Cortical thickness and VBM-DARTEL in late-life depression

Abstract

Background

Numerous studies have revealed structural brain changes in late life depression, mainly in white matter or whole lobes with few focussing just on grey matter (GM). The objective was to investigate GM changes in older depressed and similar aged healthy subjects using two different methods, cortical thickness in frontal lobe structures and voxel-based morphometry (VBM).

Methods

Sixty eight subjects participated (30 healthy comparison subjects, 38 depressed) and underwent 3T T1 MR imaging as well as clinical and cognitive assessments. Frontal cortical thickness was measured using FreeSurfer while VBM was undertaken using the DARTEL algorithm in SPM8. Group differences in cortical thickness and GM volumes were assessed using ANCOVA. Effects of cortical thickness and VBM results on cognitive and depression variables were also investigated.

Results

No significant differences in frontal lobe cortical thickness were observed between groups (F_{1, 62} ≤ 2.7, p ≥ 0.1). In addition, no significant relationships of cortical thickness on cognitive and depression scores were identified (partial correlation |r'| = 0.01–0.31, p ≥ 0.06). VBM showed that GM volumes were indistinguishable between groups but significant age effects were apparent, independent of diagnosis.

Conclusions

Results suggest that cortical GM changes in late life depression (LLD) are similar to healthy older subjects and appear to be related to age rather than cognitive or depressive symptoms. Changes to white matter and subcortical GM structures may be more relevant in explaining the underlying neurobiology of LLD.

Introduction

Late-life depression (LLD), generally taken as referring to depression in people over the age of 60, is common, associated with cognitive impairment and disability (Alexopoulos, 2005) and can be a severe disorder with high probability of relapse (Luijendijk et al., 2008). Studies have suggested that depression in older people is characterised more by clinical and imaging changes associated with cerebrovascular disease than with individuals where depression occurred earlier in life (Baldwin and O'Brien, 2002).

Evidence from structural magnetic resonance (MR) imaging has shown cortical and subcortical grey matter (GM) changes in older depressed subjects compared to similar age healthy individuals. Using region of interest (ROI) methods, volumetric reductions have been shown in frontal, temporal and parietal lobes as well as amygdala, thalamus, hippocampus and putamen (Andreescu et al., 2008). However, the ROI approach has some limitations, it does not evaluate the entire brain volume and, more importantly, cannot determine whether some lobar volume reductions are due to changes in GM, white matter (WM), or both. In contrast, voxel based morphometry (VBM) does allow separate assessment of GM and WM density changes in an unbiased way across the whole brain. Investigators have used VBM to study structural brain abnormalities in depressed older people, though findings remain inconclusive. VBM studies have shown GM reductions in right hippocampus (Bell-McGinty et al., 2002, Egger et al., 2008), right amygdala, bilateral medial orbitofrontal cortex (Egger et al., 2008), right superior frontal cortex, left postcentral cortex and right middle temporal gyrus (Yuan et al., 2008) in patients relative to healthy comparison subjects. However, there are major inconsistencies between studies in the areas affected and a more recent study challenged earlier findings, reporting no significant differences in GM density between their older depressed cohort and age matched comparison subjects (Koolschijn et al., 2010). Such discrepancies may be due to differences in samples, depression severity, illness duration and/or (antidepressant) medications as well as differences in MR acquisition and analysis methods. In addition to GM changes, these structural changes appear related to propensity to depression, representing trait rather than state markers. For example there is now good evidence from epidemiological studies that they pre-date future depressive episodes (Godin et al., 2008, Teodorczuk et al., 2007, Teodorczuk et al., 2010).

The measurement of cortical thickness (CT) is a relatively new procedure for assessing brain structure. It can provide complementary information to other image analysis techniques in understanding the neuroanatomy of various mental disorders. CT measurements allow the regional distribution and quantification of GM cortical loss to be specifically examined in contrast to gyral or lobar volumetric studies which often combine GM and WM within regional volumes. Although VBM does permit separate GM assessment, changes have been shown to be more closely related to alterations in cortical surface area and folding (Voets et al., 2008, Winkler et al., 2010). Few studies have investigated CT in mood disorders particularly in late life. Significant thinning of the corpus callosum was shown to distinguish late-onset depression from early-onset depression and comparison subjects (Ballmaier et al., 2008). Prefrontal cortical thinning has also been observed in younger individuals with bipolar disorder (Lyoo et al., 2006), while reduced CT in the right hemisphere may increase the risk of developing depressive illness (Peterson et al., 2009).

The objective of this study was to investigate GM CT in frontal lobe structures as well as whole brain GM volume changes in subjects with LLD compared to similar aged healthy comparison subjects. In accordance with previous evidence, we hypothesised that frontal cortical thinning and hippocampal GM loss would be present in LLD relative to comparison subjects, and not associated with current depressive symptom severity.