- Email: [email protected]
Ventromedial Prefrontal Cortex Processing During Emotional Evaluation in Late-Life Depression: A Longitudinal Functional Magnetic Resonance Imaging Study
Ventromedial Prefrontal Cortex Processing During Emotional Evaluation in Late-Life Depression: A Longitudinal Functional Magnetic Resonance Imaging Study Stefanie Brassen, Raffael Kalisch, Wolfgang Weber-Fahr, Dieter F. Braus, and Christian Büchel Background: Functional imaging studies using emotional stimuli have suggested a role for the ventromedial prefrontal cortex (vmPFC) in the pathophysiology of midlife depression. In contrast, the neural correlates of late-life depression (LLD), a highly prevalent but underrecognized clinical entity in which age-related brain changes might influence disease mechanisms, have not been studied in great detail. With an emotional evaluation task, we conducted a longitudinal study of vmPFC functioning in a homogeneous sample of elderly antidepressant naïve female outpatients with isolated, first diagnosed mild to moderate depressive symptoms. Methods: Neural responses of the vmPFC to the emotional evaluation of positive, negative, and neutral words were measured with functional magnetic resonance imaging (fMRI) in LLD (n ⫽ 13) and healthy older subjects (n ⫽ 13). All patients were rescanned after approximately 7 months. Results: Although there were no performance differences, compared with healthy volunteers, LLD patients showed a decreased response to negative compared with positive stimuli in the vmPFC. This altered pattern was positively correlated with symptom severity. At follow-up, the attenuated neural response in the vmPFC had “normalized,” accompanied by a significant improvement in symptoms. Conclusions: These findings indicate vmPFC dysfunction as a biological state marker of geriatric depression. Furthermore, our data underline the pathological significance of mild to moderate LLD and highlight the usefulness of functional neuroimaging for evaluating remission processes in this specific depression subtype.
Key Words: Aging, emotion, functional neuroimaging, geriatric depression, vmPFC
L
ate-life depression (LLD) is often accompanied by marked disability, diminished quality of life, caregiver burden, and mortality. Nevertheless, the most prevalent psychological disorder in old age is still under-recognized, making it a significant health care issue (1,2). Accordingly, in contrast to a broad research in midlife depression, there are surprisingly few functional studies dealing with the neural abnormalities in LLD, although it seems plausible that in the context of age-related brain changes (3,4) functional findings in younger adults might not necessarily generalize to older persons (5). The insufficient investigation of LLD patients is certainly also a consequence of the strong heterogeneity of this sample regarding medical history and comorbidity, which complicate a defined examination of the mood disorder itself. Age-related brain changes are particularly pronounced in the prefrontal cortex (3). Furthermore, recent findings indicate additional grey matter deficits (6) as well as hypoactivation during resting state (7,8) in regions of the ventromedial prefrontal cortex (vmPFC) in LLD patients compared with healthy elderly control
From the NeuroImage Nord, Institute for Systems Neuroscience (SB, RK, WW-F, CB), Department of Psychiatry and Psychotherapy (DFB), University Medical Center Hamburg-Eppendorf, Hamburg, Germany. Drs. Braus and Büchel contributed equally to this work. Address reprint requests to Stefanie Brassen, Ph.D., NeuroImage Nord, Institute for Systems Neuroscience (S10), University Medical Center Hamburg-Eppendorf, Martinistrasse 52, D-20246 Hamburg, Germany; E-mail: [email protected]. Received April 18, 2007; revised March 20, 2008; accepted March 20, 2008.
0006-3223/08/$34.00 doi:10.1016/j.biopsych.2008.03.022
subjects. The vmPFC, which comprises the orbitofrontal cortex, the rostral cingulate cortex, and the ventral medial frontal gyrus, has been associated with the evaluation and regulation of emotional events and stimuli (9) and the contextual processing of reward (10). Functional imaging studies in midlife depression revealed an abnormal neural response in this region during the perception of affective stimuli. Here, depressed patients have shown an increased response to positive (11,12) and a decreased response to negative stimuli (13,14). In addition, the few existing longitudinal studies in midlife depression have proposed vmPFC dysfunction as a state marker sensitive to remission (15–20). In contrast, the neural correlates of emotional processing and thus vmPFC functioning have not yet been studied in LLD. Given the reported structural and functional changes in this region in aging and LLD, we wondered whether a functional deficit during emotional evaluation akin to the one observed in younger depressive subjects could be demonstrated in patients suffering from isolated mild to moderate LLD and whether possible frontal dysfunctions in these patients are reversible. Because this is one of the first fMRI studies of emotional processing in LLD, we decided to examine the neural correlates of emotional appraisal, which might provide a starting point for further studies investigating other emotional processes typically affected during depression. With fMRI we assessed emotional appraisal in an explicitly homogeneous sample of elderly antidepressant-naïve female outpatients with isolated first-diagnosed mild to moderate depressive symptoms and in healthy elderly subjects. We used a valence decision task requiring the affective evaluation of emotional words, which has previously elicited vmPFC activation (21–24). Most importantly, the study was designed in a longitudinal fashion to also address the question of whether brain activation patterns represent state markers. BIOL PSYCHIATRY 2008;64:349 –355 © 2008 Society of Biological Psychiatry
350 BIOL PSYCHIATRY 2008;64:349 –355
Figure 1. Schematic display of the experimental design.
Methods and Materials Subjects Participants were 13 depressed elderly women (mean age 66.4 ⫾ 6.1 years) as well as 13 healthy age- and educationmatched women (65.6 ⫾ 6.1 years). Only patients with a clinically significant mild to moderate depressive episode (ICD10) beginning later than the age of 55, no antidepressant medication at present or in the past, and no history of neurological or other psychiatric diseases like anxiety disorders were included in the study. Pathological white matter changes were excluded with a fluid-attenuated inversion recovery (FLAIR) sequence. To exclude highly prevalent subclinical dementia in LLD (25), we additionally applied a detailed neuropsychological assessment to all participants (Supplement 1). After complete description of the study to the subjects, written informed consent complying with the local ethics committee was obtained. Objective and subjective symptom severity was assessed with the Montgomery-Asberg Depression-Rating-Scale (MADRS) (26) and the self-rated Geriatric-Depression-Scale (GDS) (27). All 13 patients were re-examined approximately 7 months (28.7 ⫾ 2.9 weeks) after their first visit. Patients were naturalistically treated— six of them with antidepressant drugs, two with behavior therapy; and five patients had not received any treatment (for details see Supplement 1). Twelve patients were clinically improved as also reflected by a significant reduction of the MADRS score; one patient showed no improvement. fMRI Paradigm. Inside the scanner, subjects were presented with 40 neutral, 40 positive, and 40 negative adjectives in randomized order (Figure 1). The trials were pseudorandomly intermixed with 40 null events (blank screen) to provide jitter. After replacement with a rating screen subjects had to indicate their emotional rating by pressing one of three buttons (positive, neutral, negative). For details of the functional imaging design see Supplement 1. www.sobp.org/journal
S. Brassen et al. Imaging. Details on imaging and statistical procedures are given in Supplement 1. In short, imaging was performed on a 3-Tesla whole-body MR-scanner (Siemens, Erlangen, Germany). Volumes of 42 transversal slices were obtained (3-mm-thickness, no gap, repetition time ⫽ 2.41 sec, echo time ⫽ 24 msec, ␣ ⫽ 80°, field-of-view 192 ⫻ 192 mm2, 64 ⫻ 64 matrix). At first-level, regressors of interest for subjects’ individually rated positive, negative, and neutral words were modeled separately with SPM2 (http://www.fil.ion.ucl.ac.uk/spm/). The 6 estimated movement parameters from the realignment stage were included as covariates. A second session was modeled for the follow-up scan. At the second level, contrast images representing differential contrasts of interest (“negative vs. neutral words,” “positive vs. neutral words,” “positive vs. negative words”) were statistically tested voxelwise in one- and two-sample t tests to assess within and between group effects within our a priori regions of interest (ROIs, see following text) with a threshold of p ⬍ .05 family-wise error (FWE) corrected. Correlation analyses were conducted with simple regression analyses implemented in SPM2 with a threshold of p ⬍ .001 uncorrected. Explorative data of within group analyses regarding the whole brain (p ⬍ .001 uncorrected) are available in Supplement 1. Areas of Interest. Given recent evidence of altered responses in midlife depression to affective material in the vmPFC, we based our correction for multiple comparisons on corresponding ROIs rather than on the whole brain. The ROIs were spheres of 10 mm radius around the following center coordinates: x ⫽ ⫺4, y ⫽ 58, z ⫽ ⫺6 (medial orbitofrontal cortex [mOFC]); ⫺3, 59, 14 (superior medial frontal cortex [sMFC]); ⫺2, 46, 10 (left); and 12, 43, 7 (left and right rostral anterior cingulate cortex [rACC]). Coordinates were chosen from previous studies reporting valence-dependent engagement of the vmPFC (21–24) and were transferred from the Talairach to Montreal Neurological Institute (MNI) space if necessary.
Results Behavioral Data At baseline, there were no significant differences in word valence categorizations, reaction times, or the different kinds of ratings changes (e.g., from negative to neutral; and vice versa) between both groups, and there were no changes in these parameters at follow-up in depressed patients (Table 1). Imaging Data Baseline. NEGATIVE VERSUS NEUTRAL. There was a more pronounced engagement of the mOFC for negative versus neutral words in control subjects. Patients showed significantly less activation to negative words in the bilateral rACC. Interaction analysis confirmed higher mOFC and right rACC activation to negative words in control subjects compared with patients (Table 2). Table 1. Rating Data
“Positive” Rating “Negative” Rating “Neutral” Rating RT “Positive” Rating RT “Negative” Rating RT “Neutral” Rating
Healthy
LLD Baseline
LLD Follow-Up
42.6 (7.3) 37.9 (2.6) 39.5 (7.6) 1.57 (.3) 1.57 (.3) 1.63 (.3)
41.6 (6.4) 37.0 (2.4) 41.4 (5.9) 1.51 (.3) 1.52 (.3) 1.55 (.3)
41.5 (3.6) 38.8 (1.5) 39.8 (3.5) 1.49 (.3) 1.54 (.3) 1.57 (.4)
Analyses of variance including a group factor revealed no differences between groups (p ⬍ .05). RT, reaction time; LLD, late-life depression.
BIOL PSYCHIATRY 2008;64:349 –355 351
S. Brassen et al. Table 2. Statistical Results of Within and Between Group Comparisons at Baseline Region
Hem
x
y
z
Z
x
Within Group: Neutral vs. Negative Healthy: Negative⬎Neutral Medial Orbitofrontal Cortex LLD: Neutral⬎Negative Rostral Anterior Cingulate Cortex
L
LLD: Positive⬎Negative Medial Orbitofrontal Cortex Superior Medial Frontal Cortex Rostral Anterior Cingulate Cortex
60
⫺3
z
Z
Group Interaction: LLD⬎Healthy (Neutral⬎Negative) 3.23a
⫺6 9 12
L R
⫺9 39 6 12 39 15 Within Group: Positive vs. Neutral
3.06b 4.25c
R
0 60 ⫺3 3 54 ⫺9 Within Group: Positive vs. Negative
3.36a 3.13a
LLD: Positive⬎Neutral Medial Orbitofrontal Cortex
Healthy: Negative⬎Positive Superior Medial Frontal Cortex Rostral Anterior Cingulate Cortex
⫺6
y
57
⫺6
36 9 39 15 No Group Interaction
3.21a 3.51a 3.49a
Group Interaction: LLD⬎Healthy (Positive⬎Negative)
R L R
3 ⫺3 9
54 51 48
18 18 15
3.59a 2.86b 2.94b
3
54
18
2.97b
L
⫺6 ⫺9 ⫺6 ⫺6 3
54 60 63 42 42
⫺9 0 6 6 6
3.84c 3.26a 4.25c 4.20c 3.40a
⫺3
60
⫺3
3.34a
3 ⫺3 ⫺6
54 63 42
⫺6 6 6
3.10a 3.56a 3.57a
L L R
The table shows peak activations in Montreal Neurological Institute space. The p values are corrected for multiple comparisons in a priori search volumes (regions of interest). Local maxima more than 4 mm apart. Hem, hemisphere; LLD, late-life depression. a p ⬍ .05. b p ⬍ .1. c p ⬍ .01 (all family-wise error corrected).
POSITIVE VERSUS NEUTRAL. There was a significantly higher activation of the mOFC for positive compared with neutral words in depressed subjects. No differences in activation for positive versus neutral were visible in control subjects in our a priori ROIs. POSITIVE VERSUS NEGATIVE. Control subjects showed a significantly stronger activation for negative compared with positive words in the sMFC and the bilateral rACC. By contrast, patients displayed a widespread stronger activation for positive compared with negative words in all regions of the vmPFC. Accordingly, all regions showed a significant group interaction (Table 2, Figure 2).
Correlations with Symptom Severity Owing to significant interaction analyses, simple regressions were applied for the contrasts neutral⬎negative and positive⬎negative in depressed patients. Reduced activation for negative compared with neutral words was positively correlated with GDS score in the mOFC (r ⫽ .78; p ⬍ .001). Increased activation for positive compared with negative words was positively correlated with GDS and MADRS scores (Figure 3). Follow-Up. At follow up, patients showed an increased activation for negative compared with neutral words in comparison with baseline in the sMFC cortex and the bilateral rACC. Furthermore, the widespread positive⬎negative differences in patients at baseline disappeared at follow-up. A direct comparison with control subjects at baseline revealed absence of patientto-control differences (Table 3, Figure 4). Normalization in the sMFC was positively correlated with MADRS score at baseline (Figure 5).
Discussion The key finding of this study is an abnormal neural response to positive and negative words in patients with mild to moderate LLD in regions of the vmPFC. In contrast to healthy subjects, patients showed a decreased response to negative compared with positive stimuli. This attenuated neural response in the vmPFC was positively associated with depression severity. After 7 months of naturalistic treatment, depressed patients’ symptoms had significantly improved and the abnormal neural pattern in the vmPFC had normalized. The normalization could be predicted by depression severity at baseline and supports the role of ventromedial prefrontal neural abnormalities as a cortical state marker of LLD. Late-life depression is an underdiagnosed depression subtype in primary care, probably because relatives and physicians often believe that mild to moderate depressive syndromes are rather normal for older people faced with an increasing impact of psychosocial adversity and a decline in physical health and cognitive abilities (2,28). We now report a biological marker in this patient group, which strongly argues against this (mis-)belief and underlines the pathological significance of the syndrome. This is further supported by an obvious analogy between our functional results and previous findings of abnormal vmPFC responses in younger patients (13). The analogy extends to an apparent reversibility of frontal changes that parallels remission processes after antidepressant medication (16,17,19,20) or cognitive behavior therapy (18) in midlife depression. Overall, this indicates similar pathophysiological mechanisms. The observed neural normalization is particularly interesting in the light of www.sobp.org/journal
352 BIOL PSYCHIATRY 2008;64:349 –355
S. Brassen et al.
Figure 2. Increased neural response to positive and decreased neural response to negative words in the ventromedial prefrontal cortex in depressed individuals compared with healthy subjects. LLD, late-life depression; rACC, rostral cingulate gyrus; mOFC, medial orbitofrontal cortex; sMFC, superior medial frontal cortex. p ⬍ .05 family-wise error corrected for small volumes.
frequent findings of frontal volume changes in LLD. Because volume changes are triggered by age (29) and are therefore probably progressive, it might have been speculated that functional alterations of the vmPFC in LLD are irreversible. Whether the observed normalization results from a lack of such volume changes in our study sample, from an independency of structural and functional brain alterations, or from compensatory processes
during remission will have to be clarified by future studies, including longitudinal structural data. We here chose to investigate emotional processing, in contrast to the few existent dynamic imaging studies in LLD (30,31), and focused our analysis on the vmPFC, which is a target structure of grey matter and metabolic changes in aging as well as LLD and which has consistently been shown to be dysfunc-
Figure 3. Scatterplots of correlations of objective (Montgomery-Asberg Depression-Rating-Scale [MADRS]; Pearson’s correlation coefficient r ⫽ .76; p ⬍ .001) and subjective (Geriatric-Depression-Scale [GDS]; r ⫽ .84; p ⬍ .001) symptom severity with increased activity for positive compared with negative words at the superior medial frontal cortex (sMFC) at baseline in depressed individuals.
www.sobp.org/journal
BIOL PSYCHIATRY 2008;64:349 –355 353
S. Brassen et al. Table 3. Statistical Results of Baseline vs. Follow-Up Comparisons in Patients Region
Hem
x
y
z
Z
Baseline⬎Follow-Up Superior Medial Frontal Cortex
L
Rostral Anterior Cingulate Cortex
L R
Superior Medial Frontal Cortex Rostral Anterior Cingulate Cortex
L L
Neutral ⬎ Negative ⫺6 51 18 ⫺9 66 9 ⫺9 48 15 3 39 12 Positive ⬎ Negative ⫺3 60 9 ⫺6 42 6 ⫺3 54 15
3.13a 2.92a 3.28b 2.92a 4.11c 3.44b 3.35b
The table shows peak activations in MNI space. The p values are corrected for multiple comparisons in a priori search volumes (ROIs). Local maxima more than 4 mm apart. Abbreviations as in Table 2. a p ⬍ .1 (all family-wise error corrected). b p ⬍ .05. c p ⬍ .01.
tional in younger depressed subjects (11,12). Through its mostly reciprocal anatomical connections with the amygdala, the hypothalamus, the striatum, and other limbic and brainstem structures (32,33), the vmPFC is theoretically in a position to process internal and external affective stimuli and to modulate affective responses. Functional studies in healthy subjects suggest that the vmPFC encodes the emotional meaning of stimuli (9). More
Figure 5. Scatterplot of the correlation between normalization in ventromedial prefrontal cortex (vmPFC) activation over time in the superior medial frontal cortex (sMFC) and symptom severity at baseline (r ⫽ .80; p ⬍ .001). Thus, a more severe depression at baseline was associated with a stronger normalization of vmPFC activation at follow-up. t0 ⫽ baseline; t1 ⫽ followup. MADRS, Montgomery-Asberg Depression-Rating-Scale.
specifically, the vmPFC has been implicated in the generation of an abstract representation of the rewarding value of a stimulus by attending to its context (34). In this line, previous findings of a double dissociation of vmPFC response to positive and negative stimuli in younger patients have been interpreted on the basis of a role of the vmPFC in the contextual appraisal of the rewarding
Figure 4. Activation changes in the ventromedial prefrontal cortex at follow-up compared with baseline for the contrast positive⬎negative in LLD patients. Normalization was found in the rACC and the sMFC. p ⬍ .05 FWE corrected for small volumes. Abbreviations as in Figure 2.
www.sobp.org/journal
354 BIOL PSYCHIATRY 2008;64:349 –355 potential of emotional stimuli (13,14,35). Keedwell et al. (13) speculated that a relatively increased response of the vmPFC to positive stimuli reflects the attempt of depressive subjects to attend more closely to the positive stimulus, owing to their reduced capacity to experience pleasure in pleasurable situations. This speculation might be supported by recent data relating increased vmPFC response to positive stimuli with anhedonia. Furthermore, the relatively decreased vmPFC response to negative stimuli in depressed patients has not only been associated with reduced negative feedback response in depression (14) but also with context- or state-dependent stimulus processing. It has been argued that the presentation of negative or sad stimuli to depressed patients who already are in a negative mood state is unlikely to be perceived as salient and to prompt an emotional response (13). By contrast, healthy normal or remitted subjects are likely to be more affected by a negative stimulus, which is discordant with the current mood state and thus requires attention and a more detailed cognitive elaboration. An alternative explanation for the observed findings in vmPFC could be based on its role in emotion regulatory processes. Animal studies suggest that the vmPFC and the amygdala inhibit each other and thus regulate emotional behavior (36). Recent human data also indicate a moderating function of vmPFC in cognitive-emotional regulation through its direct connections with lateral/dorsal PFC and with the amygdala (37,38). From this perspective, a relatively decreased vmPFC response to negative stimuli might reflect the lack of inhibitory influences on negative emotional processing performed by limbic structures like the amygdala. However, it is important to highlight that we did not observe amygdala activations for the negative stimuli (Supplement 1) and were thus unable to test a potential inverse relationship between the vmPFC and the amygdala. Most previous neurobiological studies in LLD have focused on structural changes, mainly in frontal brain regions (6,8). Their frequent finding of high-intensity lesions has been suggested as one primary neurobiological path of LLD leading to the concept of “vascular depression” in elderly persons (1). There are, however, recent data that do not support such an association (39). Notably, path analyses have revealed smaller brain volumes, particularly in frontal regions as second, distinct path leading to LLD (29). The comparability of such results is limited by the inherent heterogeneity of elderly depressed patients, especially with regard to premedication, onset of disease, previous episodes, symptom severity, and a high degree of comorbidity typically existent in old age. One potentially advantageous strategy would be to study brain abnormalities related to the isolated mood disorder before examining heterogeneous samples. In this respect, the homogeneous group of elderly outpatients examined here is particularly interesting. Our decision to study a homogeneous sample, however, also limits generalizability, for example to male populations. Yet, this approach helped us in generating a clear experimental hypothesis, which can now be tested in other subtypes of depression in aging (e.g., with and without mild cognitive impairment or cerebrovascular diseases). A second limitation lies in the naturalistic treatment of patients, which precludes any conclusions about the effects of specific treatments. Finally, we chose a very simple, low-demand task, which is easy to conduct in elderly patients yet shows robust medial frontal engagement (21). In particular, we only asked for a broad valence categorization instead of using a continuous self-relevant valence rating scale (e.g., from ⫺10 to 10). This might have occluded rating differences between patients and www.sobp.org/journal
S. Brassen et al. control subjects. Nevertheless, in our opinion this only underlines the sensitivity of vmPFC dysfunction as an early biological marker for depression. Altogether, our data point to the pathological significance of mild to moderate depressive symptoms in the elderly and thus to the need to treat these patients like younger subjects. In this context, functional imaging might be a useful tool to evaluate treatment response. The study was supported by a grant from “Forschungsförderung Medizin” F154 (SB), University Medical Center HamburgEppendorf. We thank K. Müller and F. Walde for help with MR scanning. The authors reported no biomedical financial interests or potential conflicts of interest. Supplementary material cited in this article is available online. 1. Alexopoulos GS (2005): Depression in the elderly. Lancet 365:1961– 1970. 2. Charney DS, Reynolds CF 3rd, Lewis L, Lebowitz BD, Sunderland T, Alexopoulos GS, et al. (2003): Depression and Bipolar Support Alliance consensus statement on the unmet needs in diagnosis and treatment of mood disorders in late life. Arch Gen Psychiatry 60:664 – 672. 3. Raz N, Gunning FM, Head D, Dupuis JH, McQuain J, Briggs SD, et al. (1997): Selective aging of the human cerebral cortex observed in vivo: Differential vulnerability of the prefrontal gray matter. Cereb Cortex 7:268 –282. 4. Vermeer SE, Longstreth WT Jr, Koudstaal PJ (2007): Silent brain infarcts: A systematic review. Lancet Neurol 6:611– 619. 5. Brassen S, Büchel C, Weber-Fahr W, Lehmbeck JT, Sommer T, Braus DF (2007): Structure-function interactions of correct retrieval in healthy elderly women [published online ahead of print November 17]. Neurobiol Aging doi:10.1016/j.neurobiolaging.2007.10.005. 6. Ballmaier M, Toga AW, Blanton RE, Sowell ER, Lavretsky H, Peterson J, et al. (2004): Anterior cingulate, gyrus rectus, and orbitofrontal abnormalities in elderly depressed patients: An MRI-based parcellation of the prefrontal cortex. Am J Psychiatry 161:99 –108. 7. Lesser IM, Mena I, Boone KB, Miller BL, Mehringer CM, Wohl M (1994): Reduction of cerebral blood flow in older depressed patients. Arch Gen Psychiatry 51:677– 686. 8. Vaishnavi S, Taylor WD (2006): Neuroimaging in late-life depression. Int Rev Psychiatry 18:443– 451. 9. Teasdale JD, Howard RJ, Cox SG, Ha Y, Brammer MJ, Williams SC, Checkley SA (1999): Functional MRI study of the cognitive generation of affect. Am J Psychiatry 156:209 –215. 10. Knutson B, Fong GW, Adams CM, Varner JL, Hommer D (2001): Dissociation of reward anticipation and outcome with event-related fMRI. Neuroreport 12:3683–3687. 11. Kumari V, Mitterschiffthaler MT, Teasdale JD, Malhi GS, Brown RG, Giampietro V, et al. (2003): Neural abnormalities during cognitive generation of affect in treatment-resistant depression. Biol Psychiatry 54:777– 791. 12. Mitterschiffthaler MT, Kumari V, Malhi GS, Brown RG, Giampietro VP, Brammer MJ, et al. (2003): Neural response to pleasant stimuli in anhedonia: An fMRI study. Neuroreport 14:177–182. 13. Keedwell PA, Andrew C, Williams SC, Brammer MJ, Phillips ML (2005): A double dissociation of ventromedial prefrontal cortical responses to sad and happy stimuli in depressed and healthy individuals. Biol Psychiatry 58:495–503. 14. Liotti M, Mayberg HS, McGinnis S, Brannan SL, Jerabek P (2002): Unmasking disease-specific cerebral blood flow abnormalities: Mood challenge in patients with remitted unipolar depression. Am J Psychiatry 159:1830 –1840. 15. Bench CJ, Frackowiak RS, Dolan RJ (1995): Changes in regional cerebral blood flow on recovery from depression. Psychol Med 25:247–261. 16. Kennedy SH, Evans KR, Kruger S, Mayberg HS, Meyer JH, McCann S, et al. (2001): Changes in regional brain glucose metabolism measured with positron emission tomography after paroxetine treatment of major depression. Am J Psychiatry 158:899 –905.
S. Brassen et al. 17. Mayberg HS, Brannan SK, Tekell JL, Silva JA, Mahurin RK, McGinnis S, Jerabek PA (2000): Regional metabolic effects of fluoxetine in major depression: Serial changes and relationship to clinical response. Biol Psychiatry 48:830 – 843. 18. Kennedy SH, Konarski JZ, Segal ZV, Lau MA, Bieling PJ, McIntyre RS, Mayberg HS (2007): Differences in brain glucose metabolism between responders to CBT and venlafaxine in a 16-week randomized controlled trial. Am J Psychiatry 164:778 –788. 19. Fu CH, Williams SC, Cleare AJ, Brammer MJ, Walsh ND, Kim J, et al. (2004): Attenuation of the neural response to sad faces in major depression by antidepressant treatment: A prospective, event-related functional magnetic resonance imaging study. Arch Gen Psychiatry 61:877– 889. 20. Robertson B, Wang L, Diaz MT, Aiello M, Gersing K, Beyer J, et al. (2007): Effect of bupropion extended release on negative emotion processing in major depressive disorder: A pilot functional magnetic resonance imaging study. J Clin Psychiatry 68:261–267. 21. Maddock RJ, Garrett AS, Buonocore MH (2003): Posterior cingulate cortex activation by emotional words: fMRI evidence from a valence decision task. Hum Brain Mapp 18:30 – 41. 22. Lewis PA, Critchley HD, Rotshtein P, Dolan RJ (2007): Neural correlates of processing valence and arousal in affective words. Cereb Cortex 17:742– 748. 23. Nakic M, Smith BW, Busis S, Vythilingam M, Blair RJ (2006): The impact of affect and frequency on lexical decision: The role of the amygdala and inferior frontal cortex. Neuroimage 31:1752–1761. 24. Kuchinke L, Jacobs AM, Grubich C, Vo ML, Conrad M, Herrmann M (2005): Incidental effects of emotional valence in single word processing: An fMRI study. Neuroimage 28:1022–1032. 25. Modrego PJ, Ferrandez J (2004): Depression in patients with mild cognitive impairment increases the risk of developing dementia of Alzheimer type: A prospective cohort study. Arch Neurol 61:1290 –1293. 26. Montgomery SA, Asberg M (1979): A new depression scale designed to be sensitive to change. Br J Psychiatry 134:382–389. 27. Sheikh JI, Yesavage JA (1986): Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version. Clin Gerontol 37:819 – 820.
BIOL PSYCHIATRY 2008;64:349 –355 355 28. Lebowitz BD, Pearson JL, Schneider LS, Reynolds CF 3rd, Alexopoulos GS, Bruce ML, et al. (1997): Diagnosis and treatment of depression in late life. Consensus statement update. JAMA 278:1186 –1190. 29. Kumar A, Mintz J, Bilker W, Gottlieb G (2002): Autonomous neurobiological pathways to late-life major depressive disorder: Clinical and pathophysiological implications. Neuropsychopharmacology 26:229 –236. 30. Aizenstein HJ, Butters MA, Figurski JL, Stenger VA, Reynolds CF 3rd, Carter CS (2005): Prefrontal and striatal activation during sequence learning in geriatric depression. Biol Psychiatry 58:290 –296. 31. de Asis JM, Stern E, Alexopoulos GS, Pan H, Van Gorp W, Blumberg H, et al. (2001): Hippocampal and anterior cingulate activation deficits in patients with geriatric depression. Am J Psychiatry 158:1321–1323. 32. Ongur D, An X, Price JL (1998): Prefrontal cortical projections to the hypothalamus in macaque monkeys. J Comp Neurol 401:480 –505. 33. Carmichael ST, Price JL (1995): Limbic connections of the orbital and medial prefrontal cortex in macaque monkeys. J Comp Neurol 363:615– 641. 34. Elliott R, Friston KJ, Dolan RJ (2000): Dissociable neural responses in human reward systems. J Neurosci 20:6159 – 6165. 35. Harvey PO, Pruessner J, Czechowska Y, Lepage M (2007): Individual differences in trait anhedonia: A structural and functional magnetic resonance imaging study in non-clinical subjects. Mol Psychiatry 12:703, 767–775. 36. Quirk GJ, Beer JS (2006): Prefrontal involvement in the regulation of emotion: Convergence of rat and human studies. Curr Opin Neurobiol 16:723–727. 37. Urry HL, van Reekum CM, Johnstone T, et al. (2006): Amygdala and ventromedial prefrontal cortex are inversely coupled during regulation of negative affect and predict the diurnal pattern of cortisol secretion among older adults. J Neurosci 26:4415– 4425. 38. Siegle GJ, Thompson W, Carter CS, Steinhauer SR, Thase ME (2007): Increased amygdala and decreased dorsolateral prefrontal BOLD responses in unipolar depression: Related and independent features. Biol Psychiatry 61:198 –209. 39. Rainer MK, Mucke HA, Zehetmayer S, Krampla W, Kuselbauer T, Weissgram S, et al. (2006): Data from the VITA Study do not support the concept of vascular depression. Am J Geriatr Psychiatry 14:531–537.
www.sobp.org/journal
Key Words: Aging, emotion, functional neuroimaging, geriatric depression, vmPFC
L
ate-life depression (LLD) is often accompanied by marked disability, diminished quality of life, caregiver burden, and mortality. Nevertheless, the most prevalent psychological disorder in old age is still under-recognized, making it a significant health care issue (1,2). Accordingly, in contrast to a broad research in midlife depression, there are surprisingly few functional studies dealing with the neural abnormalities in LLD, although it seems plausible that in the context of age-related brain changes (3,4) functional findings in younger adults might not necessarily generalize to older persons (5). The insufficient investigation of LLD patients is certainly also a consequence of the strong heterogeneity of this sample regarding medical history and comorbidity, which complicate a defined examination of the mood disorder itself. Age-related brain changes are particularly pronounced in the prefrontal cortex (3). Furthermore, recent findings indicate additional grey matter deficits (6) as well as hypoactivation during resting state (7,8) in regions of the ventromedial prefrontal cortex (vmPFC) in LLD patients compared with healthy elderly control
From the NeuroImage Nord, Institute for Systems Neuroscience (SB, RK, WW-F, CB), Department of Psychiatry and Psychotherapy (DFB), University Medical Center Hamburg-Eppendorf, Hamburg, Germany. Drs. Braus and Büchel contributed equally to this work. Address reprint requests to Stefanie Brassen, Ph.D., NeuroImage Nord, Institute for Systems Neuroscience (S10), University Medical Center Hamburg-Eppendorf, Martinistrasse 52, D-20246 Hamburg, Germany; E-mail: [email protected]. Received April 18, 2007; revised March 20, 2008; accepted March 20, 2008.
0006-3223/08/$34.00 doi:10.1016/j.biopsych.2008.03.022
subjects. The vmPFC, which comprises the orbitofrontal cortex, the rostral cingulate cortex, and the ventral medial frontal gyrus, has been associated with the evaluation and regulation of emotional events and stimuli (9) and the contextual processing of reward (10). Functional imaging studies in midlife depression revealed an abnormal neural response in this region during the perception of affective stimuli. Here, depressed patients have shown an increased response to positive (11,12) and a decreased response to negative stimuli (13,14). In addition, the few existing longitudinal studies in midlife depression have proposed vmPFC dysfunction as a state marker sensitive to remission (15–20). In contrast, the neural correlates of emotional processing and thus vmPFC functioning have not yet been studied in LLD. Given the reported structural and functional changes in this region in aging and LLD, we wondered whether a functional deficit during emotional evaluation akin to the one observed in younger depressive subjects could be demonstrated in patients suffering from isolated mild to moderate LLD and whether possible frontal dysfunctions in these patients are reversible. Because this is one of the first fMRI studies of emotional processing in LLD, we decided to examine the neural correlates of emotional appraisal, which might provide a starting point for further studies investigating other emotional processes typically affected during depression. With fMRI we assessed emotional appraisal in an explicitly homogeneous sample of elderly antidepressant-naïve female outpatients with isolated first-diagnosed mild to moderate depressive symptoms and in healthy elderly subjects. We used a valence decision task requiring the affective evaluation of emotional words, which has previously elicited vmPFC activation (21–24). Most importantly, the study was designed in a longitudinal fashion to also address the question of whether brain activation patterns represent state markers. BIOL PSYCHIATRY 2008;64:349 –355 © 2008 Society of Biological Psychiatry
350 BIOL PSYCHIATRY 2008;64:349 –355
Figure 1. Schematic display of the experimental design.
Methods and Materials Subjects Participants were 13 depressed elderly women (mean age 66.4 ⫾ 6.1 years) as well as 13 healthy age- and educationmatched women (65.6 ⫾ 6.1 years). Only patients with a clinically significant mild to moderate depressive episode (ICD10) beginning later than the age of 55, no antidepressant medication at present or in the past, and no history of neurological or other psychiatric diseases like anxiety disorders were included in the study. Pathological white matter changes were excluded with a fluid-attenuated inversion recovery (FLAIR) sequence. To exclude highly prevalent subclinical dementia in LLD (25), we additionally applied a detailed neuropsychological assessment to all participants (Supplement 1). After complete description of the study to the subjects, written informed consent complying with the local ethics committee was obtained. Objective and subjective symptom severity was assessed with the Montgomery-Asberg Depression-Rating-Scale (MADRS) (26) and the self-rated Geriatric-Depression-Scale (GDS) (27). All 13 patients were re-examined approximately 7 months (28.7 ⫾ 2.9 weeks) after their first visit. Patients were naturalistically treated— six of them with antidepressant drugs, two with behavior therapy; and five patients had not received any treatment (for details see Supplement 1). Twelve patients were clinically improved as also reflected by a significant reduction of the MADRS score; one patient showed no improvement. fMRI Paradigm. Inside the scanner, subjects were presented with 40 neutral, 40 positive, and 40 negative adjectives in randomized order (Figure 1). The trials were pseudorandomly intermixed with 40 null events (blank screen) to provide jitter. After replacement with a rating screen subjects had to indicate their emotional rating by pressing one of three buttons (positive, neutral, negative). For details of the functional imaging design see Supplement 1. www.sobp.org/journal
S. Brassen et al. Imaging. Details on imaging and statistical procedures are given in Supplement 1. In short, imaging was performed on a 3-Tesla whole-body MR-scanner (Siemens, Erlangen, Germany). Volumes of 42 transversal slices were obtained (3-mm-thickness, no gap, repetition time ⫽ 2.41 sec, echo time ⫽ 24 msec, ␣ ⫽ 80°, field-of-view 192 ⫻ 192 mm2, 64 ⫻ 64 matrix). At first-level, regressors of interest for subjects’ individually rated positive, negative, and neutral words were modeled separately with SPM2 (http://www.fil.ion.ucl.ac.uk/spm/). The 6 estimated movement parameters from the realignment stage were included as covariates. A second session was modeled for the follow-up scan. At the second level, contrast images representing differential contrasts of interest (“negative vs. neutral words,” “positive vs. neutral words,” “positive vs. negative words”) were statistically tested voxelwise in one- and two-sample t tests to assess within and between group effects within our a priori regions of interest (ROIs, see following text) with a threshold of p ⬍ .05 family-wise error (FWE) corrected. Correlation analyses were conducted with simple regression analyses implemented in SPM2 with a threshold of p ⬍ .001 uncorrected. Explorative data of within group analyses regarding the whole brain (p ⬍ .001 uncorrected) are available in Supplement 1. Areas of Interest. Given recent evidence of altered responses in midlife depression to affective material in the vmPFC, we based our correction for multiple comparisons on corresponding ROIs rather than on the whole brain. The ROIs were spheres of 10 mm radius around the following center coordinates: x ⫽ ⫺4, y ⫽ 58, z ⫽ ⫺6 (medial orbitofrontal cortex [mOFC]); ⫺3, 59, 14 (superior medial frontal cortex [sMFC]); ⫺2, 46, 10 (left); and 12, 43, 7 (left and right rostral anterior cingulate cortex [rACC]). Coordinates were chosen from previous studies reporting valence-dependent engagement of the vmPFC (21–24) and were transferred from the Talairach to Montreal Neurological Institute (MNI) space if necessary.
Results Behavioral Data At baseline, there were no significant differences in word valence categorizations, reaction times, or the different kinds of ratings changes (e.g., from negative to neutral; and vice versa) between both groups, and there were no changes in these parameters at follow-up in depressed patients (Table 1). Imaging Data Baseline. NEGATIVE VERSUS NEUTRAL. There was a more pronounced engagement of the mOFC for negative versus neutral words in control subjects. Patients showed significantly less activation to negative words in the bilateral rACC. Interaction analysis confirmed higher mOFC and right rACC activation to negative words in control subjects compared with patients (Table 2). Table 1. Rating Data
“Positive” Rating “Negative” Rating “Neutral” Rating RT “Positive” Rating RT “Negative” Rating RT “Neutral” Rating
Healthy
LLD Baseline
LLD Follow-Up
42.6 (7.3) 37.9 (2.6) 39.5 (7.6) 1.57 (.3) 1.57 (.3) 1.63 (.3)
41.6 (6.4) 37.0 (2.4) 41.4 (5.9) 1.51 (.3) 1.52 (.3) 1.55 (.3)
41.5 (3.6) 38.8 (1.5) 39.8 (3.5) 1.49 (.3) 1.54 (.3) 1.57 (.4)
Analyses of variance including a group factor revealed no differences between groups (p ⬍ .05). RT, reaction time; LLD, late-life depression.
BIOL PSYCHIATRY 2008;64:349 –355 351
S. Brassen et al. Table 2. Statistical Results of Within and Between Group Comparisons at Baseline Region
Hem
x
y
z
Z
x
Within Group: Neutral vs. Negative Healthy: Negative⬎Neutral Medial Orbitofrontal Cortex LLD: Neutral⬎Negative Rostral Anterior Cingulate Cortex
L
LLD: Positive⬎Negative Medial Orbitofrontal Cortex Superior Medial Frontal Cortex Rostral Anterior Cingulate Cortex
60
⫺3
z
Z
Group Interaction: LLD⬎Healthy (Neutral⬎Negative) 3.23a
⫺6 9 12
L R
⫺9 39 6 12 39 15 Within Group: Positive vs. Neutral
3.06b 4.25c
R
0 60 ⫺3 3 54 ⫺9 Within Group: Positive vs. Negative
3.36a 3.13a
LLD: Positive⬎Neutral Medial Orbitofrontal Cortex
Healthy: Negative⬎Positive Superior Medial Frontal Cortex Rostral Anterior Cingulate Cortex
⫺6
y
57
⫺6
36 9 39 15 No Group Interaction
3.21a 3.51a 3.49a
Group Interaction: LLD⬎Healthy (Positive⬎Negative)
R L R
3 ⫺3 9
54 51 48
18 18 15
3.59a 2.86b 2.94b
3
54
18
2.97b
L
⫺6 ⫺9 ⫺6 ⫺6 3
54 60 63 42 42
⫺9 0 6 6 6
3.84c 3.26a 4.25c 4.20c 3.40a
⫺3
60
⫺3
3.34a
3 ⫺3 ⫺6
54 63 42
⫺6 6 6
3.10a 3.56a 3.57a
L L R
The table shows peak activations in Montreal Neurological Institute space. The p values are corrected for multiple comparisons in a priori search volumes (regions of interest). Local maxima more than 4 mm apart. Hem, hemisphere; LLD, late-life depression. a p ⬍ .05. b p ⬍ .1. c p ⬍ .01 (all family-wise error corrected).
POSITIVE VERSUS NEUTRAL. There was a significantly higher activation of the mOFC for positive compared with neutral words in depressed subjects. No differences in activation for positive versus neutral were visible in control subjects in our a priori ROIs. POSITIVE VERSUS NEGATIVE. Control subjects showed a significantly stronger activation for negative compared with positive words in the sMFC and the bilateral rACC. By contrast, patients displayed a widespread stronger activation for positive compared with negative words in all regions of the vmPFC. Accordingly, all regions showed a significant group interaction (Table 2, Figure 2).
Correlations with Symptom Severity Owing to significant interaction analyses, simple regressions were applied for the contrasts neutral⬎negative and positive⬎negative in depressed patients. Reduced activation for negative compared with neutral words was positively correlated with GDS score in the mOFC (r ⫽ .78; p ⬍ .001). Increased activation for positive compared with negative words was positively correlated with GDS and MADRS scores (Figure 3). Follow-Up. At follow up, patients showed an increased activation for negative compared with neutral words in comparison with baseline in the sMFC cortex and the bilateral rACC. Furthermore, the widespread positive⬎negative differences in patients at baseline disappeared at follow-up. A direct comparison with control subjects at baseline revealed absence of patientto-control differences (Table 3, Figure 4). Normalization in the sMFC was positively correlated with MADRS score at baseline (Figure 5).
Discussion The key finding of this study is an abnormal neural response to positive and negative words in patients with mild to moderate LLD in regions of the vmPFC. In contrast to healthy subjects, patients showed a decreased response to negative compared with positive stimuli. This attenuated neural response in the vmPFC was positively associated with depression severity. After 7 months of naturalistic treatment, depressed patients’ symptoms had significantly improved and the abnormal neural pattern in the vmPFC had normalized. The normalization could be predicted by depression severity at baseline and supports the role of ventromedial prefrontal neural abnormalities as a cortical state marker of LLD. Late-life depression is an underdiagnosed depression subtype in primary care, probably because relatives and physicians often believe that mild to moderate depressive syndromes are rather normal for older people faced with an increasing impact of psychosocial adversity and a decline in physical health and cognitive abilities (2,28). We now report a biological marker in this patient group, which strongly argues against this (mis-)belief and underlines the pathological significance of the syndrome. This is further supported by an obvious analogy between our functional results and previous findings of abnormal vmPFC responses in younger patients (13). The analogy extends to an apparent reversibility of frontal changes that parallels remission processes after antidepressant medication (16,17,19,20) or cognitive behavior therapy (18) in midlife depression. Overall, this indicates similar pathophysiological mechanisms. The observed neural normalization is particularly interesting in the light of www.sobp.org/journal
352 BIOL PSYCHIATRY 2008;64:349 –355
S. Brassen et al.
Figure 2. Increased neural response to positive and decreased neural response to negative words in the ventromedial prefrontal cortex in depressed individuals compared with healthy subjects. LLD, late-life depression; rACC, rostral cingulate gyrus; mOFC, medial orbitofrontal cortex; sMFC, superior medial frontal cortex. p ⬍ .05 family-wise error corrected for small volumes.
frequent findings of frontal volume changes in LLD. Because volume changes are triggered by age (29) and are therefore probably progressive, it might have been speculated that functional alterations of the vmPFC in LLD are irreversible. Whether the observed normalization results from a lack of such volume changes in our study sample, from an independency of structural and functional brain alterations, or from compensatory processes
during remission will have to be clarified by future studies, including longitudinal structural data. We here chose to investigate emotional processing, in contrast to the few existent dynamic imaging studies in LLD (30,31), and focused our analysis on the vmPFC, which is a target structure of grey matter and metabolic changes in aging as well as LLD and which has consistently been shown to be dysfunc-
Figure 3. Scatterplots of correlations of objective (Montgomery-Asberg Depression-Rating-Scale [MADRS]; Pearson’s correlation coefficient r ⫽ .76; p ⬍ .001) and subjective (Geriatric-Depression-Scale [GDS]; r ⫽ .84; p ⬍ .001) symptom severity with increased activity for positive compared with negative words at the superior medial frontal cortex (sMFC) at baseline in depressed individuals.
www.sobp.org/journal
BIOL PSYCHIATRY 2008;64:349 –355 353
S. Brassen et al. Table 3. Statistical Results of Baseline vs. Follow-Up Comparisons in Patients Region
Hem
x
y
z
Z
Baseline⬎Follow-Up Superior Medial Frontal Cortex
L
Rostral Anterior Cingulate Cortex
L R
Superior Medial Frontal Cortex Rostral Anterior Cingulate Cortex
L L
Neutral ⬎ Negative ⫺6 51 18 ⫺9 66 9 ⫺9 48 15 3 39 12 Positive ⬎ Negative ⫺3 60 9 ⫺6 42 6 ⫺3 54 15
3.13a 2.92a 3.28b 2.92a 4.11c 3.44b 3.35b
The table shows peak activations in MNI space. The p values are corrected for multiple comparisons in a priori search volumes (ROIs). Local maxima more than 4 mm apart. Abbreviations as in Table 2. a p ⬍ .1 (all family-wise error corrected). b p ⬍ .05. c p ⬍ .01.
tional in younger depressed subjects (11,12). Through its mostly reciprocal anatomical connections with the amygdala, the hypothalamus, the striatum, and other limbic and brainstem structures (32,33), the vmPFC is theoretically in a position to process internal and external affective stimuli and to modulate affective responses. Functional studies in healthy subjects suggest that the vmPFC encodes the emotional meaning of stimuli (9). More
Figure 5. Scatterplot of the correlation between normalization in ventromedial prefrontal cortex (vmPFC) activation over time in the superior medial frontal cortex (sMFC) and symptom severity at baseline (r ⫽ .80; p ⬍ .001). Thus, a more severe depression at baseline was associated with a stronger normalization of vmPFC activation at follow-up. t0 ⫽ baseline; t1 ⫽ followup. MADRS, Montgomery-Asberg Depression-Rating-Scale.
specifically, the vmPFC has been implicated in the generation of an abstract representation of the rewarding value of a stimulus by attending to its context (34). In this line, previous findings of a double dissociation of vmPFC response to positive and negative stimuli in younger patients have been interpreted on the basis of a role of the vmPFC in the contextual appraisal of the rewarding
Figure 4. Activation changes in the ventromedial prefrontal cortex at follow-up compared with baseline for the contrast positive⬎negative in LLD patients. Normalization was found in the rACC and the sMFC. p ⬍ .05 FWE corrected for small volumes. Abbreviations as in Figure 2.
www.sobp.org/journal
354 BIOL PSYCHIATRY 2008;64:349 –355 potential of emotional stimuli (13,14,35). Keedwell et al. (13) speculated that a relatively increased response of the vmPFC to positive stimuli reflects the attempt of depressive subjects to attend more closely to the positive stimulus, owing to their reduced capacity to experience pleasure in pleasurable situations. This speculation might be supported by recent data relating increased vmPFC response to positive stimuli with anhedonia. Furthermore, the relatively decreased vmPFC response to negative stimuli in depressed patients has not only been associated with reduced negative feedback response in depression (14) but also with context- or state-dependent stimulus processing. It has been argued that the presentation of negative or sad stimuli to depressed patients who already are in a negative mood state is unlikely to be perceived as salient and to prompt an emotional response (13). By contrast, healthy normal or remitted subjects are likely to be more affected by a negative stimulus, which is discordant with the current mood state and thus requires attention and a more detailed cognitive elaboration. An alternative explanation for the observed findings in vmPFC could be based on its role in emotion regulatory processes. Animal studies suggest that the vmPFC and the amygdala inhibit each other and thus regulate emotional behavior (36). Recent human data also indicate a moderating function of vmPFC in cognitive-emotional regulation through its direct connections with lateral/dorsal PFC and with the amygdala (37,38). From this perspective, a relatively decreased vmPFC response to negative stimuli might reflect the lack of inhibitory influences on negative emotional processing performed by limbic structures like the amygdala. However, it is important to highlight that we did not observe amygdala activations for the negative stimuli (Supplement 1) and were thus unable to test a potential inverse relationship between the vmPFC and the amygdala. Most previous neurobiological studies in LLD have focused on structural changes, mainly in frontal brain regions (6,8). Their frequent finding of high-intensity lesions has been suggested as one primary neurobiological path of LLD leading to the concept of “vascular depression” in elderly persons (1). There are, however, recent data that do not support such an association (39). Notably, path analyses have revealed smaller brain volumes, particularly in frontal regions as second, distinct path leading to LLD (29). The comparability of such results is limited by the inherent heterogeneity of elderly depressed patients, especially with regard to premedication, onset of disease, previous episodes, symptom severity, and a high degree of comorbidity typically existent in old age. One potentially advantageous strategy would be to study brain abnormalities related to the isolated mood disorder before examining heterogeneous samples. In this respect, the homogeneous group of elderly outpatients examined here is particularly interesting. Our decision to study a homogeneous sample, however, also limits generalizability, for example to male populations. Yet, this approach helped us in generating a clear experimental hypothesis, which can now be tested in other subtypes of depression in aging (e.g., with and without mild cognitive impairment or cerebrovascular diseases). A second limitation lies in the naturalistic treatment of patients, which precludes any conclusions about the effects of specific treatments. Finally, we chose a very simple, low-demand task, which is easy to conduct in elderly patients yet shows robust medial frontal engagement (21). In particular, we only asked for a broad valence categorization instead of using a continuous self-relevant valence rating scale (e.g., from ⫺10 to 10). This might have occluded rating differences between patients and www.sobp.org/journal
S. Brassen et al. control subjects. Nevertheless, in our opinion this only underlines the sensitivity of vmPFC dysfunction as an early biological marker for depression. Altogether, our data point to the pathological significance of mild to moderate depressive symptoms in the elderly and thus to the need to treat these patients like younger subjects. In this context, functional imaging might be a useful tool to evaluate treatment response. The study was supported by a grant from “Forschungsförderung Medizin” F154 (SB), University Medical Center HamburgEppendorf. We thank K. Müller and F. Walde for help with MR scanning. The authors reported no biomedical financial interests or potential conflicts of interest. Supplementary material cited in this article is available online. 1. Alexopoulos GS (2005): Depression in the elderly. Lancet 365:1961– 1970. 2. Charney DS, Reynolds CF 3rd, Lewis L, Lebowitz BD, Sunderland T, Alexopoulos GS, et al. (2003): Depression and Bipolar Support Alliance consensus statement on the unmet needs in diagnosis and treatment of mood disorders in late life. Arch Gen Psychiatry 60:664 – 672. 3. Raz N, Gunning FM, Head D, Dupuis JH, McQuain J, Briggs SD, et al. (1997): Selective aging of the human cerebral cortex observed in vivo: Differential vulnerability of the prefrontal gray matter. Cereb Cortex 7:268 –282. 4. Vermeer SE, Longstreth WT Jr, Koudstaal PJ (2007): Silent brain infarcts: A systematic review. Lancet Neurol 6:611– 619. 5. Brassen S, Büchel C, Weber-Fahr W, Lehmbeck JT, Sommer T, Braus DF (2007): Structure-function interactions of correct retrieval in healthy elderly women [published online ahead of print November 17]. Neurobiol Aging doi:10.1016/j.neurobiolaging.2007.10.005. 6. Ballmaier M, Toga AW, Blanton RE, Sowell ER, Lavretsky H, Peterson J, et al. (2004): Anterior cingulate, gyrus rectus, and orbitofrontal abnormalities in elderly depressed patients: An MRI-based parcellation of the prefrontal cortex. Am J Psychiatry 161:99 –108. 7. Lesser IM, Mena I, Boone KB, Miller BL, Mehringer CM, Wohl M (1994): Reduction of cerebral blood flow in older depressed patients. Arch Gen Psychiatry 51:677– 686. 8. Vaishnavi S, Taylor WD (2006): Neuroimaging in late-life depression. Int Rev Psychiatry 18:443– 451. 9. Teasdale JD, Howard RJ, Cox SG, Ha Y, Brammer MJ, Williams SC, Checkley SA (1999): Functional MRI study of the cognitive generation of affect. Am J Psychiatry 156:209 –215. 10. Knutson B, Fong GW, Adams CM, Varner JL, Hommer D (2001): Dissociation of reward anticipation and outcome with event-related fMRI. Neuroreport 12:3683–3687. 11. Kumari V, Mitterschiffthaler MT, Teasdale JD, Malhi GS, Brown RG, Giampietro V, et al. (2003): Neural abnormalities during cognitive generation of affect in treatment-resistant depression. Biol Psychiatry 54:777– 791. 12. Mitterschiffthaler MT, Kumari V, Malhi GS, Brown RG, Giampietro VP, Brammer MJ, et al. (2003): Neural response to pleasant stimuli in anhedonia: An fMRI study. Neuroreport 14:177–182. 13. Keedwell PA, Andrew C, Williams SC, Brammer MJ, Phillips ML (2005): A double dissociation of ventromedial prefrontal cortical responses to sad and happy stimuli in depressed and healthy individuals. Biol Psychiatry 58:495–503. 14. Liotti M, Mayberg HS, McGinnis S, Brannan SL, Jerabek P (2002): Unmasking disease-specific cerebral blood flow abnormalities: Mood challenge in patients with remitted unipolar depression. Am J Psychiatry 159:1830 –1840. 15. Bench CJ, Frackowiak RS, Dolan RJ (1995): Changes in regional cerebral blood flow on recovery from depression. Psychol Med 25:247–261. 16. Kennedy SH, Evans KR, Kruger S, Mayberg HS, Meyer JH, McCann S, et al. (2001): Changes in regional brain glucose metabolism measured with positron emission tomography after paroxetine treatment of major depression. Am J Psychiatry 158:899 –905.
S. Brassen et al. 17. Mayberg HS, Brannan SK, Tekell JL, Silva JA, Mahurin RK, McGinnis S, Jerabek PA (2000): Regional metabolic effects of fluoxetine in major depression: Serial changes and relationship to clinical response. Biol Psychiatry 48:830 – 843. 18. Kennedy SH, Konarski JZ, Segal ZV, Lau MA, Bieling PJ, McIntyre RS, Mayberg HS (2007): Differences in brain glucose metabolism between responders to CBT and venlafaxine in a 16-week randomized controlled trial. Am J Psychiatry 164:778 –788. 19. Fu CH, Williams SC, Cleare AJ, Brammer MJ, Walsh ND, Kim J, et al. (2004): Attenuation of the neural response to sad faces in major depression by antidepressant treatment: A prospective, event-related functional magnetic resonance imaging study. Arch Gen Psychiatry 61:877– 889. 20. Robertson B, Wang L, Diaz MT, Aiello M, Gersing K, Beyer J, et al. (2007): Effect of bupropion extended release on negative emotion processing in major depressive disorder: A pilot functional magnetic resonance imaging study. J Clin Psychiatry 68:261–267. 21. Maddock RJ, Garrett AS, Buonocore MH (2003): Posterior cingulate cortex activation by emotional words: fMRI evidence from a valence decision task. Hum Brain Mapp 18:30 – 41. 22. Lewis PA, Critchley HD, Rotshtein P, Dolan RJ (2007): Neural correlates of processing valence and arousal in affective words. Cereb Cortex 17:742– 748. 23. Nakic M, Smith BW, Busis S, Vythilingam M, Blair RJ (2006): The impact of affect and frequency on lexical decision: The role of the amygdala and inferior frontal cortex. Neuroimage 31:1752–1761. 24. Kuchinke L, Jacobs AM, Grubich C, Vo ML, Conrad M, Herrmann M (2005): Incidental effects of emotional valence in single word processing: An fMRI study. Neuroimage 28:1022–1032. 25. Modrego PJ, Ferrandez J (2004): Depression in patients with mild cognitive impairment increases the risk of developing dementia of Alzheimer type: A prospective cohort study. Arch Neurol 61:1290 –1293. 26. Montgomery SA, Asberg M (1979): A new depression scale designed to be sensitive to change. Br J Psychiatry 134:382–389. 27. Sheikh JI, Yesavage JA (1986): Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version. Clin Gerontol 37:819 – 820.
BIOL PSYCHIATRY 2008;64:349 –355 355 28. Lebowitz BD, Pearson JL, Schneider LS, Reynolds CF 3rd, Alexopoulos GS, Bruce ML, et al. (1997): Diagnosis and treatment of depression in late life. Consensus statement update. JAMA 278:1186 –1190. 29. Kumar A, Mintz J, Bilker W, Gottlieb G (2002): Autonomous neurobiological pathways to late-life major depressive disorder: Clinical and pathophysiological implications. Neuropsychopharmacology 26:229 –236. 30. Aizenstein HJ, Butters MA, Figurski JL, Stenger VA, Reynolds CF 3rd, Carter CS (2005): Prefrontal and striatal activation during sequence learning in geriatric depression. Biol Psychiatry 58:290 –296. 31. de Asis JM, Stern E, Alexopoulos GS, Pan H, Van Gorp W, Blumberg H, et al. (2001): Hippocampal and anterior cingulate activation deficits in patients with geriatric depression. Am J Psychiatry 158:1321–1323. 32. Ongur D, An X, Price JL (1998): Prefrontal cortical projections to the hypothalamus in macaque monkeys. J Comp Neurol 401:480 –505. 33. Carmichael ST, Price JL (1995): Limbic connections of the orbital and medial prefrontal cortex in macaque monkeys. J Comp Neurol 363:615– 641. 34. Elliott R, Friston KJ, Dolan RJ (2000): Dissociable neural responses in human reward systems. J Neurosci 20:6159 – 6165. 35. Harvey PO, Pruessner J, Czechowska Y, Lepage M (2007): Individual differences in trait anhedonia: A structural and functional magnetic resonance imaging study in non-clinical subjects. Mol Psychiatry 12:703, 767–775. 36. Quirk GJ, Beer JS (2006): Prefrontal involvement in the regulation of emotion: Convergence of rat and human studies. Curr Opin Neurobiol 16:723–727. 37. Urry HL, van Reekum CM, Johnstone T, et al. (2006): Amygdala and ventromedial prefrontal cortex are inversely coupled during regulation of negative affect and predict the diurnal pattern of cortisol secretion among older adults. J Neurosci 26:4415– 4425. 38. Siegle GJ, Thompson W, Carter CS, Steinhauer SR, Thase ME (2007): Increased amygdala and decreased dorsolateral prefrontal BOLD responses in unipolar depression: Related and independent features. Biol Psychiatry 61:198 –209. 39. Rainer MK, Mucke HA, Zehetmayer S, Krampla W, Kuselbauer T, Weissgram S, et al. (2006): Data from the VITA Study do not support the concept of vascular depression. Am J Geriatr Psychiatry 14:531–537.
www.sobp.org/journal