واکنش پذیری عصبی ردیابی گرایش های تشدید ترس است

Abstract Recent studies on fear generalization have demonstrated that fear-potentiated startle and skin conductance responses to a conditioned stimulus (CS) generalize to similar stimuli, with the strength of the fear response linked to perceptual similarity to the CS. The aim of the present study was to extend this work by examining neural correlates of fear generalization. An initial experiment (N = 8) revealed that insula reactivity tracks the conditioned fear gradient. We then replicated this effect in a larger independent sample (N = 25). Activation in the insula, anterior cingulate, right supplementary motor cortex and caudate increased reactivity as generalization stimuli (GS) were more similar to the CS, consistent with participants’ overall ratings of perceived shock likelihood and pupillary response to each stimulus.

Introduction Paradigms that assess fear learning have provided valuable translational tools for understanding the etiology, maintenance and treatment of anxiety disorders (Milad et al., 2006 and Mineka and Oehlberg, 2008). The acquisition and extinction of conditioned fear responses involve a common neurocircuitry across species that includes the amygdala, insula, anterior cingulate cortex, hippocampus, sensory areas, and ventromedial prefrontal cortex (Büchel and Dolan, 2000, LeDoux, 2000 and Phelps et al., 2004). In addition to acquisition and extinction, there is increasing interest in fear generalization, which describes the transfer of a conditioned fear response to stimuli that are perceptually similar to the conditioned stimulus (CS). Insofar as the transfer of fear responses from threat-related stimuli to potentially innocuous cues is a common feature in anxiety disorders (Lissek et al., 2008), fear generalization may be a key learning process in the development and maintenance of pathological anxiety. Recent studies have validated laboratory-based procedures for testing fear generalization, which involves the assessment of fear responses to a CS and to generalization stimuli (GS) that vary in perceptual similarity to the CS (Hajcak et al., 2009 and Lissek et al., 2008). In these paradigms, fear responses were quantified with the fear-potentiated startle reflex, which followed a generalization gradient: the strongest startle reflex was elicited during the CS, with a steep decline corresponding to the relative decrease in similarity of the GS to the CS1 (Hajcak et al., 2009 and Lissek et al., 2008). Lissek and colleagues assessed fear generalization in a paradigm in which participants had to learn which stimulus was the CS and which were the GS. On the other hand, Hajcak and colleagues found comparable results even when participants were explicitly instructed regarding the identity of the CS and the reinforcement contingencies to the CS and GS. Despite being told explicitly which stimulus was the CS, and never being shocked following a GS, participants in the Hajcak et al. study had larger startle responses and reported greater shock likelihood as GS were more perceptually similar to the CS. Fear generalization paradigms could be useful for assessing pathological fear and risk for anxious psychopathology. For instance, patients with panic disorder exhibit a flatter fear gradient with more gradual decreases in fear response to the GS (Lissek et al., 2010). Hajcak et al. (2009) reported fear generalization deficits in a generalization paradigm as a function of variation in the brain-derived neurotrophic factor (BDNF) genotype, which has been related to both learning and anxiety-related behaviors. In the current study, we sought to extend this work by examining neural activity using fMRI in a fear generalization paradigm that we previously employed (Hajcak et al., 2009). The aim was to elucidate the brain regions associated with generalization, which have received little attention in the literature, and to examine whether reactivity in these regions exhibit a similar generalization gradient to that reported with peripheral measures of fear. These neural gradients may be useful in identifying deficits in the generalization process and may be relevant to future work on pathological anxiety (e.g., Lissek et al., 2010). In the current study, the CS was a middle-sized rectangle and the GS were six additional rectangles varying in width from the CS by ±20%, ±40% or ±60%. In an initial experiment (N = 8), we examined regions of interest (ROIs) based on neuroimaging studies of fear learning that have implicated key areas in the expression and inhibition of autonomic and behavioral fear responses ( Dunsmoor et al., 2011 and Sehlmeyer et al., 2009). These ROIs included the amygdala, insula, thalamus, caudate, anterior cingulate cortex (ACC) and ventromedial prefrontal cortex (vmPFC). We hypothesized that reactivity in one or more of these regions would demonstrate a similar gradient response to the pattern reported in previous laboratory-based studies. In a second experiment 2 (N = 25), we conducted a whole-brain analysis and obtained additional self-report ratings and physiological measures.