• frontal lobe;
  • late-life depression;
  • near-infrared spectroscopy;
  • regional cerebral blood volume;
  • temporal lobe;
  • vascular depression;
  • word fluency


  1. Top of page
  2. Abstract

Background:  Individuals with late-life depression can be divided into two categories, those with early and late-onset depression (EOD and LOD, respectively). It has been reported that LOD has more accentuated subcortical vascular lesions and frontal lobe dysfunction (hypofrontality). The aim of the present study was to examine whether LOD exhibits more prominent hypofrontality than EOD during performance of the word fluency task (WFT) under multichannel near-infrared spectroscopy (NIRS), a newly developed non-invasive functional neuroimaging technique.

Methods:  Eleven patients with EOD, 12 patients with LOD, and 13 healthy controls participated in the study. Clinical symptoms of depression were equivalent in the EOD and LOD groups. Brain magnetic resonance imaging demonstrated more robust subcortical vascular changes in LOD than EOD. The NIRS images were obtained using an ETG-4000, 52-channel NIRS system (Hitachi Medical, Tokyo, Japan). Mean changes in oxy-hemoglobin (oxy-Hb) were evaluated while the participants performed the phonemic WFT.

Results:  Healthy controls exhibited clear increases in oxy-Hb bilaterally throughout the medial to the lateral frontal cortices and the superior temporal areas during the WFT. In contrast, increases in oxy-Hb were mildly attenuated in EOD and severely attenuated in LOD in most channels. Subsequent analyses revealed that increases in oxy-Hb in LOD during the WFT was significantly poorer than in EOD in the left lateral portion of the cortex, including the dorsolateral prefrontal and the superior temporal areas. In addition, significant negative correlations were obtained between the age of onset and oxy-Hb, as well as between subcortical vascular changes and oxy-Hb in the lateral channels. These findings suggest that the higher the age of onset of depression, and the more prominent the vascular lesions, the greater the attenuation in lateral frontal and temporal activation, as indicated by NIRS.

Conclusions:  Multichannel NIRS is useful for demonstrating attenuated functional activation in the left lateral prefrontal and temporal areas in LOD and, thus, for differentiating between LOD and EOD. The NIRS findings observed may have useful clinical implications for treatment-resistant LOD.


  1. Top of page
  2. Abstract

Late-life depression is one of the most common psychiatric disorders in the elderly population. Late-life depression is a heterogeneous entity and may include different etiologic mechanisms. Individuals with late-life depression are usually divided into two categories, namely those with either early or late-onset depression (EOD and LOD, respectively).1 Although some researchers consider EOD and LOD to be phenomenologically indistinguishable,2 others emphasize clinical characteristics that may discriminate between the two.3 Patients with LOD tend to have a poorer prognosis, with resistance to treatment, higher suicidality, and a higher rate of cognitive decline or comorbid dementia compared with patients with EOD.4 The cognitive impairment observed in LOD has been regarded as ‘depression–executive dysfunction syndrome’, characterized by psychomotor retardation, reduced interest in activities, and a poor response to classical antidepressants.5 A recent meta-analysis of studies comparing neuropsychological test results between EOD and LOD revealed that patients with LOD exhibited greater reductions in processing speed and executive function than those with EOD.6 Although patients with EOD also exhibited cognitive impairment, no reduction in function specific to EOD was found. In contrast with LOD, EOD tends to be associated with genetic or familial factors, and patients with it often experience recurrence. Associations with life events and psychosocial factors are also thought to exist in EOD.7

Morphological differences in the brain have also been suggested to distinguish between EOD and LOD. These include enlarged third and lateral ventricles,8 reduced regional hippocampal volume,9–11 and callosal thinning in the genu and splenium12 in LOD. Among various morphological findings distinguishing LOD from EOD, vascular lesions observed in the subcortical structures are the most consistent and robust abnormality reported in LOD.8,13,14 Herrmann et al.15 conducted a systematic meta-analysis and concluded that LOD is characterized by more frequent and intense white matter changes, with an odds ratio over 4 for LOD compared with EOD. Greater duration of depressive symptoms, signs, and treatment did not appear to have a measurable impact on white matter signals on magnetic resonance imaging (MRI). Such outstanding subcortical lesions in LOD are consistent with the ‘vascular’ hypothesis of late-life depression. From a pathophysiological perspective, LOD overlaps the concept of ‘vascular depression’.16,17

Recent functional neuroimaging studies,18,19 including positron emission tomography (PET), single photon emission computed tomography (SPECT), and functional MRI (fMRI), have demonstrated consistent frontal lobe dysfunction (hypofrontality) in major depression. Such hypofrontality is more accentuated in LOD. Most research has focused on dysfunction of mood-related circuits involving multiple regions of prefrontal cortex and limbic structures. The regional cerebral blood flow (rCBF) of patients with unmedicated LOD exhibited significant decreases bilaterally in the anterior frontal regions.20,21 Results of previous studies show some inconsistencies: some studies have reported right–left perfusion differences, with decreases in rCBF in the anterior frontal regions more pronounced on the left side.20,22

More recently, near-infrared spectroscopy (NIRS), a non-invasive functional neuroimaging technique, has been found particularly useful for detecting hypofrontality in major depression. The NIRS technique measures regional cerebral blood volume (rCBV), as indicated by changes in concentrations of oxy-hemoglobin (oxy-Hb) and deoxy-hemoglobin (deoxy-Hb). The three major advantages of NIRS are: (i) complete non-invasiveness, enabling repeated measurements; (ii) high time-resolution of 0.1 s, enabling detailed clarification of temporal changes in rCBV; and (iii) the portability and compactness of the apparatus required, enabling measurements to be taken under natural conditions with subjects sitting on a comfortable chair. The disadvantages of NIRS include the fact that it measures Hb concentrations: (i) only as relative changes and not as absolute values; (ii) only in the cortex immediately beneath the probes and not in deeper brain structures; (iii) with low spatial resolution; and (iv) not only in the brain, but also in other surface structures, such as the skin and skull.23 The advantages and disadvantages noted above make NIRS particularly suitable for examining psychiatric patients repeatedly in the outpatient setting. Using NIRS, several studies have demonstrated clear hypofrontality in major depression during engagement in cognitive tasks, among which the phonemic word fluency task (WFT) has been used most frequently.23–26

Matsuo et al.27 examined LOD patients in remission using NIRS. Activation of the prefrontal cortex during a cognitive task was significantly less in patients compared with controls, although task performance did not differ significantly between the two groups. The LOD patients in the study of Matsuo et al.27 also exhibited a tendency towards a negative correlation between a reduction of prefrontal activation and the severity of periventricular hyperintensity (PVH). These authors concluded that prefrontal microvascular dysregulation, as demonstrated by NIRS, plays a role in the pathophysiology of functional hypofrontality in LOD. Because Matsuo et al.27 only examined LOD and not EOD patients, the question addressed in the present study was whether NIRS can reveal any differences in the pattern of brain dysfunction between patients with EOD and those with LOD. On the one hand, because the results of a number of MRI and SPECT studies have suggested the existence of subcortical vascular lesions in LOD, it is possible that NIRS will reveal more marked findings in brain pathology or hypofrontality in LOD than in EOD. Conversely, it is possible that patients with EOD will exhibit more significant functional abnormality in the brain given their repetitive episodes, long duration of disease, and long period of exposure to antidepressants.


  1. Top of page
  2. Abstract


Twenty-three patients with late-life depression were recruited at Showa University Hospital, Tokyo. All subjects were over 65 years of age and met the criteria for major depressive disorder or dysthymia according to DSM-IV.28 Eleven (mean age = 68.4 years) and 12 individuals (mean age = 70.2 years) were classified as EOD and LOD groups, respectively, based on an age of onset before or after 60 years (Table 1). The cut-off criteria of age of onset was set as 60 years based on the majority of previous publications;3,13,29 this is partly because of the increased ratio of vascular lesions observed in patients with depression over the age of 60 years. There were no differences in mean age or years of education between the two groups. Clinical symptoms of depression evaluated using the Hamilton Rating Scale for Depression (HRSD) were equivalent in the EOD and LOD groups. Individuals scoring below 24 on the Mini-Mental State Examination were excluded from the study in order to avoid concomitant dementia. All patients were taking one prescribed antidepressant at the time of the study. Medications included paroxetine (n = 6), milnacipran (n = 5), and tricyclic antidepressants (n = 12). The mean doses of these drugs were 68.8, 75, and 71.3 mg/day, respectively, as imipramine equivalent.

Table 1.  Demographic data for the two patient groups and healthy controls, and magnetic resonance imaging findings for the patient groups
 EODLODHealthy controls
  • Data are the mean ± SD. *P < 0.05 compared with the early onset depression (EOD) group.

  • The dose of antidepressants was calculated as imipramine equivalent dose (mg).

  • LOD, late-onset depression; HRSD, Hamilton Rating Scale for Depression (17 items); PVH, periventricular hyperintensity; DWMH, deep white matter hyperintensity; SCG, subcortical gray matter hyperintensity; Total, PVH + DWMH + SCG.

Number (male/female)2/93/98/5
Age (years)68.4 ± 5.670.2 ± 1.970.3 ± 4.4
Age at onset (years)53.4 ± 7.469.2 ± 1.6* 
Education (years)11.9 ± 2.312.4 ± 3.1 
Number of episodes2.5 ± 0.41.0 ± 0.0* 
Antidepressant dose77.3 ± 61.566.4 ± 26.5 
Family history of mood disorders (+/−)3/81/11 
HRSD15.4 ± 6.319.0 ± 6.8 
PVH0.1 ± 0.30.8 ± 0.8* 
DWMH0.7 ± 0.92.0 ± 0.6* 
SCG0.2 ± 0.40.6 ± 0.9 
Total1.0 ± 1.23.3 ± 1.8* 

Thirteen age- and education-matched healthy control subjects were also included in the study (mean age 70.3 years). The subjects had no history of any psychiatric disorder, including schizophrenia, mood disorder, seizures, and alcohol or other substance abuse. Individuals with neurological disorders, including cerebrovascular disease, neurodegenerative disease, and traumatic brain injury, were also excluded from the study. All patients and healthy controls were right handed. The demographic data for the EOD, LOD, and healthy control groups are given in Table 1.

The study was approved by the Ethics Committee of Showa University. All participants were informed of the purpose of the study and all provided written agreement to participate in the study.

Evaluation of subcortical vascular changes

All patients underwent brain MRI for evaluation of subcortical vascular changes. Periventricular hyperintensity (PVH), deep white matter hyperintensity (DWMH), and subcortical gray matter hyperintensity (SCG) were evaluated using the modified Fazekas criteria.13,17,30 Four-point scales were used for PVH (0, absent; 1, caps; 2, smooth halo; 3, irregular and extending into deep white matter), DWMH (0, absent; 1, punctate foci; 2, beginning of confluency of foci; 3, large confluent areas), and SCG (0, absent; 1, punctate; 2, multipunctate; 3, diffuse). Each patient's score was the sum of the scores for PVH, DWMH, and SCG (PVH + DWMH + SCG = Total).

The LOD group had significantly higher PVH, DWMH, and Total scores than the EOD group (Table 1), suggesting more prominent subcortical vascular changes in the former.


Participants were asked to sit still on a chair and to keep their eyes open while performing a task. A phonemic WFT was used in the present study. The experiment consisted of three phases: (i) a pretask baseline (30 s); (ii) activation task (WFT; 60 s); and (iii) post-task baseline (60 s). Before the start of NIRS measurements, participants were instructed in detail to generate as many words as possible during the WFT period. The phonemic WFT was introduced during the 60 s activation phase, during which initial syllables assigned were changed every 20 s (/a/,/ka/, and /sa/ in this order) in order to avoid possible empty responses and to maintain the concentration of participants on word generation. The number of correct words generated was determined as the task performance for each subject. In the pre- and post-task baseline periods, the participants were instructed to repeat a train of Japanese vowel syllables ‘/a/, /i/, /u/, /e/, /o/’. This automatic repetition of vowel production was adopted as a baseline task instead of empty silence, in order to parallel possible effects of vocalization in the baseline and activation periods. This procedure has been widely used for the Japanese WFT paradigm in NIRS experiments.23–26

Aquisition of NIRS images

The NIRS images were obtained using an ETG-4000, 52-channel NIRS system (Hitachi Medico, Tokyo, Japan). Changes in oxy-Hb, deoxy-Hb, and total Hb were measured at two different wavelengths of near-infrared light (695 and 830 nm) to compute oxy-Hb and deoxy-Hb concentrations. The probes were placed on the participants' frontal and temporal regions bilaterally. The lowest probes were positioned along the Fp1–Fp2 line according to the international 10/20 system used in electroencephalography.

The absorption of near-infrared light was measured with a time resolution of 0.1 s. The data obtained were analyzed using the ‘integral mode’ (i.e. the average of the pretask baseline determined as the mean across the last 10 s of the 30 s pretask period and the post-task baseline determined as the mean across the last 5 s of the 60 s post-task period, following Suto et al.23 and Ito et al.31). Linear fitting was applied to the data between these two baselines. Moving average methods were used to exclude short-term motion artefacts in the data analyzed.


Statistical analyses were performed using spss 11.5J for Windows (SPSS-J, Tokyo, Japan). The behavioral data (correct responses on the WFT) were analyzed using one-way analysis of variance (anova) with group as a factor (EOD, LOD, and healthy control).

For NIRS rCBV data, we focused on oxy-Hb in the present study because correlations with CBF have been shown to be stronger for oxy-Hb than for deoxy-Hb.32 In an animal study using a perfused brain rat model, oxy-Hb was also demonstrated to be the most sensitive marker of changes in CBF among oxy-Hb, deoxy-Hb, and total-Hb.33 The relative change in oxy-Hb for each channel, calculated as the mean level of oxy-Hb during the WFT activation period using the integral mode, was examined by one-way anova with group as a factor. The data were subsequently subjected to post hoc analyses using Tukey's multiple comparison test. Channels with marked artefactual contamination were excluded from personal data and the data were considered defective. In addition, Pearson product–moment correlations were computed between oxy-Hb, age of onset, age at the time of evaluation, and the clinical and behavioral variables (HRSD score and the number of words produced, respectively).


  1. Top of page
  2. Abstract

Behavioral data

There were no significant differences among the EOD, LOD, and healthy control groups in the number of correct responses on the WFT during NIRS (F2,30 = 0.14; P > 0.05; EOD = 12.5, sd = 2.8; LOD = 12.1, sd = 5.5; healthy control = 13.0, sd = 3.8).


Figure 1 shows sequential topographic maps of oxy-Hb measured at 20 s (left panel) and 50 s (right panel) after the beginning of the activation task (WFT) for the healthy control, EOD, and LOD groups. The healthy controls demonstrated clear enhancement of oxy-Hb bilaterally in the medial to lateral frontal cortices and the superior temporal areas 20 s after the beginning of the WFT; this change was more pronounced after 50 s. This increase in oxy-Hb was mildly attenuated in EOD and severely attenuated in LOD. The LOD patients exhibited only limited increases in oxy-Hb during the WFT.


Figure 1. Topographic maps of oxy-hemoglobin measured at 20 s (left panel) and 50 s (right panel) after the beginning of the activation task (word fluency test) in the healthy control, early onset depression (EOD), and late-onset depression (LOD) groups. The red, yellow, and green zones in the topographic maps indicate most activated, intermittent, and least activated areas, respectively.

Download figure to PowerPoint

Table 2 gives the results of one-way anova for each channel. A significant group effect was observed in a number of channels, suggesting differences between the patient groups and controls. For those channels in which group effect was significant, subsequent analyses using Tukey's multiple comparison tests were performed to compare EOD and LOD. Multiple comparison tests revealed that the increases in oxy-Hb in LOD during the WFT were significantly poorer than those in EOD in four channels (channels (CH) 8, 18, 41, 42; P < 0.05). Figure 2 shows P value significance maps of EOD versus LOD comparisons. Attenuation of increases in oxy-Hb in LOD was noted in the left lateral portion of the cortex, including the dorsolateral prefrontal (around CH8 and CH18) and the left superior temporal (around CH41 and CH42) areas.

Table 2.  Results of one-way analyses of variance with group as a factor (early onset depression, late-onset depression, and healthy controls) for 52 channels
ChannelF valueP valueTukey's testChannelF valueP valueTukey's test
  1. E, early onset depression; L, late-onset depression; C, healthy controls.

17.790.002E, L < C273.090.059 
25.640.008E, L < C2813.110.000E, L < C
38.270.001E, L < C2914.420.000E, L < C
47.740.002E, L < C308.470.001L < C
51.710.197 316.140.005L < C
61.370.269 3212.580.000E, L < C
75.490.009L < C3314.590.000E, L < C
87.630.002L < E, C343.080.059 
95.230.011L < C359.890.000E, L < C
100.980.387 367.700.002E,L < C
114.170.024L < C370.550.586 
123.910.030L < C389.390.001L < C
137.570.002E, L < C3914.200.000E, L < C
1413.970.000E, L < C4020.180.000E, L < C
156.100.006L < C4115.510.000L < E < C
160.830.443 4211.370.000L < E, C
175.720.007L < C437.130.003E, L < C
1810.490.000L < E, C442.860.071 
1917.550.000E, L < C453.420.044L < C
206.920.003L < C467.850.002E, L < C
211.010.375 473.740.034L < C
226.330.005L < C480.830.447 
231.450.249 496.490.004L < C
246.810.003L < C5010.540.000E, L < C
257.070.003E, L < C514.340.021L < C
264.290.022L < C527.360.002L < C

Figure 2. The upper panel shows a significance map for comparison of the late-onset depression (LOD) and early onset depression (EOD) groups. The numbered dots indicate 52 channels. The bars on the right indicate significance level in gradient color. The lower panel shows four channels projected onto the three-dimensional brain. Changes in oxy-hemoglobin concentration over time in two representative channels (CH41 and CH42) are also shown (thin line, EOD; thick line, LOD).

Download figure to PowerPoint


For the patient groups (EOD + LOD), significant negative correlations were obtained between the age of onset and oxy-Hb in six channels (CH14 r = −0.51, P < 0.01; CH18 r = −0.47, P < 0.02; CH20 r = −0.42, P < 0.04; CH30 r=−0.42, P < 0.04; CH33 r = −0.42, P < 0.04; and CH41 r = −0.44, P < 0.03), suggesting that the higher the age of onset of depression, the greater the attenuation of frontal and temporal activation, as indicated by NIRS. In contrast, there were no significant correlations between age at the time of evaluation and oxy-Hb. There were negative and positive correlations between scores of the WFT and oxy-Hb for five channels (CH25 r = −0.61, P < 0.004; CH28 r = −0.54, P < 0.01; CH32 r = 0.55, P < 0.01; CH42 r = 0.54, P < 0.01; and CH52 r = 0.48, P < 0.03). The HRSD score exhibited negative correlations with oxy-Hb for two channels (CH17 r = −0.44, P < 0.04; and CH27 r = −0.51, P < 0.01). In addition, negative correlations were observed between subcortical vascular changes (Total) and oxy-Hb for three channels (CH30 r = −0.54, P < 0.007; CH40 r = −0.53, P < 0.01; and CH50 r = −0.43, P < 0.04).


  1. Top of page
  2. Abstract

In the present study, healthy controls exhibited clear increases in oxy-Hb in most of the channels monitored during the WFT. In contrast, increases in oxy-Hb were attenuated in patients with late-life depression as a group bilaterally in the frontal and temporal regions. These findings are consistent with those of previous studies of activation in major depression using NIRS.23–26 Neuropsychological studies of patients with lesions in the left prefrontal areas have revealed impairment of word list generation, suggesting the importance of the prefrontal areas for WFT.34 Recent fMRI studies have demonstrated robust activation by the WFT in the left Brodmann's area (BA) 44/45, as well as adjacent BA46 and BA47, together with the premotor cortex (BA6) and insula.35 Previous studies using NIRS23–26 demonstrated bilateral frontal activation during the WFT in healthy controls, suggesting that not only the left, but also the right frontal lobe may play crucial roles in voluntary speech intention and/or attentional resources required for performing the WFT. Interconnected areas such as these are believed to participate in a functional language network subserving word output.35 The present study clearly demonstrated attenuation of increases in oxy-Hb in the frontal cortex in late-life depression, suggesting hypofrontality.

In addition, on multiple comparisons tests, LOD demonstrated more robust attenuation of increases in oxy-Hb than EOD in the left dorsolateral prefrontal cortex and the left superior temporal gyrus. In addition, significant negative correlations between age of onset and oxy-Hb were obtained for the channels placed in the lateral portion of the cortex. These findings show that patients with LOD are more likely to display cortical dysfunction in the lateral prefrontal to the superior temporal areas than those with EOD. Because previous studies revealed that attenuation of increases in oxy-Hb in major depression is most clearly and commonly found in the anterior medial portion of frontal cortex,23–26 it has been speculated that EOD and LOD exhibit indistinguishable impairment of medial frontal functions. Higher age of onset, rather than longer duration of illness or a large number of episodes, is linked with functional impairment in the lateral prefrontal to temporal areas of the cortex.

Because patients with LOD presented with significantly more pronounced subcortical vascular changes on MRI, the ‘vascular’ hypothesis may account for the results we obtained. This hypothesis is strongly supported by the findings of the present study, because vascular lesions as measured by MRI had a negative impact on oxy-Hb in the left prefrontal cortex (CH30, 40, and 50). The present findings are consistent with those of Matsuo et al.,27 in that in the latter study LOD patients in remission exhibited less activation in the prefrontal cortex than normal controls and negative correlations were obtained between reduced prefrontal activation and the severity of PVH. Prefrontal microvascular dysregulation as demonstrated by NIRS appears to play a role in the pathophysiology of functional hypofrontality in LOD. Similar hypofrontality in LOD was observed by Oda et al.,36 who found a negative correlation between subcortical (including basal ganglia and thalamus) vascular changes and decreases in rCBF, as determined by SPECT, in the prefrontal and temporal lobes among patients with LOD who exhibited white matter changes. Functional abnormalities in neuron networks linking subcortical and cortical areas are thought to exist in patients with LOD. Heterogeneous mechanisms have been hypothesized regarding the pathophysiology of ‘vascular depression’:16,17,37 subcortical DWMH may lead to immediate dysfunction of directly connected portions of prefrontal cortex,38 or breakdown of circuitry between the prefrontal cortex and the basal ganglia, specifically the cortico-striato-pallido-thalamo-cortical pathway, may result in dysregulation of mood and emotion.39 Either hypothesis may account for the frontal cortical dysfunction revealed by the present and previous27,36 studies.

Of note, our findings provide some evidence of temporal lobe dysfunction in LOD in addition to hypofrontality. Such temporal lobe dysfunction in LOD was reported by Ebmeier et al.,40,41 who investigated rCBF using SPECT. They found that patients with LOD exhibited decreased rCBF in the temporal areas of the cortex compared with those with EOD and emphasized the temporal lobe pathology of LOD. They also found a negative correlation between PVH in MRI and rCBF in SPECT bilaterally in the temporal lobe. Similarly, Motomura et al.42 used electroencephalograms to determine brain dysfunction in LOD and observed more frequent temporal slow waves in subjects with LOD than in normal controls. The temporal lobe dysfunction in LOD observed in the present study is consistent with these previous findings obtained using other functional imaging techniques. As suggested by other researchers, we believe that the temporal abnormalities observed in LOD are secondary to pathological small vascular lesions. It is of interest to determine whether dysfunction in the superior temporal areas is associated with medial temporal lobe (hippocampal) atrophy, which has been reported recently in LOD.9–11 However, NIRS is limited in that it can reliably measure cortical function, but not that of deep structures. To explore deep structures, including the hippocampus and limbic system, other functional imaging techniques, such as PET and fMRI must be used.

In the present study, we found that NIRS is useful to easily demonstrate fronto-temporal dysfunction in LOD, which may have some clinical implications. As suggested by Alexopoulos et al.,5 LOD individuals with fronto-temporal dysfunction may benefit from behavioral interventions, including problem-solving therapy. It is also possible that functional abnormalities in the fronto-temporal areas are suggestive of the prognosis of LOD. Baldwin et al.43 found that resistance to treatment in LOD is associated with impaired executive function, as measured using the WFT and the Stroop test. In addition, Brassen et al.44 reported considerable similarity between cognitively impaired depressed patients and Alzheimer's disease patients. Accordingly, extension of functional impairment to the temporal lobes in LOD may be a preclinical sign of developing dementia. Further research on LOD is warranted to investigate longitudinal changes with treatment. NIRS has the strong advantage of ready availability, is easily performed, and is thus useful for repeated examination in studies of prognosis.

The limitations of the present study include its relatively small sample size and the possibility of effects of the medications used by patients on NIRS findings, because all patients participating in the study were on antidepressants. Antidepressants have been reported to affect microvascular osmolarity, CBF, and brain metabolism.45 However, studies of the effects of antidepressants on brain function have yielded mixed results.46,47 In the present study, EOD and LOD patients were taking equivalent doses of antidepressants and thus medication effects should have been minimal. However, future studies with a large number of drug-naïve patients may produce findings enabling more decisive discrimination between LOD and EOD. Another limitation of the present study is that the mean age of onset in the EOD group was relatively high (53.4 years). Holroyd and Duryee48 cautioned that even patients with EOD can exhibit vascular changes. Thus, the findings of the present study may have been confounded by subcortical vascular changes in EOD. Future studies should recruit EOD patients with a younger age of onset and minimal vascular changes.


  1. Top of page
  2. Abstract

MM was supported by a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan and a Showa University Grant-in-Aid for Innovative Collaborative Research Projects. The authors thank Mr Shingo Kawasaki (Hitachi Corporation) for his help in preparing the figures.


  1. Top of page
  2. Abstract
  • 1
    Conwell Y, Nelson JC, Kim KM, Mazure CM. Depression in late life: Age of onset as marker of a subtype. J Affect Disord 1989; 17: 189195.
  • 2
    Brodaty H, Luscombe G, Parker G et al. Early and late onset depression in old age: Different aetiologies, same phenomenology. J Affect Disord 2001; 66: 225236.
  • 3
    Heun R, Kockler M, Papassotiropoulos A. Distinction of early- and late-onset depression in the elderly by their lifetime symptomatology. Int J Geriatr Psychiatry 2000; 15: 11381142.
  • 4
    Baldwin RC, Gallagley A, Gourlay M, Jackson A, Burns A. Prognosis of late life depression: A three-year cohort study of outcome and prognosis. Int J Geriatr Psychiatry 2006; 21: 5763.
  • 5
    Alexopoulos GS. Role of executive function in late-life depression. J Clin Psychiatry 2003; 64: 1823.
  • 6
    Herrmann LL, Goodwin GM, Ebmeier KP. The cognitive neuropsychology of depression in the elderly. Psychol Med 2007; 37: 16931702.
  • 7
    Grace J, O'Brien JT. Association of life events and psychosocial factors with early but not late onset depression in the elderly: Implications for possible differences in aetiology. Int J Geriatr Psychiatry 2003; 18: 473478.
  • 8
    Dahabra S, Ashton CH, Bahrainian M et al. Structural and functional abnormalities in elderly patients clinically recovered from early- and late-onset depression. Biol Psychiatry 1998; 44: 3446.
  • 9
    Ballmaier M, Narr KL, Toga AW et al. Hippocampal morphology and distinguishing late-onset from early-onset elderly depression. Am J Psychiatry 2008; 165: 229237.
  • 10
    Janssen J, Hulshoff Pol HE, De Leeuw FE et al. Hippocampal volume and subcortical white matter lesions in late life depression: Comparison of early and late onset depression. J Neurol Neurosurg Psychiatry 2007; 78: 638640.
  • 11
    Lloyd AJ, Ferrier IN, Barber R, Gholkar A, Young AH, O'Brien JT. Hippocampal volume change in depression: Late- and early-onset illness compared. Br J Psychiatry 2004; 184: 488495.
  • 12
    Ballmaier M, Kumar A, Elderkin-Thompson V et al. Mapping callosal morphology in early- and late-onset elderly depression: An index of distinct changes in cortical connectivity. Neuropsychopharmacology 2007; 33: Epub 22 August 2007; doi: 10.1038/sj.npp.1301538
  • 13
    Greenwald BS, Kramer-Ginsberg E, Krishnan RR, Ashtari M, Aupperle PM, Patel M. MRI signal hyperintensities in geriatric depression. Am J Psychiatry 1996; 153: 12121215.
  • 14
    Krishnan KR, Goli V, Ellinwood EH, France RD, Blazer DG, Nemeroff CB. Leukoencephalopathy in patients diagnosed as major depressive. Biol Psychiatry 1998; 23: 519522.
  • 15
    Herrmann LL, Lemasurier M, Ebmeier KP. White matter hyperintensities in late life depression: A systematic review. J Neurol Neurosurg Psychiatry 2007; 79: Epub 23 August 2007; doi: 10.1136/jnnp.2007.124651
  • 16
    Alexopoulos GS, Meyers BS, Young RC, Campbell S, Silbersweig D, Charlson M. ‘Vascular depression’ hypothesis. Arch Gen Psychiatry 1997; 54: 915922.
  • 17
    Krishnan KR, Hays JC, Blazer DG. MRI-defined vascular depression. Am J Psychiatry 1997; 154: 497501.
  • 18
    Baxter Jr LR, Schwartz JM, Phelps ME et al. Reduction of prefrontal cortex glucose metabolism common to three types of depression. Arch Gen Psychiatry 1989; 46: 243250.
  • 19
    Mayberg HS, Liotti M, Brannan SK et al. Reciprocal limbic–cortical function and negative mood: Converging PET findings in depression and normal sadness. Am J Psychiatry 1999; 156: 675682.
  • 20
    Navarro V, Gasto C, Lomena F, Mateos JJ, Marcos T. Frontal cerebral perfusion dysfunction in elderly late-onset major depression assessed by 99mTC-HMPAO SPECT. Neuroimage 2001; 14: 202205.
  • 21
    Ishizaki J, Yamamoto H, Takahashi T et al. Changes in regional cerebral blood flow following antidepressant treatment in late-life depression. Int J Geriatr Psychiatry 2008; 23: Epub 24 January 2008; doi: 10.1002/gps.1980
  • 22
    Kimura M, Shimoda K, Mizumura S et al. Regional cerebral blood flow in vascular depression assessed by 123I-IMP SPECT. J Nippon Med School 2003; 70: 321326.
  • 23
    Suto T, Fukuda M, Ito M, Uehara T, Mikuni M. Multichannel near-infrared spectroscopy in depression and schizophrenia: Cognitive brain activation study. Biol Psychiatry 2004; 55: 501511.
  • 24
    Matsuo K, Kato T, Fukuda M, Kato N. Alteration of hemoglobin oxygenation in the frontal region in elderly depressed patients as measured by near-infrared spectroscopy. J Neuropsychiatry Clin Neurosci 2000; 12: 465471.
  • 25
    Matsuo K, Kato N, Kato T. Decreased cerebral haemodynamic response to cognitive and physiological tasks in mood disorders as shown by near-infrared spectroscopy. Psychol Med 2002; 32: 10291037.
  • 26
    Ohta H, Yamagata B, Tomioka H et al. Hypofrontality in panic disorder and major depressive disorder assessed by multi-channel near-infrared spectroscopy. Depress Anxiety 2008 (in press).
  • 27
    Matsuo K, Onodera Y, Hamamoto T, Muraki K, Kato N, Kato T. Hypofrontality and microvascular dysregulation in remitted late-onset depression assessed by functional near-infrared spectroscopy. Neuroimage 2005; 26: 234242.
  • 28
    American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th edn. Washington, DC: American Psychiatric Association, 1994.
  • 29
    Salloway S, Malloy P, Kohn R et al. MRI and neuropsychological differences in early- and late-onset geriatric depression. Neurology 1996; 46: 15671574.
  • 30
    Fazekas F, Chawluk JB, Alavi A, Hurtig HI, Zimmerman RA. MR signal abnormalities at 1.5 T in Alzheimer's dementia and normal aging. Am J Roentgenol 1987; 149: 351356.
  • 31
    Ito M, Fukuda M, Suto T, Uehara T, Mikuni M. Increased and decreased cortical reactivities in novelty seeking and persistence: A multichannel near-infrared spectroscopy study in healthy subjects. Neuropsychobiology 2005; 52: 4554.
  • 32
    Strangman G, Culver JP, Thompson JH, Boas DA. A quantitative comparison of simultaneous BOLD fMRI and NIRS recordings during functional brain activation. Neuroimage 2002; 17: 719731.
  • 33
    Hoshi Y, Kobayashi M, Tamura M. Interpretation of nearinfrared spectroscopy signals: A study with a newly developed perfused rat brain model. J Appl Physiol 2001; 90: 16571662.
  • 34
    Vilkki J, Holst P. Speed and flexibility on word fluency tasks after focal brain lesions. Neuropsychologia 1994; 32: 12571262.
  • 35
    Perani D, Cappa SF, Tettamanti M et al. A fMRI study of word retrieval in aphasia. Brain Lang 2003; 85: 357368.
  • 36
    Oda K, Okubo Y, Ishida R et al. Regional cerebral blood flow in depressed patients with white matter magnetic resonance hyperintensity. Biol Psychiatry 2003; 53: 150156.
  • 37
    Attig E, Capon A, Demeurisse G, Verhas M. Remote effect of deep-seated vascular brain lesions on cerebral blood flow. Stroke 1990; 21: 15551561.
  • 38
    Thomas AJ, O'Brien JT, Davis S et al. Ischemic basis for deep white matter hyperintensities in major depression: A neuropathological study. Arch Gen Psychiatry 2002; 59: 785792.
  • 39
    Fujikawa T, Yamawaki S, Touhouda Y. Incidence of silent cerebral infarction in patients with major depression. Stroke 1993; 24: 16311634.
  • 40
    Ebmeier KP, Prentice N, Ryman A. Temporal lobe abnormalities in dementia and depression. A study using high resolution single photon emission tomography and magnetic resonance imaging. J Neurol Neurosurg Psychiatry 1997; 63: 597604.
  • 41
    Ebmeier KP, Glabus MF, Prentice N et al. A voxel-based analysis of cerebral perfusion in dementia and depression of old age. Neuroimage 1998; 7: 199208.
  • 42
    Motomura E, Inui K, Nakase S, Hamanaka K, Okazaki Y. Late-onset depression: Can EEG abnormalities help in clinical subtyping? J Affect Disord 2002; 68: 7379.
  • 43
    Baldwin R, Jeffries S, Jackson A et al. Treatment response in late-onset depression: Relationship to neuropsychological, neuroradiological and vascular risk factors. Psychol Med 2004; 34: 125136.
  • 44
    Brassen S, Braus DF, Weber-Fahr W, Tost H, Moritz S, Adler G. Late-onset depression with mild cognitive deficits: Electrophysiological evidences for a preclinical dementia syndrome. Dement Geriatr Cogn Disord 2004; 18: 271277.
  • 45
    Navarro V, Gasto C, Lomena F et al. Frontal cerebral perfusion after antidepressant drug treatment versus ECT in elderly patients with major depression: A 12-month follow-up control study. J Clin Psychiatry 2004; 65: 656661.
  • 46
    Nebes RD, Pollock BG, Houck PR et al. Persistence of cognitive impairment in geriatric patients following antidepressant treatment: A randomized, double-blind clinical trial with nortriptyline and paroxetine. J Psychiatr Res 2003; 37: 99108.
  • 47
    Trichard C, Martinot JL, Alagille M et al. Time course of prefrontal lobe dysfunction in severely depressed in-patients: A longitudinal neuropsychological study. Psychol Med 1995; 25: 7985.
  • 48
    Holroyd S, Duryee JJ. Differences in geriatric psychiatry outpatients with early- vs late-onset depression. Int J Geriatr Psychiatry 1997; 12: 11001106.