Sustained Neural Processing in Affective Regions Predicts Efficacy of a Computer-Based Intervention Targeting Attentional Patterns in Transdiagnostic Clinical Anxiety

Abstract

Research suggests that individuals with clinical anxiety demonstrate an attention bias toward threatening information in their environment. Attention Bias Modification (ABM) is a computer-based treatment that trains attention towards non-threatening stimuli over threatening stimuli. While alterations in initial processing of threat have been linked to responses to ABM, the impact of sustained processing in the aftermath of neutral and threatening information upon outcomes following this targeted intervention has not been well studied. Our study analyzed how sustained activity in brain regions related to cognitive and affective processing can predict who is a good candidate for ABM. Unmedicated anxious individuals assigned to the ABM condition (n=38) underwent fMRI during performance of a novel task sensitive to sustained emotional information processing. Afterward, they underwent eight ABM treatment sessions. Participants whose sustained reactivity to neutral stimuli was high in the amygdala, the left BNST, the left VLPFC, and the pgACC displayed the least improvement with ABM. These results suggest that certain anxious individuals may have difficulty distinguishing between neutral and threatening information due to an overly threat-oriented appraisal of their environment, and would thus benefit less from ABM. By studying neural predictors of success in ABM treatment and focusing on the individual differences in neural-attentional dimensions within a transdiagnostic sample of anxiety patients, we can help identify which subset of anxious patients would be good candidates for this intervention in the clinical setting. Sustained Neural Processing 3 Introduction Anxiety disorders are the most prevalent class of mental illness. A recent meta-analysis of many research studies revealed that the current global prevalence of anxiety disorders is around 7.3% (Baxter et al., 2013). In the United States alone, anxiety affects roughly 18% of the nation’s population (Kessler et al., 2005). For those affected, the excessive worry, fear, and other psychological symptoms can greatly decrease quality of life, and can cause consequential medical morbidity and disability. Clinical and subclinical forms of anxiety also represent a significant public health burden, costing the U.S. more than $42 billion a year (Greenberg et. al, 1999). Anxiety disorders share features of both fear and anxiety and can manifest as many different types. Fear is an emotional response to a perceived threat and often associated with fight or flight or escape responses. Anxiety is the anticipation of a future threat and vigilance in preparation for this future threat, which may or may not include avoidance behaviors (American Psychiatric Association, 2015). Common anxiety disorders include generalized anxiety disorder, panic disorder, specific phobias, social phobia, post-traumatic stress disorder, separation anxiety disorder, and agoraphobia. Response rates for current first-line treatments stand at only 50-70% with high rates of relapse and low rates of remission (Ballenger, 2004; Barlow et al., 2004, Hofmann and Smits, 2008; McEvoy, 2007). Only 12.7% of patients affected by an anxiety disorder receive adequate treatments such as cognitive-behavioral therapies (CBT) or pharmacotherapy, and disorder prevalence rates remain high (Wang et al., 2015). Given the prevalence of these disorders, these observations emphasize the need to continue developing new and refining old treatment approaches to increase patient access and reduce costs. Attention Bias Modification (ABM) is a novel computer-based treatment approach that Sustained Neural Processing 4 offers several benefits over current first-line treatments including cost-effectiveness, ease of dissemination, and low patient burden (Price et al., 2016). ABM is designed to target a wellreplicated and studied observation in anxiety: selective attention to threat. A large body of research has established that anxious individuals as a trans-diagnostic group exhibit an attentional preference toward threatening information, or an attentional bias (AB). For example, researchers have demonstrated AB in anxious patients by using an emotional version of the Stroop task. Participants are asked to name the color in which words are printed, and the results show that anxious individuals are slower to name colors of words associated with concerns relevant to their clinical condition, indicating AB to the threatening content of the words (Williams, Matthews, & MacLeod, 1996). The most frequently used method that has been used to study anxiety-linked AB has been the attentional probe assessment task (MacLeod, Mathews, & Tata, 1986). In this visual attention task, threatening and neutral stimuli in the form of words or images are briefly and simultaneously presented in two different areas on a screen. This is followed by a small probe in the location of either of the two prior stimuli, and the participant’s speed to make a response to the probes in each location is recorded. Anxious individuals demonstrate a quicker response to probes appearing in the location of the threatening stimuli than the neutral stimuli, indicating AB towards threatening stimuli (Bar-Haim et al, 2007; MacLeod et al., 1986). The observation that AB towards threat was present in those suffering from anxiety disorders invited speculation on the potential causal role of AB on anxiety. ABM treatment was developed to therapeutically exploit the potential causal role of AB upon anxiety. ABM treatment seeks to modify AB and train patients to attend to non-threatening stimuli Sustained Neural Processing 5 preferentially over threatening stimuli in the initial stages of threat processing. The treatment includes variants of the original attentional probe assessment task reconfigured to encourage attentional change (MacLeod, Clark, 2015). In ABM, across repeated training sessions, a probe is systematically placed in the location of neutral stimuli to shape attention through practice, thereby training participants to selectively attend away from threatening information. If AB does have a causal role in anxiety, then reduction of AB towards threat should also reduce anxious symptoms. In two foundational ABM studies conducted in 2002, MacLeod et al. found that ABM was successful in both modifying attentional bias and that this modification of attentional selectivity did influence stress reactivity (MacLeod et al., 2002). After these studies, a continually growing literature suggests that attention modification interventions are effective in reducing anxiety symptoms. In a study in 2009, individuals with generalized anxiety disorder (GAD) were recruited to complete ABM or a sham/control version of the training, in which attentional patterns are not shaped either toward or away from threat. Participants who completed ABM (but not sham) reported a decrease in anxiety both through self-reports and interview measures (Amir et al, 2009). Another study in 2011 used ABM to train attention away from threat among anxious youth or children. After only 4 weeks, many youths reported a significant decrease in anxiety and no longer met criteria for an anxiety diagnosis (Rozenman, Weersing, & Amir, 2011). In another study in 2012 with an attention training program on participants with generalized social phobia, those who were trained to attend to nonthreatening cues demonstrated reductions in self-reported and physiological measures such as skin conductance of social anxiety (Heeren et al, 2015). These results are consistent with the hypothesis that attention bias plays a causal role in anxiety. However, subsequent meta-analyses studying ABM treatment Sustained Neural Processing 6 show that its potential beneficial effects on anxiety are inconsistent across individuals and studies (Price et al., 2016). Thus more research is necessary to understand which anxious patients will most likely benefit from ABM and why they do, since anxiety disorders affect a heterogeneous group of individuals. To understand which anxious patients are most likely to benefit, one factor that may be important to consider is the timeframe over which a given anxious individual exhibits attentional bias towards threat. As a group, anxious individuals exhibit threat vigilance during initial stages of processing (e.g. 16-500 ms after stimulus onset) (Bar-Haim et al., 2007). However, some anxious patients also exhibit sustained threat processing involving perseverative attention— worry and rumination— about the information even after the stimulus has been removed. The sustained or perseverative processing in the aftermath of neutral and threatening information may be important and potentially impact outcomes following ABM. Thus while alterations in initial processing of threat have been linked to responses to ABM (Amir et al, 2011; Kuckertz et al., 2014; Price et al., 2016), we currently know nothing about how sustained patterns of threat processing might impact outcomes following this targeted intervention. To date, ABM studies have largely focused on group-level observations of whether anxious patients, as a group, benefit from ABM, which can mask considerable within-group heterogeneity linked with ABM treatment outcome. Thus an individual differences approach that examines initial as well as sustained threat processing mechanisms may capture critical, clinically relevant information. Additionally, previous studies have largely focused on establishing ABM efficacy in narrow diagnostic categories. A transdiagnostic approach where Sustained Neural Processing 7 patients across multiple diagnostic categories are recruited to empirically derive common psychological mechanisms behind anxiety disorders has become increasingly viewed as important in advancing the field of psychiatry by better representing the real-world clinical patient population (Insel et. al, 2010). Consistent with this viewpoint, an important que