Neural correlates of inhibition of socially relevant stimuli in adults with autism spectrum disorder

Highlights

• Response inhibition was examined in adults with autism spectrum disorder (ASD).

• A Go/NoGo paradigm was used during functional magnetic resonance imaging.

• Control adults recruited brain regions involved in inhibition as predicted.

• Adults with ASD evoked weaker activation and recruited visual association cortices.

• Atypical brain activation may underlie deficits in adults with ASD.

Abstract

Adults with autism spectrum disorder (ASD) can demonstrate difficulties with inhibiting inappropriate social responses. Presently, little research has utilized socially relevant stimuli to explore the modulatory effects of emotion on cognitive control in this population. To assess neural mechanisms of inhibiting social stimuli, we presented images of happy or sad facial expressions in a Go/NoGo task to unmedicated adults with ASD and to controls during functional magnetic resonance imaging (fMRI). Groups did not differ on behavioral measures. Brain activation in response to NoGo vs. Go trials revealed differing regional patterns of activation within groups. Controls recruited brain regions involved in inhibition (dorsal- [DLPFC] and ventro-lateral prefrontal cortices [VLPFC], anterior cingulate cortex [ACC]), response suppression (parietal lobe), interoceptive awareness (insula), and also the fusiform and middle temporal gyri. Adults with ASD only recruited the VLPFC and right fusiform gyrus, and weakly activated the ACC and insula. Between-group comparisons indicated that controls activated the DLPFC, while adults with ASD relied on the VLPFC and the fusiform gyrus to inhibit responses. Adults with ASD may have relied more on visual association cortex, possibly as a means of recruiting additional neural processes that could act as a compensatory mechanism.

Introduction

Core features of autism spectrum disorder (ASD) include difficulties with language, social communication and repetitive behaviors, which are believed to result from a combination of genetic and environmental factors that disrupt brain development. Brain atypicalities have been associated with impaired frontal lobe functioning such as executive dysfunction, which is highly prevalent in this population (Dawson, 2008) and is apparent during childhood and into adulthood (Bennetto et al., 1996, Luna et al., 2007, Ozonoff et al., 2004, Volkmar et al., 2005).

One aspect of executive function with which adults with ASD have difficulty is the inhibition of inappropriate behaviors as shown in some behavioral studies of inhibition (Geurts et al., 2004, Mosconi et al., 2009, Noterdaeme et al., 2001, Ozonoff et al., 1994). However, other evidence suggests that individuals with ASD do not have widespread inhibition deficits but only for particular tasks. For example, Christ and colleagues (Christ et al., 2007) reported that children with ASD were impaired on two out of three inhibition tasks, adjusting for IQ, age and processing speed. In social settings, impaired inhibition can lead to more subtle inappropriate behaviors or verbalizations. During social interactions it is necessary to select and initiate proper responses while inhibiting those that are inappropriate; this inhibitory control is a fundamental feature of executive processing. Despite the importance of studying response inhibition in social domains, the majority of neuroimaging studies to date have focused on non-social aspects of response inhibition in individuals with ASD.

Response inhibition is often studied using the Go/NoGo task (Donders, 1969/1868), which requires the execution or inhibition of a motor response that is cued by a Go (respond) or NoGo (do not respond) stimulus, respectively. Tasks are most often time-sensitive and the instruction to respond as quickly as possible can create an overriding tendency to react that must be inhibited when cued by a NoGo stimulus. While this is an everyday occurrence, it involves several cognitive processes such as attention, stimulus identification and discrimination, maintenance, motor response, inhibition of an inappropriate response and error monitoring.

The cerebral mechanisms of response inhibition have been examined in animal models (Parker et al., 2005, Roberts and Wallis, 2000, Thompson et al., 1985), imaging studies in healthy subjects (Braet et al., 2009, Duann et al., 2009, Fassbender et al., 2009, Goghari and MacDonald, 2009, Goya-Maldonado et al., 2010, Hu and Li, 2011, Leon-Carrion et al., 2010, Mitchell et al., 2008, Thoma et al., 2008, Zheng et al., 2008) and in neurological (Amanzio et al., 2011, Baglio et al., 2011, Cilia et al., 2011) and neurodevelopmental disorders including ASD (Agam et al., 2010, Lee et al., 2009). While a number of different tasks have been used to explore response inhibition, evidence from functional imaging studies with healthy participants has suggested strong roles for the pre-supplementary motor area (SMA), anterior cingulate cortex (ACC), the inferior parietal lobule, precuneus, the insula, middle temporal gyrus, fusiform gyrus, association occipital cortex (Simmonds et al., 2008), and also the prefrontal cortex (PFC) (Ridderinkhof et al., 2004), in both dorsal (Kadota et al., 2010, MacDonald and Carter, 2003) and ventral subdivisions (Leung and Cai, 2007, Rubia et al., 2003). Both PFC regions have extensive connections with sensory and motor regions involved in the selection of the behavior that needs to be inhibited (Tanji and Hoshi, 2008). In ASD, dysregulation of frontal lobe function and connectivity is believed to underlie in part their poorer inhibition (Dumontheil et al., 2008). Functional neuroimaging studies of response inhibition in adults with ASD have indicated atypical activation in the prefrontal cortex relative to controls, despite similar behavioral performance (Kana et al., 2007, Schmitz et al., 2006), suggesting that the ASD group may rely less on this brain region to complete the task.

Neuroimaging studies with individuals that have ASD using variations of response inhibition tasks have reported somewhat inconsistent results (Goldberg et al., 2011, Schmitz et al., 2006, Thakkar et al., 2008). For example during antisaccade tasks, one study reported increased activity in the cingulate cortex (Thakkar et al., 2008), while another reported decreased activity in this region (Agam et al., 2010). Additionally, studies employing more complex response inhibition tasks have reported increased activation in brain regions such as the ACC, insula, ventrolateral and medial prefrontal cortices, and in areas of the parietal and temporal cortices (Goldberg et al., 2011, Schmitz et al., 2006). In contrast, using a simpler inhibition task, decreased activity was noted in the ACC, insula, ventrolateral prefrontal cortex (VLPFC), and the premotor cortex (Kana et al., 2007). Differences in brain activation patterns across response inhibition tasks in ASD could be attributed to alterations in brain functions (e.g. motor and visual) being regulated by inhibitory brain mechanisms. Additionally, response inhibition tasks with high working memory load are particularly difficult for individuals with ASD (Goldberg et al., 2002, Goldberg et al., 2005, Luna et al., 2007, Minshew et al., 1999, Ozonoff and Jensen, 1999). In turn, activation in brain regions involved in cognitive control may be differentially activated during these more complex tasks.

A key component to response inhibition is modulation by emotional and cognitive control. While emotional neurophysiological processes have largely been elucidated, less research has explored emotion–cognition interactions (Dolan, 2002, Ochsner and Phelps, 2007). These studies in typical individuals have indicated that limbic brain regions work in concert with prefrontal areas to perform emotional response inhibition (Beauregard et al., 2001, Goldstein et al., 2007). Emotional inhibition can be particularly challenging for individuals with ASD as they can have difficulty with recognizing affect conveyed by social stimuli (Adolphs et al., 2001, Boraston et al., 2007, Castelli, 2005, Dalton et al., 2005, Dawson et al., 2004, Golarai et al., 2006, Hadjikhani et al., 2006, Howard et al., 2000). Few studies have addressed response inhibition of emotional stimuli in ASD, although a study with children (aged 8–13 years) used emotional faces in a Go/NoGo task and reported that those with ASD responded more slowly than controls only when stimulus presentation slowed (Geurts et al., 2009). This finding was interpreted to indicate that the ASD group had normal response inhibition, using that task, but perhaps their attention decreased as the facial stimuli became less salient. Motivation to attend to emotional stimuli may be lower in children with ASD compared to typically developing controls due to atypical affective processing. The authors argued that their results might be influenced by age, as a previous study that used neutral stimuli in a Go/NoGo task found that adults with ASD had more difficulty with faster stimulation rates (Raymaekers et al., 2004) compared to a study with children (Raymaekers et al., 2006). Whether adults with ASD also have difficulty inhibiting socially relevant stimuli remains unclear, and the brain-behavior correlates have not been determined.

Thus, in the present study, we assessed inhibition to emotional socially relevant stimuli and the underlying neural correlates using fMRI, in adults with and without ASD. The ASD sample was screened for co-morbid psychiatric disorders and was not taking medication during the time of study. We utilized a Go/NoGo task with minimal working memory load, which employed images of happy or sad faces. Non-social stimuli were not included in the paradigm as brain regions that underlie these processes have been well established and the purpose of the current experiment was to examine response inhibition of social stimuli. This paradigm was selected to ensure that performance on the task would be similar between the groups and that the underlying brain activity would reflect response inhibition and not be biased by heavy attentional demands. We hypothesized that the controls would produce activation in brain regions involved in response inhibition including the pre-SMA, ACC, the inferior parietal lobule, precuneus, the insula, middle temporal and fusiform gyri and the prefrontal cortices. We predicted that the ASD population would show reduced activation in these brain regions.