Alterations of regional blood flow in depressed individuals have been consistently observed. Most studies suggest that flow in the frontal lobe (particularly the prefrontal area) is reduced, but the involvement of other areas remains controversial. Discrepancies between investigations may be due not only to differences in scan methodology (tracer, imaging modality, and data analysis) but also to patient selection and heterogeneity of the mechanisms underlying major depressive disorder. Treatment-resistant depression may be associated with perfusion reductions in a greater number of areas than treatment-responsive depression (Nagafusa et al. 2012). Since the prefrontal cortex has been linked to selective attention, short-term memory, emotion, and volition, reductions of flow in this area may be related to losses of attention, mood changes, and psychomotor inhibition in subjects with depression. Most imaging findings are in agreement with the hypothesis that late-life depression relies on dysfunctioning of the frontal lobe. Frontal dysfunction has been reported not only in primary depression but also in depression associated with neurodegenerative diseases such as Parkinson, Huntington, and Alzheimer’s disease.

Several investigators have examined whether the magnitude or regional extent of flow changes is correlated with the severity of depressive symptoms. In a few published articles, flow values and scores on depression scales were significantly and inversely correlated (Bonne et al. 1996; Kang et al. 2012; Liao et al. 2003; Sabbagh et al. 1997; Sackeim et al. 1990). However, in other studies no significant correlation between flow and depression severity was observed (Awata et al. 1998; Dolan et al. 1994; Galynker et al. 2000; Navarro et al. 2001, 2002; Philpot et al. 1993; Vercelletto et al. 2002), although flow was sometimes correlated with other phenomena, such as psychosis, anxiety, negative symptoms, or intellectual deficits. Combination of the data of an initial scan with data of a second scan made after an interval of at least 1 year may provide meaningful information. Longitudinal decreases of flow can indicate refractoriness and chronification of depression (Awata et al. 1998) or be related to higher depression scores (Dotson et al. 2009).

Many studies have focused on persistence or reversibility of flow deficits during treatment, which comprised either electroconvulsive therapy (ECT) or administration of antidepressant drugs. Increases of regional cerebral blood flow were noticed in responders but were absent or nonsignificant in nonresponders to the applied therapy (Awata et al. 2002; Bonne et al. 1996; Halloran et al. 1999; Ishizaki et al. 2008; Kimura et al. 2003; Navarro et al. 2002, 2004a; Palhagen et al. 2009). The ratio of perfusion in the left anterior frontal cortex and cerebellum at baseline may be predictive of treatment outcome, low values being associated with a greater risk of therapy resistance, particularly if age of onset and duration of index episode are included as co-variables (Hanada et al. 2013; Navarro et al. 2004b). Complete reversal of the initial perfusion abnormalities was observed in some studies during successful therapy (Navarro et al. 2002, 2004a), but other researchers found both reversible and persistent perfusion deficits (Awata et al. 2002; Ishizaki et al. 2008), suggesting that flow reductions in certain brain regions are disease state-related, whereas deficits in other areas may reflect traits underlying vulnerability to depression.

Some studies reported gender differences in flow patterns associated with late-life depression. Curran et al. (1993)) detected perfusion deficits in anterior cingulate and frontal cortex only in male patients. Lesser et al. (1994) observed reduced flow in orbitofrontal and temporal areas of depressed individuals which were more striking in men than in women. Dotson et al. (2009) found more widespread decreases of flow in elderly depressed males than in females. These findings may be related to the fact that clinical depression is associated with greater decreases in frontal volumes in men than in women (Lavretsky et al. 2004) and depressive symptoms are associated with an increased risk for dementia in men but not in women (Dal Forno et al. 2005; Fuhrer et al. 2003). Most published imaging studies involved subject groups consisting of individuals from both sexes. It would be interesting to examine longitudinal blood flow changes associated with subclinical and clinical depression in a sex-specific manner.

It is a pity that deficiencies of cerebral blood flow have only rarely been linked to structural findings obtained with MRI. Particularly in neurodegenerative disease, two different mechanisms may contribute to relative hypoperfusion: (1) an actual loss of tissue in the target region and (2) a reduced function of existing tissue. The regional distribution and number of white matter hyperintensities in T2-weighted MRI images may reflect cerebral small vessel disease and can be compared to the regional pattern of hypoperfusion in the brain of patients with late-life depression (Ebmeier et al. 1997, 1998; Kimura et al. 2003; Lesser et al. 1994; Vardi et al. 2011). The term “vascular depression” has been coined to describe a subtype of depression which occurs in the context of cerebrovascular disease (Alexopoulos et al. 1997a, b; Krishnan et al. 1997). The response of regional blood flow to successful therapy in depressed individuals may depend on the underlying pathophysiology. Whereas perfusion deficits in nonvascular depression can disappear completely during remission (Navarro et al. 2002, 2004a), rCBF in the frontal lobe of subjects with vascular depression may remain subnormal both in the depressed and remitted states (Kimura et al. 2003).

Only a few reports have examined flow differences between individuals with early onset and late-onset depression. Initial studies found no significant effect of age at onset upon rCBF (Curran et al. 1993; Philpot et al. 1993), although patients with late-onset depression tended to have lower relative flow in affected brain areas (Lesser et al. 1994). Later reports have suggested that late-onset depression is associated with more perfusion abnormalities, particularly in the left temporal lobe, than early-onset depression and also with more periventricular white matter changes in MRI (Ebmeier et al. 1997, 1998). Thus, late-onset depression may be associated more frequently with cerebral small vessel disease.

Flow studies have supported the concept of a continuum of severity of depressive syndromes. Subthreshold depressive symptoms are associated with relatively small alterations in regions implicated in clinical depression (Dotson et al. 2009). SPECT with automated, semiquantitative techniques of data analysis can discriminate Alzheimer’s dementia from depression with cognitive impairment, but may not be accurate enough to differentiate Alzheimer’s dementia from mild cognitive impairment, or mild cognitive impairment from depression with cognitive impairment (Staffen et al. 2009). Early diagnosis of degenerative dementias may require the combination of SPECT with other molecular imaging techniques.