The present study found that individuals with MDD demonstrated a volumetric decrease in the right rACC (BA 32) as compared to controls. This finding is supported in part by previous studies, which showed that grey matter volume within other sub-regions of the ACC is reduced in individuals with MDD 
. Notably, our findings demonstrate linkages between decreased rACC volume in the disorder and both salivary cortisol and early childhood maltreatment. These data extend findings from preclinical studies suggesting that observed volumetric decreases in the ACC may be a consequence of prolonged exposure to glucocorticoids resulting from chronic stress 
Localization of the decrease in anterior cingulate volume to the rostral subdivision is particularly relevant, as this region has been strongly implicated in the pathophysiology of depression in previous studies 
. Additionally, several fMRI studies using working memory and attention tasks have reported that individuals with current or remitted MDD exhibit increased rACC activity in order to match the same level of performance as control subjects 
. This suggests that the rACC may be less efficient in individuals with MDD as compared to controls, which may result from altered ACC morphology. Consistent with this interpretation is a recent animal study that found reduced working memory performance and decreased connectivity between the hippocampus and ACC following exposure to chronic stress, which was accompanied by atrophy in laminas I and II of the ACC 
Prior reports of decreased cingulate volume have emphasized changes in the left hemisphere (14, 17), although some reports have suggested decreased cingulate volume bilaterally 
. Additionally, a recent meta-analysis of volumetric changes in depression demonstrated that both right and left anterior cingulate volumes are reduced in the disorder 
. In our study, although we emphasize the volumetric decrease in right anterior cingulate, which remained significant after correcting for multiple comparisons, a small cluster in the left hemisphere was also marginally significant.
The observed relationship between reduced rACC volume and a history of early adverse events is consistent with prior findings that have revealed a relationship between chronic and repeated stress and anterior cingulate structure 
. The correlation between elevated cortisol levels and reduced rACC volume among depressed individuals is also consistent with results from animal studies regarding the role of mPFC in HPA axis negative feedback regulation. It is noteworthy that only the average of all seven cortisol samples was correlated with rACC volume, while the average of the morning samples was not. This suggests that volumetric decreases in the rACC are not specifically linked to peak cortisol activity; rather, rACC volume appears to be more closely related to sustained glucocorticoid exposure.
Our findings suggest that chronic stress subsequent to childhood maltreatment may serve to initiate glucocorticoid-related injury to the ACC. This damage may impair cortico-limbic circuits involved in emotion regulation; in addition, insult to the ACC may diminish its ability to exert negative feedback control over future HPA activity. Together, these two outcomes may result in poor regulation of stress, and could play a role in both the initiation of depression and increased vulnerability to recurrence. A recent longitudinal study also suggests that decreased volume in ACC in individuals with depression may result from stress 
. Further research will be required to clarify the temporal relationships between early adverse events, increased HPA activity and structural integrity of the ACC.
We did not find any group differences in the hippocampus. This may result from heterogeneity of important clinical variables within our sample, including the number of episodes, duration of illness and severity of early life trauma. Prior studies that have identified hippocampal decreases associated with MDD have often reported that the extent of hippocampal damage is associated with the duration of illness 
particularly when it is untreated 
(cf Campbell and MacQueen, 2006 for a review 
). In contrast, the number of previous episodes in our sample ranged from none to four or more. Similarly, Vythilingam et al found volumetric decreases in individuals with both MDD and a history of severe child abuse, but not MDD alone 
. In our sample, the severity of early adverse events varied from none to severe. This heterogeneity in the severity of early life stress may partially explain why we failed to observe group differences in hippocampal volume. Finally, it should be mentioned that voxel-based morphometry is not the ideal method for investigation of the hippocampus, as the anatomical complexities of this structure make it very difficult to accurately segment grey and white matter tissue classes using this technique.
Several limitations in the present study warrant mention. First, we did not find any significant differences using a whole-brain analysis after correcting for multiple comparisons, suggesting that where volumetric differences occurred in our depressed subjects, the effects sizes were only small to moderate. An additional limitation is the use of the VBM method, which is susceptible to normalization and segmentation errors. This issue is compounded by our relatively small sample size, as VBM is best suited for a sample size of 25 subjects or more per group. It should also be noted that this method is particularly susceptible to errors in the evaluation of hippocampal volume, and manual segmentation remains the preferred method for this region. Further, patients with MDD were asked to evaluate their history of traumatic childhood experiences while they were in the acute phase of depression, which may have influenced their memory for events. Our study was limited by its reliance on salivary cortisol as the only measure of HPA activity, as opposed to other forms of assessment of HPA function such as the dexamethasone suppression test or the corticotropin releasing hormone (CRH) test. Finally, the complete neurobiological mechanisms by which elevated cortisol precipitates structural damage in the ACC are likely to involve additional variables that were not evaluated in the present study.
The present study replicates several previous reports suggesting the MDD is associated with decreased grey matter volume in the anterior cingulate. In addition, this study suggests that the extent of grey matter loss is related to both a history of early adverse events and circulating cortisol levels. These data further implicate the rostral cingulate as a key region in the regulation of HPA activity and the pathophysiology of MDD.