fMRI Correlates of White Matter Hyperintensities in Late-Life Depression
2011; American Psychiatric Association; Volume: 168; Issue: 10 Linguagem: Inglês
10.1176/appi.ajp.2011.10060853
ISSN1535-7228
AutoresHoward Aizenstein, Carmen Andreescu, Kathryn Edelman, Jennifer L. Cochran, Julie C. Price, Meryl A. Butters, Jordan F. Karp, Meenal J. Patel, Charles F. Reynolds,
Tópico(s)Advanced MRI Techniques and Applications
ResumoBack to table of contents Previous article Next article New ResearchFull AccessfMRI Correlates of White Matter Hyperintensities in Late-Life DepressionHoward J. Aizenstein, M.D., Ph.D., Carmen Andreescu, M.D., Kathryn L. Edelman, B.A., Jennifer L. Cochran, Ph.D., Julie Price, Ph.D., Meryl A. Butters, Ph.D., Jordan Karp, M.D., Meenal Patel, B.S., and Charles F. Reynolds III, M.D.Howard J. AizensteinFrom the Departments of Psychiatry, Bioengineering, Neurology, Neuroscience, and Radiology at the University of Pittsburgh, Pittsburgh., M.D., Ph.D., Carmen AndreescuFrom the Departments of Psychiatry, Bioengineering, Neurology, Neuroscience, and Radiology at the University of Pittsburgh, Pittsburgh., M.D., Kathryn L. EdelmanFrom the Departments of Psychiatry, Bioengineering, Neurology, Neuroscience, and Radiology at the University of Pittsburgh, Pittsburgh., B.A., Jennifer L. CochranFrom the Departments of Psychiatry, Bioengineering, Neurology, Neuroscience, and Radiology at the University of Pittsburgh, Pittsburgh., Ph.D., Julie PriceFrom the Departments of Psychiatry, Bioengineering, Neurology, Neuroscience, and Radiology at the University of Pittsburgh, Pittsburgh., Ph.D., Meryl A. ButtersFrom the Departments of Psychiatry, Bioengineering, Neurology, Neuroscience, and Radiology at the University of Pittsburgh, Pittsburgh., Ph.D., Jordan KarpFrom the Departments of Psychiatry, Bioengineering, Neurology, Neuroscience, and Radiology at the University of Pittsburgh, Pittsburgh., M.D., Meenal PatelFrom the Departments of Psychiatry, Bioengineering, Neurology, Neuroscience, and Radiology at the University of Pittsburgh, Pittsburgh., B.S., and Charles F. Reynolds IIIFrom the Departments of Psychiatry, Bioengineering, Neurology, Neuroscience, and Radiology at the University of Pittsburgh, Pittsburgh., M.D.Published Online:1 Oct 2011https://doi.org/10.1176/appi.ajp.2011.10060853AboutSectionsView articleAbstractPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinked InEmail View articleAbstractObjective:This study tests whether or not the structural white matter lesions that are characteristic of late-life depression are associated with alterations in the functional affective circuits of late-life depression. This study used an emotional faces paradigm that has been shown to engage the affective limbic brain regions.Method:Thirty-three elderly depressed patients and 27 nondepressed comparison subjects participated in this study. The patients were recruited through the NIMH-sponsored Advanced Center for Interventions and Services Research for the Study of Late-Life Mood Disorders at the University of Pittsburgh Center for Bioethics and Health Law. Structural and functional MRI was used to assess white matter hyperintensity (WMH) burden and functional magnetic resonance imaging (fMRI) blood-oxygen-level-dependent (BOLD) response on a facial expression affective-reactivity task in both elderly participants with nonpsychotic and nonbipolar major depression (unmedicated) and nondepressed elderly comparison subjects.Results:As expected, greater subgenual cingulate activity was observed in the depressed patients relative to the nondepressed comparison subjects. This same region showed greater task-related activity associated with a greater burden of cerebrovascular white matter change in the depressed group. Moreover, the depressed group showed a significantly greater interaction of WMH by fMRI activity effect than the nondepressed group.Conclusions:The observation that high WMH burden in late-life depression is associated with greater BOLD response on the affective-reactivity task supports the model that white matter ischemia in elderly depressed patients disrupts brain mechanisms of affective regulation and leads to limbic hyperactivation.The primary model of the neurobiology of late-life depression is the vascular depression hypothesis (1). The vascular depression model posits that vascular changes, primarily ischemic changes in the white matter, are a significant factor in the pathogenesis of late-life depression. This model has been supported by several neuroimaging studies that have demonstrated higher white matter hyperintensity (WMH) ratings in patients with late-life depression relative to nondepressed elderly comparison subjects (2, 3). Another study has shown that indicators of the vascular depression phenotype (e.g., WMHs, executive function, and age at onset) are significant predictors of treatment response (4). Several studies have suggested that the localization of WMHs in late-life depression may be specific to the frontal and subcortical regions (5–7), thus disrupting critical frontal-subcortical affect-regulating circuits.Over the past decade, functional imaging, primarily positron emission tomography (PET) and functional MRI (fMRI), has been used to develop and test models of the functional neuroanatomy of depression (8–10). These studies have primarily used participants with midlife major depression and have consistently observed rostral and ventral brain hyperactivation in the depressed patients relative to nondepressed comparison subjects. The hyperactivation has been observed at rest in studies of cerebral metabolism ([18F]fluorodeoxyglucose [FDG] PET [9]) and also with blood-oxygen-level-dependent (BOLD) fMRI in response to affective tasks (11). The particular fMRI tasks have varied, but a common approach is to have participants view faces that have emotional expressions. This task seems to be particularly effective at engaging the anterior and ventral structures, which include limbic regions such as the amygdala, the rostral cingulate, the medial prefrontal cortex, and the anterior insula. Hariri et al. (12) reported that this emotional faces task is particularly effective at eliciting BOLD fMRI activity in the amygdala and in ventral limbic regions. Studies of midlife depression have also identified hypoactivation in the dorsal structures associated with cognitive control, including the dorsal anterior cingulate and the dorsolateral prefrontal cortex. According to the functional neuroanatomical models of depression (10), depression can be described as altered patterns of neural regulation between the dorsal-cognitive and ventral-affective components of an affect regulation system.This functional neuroanatomical model of depression is further supported by treatment studies of midlife depression that have shown that ventral hyperactivity responds to treatment. In a study of 11 depressed patients using an emotional faces task, the successful treatment of depression was correlated with a decrease in amygdala activity from pretreatment to posttreatment assessments (4). Other studies have reported similar findings in the subgenual cingulate (9). This greater activity in pretreatment midlife depression has served as the basis of innovative studies examining treatment-resistant depression (13).Functional imaging studies of late-life depression have replicated some aspects of the pattern observed in midlife depression. Some studies of late-life depression, but not all (14), reported decreased dorsal cognitive activation (15, 16) and greater ventral limbic activation (15). A relatively small number of functional imaging studies of late-life depression, with relatively small sample sizes (e.g., N=13 per group [16]), have been conducted compared to the number of larger studies reporting structural imaging findings in late-life depression (4). The structural imaging studies have shown a significant correlation between late-life depression and white matter ischemic changes, which are visualized as WMH lesions on MRI (6).Our study relates the white matter structural changes in late-life depression to the functional models of depression that have been primarily developed through studies of midlife depression. In particular, are the white matter lesions of late-life depression associated with the same ventral limbic hyperactivity that is observed in midlife depression? Our study used a standard fMRI task (the faces and shapes task [17]) that has been shown to activate the expected limbic structures (18). WMH burden was assessed using an automated segmentation of T2-weighted fluid attenuated inversion recovery (FLAIR) images. In this study we focused on overall global WMH burden rather than using regional WMH measures from the various regions and white matter tracts implicated in affective processing. The choice of whether to use global or regional WMH measures depends on whether the hyperintensities are viewed as a summary biomarker of overall global white matter disease (i.e., a global biomarker) or whether the WMHs are viewed as an indicator of localized white matter damage to the tracts intersected by WMHs (i.e., localized WMHs) instead. Evidence supports both the global biomarker view and the localized WMHs view (6, 19). Our choice of using global WMHs as our primary marker in this study was motivated by the global biomarker view and our intent to limit the number of independent variables.MethodParticipantsWe recruited patients with nonpsychotic, unipolar major depressive disorder from ongoing open-label treatment studies of late-life major depression (20, 21) at the University of Pittsburgh's Advanced Center for Interventions and Services Research for the Study of Late-Life Mood Disorders. Depressed patients had an fMRI scan before starting antidepressant medication, and they were asked to participate in another scan 12–16 weeks after starting treatment. (Analyses of the follow-up scans will be presented in a separate report.) Nondepressed elderly comparison subjects were recruited from the community and from the healthy volunteer registry of the University of Pittsburgh Alzheimer Disease Research Center. After receiving a complete description of the study, each participant provided written informed consent. All participants were paid $50 to take part in the study.All participants were assessed with the Structured Clinical Interview for DSM-IV (22). Other than major depressive disorder and anxiety disorders, all axis I disorders were used as exclusion criteria. Other exclusion criteria were a history of stroke or significant head injury, Alzheimer's disease, Parkinson's disease, or Huntington's disease. We chose to include participants with anxiety disorders because of the high prevalence (48%) of anxiety disorders in individuals with late-life depression (23). Individuals were excluded if they had taken psychotropic medications during the 2 weeks before imaging, but other medications were acceptable because their use is common in the elderly population.ProceduresFaces and shapes affective reactivity task.The faces and shapes fMRI task (illustrated in Figure 1) has been used extensively to explore the neural circuitry of affective reactivity (17). One study observed that subjects engage in spontaneous emotional regulation when exposed to this particular task (24). In this task, participants are required to select one of two facial expressions (either angry or afraid) that matches a simultaneously presented target expression. As a control task, the participants match one of two geometric shapes with a simultaneously presented target shape. The task involves five blocks of the shape-matching task alternating with four blocks of the face-matching task. Each block lasts 30 seconds (six trials that last 5 seconds each). On the faces trials, 12 different images derived from a standard set of pictures of facial affect are used, six per block and three of each gender (25). Stimulus presentation and response recording was controlled using the E-Prime software package (Psychology Software Tools, Inc., Pittsburgh, 2002).FIGURE 1. Faces and Shapes fMRI Task Used to Examine Neural Circuitry in a Study of White Matter Hyperintensities in Late-Life Depressionaa The faces and shapes block design task involves six trials per block and nine blocks: five blocks of shapes and four blocks of faces.Data acquisition.Imaging data were collected with a 3-T Siemens Trio TIM scanner located in the MR Research Center at the University of Pittsburgh. For functional image alignment we used a T2-weighted sequence (TR=3,000 msec, TE=101 msec, in-plane resolution=1 mm×1 mm, 3-mm slice thickness, 48 slices). To assess for small vessel ischemic changes, we used a T2-weighted FLAIR sequence (TR=9,002 msec, TE=56 msec [effective]; TI=2,200 msec, number of excitations=1) using an interleaved acquisition (48 slices, 3-mm slice thickness, no gap). T2*-weighted BOLD acquisition was done using a gradient-echo echo planar imaging sequence: TR=2,000, TE=32, matrix=128×128×29, voxel size=2×2×3 mm3, oblique axial acquisition (parallel to anterior commissure-posterior commissure line), integrated parallel acquisition techniques=2. The most inferior functional scan was below the most inferior aspect of the temporal lobes.Imaging analysis.We used an automated WMH localization and segmentation method to compute the normalized WMH volumes (7). For each participant, the calculated WMH volume was normalized with the overall brain volume. The automated WMH segmentation method (7) is an iterative algorithm that involves an automated selection of "seeds" of possible WMH lesions and fuzzy connectedness, which clusters voxels based on their adjacency and affinity,to segment WMH lesions around the seeds (26). The fully automated WMH segmentation system was implemented in C++ and ITK. (The WMH matter extraction algorithm is available at http://www.gpn.pitt.edu.)To assess the relationship of WMHs on the fMRI BOLD response pattern in late-life depression, we first identified the median normalized WMH volume among the late-life depression patients. This split was used to define a group with a low WMH burden (WMH burden < the median) and a group with high WMH burden (WMH burden ≥ the median) (Figure 2).FIGURE 2. Histograms of White Matter Hyperintensity (WMH) Burden in Late-Life Depression Patients and in Nondepressed Elderly Comparison Subjectsaa T2-weighted FLAIR (fluid attenuated inversion recovery) images from late-life depression patients (top panel) and nondepressed elderly comparison subjects (bottom panel) in the low- and high-burden groups are shown to illustrate the degree of WMH burden.fMRI Preprocessing and Group AnalysisFunctional imaging data were analyzed with SPM5 (Wellcome Department of Imaging Neuroscience, London, 2007) implemented in MatLab (MathWorks, Natick, Mass.). For each participant, all echo planar imaging volumes were realigned to the first echo planar imaging volume, and a mean image of the realigned volumes was created. The realigned images were coregistered to the anatomical T1-weighted image. To normalize the anatomical image as well as the echo planar images to a standard SPM reference system, the following procedure was applied. First, the anatomical image as well as a representative template image (from the Montreal Neurological Institute) was segmented into gray matter, white matter, and CSF. Then the anatomical gray matter image was normalized to the gray matter of the Montreal Neurological Institute brain. Subsequently, the derived normalization parameters were applied to the echo planar images, which were resampled to a voxel size of 2 mm×2 mm×3 mm and smoothed with a 10-mm wide Gaussian kernel to account for greater morphologic variability in an elderly sample (27). All statistical analyses were performed in the context of the general linear model (28).Movement parameters derived from realignment were assessed, and no participant with excessive motion (>4 mm or >4°) was identified. We also added movement as a covariate of no interest to correct for confounding effects induced by head movement. The primary contrast of interest (faces > shape condition) was first estimated for each participant individually (first-level analysis) and then used in a second-level random-effects analysis to assess the group-level effects.The fMRI analysis was performed on 60 participants. The second-level analysis initially included all of them in order to identify the main effect of the task. Second-level analyses were also done between groups (depressed patients relative to nondepressed comparison subjects) to identify the effect of depression, and in the depressed and comparison groups to identify the effect of WMHs. High and low WMH burden was designated by a median split of normalized WMH volume.The second-level analysis over all participants was performed using a one-sample two-tailed t test. The between-group analyses were modeled using two-sample two-tailed t tests. In each group, the mean group maps were estimated using the difference between BOLD signal during faces and comparison conditions. We tested between-group differences in the BOLD signal during the task, and group differences were measured at each voxel and restricted to the gray matter. We set a t value threshold at a minimal voxel entry value of t>2.6618 (df=59) for main effects, t>2.6633 (df=58) for depressed patients relative to comparison subjects, t>2.7564 (df=29) for high relative to low WMH burden in the depressed group, and t>2.7874 (df=25) for high relative to low WMH burden in the comparison group (p 2.6618, df=59, p shapes is shown. The crosshairs point out the robust amygdala BOLD response with this task.The amygdala bilaterally, the insula, and the rostral cingulate approached significance with family-wise error rate with a small-volume correction. Table 2 and Table 3 show the family-wise error corrections for the a priori regions of interest for the main effect of task contrast as well as for the comparisons with depression and WMH burden. These tables include values that meet a threshold of p Nondepressed Comparison SubjectsRegionMNI Coordinatest (df=59)bCluster Sizepct (df=58)bCluster SizepcRostral cingulate2, 26, –33.6150.0032.7430.037Amygdala22, –4, –18; –22, –6, –186.901380.0042.7110.087Insula–30, 20, –3; 36, 22, 84.703610.001———a Table includes all values that meet a threshold of p Low WMH BurdenDepressed PatientsNondepressed Comparison SubjectsDepressed and Comparison InteractionRegiont (df=29)aCluster Sizepbt (df=25)aCluster Sizepbt (df=56)aCluster SizepbRostral cingulate3.481000.004———2.89390.040Amygdala—————————Insula3.0941<0.001———2.8720 shapes for each participant were used to compare the difference in BOLD response between the depressed and nondepressed groups. The results of this random-effects between-group analysis is shown in Figure 4. We set the threshold for the two-sample t test map at a voxel-wise t>2.6633, df=58, p shapes contrast maps were compared using a random-effects between-group analysis, comparing those with a low WMH burden to those with a high WMH burden. The results of the two-sample t test for the depressed and comparison groups are shown in Figure 5. There is a more robust effect of WMH burden in the depressed group relative to the nondepressed group. In the depressed group, the rostral cingulate shows greater BOLD response in those late-life depression patients with high WMH. There are no areas that show significantly greater activity in the low-WMH-burden patients (at t>2.744, df= 31, p<0.005, uncorrected). In addition to the rostral cingulate regions of interest, the insula also shows greater BOLD response in the high-WMH-burden participants. The amygdala showed higher BOLD response with greater WMHs, but the association did not reach statistical significance. Results of fMRI based on the regions of interest are summarized in Table 2. These results suggest that the association of WMH burden with BOLD activity appears to be specific for elderly patients with depression. This supports the view that the burden of WMHs is associated with the functional changes in elderly patients with clinical depression, despite the fact that white matter damage may also occur in nondepressed elderly people.FIGURE 5. High Compared With Low White Matter Hyperintensity (WMH) Burden in Late-Life Depression Patients and Nondepressed Elderly Comparison Subjectsaa Brain activation map shows areas with greater BOLD response in high WMH burden compared with low WMH burden in late-life depression patients (top) and nondepressed elderly comparison subjects (middle), as well as the interaction between the two (bottom). The crosshairs point out the rostral (subgenual) cingulate region.DiscussionWe examined the association of WMH burden with the pattern of functional BOLD response during affective processing in late-life depression. The primary aim of this study was to test whether the often-reported white matter changes in late-life depression are associated with functional brain changes during affective processing. Our results show that the late-life depression patients showed a pattern of greater limbic fMRI activity relative to nondepressed elderly comparison subjects. This pattern was accentuated in the late-life depression patients with the highest WMH burden. Moreover, a significant interaction between depression and WMH burden was observed with respect to their association with fMRI activity, indicating that the association of WMHs with greater limbic fMRI activity is particularly pronounced in the depressed group. This cross-sectional study shows association, rather than causation, but it nonetheless supports the vascular depression hypothesis that ischemic white matter changes contribute to the pathogenesis of late-life depression. One possible interpretation is that the white matter changes in late-life depression disrupt the ability of the prefrontal cortical structures to modulate the hyperactive limbic structures through altered connectivity.In this study, the difference in WMH burden in the late-life depression patients relative to comparison subjects fell short of significance. This was somewhat unexpected, as higher WMH burden in late-life depression has been reported in many (but not all [5]) previous reports. We believe the observation of WMH differences between the depression and comparison groups depends on factors such as how closely subjects are matched on cerebrova
Referência(s)