The treatment of major depression in late life (LLD) is an important health problem with a large and growing number of affected individuals. A significant subset of patients with LLD, particularly those with cerebrovascular risk factors (VRF)s, have subcortical gray matter (SCGMH) or frontal deep white matter (FDWMH) hyperintensities on brain mRI scans. Naturalistic studies suggest that such changes may be associated with poor acute and long-term antidepressant treatment response. However, the prognostic significance of brain MRI findings has not been evaluated systematically in prospective treatment trials. Similarly, studies have indicated that LLD patients frequently have deficits in Initiation/Perseveration (IP) and other tests of frontal lobe function. An elegant longitudinal study with nortriptyline has demonstrated that frontal executive dysfunction is associated with poor acute response as well as a greater risk for relapse/recurrence in LLD. However, the generalizability of this important finding to other classes of antidepressants, such as the widely used SSRIs, is not known. The interaction between frontal brain lesions and executive function deficits in LLD is also not fully studied. The proposed study will be the first large-scale prospective trial to simultaneously test the effects of both specific neuroanatomical and neuropsychological factors on course of response, remission rate as well as on broader measures of health outcomes in LLD. We propose to conduct a collaborative 12-week acute treatment trial with sertraline in 320 elderly (age greater than 60) patients with major depression at Duke University Medical Center and Washington University Medical Center. In a sample chosen to balance generalizability and scientific rigor, we will test the hypotheses that: 1) greater lesion severity on MRI will predict reduced treatment response, and 2) greater lesions severity on MRI willbe correlated with frontal executive dysfunction. In secondary aims, we will test the effects of lesion volume/location and frontal executive dysfunction on treatment response. In addition, we will test the effect of interactions of MRI-lesions, vascular risk factors and executive dysfunction on treatment response. The proposed trial will provide information of value to the practitioner and also provide data to build a more comprehensive model of the prognostic impact of frontal-subcortical brain changes on the outcome of late-life depression.