Referee 1's Review:

The authors present novel data on a volumetric/shape analysis study of the hippocampus in late-life depression. They report a reduced left hippocampus size in elderly depressed subjects compared with a control group. This effect seems to be confined to currently depressed subjects. Their shape analysis yielded group differences in the mid-body of the left hippocampus, again only in the non-remitted subgroup.

The manuscript is concise, focused and well-written. With some exceptions (see below), the relevant literature has been covered. The results are an interesting contribution to the growing body of research concerning volumetric changes in major depression. Some of the strengths include the shape-analysis approach, a reasonable sample size, exclusion of co-morbid patients, and employment of a screening measure for dementia (MMSE). Investigating late-life depression represents a good approach to examine volumetric changes in major depression since such changes seem to accumulate with a longer course of disease and might therefore be very subtle in younger depressed subjects. However, there were some questions and methodological issues that deserve reconsideration:

Major points:

1) A major shortcoming of the investigation is the rather simplistic clinical characterization of the depressed subjects. The authors themselves state in the introduction that hippocampus volume in depression seems to be influenced by the clinical course of disease including some important variables such as age of onset, number of episodes, life-time hospitalization, cumulative time depressed, cumulative time under antidepressant medication, and type/dose of current medication. Particularly, given the proposed neurotoxicity hypotheses and the potential neuroprotective effects of antidepressant medication, these variables should be more important than the current clinical depression level assessed with the MADRS scale. At least depression duration (Sheline et al., 1999) and medication should be considered. There is some evidence that particularly (life-)time depression without medication predicts hippocampal volume loss (Sheline et al., 2003). It would be very helpful to have some of these variables investigated - this would strengthen the manuscript tremendously.

2) As the authors state correctly, non-corrected p-values should be interpreted with extreme care - even more since the authors did not provide an a-priori hypothesis concerning particular locations of shape changes. The authors should include a strong statement concerning the exploratory nature of their analysis.

3) Were the remitted and non-remitted subjects comparable with respect to socio-demographic and clinical variables? The authors could include a table for all sociodemographic and available clinical variables for controls, non-remitted depressed, remitted depressed, and the whole depressed sample.

4) Their conclusion that „It appears that remission of depression is associated with correction of hippocampal shape and volumetric differences seen in depressed subjects who are currently symptomatic" (p. 16, last sentence) is not supported by the data (rather a confusion of correlation and causality). In fact, it seems far more likely that subjects with reduced hippocampus size show a worse treatment response and are not remitted therefore (Hsieh et al., 2002).

5) The authors did not comment on recent evidence, that hippocampus volume in depression seems to be highly affected by BDNF genotype (Frodl et al., 2007). The authors might want to include a limitation statement that their sample has not been investigated for some relevant genetic variants (e.g., BDNF or 5-HTTLPR; Taylor et al., 2005).

Minor points:

1) The authors state that the hippocampus is involved in mood regulation. This seems to be a very specific statement - I would recommend rephrasing it to emotion processing rather than regulation.

2) There seems to be a mistake in the references (reference 2 has nothing to do with hippocampus or depression)

3) I realize that dichotomizing subjects into a remitted and a non-remitted group based on current MADRS score is an intuitive approach. However, it ignores about 2/3 variance of this variable. Why not including current MADRS score as a continuous covariate?

3) Were gender and age also included as covariates in the shape analysis (as in the volumetric comparison)?

4) The two groups were not matched for age, sex, handedness and education levels. Although age and sex were included as covariates, this is a limitation that should be acknowledged (as already done with respect to age differences).

References:

Frodl T, Scheuerecker J, Albrecht J, Kleemann AM, Müller-Schunk S, Koutsouleris N, Möller HJ, Brückmann H, Wiesmann M, Meisenzahl E. Association of the brain-derived neurotrophic factor Val66Met polymorphism with reduced hippocampal volumes in major depression. Arch Gen Psychiatry. 2007 Apr;64(4):410-6.

Hsieh MH, McQuoid DR, Levy RM, Payne ME, MacFall JR, Steffens DC. Hippocampal volume and antidepressant response in geriatric depression. Int J Geriatr Psychiatry. 2002 Jun;17(6):519-25.

Sheline YI, Gado MH, Kraemer HC. Untreated depression and hippocampal volume loss.
Am J Psychiatry. 2003 Aug;160(8):1516-8.

Sheline YI, Sanghavi M, Mintun MA, Gado MH. Depression duration but not age predicts hippocampal volume loss in medically healthy women with recurrent major depression.
J Neurosci. 1999 Jun 15;19(12):5034-43.

Taylor WD, Steffens DC, Payne ME, MacFall JR, Marchuk DA, Svenson IK, Krishnan KR. Influence of serotonin transporter promoter region polymorphisms on hippocampal volumes in late-life depression. Arch Gen Psychiatry. 2005 May;62(5):537-44.

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.