Elderly depression and diffusion tensor imaging


Dr. Masatoshi Takeda, MD, PhD, Department of Psychiatry, Osaka University Graduate School of Medicine, D3, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan. E-mail: mtakeda@psy.med.osaka-u.ac.jp

The diagnosis of depression changed significantly after the introduction of the operative diagnostic systems DSM-4 and ICD-10.1 For a long time, the traditional classification system of mood disorders was used to categorize depression into subtypes (endogenous, neurotic, or reactive) depending on the specific set of symptoms and etiologic factors. Once a subtype had been assigned to a depressed patient, this subtype diagnosis was used to select suitable intervention modalities (pharmacotherapy, psychotherapy, or environmental adjustment). This scenario may be an oversimplification of the process in an actual clinical setting, but subtype classification of mood disorders was used as a guide in psychiatric treatment. The DSM has broadened the range of depression regardless of etiology; however, it is also true that many clinicians are not totally happy with the reality that all depressed patients are grouped together under the broad concepts of ‘major’ or ‘minor’ depression.2


Before the introduction of the DSM, a lower prevalence of depression in the elderly was reported in Japan, with a national survey after 1995 reported the prevalence of major depression in the elderly at 6%, which is in the range of reported for other countries (2.0–6.1%). Currently, the prevalence of depression in Japanese elderly patients (both minor and major depression) is 15%, which is similar to that in other countries (4.5–26.9%).

There are several features that differentiate depression in the elderly from depression in younger patients: (i) many elderly patients often have somatic and/or psychiatric diseases; (ii) the course is sometimes atypical; and (iii) elderly patients are more affected by environmental or psychological factors.2


In the clinical setting, depression in the elderly needs to be differentially diagnosis from neurodegenerative diseases causing dementia,3 including Alzheimer's disease,4 diffuse Lewy body disease,5 frontotemporal dementia,6 and Parkinson's disease7, but this is sometimes difficult, even with neuropsychological,8,9 biochemical,10–12 electrophysiological,13,14 or brain imaging examinations.15–18 For dementia patients with depression, effective pharmacological4 and non-pharmacological19–22interventions are sometimes required to control agitation, irritation, negativism, and violent and suicidal behaviors.11


Elderly patients with depression show clinical features that differ from those seen in younger patients with depression, such as: (i) less depressive mood; (ii) more agitation and irritation; (iii) greater hypochondria; (iv) greater apathy and less motivation; and (v) a higher incidence of lacunes or cerebrovascular abnormalities, as determined by brain imaging.

Agitated depression is often seen in the elderly, who may not exhibit an obvious depressive mood, rather appearing agitated and/or irritated. These types of elderly depressive patients are prone to suicide and self-harm, and clinicians should pay close attention to any abnormal behaviors.23 Some elderly depressed patients complain of their physical rather than psychiatric symptoms. They are often obsessed with minor physical ailments and regard their physical illness with a pessimistic prognosis (masked depression). Underlying the pathological mechanism of pseudodementia is elderly depression. Those patients with pseudodepression exhibit marked disturbances in activities of daily life (ADL), but their performance on memory tests is relatively well preserved. Impairments in their ADL are probably due to the depressive mood inhibiting spontaneous activity.24 Conversely, it should be kept in mind that there are dementia patients who exhibit depressive symptoms as the initial signs of dementia. Elderly depressed patients easily develop delusions and psychotic features (one of the characteristics of elderly depression), which may develop into Cotard syndrome with severe nihilistic delusions.


Epidemiological studies have identified risk factors for elderly depression as follows: (i) female gender; (ii) comorbidity with physical illnesses; (iii) cognitive impairment; (iv) impairments of ADL; (v) social isolation; and (vi) a past history of depression.

Elderly people experience many losses, including a loss of health, loss of social role, loss of financial independence, and loss of human relationships. All these losses make the elderly more vulnerable to depression.

Recently, however, vascular factors have been proposed to be more important risks for the development of elderly depression. It has been noted for many years that elderly depression is associated with cognitive impairment, ischemic brain regions, more impaired ADL, less genetic load in a family, and fewer personality changes. These features of elderly depression make it different from depression in younger people. It is also well recognized that elderly depressed patients with white matter high-intensity areas by T2-weighted magnetic resonance imaging (MRI) are less likely to recover. Ischemic lesions between the striatum and the frontal lobe, as well as periventricular ischemic lesions, are specifically found in elderly depression. Considering these findings, the concept of vascular depression was proposed25,26 and it is now widely accepted that depression will be caused by vascular lesions in the white matter and deep cerebral basal ganglia. Multiple infarctions of the lacunae may be related to the occurrence of depression in the elderly. It has been hypothesized that the white matter tract between the striatum and the frontal lobe is closely related to depression and cognitive impairment in the elderly.27


Depression is observed in 20–60% of post-stroke patients, which is much higher than the rate in the general population. Although it is conceivable that patients with stroke may show depressive response after life-threatening events, stroke itself is now regarded as a cause of depression based the following findings. First, the frequency of depression in stroke patients was much higher than in patients after orthopedic surgery, which results in a similar level of physical immobilization.28 Second, the prevalence of depression was higher in patients after or carotid artery embolism than in patients with peripheral arterial disorders.29 Of interest, there rates of depression are comparable between patients with myocardial infarction and those with stroke.30 Thus, depression is considered to be caused by insufficiency of relatively large arterial vessels in the central nervous system.

Post-stroke depression

Post-stroke depression (PSD) was proposed by Robinson and Price in 1982.31 According to Robinson and Price, post-stroke patients exhibit both major (19.3%) and minor (18.5%) depression. The neuropsychiatric profile of PSD is quite similar to that in depressive patients in general, except that PSD patients have higher slowness in and lower social anxiety and loss of interest scores. The depressive symptoms seen in PSD are related to the loci of infarction. In earlier studies, Robinson reported that 60% of PSD patients had lesions in their left frontal area. Lesions near the left frontal pole are more often associated with more severe depression, highlighting the relevance of neural circuits in the deep white matter in the left frontal cortex32

Vascular depression

Many clinicians recognize specific features of elderly depression compared with features of depression seen in young adults. Taking into account brain imaging data, the concept of vascular depression has been developed by Alexopoulos and Krishnan.25 Alexopoulos et al. proposed clinically defined vascular depression,33 whereas Krishnan et al. proposed the concept of MRI-defined vascular depression.26 Patients with hyperintensities on brain images respond less well to anti-depressant treatment, tend to exhibit greater cognitive impairment after recovery from depression, have a greater relapse rate, and tend to transit into chronic depression.


There are many findings supporting the notion that vascular insufficiency could be a risk factor for depression. Subjects at risk of vascular disorder tend to exhibit depressive symtoms. Elderly patients with depression are frequently found to have deep white matter hyperintensities (DWMH) on T2-weighted MRI and hyperintensities in the periventricular area. A higher frequency of artheriosclerosis and ischemic lesions is observed in elderly patients with depression at autopsy. In addition, higher rates of depression have been confirmed for patients with coronary artery insufficiency, after stroke, in those with diabetes mellitus, and in those with hypertension. These observations may imply that cerebral insufficiency causes depression.

In a prospective study of nursing homes residents, a higher incidence of depression was reported in subjects with vascular risk factors (n= 100).34 In another study, the incidence of depression was investigated in 670 elderly people who were divided into three groups: (i) stroke patients; (ii) individuals with vascular risk factors; and (iii) normal control subjects. The incidence of depression was 36.4 % in stroke patients, 35.2% in individuals with vascular risk factors, and 28.7% in control subjects. The weight of risk factors and the incidence of depression were correlated among the subjects with vascular risk factors.35


Conversely, vascular disorders, including coronary ischemic disorders,36,37 and cerebral infarction38–42 have been shown to be risk factors for the development of depression. Of the risk factors for cerebral vascular disorders (depression, hypertension, smoking, diabetes, carotid artery stenosis, hyperlipidemia, atrial fibrillation, and lifestyle) recognized by the American Heart Association, depression is claimed to be the strongest risk factor.

The Rotterdam study43 investigated the relationship between depression and atherosclerosis in elderly people aged 69 years old and older. The findings showed that the size of carotid plaques and the level of calcification of the coronal artery were significantly greater in individuals with a past history of depression, indicating that depression could be a risk for arteriosclerosis in the elderly.43


Findings from earlier brain imaging studies have demonstrated non-specific findings, such as ventricular enlargement, an increase in the ratio of ventricles to brain tissue, increased cerebrospinal fluid (CSF) volume, and enlargement of the cerebral sulci, in patients with major depression, especially the elderly.

Decreased gray matter volume and decreased blood flow have been reported in the subgenual prefrontal cortex of depressed patients (Drevets et al. 1997). Furthermore, decreased volume of the left caudate has been reported in elderly patients with depression (Greenwald et al. 1997).

In an MRI study of elderly patients with depression, hyperintensities on T2-weighted MRI were observed in three areas, namely periventricular hyperintensities (PVH), DWMH, and subcortical hyperintensities (SCH). Basal ganglia hyperintensities are more frequent with late-onset depression than in early onset depression and are related with a higher frequency of delirium after treatment with antidepressants or electroconvulsive therapy.


Diffusion tensor imaging (DTI) assesses the microscopic diffusion of water, allowing evaluation of the integrity and bundle coherence of brain white matter tracts.44 The direction of water diffusion is limited to alignment with the myelinated axons, and this tendency to diffuse in one direction as opposed to all others is termed anisotropy. Anisotropy is calculated from the three eigenvalues, namely diffusion along the principal direction (λ1) and radial diffusion (λ2 and λ3) perpendicular to axonal walls.45

Fractional anisotropy (FA) values range between 0 and 1 and refer to the degree of anisotropy of diffusion, where 0 indicates completely random motion and 1 in dicates maximal directional selectivity. Fractional anisotropy can be used to determine the structure of axonal cell membranes and myelin sheaths.46 A decreased tendency to diffuse along the principal direction (λ1) of the fiber, axial diffusivity (AD), suggests axonal loss or a reduction in bundle coherence; an increase in radial diffusivity (RD; the average of λ2 and λ3) suggests disrupted myelination.47 Mean diffusivity (MD), the direction-independent average diffusion of λ1, λ2, and λ3, is a measure of the average molecular motion, independent of tissue directionality. In white matter pathologies involving disruption of myelin sheaths, the reduction in FA is usually paralleled by an increase in MD.48 High FA values are observed in heavily myelinated tracts and a reduction in FA is observed in parallel with the progression of several degenerative white matter disorders involving axonal loss or myelin sheath destruction, such as multiple sclerosis, leukoaraiosis, dementia, and infections.

Reduced anisotropy has been reported in the frontal49 and temporal lobe50 of elderly depressive patients. The reduction in FA in these regions is correlated with the severity of the depression50 and poor responses to anti-depressant treatment.51,52

White matter microstructural integrity has been investigated in with non-geriatric depressive patients using tract-based spatial statistical DTI analysis.53 In that study, 29 non-geriatric patients with major depression (11 with melancholic and 18 with non-melancholic depression) and 39 healthy controls were evaluated. The authors reported that melancholic patients exhibited an average reduction of 7.8% FA over white matter regions associated with the limbic system, dorsolateral prefrontal cortex, thalamic projection fibers, corpus callosum, and other association fibers.53 These FA deficits were also associated with decreased AD and increased RD in these white matter regions, suggesting a pattern of decreased myelination or other degenerative changes.

In another DTI voxel-based analysis of 23 single-episode, medication-naïve, major depressive patients and 21 healthy controls, decreased AD and increased RD were found in the right superior longitudinal fasciculus and the right middle frontal and left parietal white matter in patients with major depression compared with the healthy controls.54 These results may indicate decreased myelination or a decrease in the number of normally myelinated fibers in these areas.


Elderly depression differs in many respects to depression in younger patients, including symptomatology, natural course, risk factors, and possible involvement of vascular factors. Recent findings from MRI and DTI analysis strongly indicate disrupted connectivity between the limbic area and the prefrontal cortex. Further studies may elucidate the underlying mechanisms of elderly depression.