Neural correlates of anxiety sensitivity in panic disorder: A functional magnetic resonance imaging study

Highlights

• Anxiety sensitivity is a characteristic feature of panic disorder.

• Anxiety sensitivity positively correlates with neural responses to emotional faces.

• Correlations were observed in insula and anterior cingulate cortex.

• Insula and anterior cingulate cortex involved in processing of threat related stimuli.

Abstract

Panic disorder has been associated with dysfunctional neuropsychological dimensions, including anxiety sensitivity. Brain-imaging studies of the neural correlates of emotional processing have identified a network of structures that constitute the neural circuitry for emotions. The anterior cingulate cortex (ACC), medial prefrontal cortex (mPFC) and insula, which are part of this network, are also involved in the processing of threat-related stimuli. The aim of the study was to investigate if neural activity in response to emotional stimuli in the cortico-limbic network is associated to anxiety sensitivity in panic disorder. In a sample of 18 outpatients with panic disorder, we studied neural correlates of implicit emotional processing of facial affect expressions with a face-matching paradigm; correlational analyses were performed between brain activations and anxiety sensitivity. The correlational analyses performed showed a positive correlation between anxiety sensitivity and brain activity during emotional processing in regions encompassing the PFC, ACC and insula. Our data seem to confirm that anxiety sensitivity is an important component of panic disorder. Accordingly, the neural underpinnings of anxiety sensitivity could be an interesting focus for treatment and further research.

Introduction

Panic disorder (PD) has been associated with several dysfunctional neuropsychological dimensions, among which anxiety sensitivity seems to be a unique feature of the disorder. Amplification of somatic sensations with subsequent dysfunctional cognitive appraisal, and a significant bias toward a danger-related and catastrophizing interpretation, is considered a core neuropsychological construct in anxiety disorders including panic disorder (Clark, 1986). Anxiety sensitivity (AS) refers to a tendency to fear anxiety-related sensations (Reiss and McNally, 1985), and it is characterised by fear amplification in response to stimuli that elicit anxiety (Reiss, 1991). High AS individuals maintain the belief that experiencing anxiety or fear will cause severe negative consequences (i.e., heart attack), are hypervigilant to stimuli that signal anxiety (i.e., increased heart rate), worry about becoming anxious, and avoid stimuli that provoke anxiety (Reiss et al., 1986, McNally, 2002). AS has been associated with PD (Foot and Koszycki, 2004, White et al., 2006, Naragon-Gainey, 2010), and the level of AS is greater among individuals with anxiety disorders in general (i.e., PD, social phobia, specific phobia, generalized anxiety disorder, obsessive-compulsive disorder, post-traumatic stress disorder, and agoraphobia without panic) as compared with nonclinical controls (Olatunji and Wolitzky-Taylor, 2009). However, prospective studies have shown that AS specifically predicts the onset of panic (Benitez et al., 2009) and that PD patients significantly differ from other anxiety disorders in AS levels, suggesting unique features of AS in PD (Olatunji and Wolitzky-Taylor, 2009).

Different studies have identified deficits regarding emotion processing in PD (Galderisi et al., 2008, Cucchi et al., 2012), including the tendency to interpret ambiguous stimuli as threatening, and a bias toward somatic concerns in stressful situations (Clark, 1986). Brain-imaging studies of the neural correlates of human emotional processing, together with structural data from human and animal research, identified a network of structures that constitute the neural circuitry for emotions. The circuit includes the amygdala, insula, cingulate and prefrontal cortex (PFC), which interact to identify the emotional significance of the stimuli and to generate and regulate affective states (Phillips et al., 2003a, Phillips et al., 2003b). This network can be studied with functional magnetic resonance imaging (fMRI) paradigms, which can be used to examine the effective connectivity of the amygdala and its activation in response to negative facial expressions (Stein et al., 2007a). Increased activation of regions pertaining to this network has been observed across a range of anxiety diagnoses (post-traumatic stress disorder, social anxiety disorders, specific phobia and panic disorder) (Sakai et al., 2005; Britton and Rauch, 2008), and recruitment of these regions appears to be central to different self-preservation behaviours (i.e., frozen with fear; fight or flight). Moreover, the insula and the ACC, together with midbrain periaqueductal gray matter, have been suggested to be involved in the pathophysiology of panic disorder (Graeff and Del-Ben, 2008). However, responses in this circuit have varied depending on the paradigm used and the specific state explored. Hyperactivity of the amygdala and hypoactivity of the ACC have been found in anticipatory anxiety of a chemical challenge that may trigger panic attack (Boshuisen et al., 2002), while increased activity of the ACC has been observed during imagination of anxiogenic situations (Bystritsky et al., 2001). Hyperactivity of the amygdala has also been reported during emotional conflict (Chechko et al., 2009) and spontaneous panic attacks (Pfleiderer et al., 2007, Dresler et al., 2011), while hypoactivation has been shown in response to angry, fearful, happy and neutral faces (Pillay et al., 2006, Demenescu et al., 2013). A fear-conditioning task led to increased activity in the amygdala, subgenual cingulate, and midbrain structures (Tuescher et al., 2011), while activation in the frontal cortex has been observed for differential conditioning (Lueken et al., 2014). Finally increased activity in the prefrontal cortex, hippocampus and cingulate cortex has been reported (Ball et al., 2013, Dresler et al., 2013).

Fear-conditioning paradigms in panic disorder lead to increased activation in this network during a safety condition, suggesting that a network that signals fear is incorrectly activated (Tuescher et al., 2011). Changes in brain activation in cortico-limbic structures follow successful treatment with cognitive behavioural therapy (CBT), in particular, an inhibitory functional coupling between the ACC and the amygdala was found to be associated with response to treatment (Lueken et al., 2013).

Abnormalities in activity in these brain areas may then represent the core neural correlates of the psychopathology of panic disorder (Holzschneider and Mulert, 2011). Contrasting results have been reported in relation to the neural correlates of AS in healthy subjects. AS has been positively associated with activity in the insula (Stein et al., 2007b, Killgore et al., 2011) and ACC (Ball et al., 2012) during emotional processing, while no correlation between AS and brain activation has been observed during affective uncertainty processing (Simmons et al., 2008). A negative association between AS and neural responses to fearful faces compared with neutral faces has been found in the inferior parietal cortex, prefrontal cortex, insula, and ACC (Schafer et al., 2009).

Only one study, to our best knowledge, investigated the neural correlates of AS in panic disorder, and that study found a positive correlation between AS and brain activation in the midbrain during fear conditioning (Lueken et al., 2014).

Because PD is clinically characterized by AS and that, from a neuroimaging perspective, previous studies showed functional abnormalities in brain regions related to this construct, the aim of the study was to investigate whether neural activity in response to emotional stimuli in the cortico-limbic network was associated to AS in a sample of patients affected by PD.