The interference of operant task performance by emotional distracters: An antagonistic relationship between the amygdala and frontoparietal cortices

Abstract

This fMRI study investigates neural activity associated with the interfering effects of emotional distracters. While in the scanner, participants made simple motor responses to target stimuli that were preceded and followed by positive, negative, or neutral images. Despite instructions to disregard the pictorial images, participants were slower to respond in the presence of positive or negative relative to neutral distracters, and significantly slower for negative relative to positive distracters. Enhanced activity in the amygdala and visual cortex was evident during trials that included positive and negative distracters. In contrast, increased activity in inferior frontal gyrus (BA 47) was only observed during trials that involved negative distracters. Connectivity analysis showed that activity in right amygdala correlated with activity in cingulate gyrus, posterior cingulate, middle temporal cortex, and was negatively correlated with activity in lateral superior frontal gyrus, middle frontal/orbital gyrus, and parietal cortex. The pattern of neural activity observed was interpreted within the framework of current cognitive models of attention. During a task demonstrating behavioural interference in the context of emotional distracters, increased activity in neural regions implicated in emotional processing (the amygdala) was associated with reduced activity in regions thought to be involved in exerting attentional control over task-relevant sensory representations (a frontoparietal network).

Introduction

The ability to quickly orient attention to positive or negative emotional cues can enhance a range of complex behaviors. However, an unhealthy preoccupation with emotional cues can hinder functioning. For example, a cognitive bias for negative material is thought to contribute to mood and anxiety disorders (Beck et al., 1979, Mathews and MacLeod, 1994, Bradley and Mogg, 1994). At the other end of the spectrum, impoverished processing of emotional material is seen as a key feature of psychopathic disorder (Blair, 2004, Hare, 1991, Mitchell et al., 2006). Adaptive responding therefore requires striking a balance between attending to emotional representations when they are relevant, and minimizing their influence when they are not. However, we know relatively little about the neural regions involved when goal-directed behaviors are influenced by negative distracters. Even less is known about the neurocognitive systems important for regulating the impact of positive distracters. One method of indexing the level of interference generated by emotional distracters on goal-directed behavior is through the use of the “Emotional Interrupt Task,” which measures the impact of irrelevant emotional stimuli on the speed of operant responding. The purpose of the present study was to pair this task with fMRI to delineate the neural activity associated with operant responding influenced by positive and negative distracters. Specifically, we examined whether there would be an antagonistic relationship between neural activity associated with emotional versus operant processes.

The biased competition model of attention predicts a common neurocognitive pathway for the emotional interference of simple operant behavior, and that of more complex forms of executive function (e.g., memory or cognitive control). This model suggests that attending to one stimulus or class of stimuli decreases the availability of cognitive resources available for others (Desimone and Duncan, 1995). In this way, stimuli compete for neural representation and control over behavior. Stimuli that are relevant to ongoing behavior can be “selected” for processing through the influence of executive attention mechanisms, thereby reducing the impact of distracters. However, other factors such as stimulus salience or emotion also play a role in selection. The competitive advantage that emotional stimuli have for attention is thought to result from reciprocal functional connections between the amygdala, occipital, and temporal cortices (Morris et al., 1998, Pessoa et al., 2002, Vuilleumier et al., 2004). Specifically, the amygdala is thought to be activated by valenced stimuli, and subsequently to strengthen the representation of emotional stimuli in sensory representation areas of temporal and occipital cortex (Morris et al., 1998, Pessoa et al., 2002, Vuilleumier, 2005).

The Emotional Interrupt Task provides a test of the biased competition model in that it involves operant responding that can be influenced by both emotional processes, and top–down executive attention processes. The data concerning the impact of emotional distracters on behavior have largely been derived from studies involving two behavioral paradigms: (1) tasks involving the emotional interference of executive function (i.e., higher order cognitive function such as working memory or Stroop task performance) and (2) “conditioned suppression” paradigms. With regard to the impact of emotion on executive function, it has been shown that the presentation of a CS+ that predicts an aversive auditory stimulus interferes with executive task performance (Salgado et al., 2000). At the neural level, increased amygdala activity and decreased lateral prefrontal and parietal cortex activity is associated with the emotional interference of working memory (Dolcos and McCarthy, 2006). Interference of Stroop task performance by positive and negative distracters is associated with increased amygdala activity and compensatory recruitment of a frontopolar region of middle frontal cortex (Blair et al., 2007). Studies have also shown that increased processing load reduces emotional distracter-related activity in the amygdala, superior temporal gyrus, and ventromedial prefrontal cortex (Mitchell et al., 2007, Pessoa, 2005). Taken together, these studies suggest that an antagonistic relationship exists between regions of prefrontal cortex and the amygdala. However, relatively little work has specifically tested this. Furthermore, whereas these studies explore the impact of emotion on more complex executive operations, less is known at the neural level about how emotion interferes with more basic forms of operant behavior.

Conditioned suppression paradigms model the interference of basic operant behavior by emotional stimuli. Early studies demonstrating conditioned suppression revealed that presenting a conditioned stimulus (CS+) predicting shock interferes with operant responding in rats (Bouton and Bolles, 1980, Estes and Skinner, 1941). Conditioned suppression is neurally dissociable from freezing. Lesions of the periaqueductal gray area impair freezing but not conditioned suppression (Amorapanth et al., 1999). However, lesions of the amygdala do interfere with conditioned suppression (Killcross et al., 1997, Lee et al., 2005). Furthermore, whereas freezing occurs in response to threatening or aversive stimuli (Blanchard et al., 1977), conditioned suppression also occurs to a CS+ that predicts a rewarding outcome (Schindler et al., 1999). Although the interference of operant behavior by complex emotional stimuli has been demonstrated behaviorally in humans (Hartikainen et al., 2000, Mitchell et al., 2006, Tipples and Sharma, 2000), its neural correlates remain unclear.

The current fMRI study involves a laboratory task that, like measures of conditioned suppression used in animal studies, uses emotional distracters to interfere with basic goal-directed behavior. In line with current models of attentional competition (Desimone and Duncan, 1995), and cognitive control (Botvinick et al., 2004), we predicted that positive and negative distracters would activate the amygdala, be more strongly represented in sensory representation areas, and interfere with operant behavior. We also predicted that a mutually antagonistic relationship (negative connectivity) would exist between activity in regions implicated in attributing emotional salience to stimuli (i.e., the amygdala), and those thought to be involved in attentional selection (i.e., lateral prefrontal cortex and parietal cortex).