A3 3..3

In the absence of robust biological markers of psychiatric disease, Phillips (pp. 1–3) reviews data from neuroimaging studies and concludes that this approach offers the opportunity to examine the neural networks subserving the cognitive and emotional processes that are dysfunctional in psychiatric disorders. In this editorial, she reviews recent advances in using neuroimaging technology to predict diagnosis, prognosis and therapeutic response to treatment. A practical application of these approaches is evident in three papers in the Journal reporting imaging findings in depressive illness. Cole and colleagues (pp. 33–39) found that there are white matter changes in depressive illness which correlate with increasing severity of depressive symptoms. The most prominent changes were evident in the corpus callosum and superior longitudinal fasciculi. Late-life depression is associated with changes in the connections between frontal and subcortical and limbic areas. Sexton et al (pp. 46–51) found widespread change in white matter tracts in their participants with late-life depression. Changes in frontal tracts were associated with older age at onset, supporting a vascular aetiology of illness, whereas younger age at onset correlated with hippocampal change and a longer duration of illness. The authors suggest that these changes support a glucocorticoid cascade hypothesis. Firbank and colleagues (pp. 40–45) performed a longitudinal study of white matter changes and late-life depression, reporting that progression in white matter changes, over a 3-year period, was associated with depressive illness in year 3. The authors suggest that their data support the vascular depression hypothesis and encourage measures to control vascular risk factors, such as antihypertensive treatment, that have been demonstrated to improve blood pressure control but also shown to reduce the progression of white matter changes.