Neural predictors and mechanisms of cognitive behavioral therapy on threat processing in social anxiety disorder

Abstract

Cognitive behavioral therapy (CBT) is “gold standard” psychotherapy for social anxiety disorder (SAD). Cognitive models posit that preferential processing of threat mediates excessive forms of anxiety, which is supported by exaggerated amygdala, insula, and cortical reactivity to threatening socio-emotional signals in SAD. However, little is known about neural predictors of CBT success or the mechanisms by which CBT exerts its therapeutic effects. Functional magnetic resonance imaging (fMRI) was conducted during responses to social signals of threat (fearful/angry faces) against positive signals (happy faces) in 14 patients with SAD before and after 12 weeks of CBT. For comparison, 14 healthy control (HC) participants also underwent two fMRI scans, 12 weeks apart. Whole-brain voxel-wise analyses showed therapeutic success was predicted by enhanced pre-treatment activation to threatening faces in higher-order visual (superior and middle temporal gyrus), cognitive, and emotion processing areas (dorsal anterior cingulate cortex, dorsomedial prefrontal cortex). Moreover, a group by time interaction was revealed in prefrontal regions (dorsomedial, medial gyrus) and insula. The interaction was driven by relatively greater activity during threat processing in SAD, which significantly reduced after CBT but did not significantly predict response to CBT. Therefore, pre-treatment cortical hyperactivity to social threat signals may serve as a prognostic indicator of CBT success in SAD. Collectively, CBT-related brain changes involved a reduction in activity in insula, prefrontal, and extrastriate regions. Results are consistent with cognitive models, which associate decreases in threat processing bias with recovery.

Prefrontal Reactivity to Social Signals of Threat as a Predictor of Treatment Response in Anxious Youth

“…Precentral and postcentral gyri are involved in somatosensory processing and voluntary movement and have also been shown to activate during emotion regulation (Buhle et al, 2014;Goldin et al, 2008;Kohn et al, 2014). Though these regions have previously been related to treatment response in anxious adults (Klumpp et al, 2013;Phan et al, 2013), very …”

“…Consistent with this, activation in the amydala, involved in signaling the presence of threat, has been shown to predict response to treatment in anxiety (McClure et al, 2007a;Whalen et al, 2008); however, results have been inconsistent. Less amygdala reactivity in adults with GAD predicted better response to venlafaxine (Whalen et al, 2008), yet greater amygdala activation predicted better response to CBT or fluoxetine in anxious youth (McClure et al, 2007a), and other studies have failed to find amygdala effects (Doehrmann et al, 2013;Klumpp et al, 2013;Nitschke et al, 2009). Other work has indicated that greater activation to threatening faces in visual processing regions (ie, occipital and temporal gyri) involved in detecting and appraising emotional cues predicted better response to CBT among adults with social anxiety disorder (Doehrmann et al, 2013;Klumpp et al, 2013), possibly indicating greater initial threat reactivity which can be modified by treatment.…”

“…For example, greater ACC activation during threat processing predicted better response to medication and CBT for adults with GAD (Nitschke et al, 2009;Whalen et al, 2008) and social anxiety disorder (Klumpp et al, 2013). Greater dorsolateral PFC (dlPFC) activation during emotion regulation predicted response to treatment in adults with GAD or panic disorder (Ball et al, 2014b), and dlPFC and ventrolateral PFC (vlPFC) activation to threat predicted CBT response in socially anxious adults (Doehrmann et al, 2013).…”

“…Given the limited existing work on predictors of treatment response in anxious youth, analyses examined whole-brain activation to emotional faces to identify predictors of change in anxiety severity following treatment. As previous fMRI work in anxious adults has indicated that brain activation during threat processing predicts response to treatment (eg, Ball et al, 2014b;Doehrmann et al, 2013;Klumpp et al, 2013), our primary analyses focused on responses to threatening (ie, angry and fearful) faces, with secondary analyses examining response to happy faces. To maximize power, our primary analyses examined predictors across treatment types, and additional post-hoc analyses examined whether effects were moderated by treatment or clinical characteristics.…”

Hjärnan formas av psykologisk behandling

“…Klumpp and colleagues showed emotional faces to SAD individuals before and after treatment and found decreased responsivity in the insula, the dmPFC and the mOFC after CBT (Klumpp, Fitzgerald, & Phan, 2013a). Combined Internet-delivered CBT and escitalopram treatment for SAD dampened amygdala responsivity to emotional faces (Gingnell et al, 2016), furthering the support for the notion that the amygdala may be a common anxiolytic pathway for various SAD treatments (Furmark et al, 2002).…”