Differences in corpus callosum (CC) morphology¹ and microstructure² have been implicated in late-life depression, however the relationship between macrostructure and microstructure is rarely investigated in the same region using complementary techniques. This is necessary to disentangle genuine microstructural alterations from partial volume effects arising from macrostructural differences³. In this study we used T1 anatomical information in combination with two different diffusion MRI techniques to investigate CC structure in late-life depression.3T (Philips) 3DTFE T1 MRI data were collected from 51 currently depressed patients with late-life depression and 52 healthy controls (voxel-size 0.98 x 0.98 x 1.2mm); diffusion MRI (dMRI) data were obtained on a subset of 45 patients and 32 controls (b800mm/s2, 45 directions, voxel-size 2.2 x 2.2 x 2.2mm). The CC mid-sagittal slice was defined in each subject's native T1 image using an automated pipeline⁴ and parcellated into eight callosal sub-regions: rostrum, genu, body (anterior, mid and posterior), isthmus, splenium (ventral and dorsal) (fig 1). The T1 and dMRI data were diffeomorphically coregistered using ANTS and all measurements were obtained in T1 space. The following parameters were obtained from the whole CC and each subcallosal region for group comparison: thickness and area, estimated using Adamson et al⁴; fractional anisotropy (FA) and radial diffusivity (RD), estimated following motion and distortion correction using ExploreDTI⁵; and apparent fibre density (AFD)⁶ using MRTrix⁷. AFD was calculated following bias correction; WM and CSF response functions were calculated in all controls in single fibre and pure CSF masks, then averaged. Multi-tissue CSD⁸ with average WM and CSF responses was performed on all subjects. AFD was calculated as the integral across the entire sphere and intensity normalised using the median b0 obtained from a whole brain mask. MANCOVA analyses were used to identify group differences in each parameter, with diagnosis as fixed factor and age and total intracranial volume (tiv) as covariates. Exploratory ANCOVA analyses were used to assess main effects of late depression onset, psychosis and white matter hyperintensity (WMH) lesion load (estimated using⁹). Bonferroni adjusted p-values were deemed significant at p<0.05There were no group differences with respect to age (t=-0.920,p=0.360), gender (χ²=0.095, p=0.830), tiv (t=0.166,p=0.0889), or WMH lesion load (t=-1.253,p=0.213). We detected a decrease in area in the mid- F=4.915,p=0.029, η²=0.047 and posterior body (F=4.977,p=0.028, η²=0.048), but no differences in mean thickness between patients and controls in any region (table 1). There were no regional differences in FA or RD, whereas mean AFD was significantly decreased in patients across the whole CC (median AFD, F=4.40,p=0.039,η²=0.058), CC body (F=<5.849,p<0.022,η²>0.073) and ventral splenium (F=5.767,p=0.019, η²=0.077) (table 1, figure 2). There were no associations between callosal area or AFD and the presence of psychotic features, late-onset of depression or depression severity (MADRS).In this study we found that corpus callosum macrostructure was largely preserved, with only regional area reduction in the mid and posterior callosal body in late-life depression. This is line with other studies¹. There was some evidence of microstructural alterations throughout the CC, particularly in the callosal body and ventral splenium, and that AFD was more sensitive than DTI metrics for detecting these differences. This is in contrast to other studies which detected reduced FA in patients². As we did not detect significant differences in thickness or DTI measures, it remains plausible that in other studies, changes in DTI measures are enhanced by macrostructural partial volume effects³. The lack of group differences in WMH and a lack of association between WMH or late onset of depression and CC measures suggests that different kinds of white matter pathology may independently contribute to late life depression.Late-life depression is associated with subtle, independent macro and microstructural changes in the corpus callosum, which are unrelated to depression subtype or illness severity. Combining anatomical and diffusion data provides greater insight into structural MRI changes than using either technique in isolation