To the Editor: We agree with the comments of Dr. Tebartz van Elst et al. that the finding of larger volume of the amygdala in patients with depression is worthy of comment. Our finding of a 25% larger mean amygdala volume (23% on the left and 27% on the right) in patients with treated unipolar depression seemed difficult to explain at first and was not hypothesized a priori; however, as Dr. Tebartz van Elst et al. point out, there are now four reports of enlarged amygdala volumes in patients with various affective disorders, including bipolar disorder, temporal lobe epilepsy with comorbid dysthymia, and unipolar major depression in remission (our report). Dr. Tebartz van Elst et al. mention several potential explanations for a greater amygdala volume in depression, including the possibility that this is a risk factor for depression, which certainly should be explored through genetic studies. They also indicate that a greater amygdala volume could represent a neuroanatomical correlate of depression.

Dr. Tebartz van Elst et al. suggest that greater blood flow or greater neuronal number may explain larger amygdala volume. Although a greater blood flow in depression has been reported in at least one study, there are a larger number of studies that have not shown this. In any case, it is not clear how a larger blood flow would translate into greater volume. Although the potential for neurogenesis has been reported for the hippocampus, we are not aware of any studies related to the capacity for neurogenesis in the amygdala. The role of the amygdala in emotional processing suggests the possibility that greater demands on the amygdala in patients with depression result in structural changes, possibly related to plastic changes in dendritic branching or neuronal morphology in the amygdala. Dr. Tebartz van Elst et al. mention possible functional correlates of greater amygdala volume, including regulation of corticotropin-releasing hormone (CRH) release by the amygdala.

The amygdala primarily regulates extrahypothalamic release of CRH, which may explain the findings of elevations of CRH in CSF in depression, although that would not explain hypercortisolemia per se. We found no pattern of relationship between amygdala volume and plasma cortisol level, although there was a modest but nonsignificant relationship between a higher cortisol level and a smaller left hippocampal volume (r=–0.31, df=9, p=0.38). However, our patients had treated depression, and hypercortisolemia has been reported only with current episodes of depression. Future studies should look at cortisol-amygdala relationships in untreated depression. The amygdala and hippocampus also have important interconnections, as suggested by Dr. Tebartz van Elst et al., and we found a modest but nonsignificant relationship between a greater right amygdala volume and a smaller right hippocampal volume in patients (r=–0.35, df=15, p=0.19) but no pattern of relationship in comparison subjects. Future studies are indicated to replicate and extend the finding of greater amygdala volume in depression and the relationship between amygdala volume and number of depressive episodes.