social motivation hypothesis of autism posits that core deficits arise from an early failure to attach reward value to social stimuli. Despite converging evidence that important functional interactions exist between oxytocin and dopaminergic reward circuits, no studies have directly examined the effects of oxytocin on reward processing in ASD. Previous research has shown that, relative to typically developing children, those with ASD exhibit impairment in reward-related learning as well as reduced frontostriatal response to social reward (Scott-Van Zeeland et al., 2010).

Objectives: This study examined the acute effects of oxytocin on socially rewarded learning and the associated neural circuitry in ASD.

Methods: In a double-blind, placebo-controlled crossover study, 17 adults with ASD underwent functional magnetic resonance imaging (fMRI) after receiving a single dose (24 IU) of intranasal oxytocin or placebo on 2 days separated by 3-5 weeks. The fMRI paradigm, adapted from Scott-Van Zeeland and colleagues, was a probabilistic learning task that required participants to view fractal-like images and classify them into “Group 1” or “Group 2”. Social feedback (e.g., a smiling or sad face for correct and incorrect reward trials, respectively) was provided after each response to guide the learning of stimulus-response associations.

Results: Although the task was designed to allow for improvement in classification accuracy, participants’ performance remained at chance levels over the course of the placebo scan, suggesting a deficit in implicit learning consistent with Scott-Van Zeeland et al. Furthermore, we found no effect of oxytocin on implicit learning. At the neural level, social feedback was associated with activity in “social brain” regions, such as the fusiform gyrus bilaterally, the orbitofrontal cortex, and the inferior frontal gyrus (Brodmann’s Area 44), following both oxytocin and placebo administrations. No selective effects were observed for oxytocin relative to placebo in response to social feedback across both correct and incorrect trials. However, when examining the neural response to positive vs. negative social feedback (i.e., happy vs. sad faces), oxytocin yielded significantly greater activity than placebo in the striatum, the medial prefrontal cortex, and the fusiform gyrus.

Conclusions: These findings are consistent with recent work in healthy controls showing that oxytocin may selectively bias attention to positive social cues (Domes et al., 2013). While some studies suggest that oxytocin results in enhanced social brain activity more broadly in ASD, we observed a selective effect of oxytocin in response to happy faces following rewarded trials. The lack of improvement in implicit learning following oxytocin could reflect task difficulty or insufficient dosing or length of treatment. Future research should examine the effects of different doses of oxytocin, as well as sustained and augmentative treatment effects.