IMPAIRED COGNITIVE FUNCTION is a common and disabling feature of depression, both during an acute depressive episode and in the remitted state. A number of studies in recent years have used functional magnetic resonance imaging (fMRI) to study brain activation associated with cognitive activity in patients with depression. Although these studies have provided conflicting results due to the fact that most studies have used different cognitive tasks,1–6 we reported previously that depressed patients showed significantly attenuated activations in the left prefrontal cortex (PFC) during a verbal fluency task, when compared to control subjects.7 In the present study we used the same activation paradigm to investigate changes in brain activation in remitted patients with major depression.
The aim of the present study was to investigate whether a functional abnormality in the left prefrontal cortex observed in patients with major depression performing a verbal fluency task is present after remission of depression. Functional magnetic resonance imaging was used to study changes in cerebral blood oxygenation in eight remitted patients with major depression and 10 healthy control subjects during a verbal fluency task. Compared to the control subjects, the patients had a reduced response in the left prefrontal cortex (middle frontal gyrus, Brodmann area 10). These findings suggest the presence of dysfunction in the left prefrontal cortex during remission in major depression.
The study participants were eight remitted patients with major depression (six men, two women, aged 35–55 years) and 10 healthy volunteers (eight men, two women, aged 34–57 years) with no history of neurological or psychiatric illness. All subjects were right-handed according to the Edinburgh Handedness Inventory. All of the patients were inpatients at Hiroshima University Medical Hospital, diagnosed by three experienced clinicians as having major depressive disorder, but no other major disorders according to DSM-IV criteria.8 Remission was defined as ≤7 on the 17-item Hamilton Rating Scale for Depression (HRSD).9 Their mean score on the HRSD was 5.3 ± 1.4. The duration of the remitted phase in the patients prior to the study was 8.4 ± 3.3 days. All were medicated as follows: four patients were taking 200 mg fluvoxamine, and the other four were taking either 50 mg fluvoxamine, 150 mg clomipramine, 225 mg amoxapine, or 20 mg paroxetine. Seven of the patients and all of the healthy volunteers participated in the previous study.7 The examinations were conducted under a protocol approved by the Ethics Committee of Hiroshima University School of Medicine. After complete description of the study to the subjects, written informed consent was obtained.
The verbal fluency activation paradigm, echo planar imaging acquisition, image processing, measurements of verbal fluency test performance, and data analysis were performed according to methods described previously.7
No significant differences were detected between the patients and the control subjects for verbal fluency performance (patients: mean, 21.8 ± 3.7 words; control subjects: mean, 25.8 ± 8.3 words; P = 0.220).
For both groups, verbal fluency task resulted in the significant activation of the left PFC (P < 0.001 uncorrected on the single voxel level and P < 0.05 corrected on the cluster level; Fig. 1a,b). The control group also had significant activation in the cingulate cortex and thalamus, while the depressed group did not (P < 0.001 uncorrected on the single voxel level and P < 0.05 corrected on the cluster level; Fig. 1a,b). A direct comparison on two-sample t-test at each voxel of the brain activation in the two groups showed that the control group had significantly greater activations than the patients in a small portion of the left PFC (middle frontal gyrus, Brodmann area 10; P < 0.001 uncorrected on the single voxel level, extent threshold of 10 voxels; Fig. 1c).
Within this verbal fluency neural network, brain activation in the left PFC remained impaired in patients in remission during the short observation period in spite of their clear clinical improvement. In our previous study there was no significant difference in this area of activation between depressive patients in recovery (defined as maintaining a score of ≤7 on the 17-item HRSD for >3 months) and healthy volunteers.10 This result suggests that the brain activity may take longer to return to a normal level than the observed mood improvement.
Results of recent studies suggest that functional neuroimaging is a more sensitive assay of cognitive processing than behavioral measures.11 Therefore, the neurophysiological problems identified in the present study may reflect subtle cognitive deficits of remitted patients. The present findings, however, are limited by the relatively small group size and potential medication effects. A second limitation is the short period of the remitted phase in the present patients. Although they were asymptomatic at examination, they fulfilled criteria for partial remission on DSM-IV-TR. Further longitudinal studies using larger numbers of unmedicated subjects are required to elucidate the neurophysiological abnormalities in major depression.
This study was partly supported by KAKENHI (20790841).