Background: Selective serotonin reuptake inhibitors (SSRIs) are commonly accepted as the first line pharmacological therapy for anxiety disorders and depression. However, there is a high percentage of patients that fail to achieve satisfactory response with SSRI treatments. The neural mechanisms underlying effective and ineffective outcome with SSRIs are not well characterized. The amygdala has dense serotonergic innervation, and studies have suggested the amygdala to be a crucial brain target for SSRI treatment. This study aimed at investigating differences in amygdala responsivity between responders and nonresponders to SSRI treatments in patients with social anxiety disorder (SAD).

Methods: Stress-related regional cerebral blood flow (rCBF) was measured in SAD patients (n=35) with 15O-water positron emission tomography (PET) during public speaking before and after 6-8 weeks of treatment with citalopram or paroxetine. Response rate was determined by the Clinical Global Impression-Improvement scale.

Results: Within-group comparisons revealed reduced rCBF response bilaterally in the amygdala in responders (n=20) as well as in nonresponders (n=15). Between-group contrasts revealed a greater amygdala attenuation in responders (>nonresponders) in the left basolateral/basomedial (x-16, y-6, z-14, Z=1.66, Puncorr=0.024) and right ventrolateral subregions (x26, y-4, z-26, Z=2.12, Puncorr=0.009). However, greater rCBF attenuation in nonresponders (> responders) was observed in the left lateral amygdala (x-28, y-6, z-14, Z=2.38, Puncorr=0.005).

Conclusions: Lowered amygdala responsivity does not seem to be exclusively related to clinical improvement in anxiety patients. In accordance with animal literature, our data suggest that amygdala subregions are functionally heterogeneous with regards to anxiolysis.