Group cognitive behavioral therapy modulates the resting-state functional connectivity of amygdala-related network in patients with generalized social anxiety disorder

In this longitudinal study, we used fMRI to investigate changes of rsFC in patients with GSAD before and after 8 weekly group CBT treatment. Based on a standard clinician measure, the majority of patients improved following treatment. As we hypothesized, GSAD patients showed aberrant amygdala-related network and the short-term group CBT could normalize the abnormal hyper-connectivity of amygdala-prefrontal network. Furthermore, the changes of the left amygdala-dACC connectivity were associated with the clinical improvement (changes of the HAMA scores) in GSAD patients after treatment.

The most important finding of the current study is the abnormally elevated amygdala-dACC connectivity in GSAD patients, which is normalized along with clinical symptom improvement after group CBT. The dACC is densely and reciprocally connected with amygdala [35] which mostly sends projections to amygdala [36]. Moreover, joint activation pattern of amygdala and dACC have been identified in many neuroimaging studies on emotion [37, 38]. It has been previously reported that the dorsal–caudal regions of the ACC and mPFC are involved in the appraisal and expression of negative emotions, whereas the ventral–rostral portions of the ACC and mPFC play a regulatory role in negative emotion [39, 40]. The authors concluded that the amygdala-dmPFC/ACC connectivity is associated with “negative emotion generation” [39]. Abnormal amygdala-dACC connectivity has been previously reported in association with anxiety [13, 40]. Considering the above mentioned findings, and in conjunction with our results, the abnormal hyper-connectivity of amygdala-dACC connectivity could attribute an exaggerated significance to possible threat stimuli from the internal and the external environment, leading to a high level of anxiety in patients with GSAD. Besides, present study provide further evidence to support the 8 weeks of group CBT exert an effect on normalizing the amygdala–dACC connectivity along with the improvement of anxiety symptom. Since CBT could alleviate the predisposition for excessive fear and anxiety in response to environmental stimuli [41] by effective strategies such as relaxation training, in vivo exposure and social skills training, such therapy could relief the abnormal elevated connectivity between amygdala and dACC. Therefore, our finding not only provided robust evidence for the key role of the amygdala-dACC connectivity in GSAD pathogenesis, but also provide a potential biomarker to monitor the treatment effect of CBT in GSAD patients.

Besides, we also found that the short-term group CBT reduced the abnormally elevated connectivity of the left amygdala with the left dmPFC at baseline. In fact, the abnormal elevated amygdala-dmPFC connectivity has been reported during self-directed criticism in patients with generalized social phobia [42]. Both the dmPFC and the amygdala are regions associated with fear expression in response to stimuli from the external world and a positive amygdala-dmPFC connectivity was associated with hypervigilance [14]. Thus the observed impaired amygdala–dmPFC connectivity in our study may reflect the tendency to attribute an exaggerated significance to possible self-relevant stimuli from the processed threat cues in patients with GSAD [42, 43]. And the decreased coupling of these two regions after cognitive restructuring in group CBT could underlie this modulated knowledge of the self.

Beyond the above findings, we found decreased connectivity between the left amygdala and the right putamen in GSAD after treatment. The putamen is the one of the sectors of the striatum. The striatum has been proved to be the “emotion guarder”, which was an important terminal to receive the sensory and emotional information from the prefrontal areas [44]. Thus the decreased connectivity between the left amygdala and the right putamen in SAD may represent the short-term recovery process of the negtive emotions in social situations.

Another thing should be noted is that the altered connectivity only involved the left amygdala. This phenomenon could be explained by the abundantly reported lateralization of the human amygdala. The left amygdala was more often activated than the right amygdala, irrelevant to stimulus type [45]. Our finding of left amygdala-dominant brain networks at resting-state further suggested the asymmetric interhemispheric information transfer, in line with the left hemifield of amygdala advantage for fear processing in task-related studies [46].

The findings of the present study should be considered in the context of several limitations. First, the study design lacked re-test data of the healthy control group and a placebo or wait-list control. Thus, the neural and clinical findings cannot be causally attributed to group CBT and could be related to a number of plausible factors not related to the treatment, such as the natural course of the illness over the 8-week period and placebo/expectancy effects. Second, four of the GSAD patients were on SSRI medication at the time of study participation and the medications were reported to be ineffective in the treatment of this illness. However, the effects of medication should not be ignored in explaining the results. A recent study showed mediating effects of acute SSRI intake on prefrontal-amygdala effectivity connectivity [47], although the patients withdrawed medication at least two weeks before the baseline MRI scan in present study, our findings might be related to the medications. Third, although from previous studies we know that a group of 14 participants were sufficient to detect significant treatment effects measured with fMRI [48], the limited sample size (n?=?15 GSAD subjects) increased the risk for false negatives and constrains the test for relationships between rsFC and treatment effects. These findings, however, may guide subsequent fMRI studies of treatment. Finally, there are a number of studies suggesting distinct functional connectivity patterns for the amygdala subregions such as the superficial and centromedial amygdala when choosing these subnuclei as seed [49, 50]. However, accurate delineation in the human brain is still under debate [51] and such approach has a major limitation as the amygdala is peculiarly susceptible to image distortion, normalization failure and draining vein effects [52]. Thus we chose a more conservative approach based on functional results of the whole amygdala region using the AAL template according to most human literature for investigation of the amygdala-related network.