ORIGINAL ARTICLES
    
    The Neural Correlates of Anhedonia in Major Depressive Disorder
    Paul A. Keedwell, Chris Andrew, Steven C.R. Williams, Mick J. Brammer, and Mary L. Phillips
    Background: Anhedonia is a relative lack of pleasure in response to formerly rewarding stimuli. It is an important diagnostic feature of major depressive disorder (MDD), and predicts antidepressant efficacy. Understanding its neurobiological basis may help to target new treatments and predict treatment outcomes. Using a novel paradigm, we aimed to explore the correlations between anhedonia severity and magnitude of neural responses to happy and sad stimuli in regions previously implicated in studies of human reward processing and depressive anhedonia. Methods: Neural responses to happy and sad emotional stimuli (autobiographical prompts and mood congruent facial expressions) were measured using blood oxygen level dependent (BOLD) functional magnetic resonance imaging in twelve MDD individuals with varying degrees of anhedonia. Results: In response to happy stimuli, anhedonia, but not depression severity per se, was positively and negatively correlated with ventromedial prefrontal cortex (VMPFC) and amygdala/ventral striatal activity, respectively. State anxiety independently contributed to a VMPFC-subcortical dissociation of response to happy (but not sad) stimuli, which was similar, but different, to anhedonia. Conclusions: These findings suggest that anhedonia and state anxiety are associated with dysfunction within neural systems underlying the response to, and assessment of, the rewarding potential of emotive stimuli in MDD, and highlight the importance of employing a symptom-dimension-based approach in the examination of the neurobiology of depression. Key Words: Major depression, fMRI, reward, mood induction, striatum, prefrontal cortex prefrontal cortex may respond to aversive rather than rewarding stimuli (O’Doherty et al 2001; Small et al 2003). Dysfunction within this neural system has been demonstrated in animal models of depression, and is reversed by antidepressants (Naranjo et al 2001). Furthermore, recent findings in individuals with MDD suggest supersensitive behavioural and pharmacological responses to d-amphetamine (Rampello et al 2000). Other studies have demonstrated that depressed subjects are less motivated to maximise their earnings in a monetary reward paradigm than healthy controls, with no group differences in response to punishment (losing money)(Henriques and Davidson 2000). There has, however, been relatively little study of the neural basis of anhedonia in humans. In one neuroimaging study, negative correlations were reported between anhedonia severity and response in subcortical regions, including the ventral striatum, and positive correlations with VMPFC activity (Dunn et al 2002). Here, individuals were engaged in a cognitive task rather than a reward-related task. Another study examined depressed individuals’ responses to pleasant stimuli from a standardized series, and demonstrated a pattern of increased VMPFC response in depressed individuals (Mitterschiffthaler et al 2003). A similar finding was observed using positive picture-caption pairs in MDD individuals (Kumari et al 2003). No studies to date have specifically examined neural responses to personally-relevant, rewarding stimuli in anhedonically-depressed individuals. One study has measured neural responses to sad autobiographical memory scripts in depressed individuals, demonstrating a decreased VMPFC response (Liotti et al 2002), but responses to happy memories were not examined. Previous work by this research team demonstrated a double dissociation of VMPFC response to happy and sad stimuli in depressed and healthy individuals: increased and decreased responses to happy and sad stimuli respectively in MDD, compared to neutral stimuli, but a reversed pattern of response in healthy volunteers (Keedwell et al 2003, Keedwell et al, in press). One interpretation of this finding is that the increased activity seen in this area in MDD to happy stimuli occurred as a result of attending more closely to the happy stimuli, in an attempt to get into happy mood. Sad BIOL PSYCHIATRY 2005;58:843– 853 © 2005 Society of Biological Psychiatry
    
    T
    
    he two most widely used diagnostic systems (ICD-10 (World Health Organization 1992) and DSM-IV (American Psychiatric Association 1994)) identify sadness, anergia and anhedonia as consistently important diagnostic features of MDD. However, only anhedonia, a relative failure to obtain pleasure from activities, or stimuli, previously experienced as rewarding (Fawcett et al 1983), may be both necessary and sufficient for a diagnosis (American Psychiatric Association 1994; Snaith 1993), and its presence has been shown to be predictive of antidepressant response (Klein 1974). The neural system underlying reward is increasingly welldefined in humans: dopamine release in the ventral caudate and putamen is correlated with euphoric response to dextro-amphetamine (Drevets et al 2001), and in neuroimaging studies cocaineinduced euphoria (Breiter et al 1997), monetary reward (Knutson et al 2001; O’Doherty et al 2001), pleasurable responses to music (Blood and Zatorre 2001), and viewing attractive facial expressions (Senior 2004), have all been associated with activity within nucleus accumbens, ventral caudate and ventral putamen. The ventral striatum may be involved in the anticipation and generation of motor responses associated with future rewards (Elliott et al 2004; Knutson et al 2001).The ventromedial prefrontal cortex (VMPFC) has been implicated in the generation of an abstract representation of the rewarding value of a stimulus by attending to its context (Elliott et al 2000), and the learning of contingencies based on the outcome of a rewarding situation (Knutson et al 2001). By contrast, lateral areas of the ventral
    
    From the Department of Psychological Medicine (PAK, MLP), Neuroimaging Research Group (CA, SCRW), and Department of Biostatistics (MJB), Institute of Psychiatry, London, United Kingdom. Address reprint requests to Dr. Paul Keedwell, M.B. Ch.B. M.R.C.Psych., Box 69, Section of Neuroscience and Emotion, Department of Psychological Medicine, Institute of Psychiatry, Decrespigny Park, London SE5 8AF, United Kingdom; E-mail: p.keedwell@iop.kcl.ac.uk. Received December 16, 2004; revised May 3, 2005; accepted May 10, 2005.
    
    0006-3223/05/$30.00 doi:10.1016/j.biopsych.2005.05.019
    
    844 BIOL PSYCHIATRY 2005;58:843– 853
    stimuli, on the other hand, were more congruent with the prevailing mood, and hence greater attention did not need to be paid to these stimuli. Given that anhedonia is a core feature of depression, one might predict that this abnormal increase in VMPC in MDD to happy stimuli would be particularly apparent in more anhedonic individuals. In the present study, we employed sad and happy autobiographical memory scripts to determine the extent to which depressive anhedonia is associated with an abnormal pattern of response in neural regions implicated in the limited number of studies specifically examining neural responses to pleasure in depressed individuals and the larger number of studies examining reward processing in healthy individuals. Based on previous research findings, we predicted a relatively greater VMPFC response to happy compared with sad stimuli, and positive and negative correlations between anhedonia severity and responses in VMPFC and ventral striatum, respectively, to happy stimuli (but not sad stimuli). For comparison, we also explored correlations between neural responses to these stimuli and other clinical measures, including depression severity per se, state anxiety, and on-line subjective mood response.
    
    P.A. Keedwell et al memories suggests that the probability of recall of happy events tends to be lower for depressed individuals than for normal individuals (Clark and Teasdale 1982; Lloyd and Lishman 1975). However, attempts were made to ensure recalled events were contextually specific; happy and sad events were similar in their detail and contextual specificity in both groups. No significant differences in subjective pleasure ratings were observed for positive life events reported by depressed subjects (t 1.2, df 9, p .28) and an age and sex matched control group of healthy volunteers recruited from a previous study (Keedwell et al, in press), as assessed by independent raters on a five point Likert scale. (Independent raters were ten healthy volunteers not otherwise involved in the study, blind to the source of the events. Events were presented in random order). Memory prompts, of up to 8 seconds duration, derived from the above procedure, were pre-recorded using standard digital software. They were designed to be played prior to the presentation of mood congruent facial expressions. Prompts took the form “Use the following sad faces to remember how you felt at your father’s funeral when your mother cried.” Facial expressions (100% happy, 100% sad or neutral) were selected from a standardized series (Ekman and Friesen 1976). Mood Provocation Paradigm. All subjects participated in two separate 6-minute experiments: in one experiment they were exposed to sad and neutral stimuli alternating with each other, and in the other happy and neutral stimuli alternated. There were ten 36-second alternating blocks in each experiment (see Figure 1). The order in which the 2 experiments were conducted was counterbalanced, as was the order of the emotional and neutral conditions within each experiment. Individuals heard the prerecorded memory prompts through headphones prior to viewing 8 different facial identities, all displaying moodcongruent (happy, sad or neutral) facial expressions (each stimulus duration 2.5 seconds). The latter were employed in order to help maintain attention on the task of mood induction over a 20 second period. Facial identities were presented in pseudo-randomized order to minimise habituation effects. Novel scripts were employed for each of the five ‘emotional’ or neutral experimental blocks. Individuals were then asked to give an oral subjective rating of mood from 5 (very sad) through zero (for neutral) to 5 for very happy. Similar techniques, convolving oral and visual stimuli to induce mood has been used successfully previously (Phan et al 2001; Schneider et al 1997). The interval between presentation of emotional and neutral stimuli was 16 seconds. This period was limited by consideration of the negative effects of prolonged experimentation, including fatigue, boredom, inattention and habituation. A previous off-line study observed an 80% reduction in aversive mood ratings 16 seconds after mood induction (Garrett and Maddock 2001). Hence it was anticipated that mood ratings to emotional stimuli would significantly reduce before presentation of the neutral stimuli. Image Acquisition Gradient-echo echoplanar images were acquired on a 1.5-T MRI system (GE Signa Neuro-optimized MR system; General Electric, Milwaukee, Wisconsin) at the Maudsley Hospital. One hundred T2*-weighted whole-brain volumes depicting blood oxygen level– dependent contrast 47 and consisting of 16 axial sections oriented according to the bicommissural plane (thickness, 7 mm; .7-mm gap) were acquired during 6 minutes for each of the 2 experiments (repetition time, 2.0 seconds; echo time, 40 milliseconds; field of view, 24 cm; flip angle, 70; 64 64 matrix).
    
    Methods and Materials
    Participants Twelve ICD-10 (World Health Organization 1992)-diagnosed individuals with MDD (mean age 43, SD 9.8) were recruited from the inpatient and outpatient departments of the South London and Maudsley National Health Service Trust, London. Exclusion criteria included cognitive deficits, history of brain injury or structural brain abnormality, and an additional Axis 1 diagnosis. There were eight females and four males. Two female individuals were left-handed, The remaining 10 individuals were righthanded. Eleven depressed subjects were taking antidepressants: these included venlafaxine 75–375 mg (n 4), citalopram 40 mg, lofepramine 210 mg, sertraline 100 mg, trazadone 300 mg, fluoxetine 40 mg, mirtazepine 45 mg, and dothiepin 150 mg. The study was approved by the South London and Maudsley Ethics Committee. All individuals gave valid, informed consent to participate. Measures MDD individuals completed the Beck Depression Inventory (BDI; Beck et al 1961), and a modified form of the self-rated Fawcett-Clarke Pleasure scale (FCPS; Fawcett et al 1983), which provided a quantitative measure of anhedonia. They also completed a measure of state anxiety just prior to scanning (the Spielberger State Anxiety Inventory, STAI; Spielberger 1983). Procedure Stimuli. All individuals completed a life events questionnaire, asking them to describe five positive, five negative and five neutral previous life experiences that had made them feel particularly happy, sad or emotionally indifferent, respectively. Personal memories have been shown to be reliable inducers of mood in clinical settings (Beck 1967) and in scanning environments (Drevets et al 2001). They have the advantage over standardized stimuli, (like International Affective Picture System (IAPS); Lang et al 1997), of being meaningful to each individual, respecting individual differences in the emotional meaning attached to similar events (Brown and Harris 1978). This is the first study of its kind to employ happy memories as mood inducing stimuli in depressed individuals. Previous work on emotional www.sobp.org/journal
    
    P.A. Keedwell et al
    
    BIOL PSYCHIATRY 2005;58:843– 853 845
    
    Figure 1. Experimental design.
    
    This echoplanar image data set provided almost complete brain coverage. Ten whole-brain volumes were acquired during each stimulus presentation block. Each stimulus block was followed by (1) an 8-second period of complete silence during which subjects were asked to rate their mood and (2) an additional 8-second period during which the subjects listened to a sound file containing instructions pertinent to the next stimulus block. Four “dummy volumes” were excited during this 8-second period by means of exactly the same radiofrequency envelope and gradient section selection parameter, with the same repetition time of 2 seconds to allow the magnetization to reach an equilibrium amplitude before the next period of data acquisition. The frequencyencoding gradient was turned off during this period to minimize acoustic noise and ensure that the instructions were heard clearly by the subjects. The 4 dummy volumes were later discarded from the time series. Individual brain activation maps were co-registered to a “whole head” gradient-recalled echo planar imaging scan of superior spatial resolution acquired on all subjects. This structural scan had the following acquisition parameters: echo time,
    
    40 milliseconds; repetition time, 3000 milliseconds; field of view, 24 cm; image resolution, 128 x 128; number of sections, 43; section thickness, 3.0 mm; intersection gap, .3 mm; number of signal averages, 8. Functional MRI Data Analysis Individual Maps. Data were first realigned (Bullmore et al 1999) to minimize motion-related artifacts and smoothed by means of a gaussian filter (full width at half maximum, 7.2 mm). Responses to the experimental paradigms were then detected by time-series analysis using gamma variate functions (peak responses weighted between 4 and 8 seconds) convolved with the experimental design to model the blood oxygen level– dependent response. A goodness-of-fit statistic and a measure of the mean power of neural response (the sum of squares [SSQ] ratio) was computed at each voxel. This was the ratio of the sum of squares of deviations from the mean intensity value due to the model (fitted time series) divided by the sum of squares due to the residuals (original time series minus model time series). To sample the distribution of SSQ ratio under the null hypothesis that observed values of SSQ ratio were not determined by
    
    Table 1. Clinical Variables in Depressed Individuals: Mean Values and Correlations Sad Rating Sad Rating BDI Anhedonia STAI Neg Cog Mean (SD) r Sig. (2-tailed) r Sig. (2-tailed) r Sig. (2-tailed) r Sig. (2-tailed) r Sig. (2-tailed) 1.000 . .044 .892 .851a .000 .018 .957 .163 .612 3.2 (1.6) BDI .044 .892 1.000 . .011 .974 .042 .897 .984a .000 33.5 (11.2) Anhedonia .851 .000 .011 .974 1.000 . .168 .601 .137 .672 63.3 (25.7)
    a
    
    STAI .018 .957 .042 .897 .168 .601 1.000 . .077 .811 40.2 (17.7)
    
    Neg Cog .163 .612 .984a .000 .137 .672 .077 .811 1.000 . 13.1 (6.3)
    
    Sad rating mean subjective rating of sad mood in response to sad stimuli subtracted from mean value for neutral stimuli (no significant correlations with happy mood ratings were found). Anhedonia, Score on modified Fawcett-Clarke Pleasure Scale; BDI, Beck Depression Inventory Score; STAI, Score on Spielberger State Anxiety Inventory; Neg Cog, Negative cognition Score derived from BDI (see text). a Correlation is significant at the .01 level (2-tailed). r correlation coefficient.
    
    www.sobp.org/journal
    
    846 BIOL PSYCHIATRY 2005;58:843– 853
    Table 2 (A). Positive Correlations Between Clinical Variables and Neural Responses to Happy (vs. Neutral) Stimuli (p Area Description (with Brodmann Areas) Anhedonia L middle temporal gyrus (BA21) Bilateral anterior cingulate gyrus BA24/32 R posterior cingulate BA31 Bilateral VMPFC (medial frontal gyrus) BA10 Bilateral dorsomedial PFC (superior temporal gyrus) BA9/10 R dorsomedial PFC (superior frontal gyrus) BA9/10 Bilateral cerebellum L orbitofrontal cortex (medial frontal gyrus) BA11 Dorsomedial PFC (medial frontal gyrus) BA 32 L superior temporal gyrus (BA22) BDI Left dorsomedial PFC BA24/32 Left post central gyrus BA2 Anxiety Bilateral VMPFC BA10 R orbitofrontal cortex Bilateral dorsomedial PFC (superior frontal gyrus) BA10 L superior temporal gyrus BA21 Subjective Happy Ratings Bilateral putamen Bilateral uncus L precentral gyrus R cerebellum L insula R precentral gyrus BA6 L dorsolateral BA9 R ventrolateral PFC (inferior frontal gyrus) BA47 L middle cingulate gyrus Cluster Size (total) Cluster Size ( .001)
    
    P.A. Keedwell et al
    
    Talairach Coordinates x, y, z)
    
    23 15 13 10 8 7 7 3 2 2 6 2 35 16 12 8 14 7 6 6 5 4 3 2 2
    
    2 ( 54 11 2); 2 ( 54 11 7); 2 ( 51 4 13); 5 ( 43 0 7 ( 61 26 13); 5 ( 47 4 24) 6 (4 41 9); 2 ( 7 41 2); 4 (4 41 4); 3 (7 44 2) 8 (4 33 31); 5 (0 33 26) 4 (4 56 4); 3 ( 7 44 7); 3 (4 56 2) 4 (7 56 9); 2 ( 4 56 20); 2 (4 41 20) 7 (11 52 15) 3 (7 67 7); 2 ( 51 4 29); 2 (7 81 18) 3 ( 7 44 13) 2 (14 22 37) 2 ( 54 11 4) 4 ( 25 4 42); 2 ( 22 2 ( 40 33 53) 7 42)
    
    18);
    
    12 (0 48 4); 8 (14 56 4); 6 (18 52 7); 5 (0 48 6 (18 48 13); 6 (32 48 18); 4 (7 52 18) 10 (18 56 15); 2 ( 4 59 15) 8 ( 61 22 2)
    
    7); 2 (4 56
    
    2); 2 (18 59
    
    2)
    
    4 ( 22 4 2); 4 ( 29 0 4); 3 ( 22 7 9); 3 (22 11 5 ( 14 11 24); 2 (22 4 18) 2 ( 43 11 48); 2 ( 40 7 37); 2 ( 43 7 42) 4 (7 81 29); 2 (7 85 35) 3 ( 29 0 15); 2 ( 29 22 15) 4 (32 0 37) 3 ( 32 0 42) 2 ( 22 11 18) 2 ( 18 22 26)
    
    7)
    
    experimental design (with minimal assumptions), the time series at each voxel was permuted by a wavelet-based resampling method (Breakspear et al 2003; Bullmore et al 2001). This process was repeated 10 times at each voxel to produce the distribution of SSQ ratios under the null hypothesis. Voxels activated at any desired level of type I error can then be determined by obtaining the appropriate critical value of SSQ ratio from the null distribution. Individual brain activation maps were produced for each subject for each experiment versus the neutral condition. Group Maps. To extend inference to the group level, the observed and randomized SSQ ratio maps were transformed into standard space (Talairach and Tournoux 1998) by a 2-stage process (Brammer et al 1997) using spatial transformations computed for each subject’s high-resolution structural scan. Once the statistic maps were in standard space, a generic brain activation map (GBAM) was produced for each experimental condition by using a wavelet-permutation method (Brammer et al 1997). Within group comparisons of response to different stimuli were carried out using an analysis of covariance method as described by Bullmore et al (1999). Statistical thresholds were set at a level (voxelwise p value of .05 and a clusterwise p value of .005) which ensured less than 1 false positive cluster. Partial Correlation Analyses. The Pearson product moment correlation coefficient was calculated at each voxel between the standardized power of the fMRI response (SSQ ratio) and the behavioral variable for each subject. The null distribution of correlation coefficients was then computed by randomly permuting group membership (see previous section) and recomputing www.sobp.org/journal
    
    the correlation coefficient in an analogous fashion to that used for computation of group differences. Cluster level maps of significant correlations were then computed as described by Bullmore et al (1999). The voxelwise and clusterwise p values were set at .05 and .001, respectively, ensuring total number of false positives close to zero. For each of the significant clusters identified by the above method, we next extracted the SSQ ratio of each participant and conducted a series of partial correlation analyses with the relevant questionnaire measures. Initial correlation results included correlations with responses to both emotional and neutral stimuli, which were difficult to interpret. Hence, we eliminated any voxels in the correlation maps which did not correspond with voxels highlighted in the corresponding GBAMs of response to emotional (versus neutral) stimuli. This effectively restricted our results to emotional stimulus-related neural responses only.
    
    Results
    Clinical Variables Depressed individuals were in the moderate to severe range of depression severity according to the BDI and had a wide range of anhedonia scores. There were no significant correlations between anhedonia, depression severity and state anxiety (See Table 1). Subjective Mood Ratings Mean subjective mood ratings were significantly increased during the mood conditions when compared with the neutral
    
    P.A. Keedwell et al
    
    BIOL PSYCHIATRY 2005;58:843– 853 847
    
    Figure 2. (A) Positive correlations (p .001) between anhedonia and (i) VMPFC responses to all stimuli over entire epoch and (ii) VMPFC responses which were greater to happy stimuli than neutral stimuli (intersection with happy neutral GBAM, see text). Scatter plots show anhedonia score on y axis, and SSQ values on x axis, the latter extracted for each individual from a GBAM cluster overlapping with the correlation map (see text). (B) Negative correlations (p .001) between anhedonia and (i) subcortical responses to all stimuli over entire epoch and (ii) subcortical responses which were greater to happy stimuli than neutral stimuli (intersection with happy neutral GBAM, see text). Scatter plots show anhedonia score on y axis, and SSQ values on x axis, the latter extracted for each individual from a GBAM cluster overlapping with the correlation map (see text). Images are nonradiological. Coordinates given are Talairach Coordinates (x, y, z). VMPFC, ventromedial prefrontal cortex; GBAM, generic brain activation map; SSQ, sum of squares.
    
    conditions across all depressed individuals (F 65.7, p .000, df 1, 11). The mean change was 3.2 points, SD 1.6 for sad, 3.3 points, SD 1.7 for happy. There was no interaction between experiment type (happy-neutral or sad-neutral) and change in subjective ratings as a result of mood induction (F .23, p .882, df 1, 11). Generic Brain Activation Maps Numerous brain regions were consistently activated more by the emotional than neutral stimuli, including the insula, primary visual cortex, cerebellum, superior temporal gyrus and dorsal prefrontal cortex. Full details (coordinates, cluster size) of the group activation maps are available on request. Neural Responses To Happy Versus Sad Stimuli Group maps of response versus neutral (GBAMs) were performed in the first instance for sad and happy stimuli. Areas activated more by happy stimuli than neutral stimuli included bilateral orbitofrontal cortex (BA11), R posterior cingulate gyrus (BA31), L thalamus, L middle and L superior temporal gyrus (BA21,22), R hippocampus, bilateral cerebellum (BA71), bilateral frontal pole (BA10), R insula, R rostral anterior cingulate gyrus (BA16), bilateral medial frontal lobe (BA32) and R ventrolateral prefrontal cortex (BA47). Areas activated more by sad stimuli than neutral stimuli included R insula, R superior temporal gyrus (BA22), R inferior frontal (broca’s) area (BA44), L primary visual cortex (BA19), R cerebellum, L precuneus (BA7), and R supramarginal gyrus (BA40). Further information is available from the authors on request. The main focus of this study was to examine
    
    correlations between these responses and clinical dimensions rather than the neural responses per se. A second order of analysis (within subject comparison) was performed to detect those neuronal responses significantly differentiated by the different stimuli at the whole brain level (cluster-wise significance p .005). Neural responses were significantly greater to happy compared with sad stimuli in the R VMPFC, BA10/32, (cluster size 237, Talairach (Talairach and Tournoux 1998) coordinates 10, 50, 1), extending bilaterally, ventrally to BA11 and dorsally to BA9. Neural responses to sad compared with happy stimuli were significantly greater in the right inferior frontal gyrus, BA44 (cluster size 82; Talairach coordinates 47, 16, 7), extending to BA47. Further analysis revealed that there were relative increases and decreases in responses to happy and sad stimuli, respectively, in the VMPFC. Correlations Between Neural Responses and Clinical Variables There were no significant correlations between the clinical ratings of anhedonia, BDI and STAI (see Table 1). Separate correlational analyses were therefore performed between each of these clinical ratings and neural responses as indicated in the Methods. Primary Analyses – Anhedonia Happy versus Neutral Stimuli. Intersections of correlation maps and the group map (GBAM) of responses to happy versus neutral stimuli revealed positive correlations between anhedonia and neural responses in bilateral orbitofrontal cortex (BA11), right VMPFC (BA10), left middle temporal gyrus (BA21) and right www.sobp.org/journal
    
    848 BIOL PSYCHIATRY 2005;58:843– 853
    
    P.A. Keedwell et al
    
    Table 2 (B). Negative Correlations between Clinical Variables and Neural Responses to Happy (vs. Neutral) Stimuli (p .001) Area Description (with Brodmann Areas) Anhedonia L insula R cerebellum R occipital cortex Bilateral putamen R posterior cingulate gyrus R amygdala R lateral sulcus BA38/47 L anterior caudate L middle cingulate gyrus BDI Bilateral thalamus Bilateral dorsal cingulate gyrus (BA24) L pornix L amygdala Bilateral caudate nucleus L middle temporal lobe BA21 L inferior temporal lobe BA 28 Substantia nigra L putamen L nucleus accumbens L uncus L superior temporal lobe BA 38 Posterior cingulate BA31 Anxiety Left putamen Left amygdala Subjective Happy Ratings Bilateral ventral anterior cingulate gyrus BA24 R VMPFC BA10 R orbitofrontal cortex R posterior cingulate gyrus R orbital gyrus BA19 Cluster Size (total) Cluster Size ( Talairach Coordinates x, y, z)
    
    9 8 5 5 4 2 2 2 2 22 14 8 6 5 4 4 3 2 2 2 2 2 9 2 36 14 8 3 2
    
    4 ( 29 4 15); 3 ( 29 22 15); 2 ( 32 4 6 (18 89 29); 2 (7 74 35) 5 (22 67 15) 3 ( 25 4 9); 2 (18 13 1) 2 (14 22 31); 2 (14 22 26) 2 (22 4 13) 2 (25 15 13) 2 ( 12 19 4) 2 ( 7 4 31)
    
    7)
    
    15 (14 22 15); 7 ( 7 11 9) 4 (4 15 37); 3 ( 4 0 37); 4 (4 37 20); 3 (11 7 37) 8 ( 4 11 15) 6 ( 25 4 18) 3 (18 4 20); 2 ( 4 7 2) 2 ( 36 4 13); 2 ( 40 0 29) 4 ( 32 4 24) 3 ( 11 22 7) 2 ( 22 11 2) 2 ( 7 7 7) 2 ( 29 4 29) 2 ( 36 0 7) 2 (0 37 26) 9 ( 25 2 ( 25 7 7 2) 13) 2); 8 (4 41 9)
    
    14 (0 37 4); 14 (0 41 14 (4 44 7) 3 (7 48 13); 3 (4 48 3 (4 44 31) 2 (32 74 26)
    
    18); 2 (32 41
    
    13)
    
    BA, Brodmann Areas; VMPFC, ventromedial prefrontal cortex; BDI, Beck Depression Inventory Score; R, right; L, left.
    
    ventral anterior cingulate gyrus (BA24/32; see Table 2A; Figure 2A). For all neural activity across the epoch (happy and neutral stimuli), positive correlations were found with activity in a larger area of the VMPFC, extending to a larger area of the anterior cingulate, and the orbitofrontal cortex (see Figure 2A). Negative correlations were observed between anhedonia severity and neural responses in left insula, left anterior caudate, bilateral putamen, right amygdala, and left visual cortex (BA18; Table 2B; Figure 2B). For all neural activity across the epoch (happy and neutral stimuli), negative correlations were found with activity in larger areas of the striatum, including more extensive involvement of the bilateral anterior caudate and bilateral putamen (see Figure 2B). Sad versus Neutral Stimuli. Intersections of correlation maps and the group map (GBAM) of responses to sad versus neutral stimuli, positive correlations were observed between anhedonia severity and neural responses in the right ventral anterior cingulate (BA24), right middle temporal gyrus (BA21) and right superior temporal gyrus (BA40; see Table 3A). Negative correlations were observed between anhedonia severity and neural responses in the right inferior frontal gyrus (BA45) and left inferior frontal gyrus (BA44; see Table 3B). www.sobp.org/journal
    
    Secondary Analyses – Other Clinical Variables Positive and negative correlations between the other clinical variables (including BDI score, anxiety and subjective mood ratings) and neural responses to sad and happy stimuli were also analysed. Intersections of correlation maps and the group map (GBAM) of responses to emotional versus neutral stimuli were computed as above. The results are shown in Tables 2–3 and Figures 3 and 4 (where indicated). Happy versus Neutral Stimuli. With regard to happy versus neutral stimuli, global depression severity was correlated positively only with dorsal prefrontal cortex responses, but negatively with responses within regions associated with reward processing, including putamen, caudate nucleus, nucleus accumbens and amygdala (see Figure 3A). State anxiety correlated positively with VMPFC responses (Figure 3B) and negatively with putamen responses. Happy mood ratings were negatively correlated with neural responses in VMPFC regions similar to those positively correlated with anhedonia (Figure 2A and Figure 4), and positively correlated with responses within putamen and insula. Sad versus Neutral Stimuli. With regard to sad stimuli, no significant correlations, positive or negative, were demonstrated
    
    P.A. Keedwell et al
    
    BIOL PSYCHIATRY 2005;58:843– 853 849
    .001) Talairach Coordinates x, y, z)
    
    Table 3 (A). Positive Correlations between Clinical Variables and Neural Responses to Sad (vs. Neutral) Stimuli (p Area Description (with Brodmann Areas) Anhedonia R middle temporal gyrus BA21 R putamen R supramarginal gyrus BA40 BDI R ventrolateral PFC (inferior frontal gyrus) BA47 R postcentral gyrus BA40 R superior temporal gyrus BA42 R dorsolateral PFC (middle frontal gyrus) BA45/4 L posterior cingulate gyrus BA29/BA30 Bilateral Inferior parietal lobule BA40 L subthalamic nucleus R precentral gyrus BA6 L superior temporal gyrus BA22 L thalamus L hippocampus Anxiety Bilateral cerebellum R insula Bilateral inferior parietal lobule BA36/40 R precuneus BA7/31 L Dorsal cingulate gyrus (BA23/24/31) Bilateral thalamus R paracentral lobule BA7 L postcentral gyrus L Medial frontal gyrus BA6 L parahippocampal gyrus L precuneus BA7 R middle temporal gyrus BA21 R dorsomedial PFC BA31 SAD Subjective Mood Ratings R superior temporal gyrus BA21 L post central gyrus BA40 R lingual gyrus BA19 L caudate nucleus Cluster Size (total) Cluster Size (
    
    5 4 2 20 17 14 9 4 4 3 3 2 2 2 58 23 23 21 17 12 7 6 5 2 2 2 2 7 4 4 3
    
    3 (43 0 24); 2 (43 0 4 (32 4 4) 2 (51 52 26)
    
    18)
    
    13 (47 22 13); 5 (51 19 2); 2 (51 30 7) 17 (54 22 15) 9 (54 30 9); 5 (54 15 9) 2 (43 41 9); 3(43 41 15); 2 (47 41 4); 2 (54 15 4) 2 ( 4 48 9); 2 ( 4 48 15) 2 ( 36 44 42); 2 (51 30 31) 3 ( 18 15 7) 3 (47 22 37) 2 (58 4 4) 2 ( 18 19 15) 2 ( 18 19 13) 23 (22 67 29); 17 (22 70 35); 7 (25 63 40) 6 ( 11 48 3 ( 7 63 35); 2 (4 59 40) 12 (43 4 2); 11 (43 4 4) 9 (43 33 37); 6 (43 30 31); 5 (25 30 24); 3 ( 36 33 31) 9 (0 48 48); 8 (11 63 31); 2 ( 25 48 37); 2 ( 14 44 42) 3 (7 0 31); 7 ( 4 22 31); 7 ( 11 26 37) 6 (7 22 15); 2 ( 18 15 15); 2 ( 4 11 9); 2 ( 4 11 15) 7 (4 41 53) 6 ( 40 15 26) 5 ( 14 11 53) 2 ( 14 22 18) 2 ( 18 41 31) 2 (40 7 18) 2 (14 19 42) 5 (58 11 4); 2 (47 4 ( 58 22 15) 4 (11 56 2) 3 ( 14 7 20) 7 2) 35);
    
    Table 3 (B). Negative Correlations between Clinical Variables and Neural Responses to Sad (vs. Neutral) Stimuli (p .001) Area Description (with Brodmann Areas) Anhedonia R dorsolateral PFC (inferior frontal gyrus) BA45 L dorsolateral PFC (inferior frontal gyrus) BA44 BDI – Anxiety R visual cortex BA18 R precuneus BA7 R lingual gyrus BA17 Medial cerebellum SAD Subjective Mood Ratings R dorsolateral PFC BA40/44/45 R ventrolateral PFC BA47 Cluster Size (total) Cluster Size ( Talairach Coordinates x, y, z)
    
    22 2 – 6 4 2 2 28 2
    
    11 (47 19 9); 7 (47 26 4); 4 (47 22 15) 2 ( 47 0 26) – 6 (7 85 7) 2 (11 70 42); 2 (7 2 (7 85 13) 2 (0 81 18)
    
    74 37)
    
    13 (51 15 9); 8 (36 4 15); 9 (47 22 4); 6 (47 22 15) 2 (43 33 7)
    
    PFC, prefrontal cortex; BA, Brodmann areas; R, right; L, left.
    
    www.sobp.org/journal
    
    850 BIOL PSYCHIATRY 2005;58:843– 853
    
    P.A. Keedwell et al
    
    Figure 3. Correlations (at p .001) between BDI and anxiety and neural responses which were greater to happy stimuli than neutral stimuli (intersection with happy neutral GBAM, see text): (A) negative correlations between BDI and subcortical responses, and (B) positive correlations between anxiety and VMPFC responses. Scatter plots show clinical variable scores on y axis, and SSQ values on x axis, the latter extracted for each individual from a GBAM cluster overlapping with the correlation map (see text). BDI, Beck Depression Inventory; GBAM, generic brain activation map; VMPFC, ventromedial prefrontal cortex; SSQ, sum of squares.
    
    between depression severity or state anxiety and responses in regions associated with reward processing. However, BDI score was positively correlated with response within ventrolateral prefrontal cortex, BA47. State anxiety scores were positively correlated with responses in the right insula. Positive correlations were also seen between both depression and state anxiety and responses in hippocampus, parahippopcampal gyrus, and several occipitotemporal cortical regions. Positive correlations were observed between subjective ratings of sad mood and neural responses to sad stimuli in dorsal cortical areas and a small area of the left caudate nucleus. Negative correlations with subjective ratings of sad mood were observed in ventrolateral (BA47) and dorsolateral (BA44/45) prefrontal areas. A summary of the findings for the main regions of interest is shown in Table 4. Post-Hoc Analyses: Medication Effects All but one of the depressed individuals were taking antidepressant medication. They were ranked by ordinal score of relative dose potency, using Sackheim’s method of calculating
    
    dose equivalents for different preparations (Sackheim 2001). Using the median split, high and low dose subgroups were created (n 6 for each subgroup). High and low dose groups did not differ significantly on clinical variable scores: anhedonia (t 1.9, p .86, df 10), anxiety (t 1.1, p .29, df 10) and BDI (t .71, p .50, df 10). Statistical comparisons were performed on the neural responses of these subgroups to happy-neutral and sad-neutral stimuli using the same method as was employed for comparing neural responses to happy versus sad stimuli. This produced two between -subgroup maps of medication effect for each experiment (happy-neutral and sadneutral). Increased VMPFC responses to both happy and sad stimuli (compared with neutral) were demonstrated by the high medication subgroup (Figure 5).
    
    Discussion
    To our knowledge, this is the first study to examine the neural basis of anhedonia in MDD by correlating anhedonia severity with responses to personally-relevant happy and sad stimuli.
    
    Figure 4. Negative correlations (p .001) between subjective happy mood ratings and VMPFC responses which were greater to happy stimuli than neutral stimuli (intersection with happy neutral GBAM, see text). Scatter plots show mood ratings on y axis, and SSQ values on x axis, the latter extracted for each individual from a GBAM cluster overlapping with the correlation map (see text). VMPFC, ventromedial prefrontal cortex; GBAM, generic brain activation map; SSQ, sum of squares.
    
    www.sobp.org/journal
    
    P.A. Keedwell et al
    Table 4. Summary of Neural Correlates of Anhedonia and Other Clinical Variables Happy Anhedonia Putamen Caudate Nucleus accumbens Insula Amygdala VMPFC BA10/24ca VMPFC: BA11b VLPFC ve (Bil) ve (L) ve (L) ve (R) ve (Bil) ve (L) Anxiety ve (L) Hap rating ve (Bil) BDI ve (L) ve (Bil) ve (L) ve (L) ve (Bil) ve (R) ve (R)
    
    BIOL PSYCHIATRY 2005;58:843– 853 851
    
    Sad Anhedonia ve (R) ve (L) ve (R) Anxiety Sad rating BDI
    
    ve (L) ve (L) ve (Bil) ve (R)
    
    ve (R)
    
    Neural correlates within regions of interest (for neural responses to happy and sad stimuli, contrasting with neutral stimuli, previously highlighted on Generic Brain Activation Maps of response contrast; ve, negative correlation; ve, positive correlation; L, Left; R, Right; Bil, Bilateral; all at p .001, see text). VMPFC, ventromedial prefrontal cortex; BA, Brodmann area; BDI, Beck Depression Inventory. a Medial frontal pole/anterior cingulate (Zcoordinate 4). b Orbitofrontal cortex.
    
    Subjective mood ratings indicated that changes in mood occurred in all participants in response to the emotional stimuli compared with neutral stimuli. Examination of neuroimaging data revealed relatively greater VMPFC response to the happy stimuli compared with sad stimuli. This finding is consistent with previous studies demonstrating increases and decreases in VMPFC activity in response to positive and negative stimuli, respectively, in depression (Kumari et al 2003; Liotti et al 2002; Mitterschiffthaler et al 2003). The main finding was that positive and negative correlations were demonstrated between anhedonia severity and responses to happy stimuli predominantly within the VMPFC and striatal structures, respectively, in support of our prediction, and consistent with earlier findings revealed in individuals with MDD during performance of a cognitive task (Dunn et al 2002) (negative correlations with insula response were also found in this and the previous study). Also consistent was our finding of a negative correlation between happy mood ratings (change from neutral to happy) and neural responses in similar VMPFC regions, and a positive correlation between happy mood ratings and responses within putamen and insula to happy stimuli. Hence, those who found it more difficult to feel happy in response to the happy stimuli (i.e.
    
    had smaller changes compared to neutral stimuli - an anhedonic response by definition) displayed a greater VMPFC response, whereas those who did feel happier in response to the happy stimulus had greater striatal responses. These results suggest that anhedonic individuals attended more closely to the rewarding stimulus in an attempt to get into happy mood, consistent with previous literature on the role of the VMPFC in attending to the rewarding context of potentially rewarding stimuli. On line happy ratings were negatively correlated with anhedonia, but this did not reach significance. Findings can not be explained by greater resting state VMPFC activity per se in more anhedonic individuals because the analysis compared neural responses to happy and neutral stimuli within individuals. It was this contrast that was then correlated with anhedonia severity. These findings suggest a dissociation of function between the VMPFC and striatum in response to happy stimuli, in anhedonically depressed individuals. We know that they have reciprocal connections, and are both involved in the processing of reward. The level of the primary abnormality within this system remains to be defined: the ‘hypofrontality’ hypothesis of depression suggests that the primary deficit may be in the VMPFC, but the VMPFC could be compensating for an under-active subcortical/ striatal response. Findings of reduced caudate, amygdala and
    
    Figure 5. Medication effects. Comparisons of low and high dose antidepressant medication shown in red (p .005). Radiological axial slices. (A) Happy mood condition. High dose Low dose. L VMPFC BA10/32 (449 voxels, 10, 47, 4); R superior temporal gyrus BA 12/38 (93 voxels, 44, 15, 11); L superior temporal gyrus BA22 (35 voxels, 49, 12, 1). (B) Happy mood condition. Low dose High dose. Bilateral cerebellum (131 voxels, 19, 76, 24; 23, 81, 21); R VMPFC BA10 (58 voxels, 26, 57, 4), L DMPFC (middle frontal gyrus, 41 voxels, 31, 36, 35). (C) Sad mood condition. High dose Low dose. L VMPFC BA32/10 (108 voxels, 21, 48, 2); R orbitofrontal cortex BA11 (67 voxels, 11, 40, 16); R superior temporal gyrus BA38 (72 voxels, 51, 4, 10); R post-central gyrus BA2 (130 voxels, 50, 21, 32); L posterior cingulate gyrus BA31 (130 voxels, 5, 35, 39). (D) Sad mood condition. Low dose High dose. R dorsomedial PFC BA9 (not shown, 71 voxels, 42, 22, 27); L dorsomedial PFC BA9 (not shown, 58 voxels, 32, 34, 35) VMPFC, ventromedial prefrontal cortex; BA, Brodmann Area; L, left; R, right.
    
    www.sobp.org/journal
    
    852 BIOL PSYCHIATRY 2005;58:843– 853
    putamen volumes in MDD (Beyer and Krishnan 2002) support the latter interpretation. No correlations were found between anhedonia and responses to sad stimuli in these reward system areas, despite the fact that subjects reported sad memories to be aversive rather than merely nonrewarding. Hence, our findings indicate that anhedonia in depressed individuals is associated with increased VMPFC and decreased striatal responses to previously rewarding, but not aversive, events. Secondary analyses revealed that global depression severity was correlated positively only with dorsal prefrontal cortex responses to happy stimuli, but negatively with responses within regions associated with reward processing, including putamen, caudate nucleus, nucleus accumbens and amygdala, in common with anhedonia. The latter finding could be explained, at least in part, by the inclusion within the BDI of items measuring anhedonia and psychomotor retardation. Again, it is consistent with findings of striatal and amygdala structural abnormalities in MDD (Beyer and Krishnan 2002). State anxiety correlated positively and negatively with VMPFC and subcortical (L putamen and L amygdala) responses, respectively, demonstrating some similarities with anhedonia’s neural correlates. It seems intuitively true that high levels of anxiety would be incompatible with pleasurable experiences. However, there was no positive correlation between state anxiety and anhedonia scores. Furthermore, there were differences in laterality (anhedonia being negatively correlated with right, but not left, amygdala response), and anxiety was not negatively correlated with other areas of the striatum associated with reward, including more ventral putamen, caudate, or nucleus accumbens. Hence, these findings suggest that anhedonia and state anxiety independently contribute to a dissociated VMPFC-subcortical pattern of response to happy stimuli in depressed individuals, and recruit different areas of the striatum. More research is required to further delineate the functions of different parts of the human ventral striatum, and to determine how dysfunction in discrete areas may correlate with different dimensions of psychiatric disorder. With regard to sad stimuli, no significant correlations, positive or negative, were demonstrated between depression severity or state anxiety and responses in regions associated with reward processing. However, interesting findings emerged, which were consistent with previous literature. BDI score (which we found to be highly correlated with the sub-score for negative thinking) was positively correlated with response within BA47, an area within which activity has previously been shown to be highly correlated with the ‘negative cognitions’ factor of the BDI following a factor analysis (Dunn et al 2002). Also, increases in ventrolateral PFC to sad distracters have been demonstrated in depressed individuals during an affective go-no-go task (Elliott et al 2004). State anxiety scores were positively correlated with responses in the right insula, an area known to be involved in the monitoring of arousal and internal states (Critchley et al 2000). Positive correlations were also seen between both depression and state anxiety and responses in hippocampus and parahippopcampal gyrus, areas important for the recall of emotional memories (Smith et al 2004), and several occipitotemporal cortical regions previously demonstrated to be involved in the response to facial expressions (Haxby et al 2002; Phan et al 2001). This suggests heightened recall of sad memories in more depressed and more anxious individuals, and also a greater response to sad faces in regions involved in visual object processing. The latter finding is consistent with results from a www.sobp.org/journal
    
    P.A. Keedwell et al study examining covert neural responses to sad faces alone in MDD (Surguladze et al 2005). The main limitation of this study is that all but one of the MDD individuals were taking antidepressants. Medication dose may have contributed to our patterns of findings in the VMPFC in depressed individuals but can not explain all of the findings. Firstly, high and low dose medication was associated with increases in VMPFC to happy stimuli. Furthermore, medication dose did not explain increased and decreased responses in VMPFC to happy and sad stimuli, respectively, within depressed individuals. Finally, medication dose was not positively associated with scores on measures of anhedonia (or other clinical variables): clinical variable scores were not significantly different in high and low dose groups. We have highlighted the importance of examining the relationship between specific symptom dimensions and patterns of abnormal response within the reward system (and other mood regulatory areas) to ecologically valid stimuli in MDD. Future studies will further define the sites of abnormality with these functional systems, elucidate reciprocal relationships within them, and determine the extent to which investigations of their function in MDD are predictive of clinical outcome after specific pharmacological and psychological treatments.
    
    This study was presented in part at the 58th Annual Meeting of the Society of Biological Psychiatry; May 15, 2003; San Francisco, California. This research was supported by the Goldsmith Charitable Foundation.
    American Psychiatric Association (1994): Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (DSM-IV). Washington DC: American Psychiatric Association. Beck AT (1967): Depression: Clinical, Experimental and Theoretical Aspects. New York: Harper and Row. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961): An inventory for measuring depression. Arch Gen Psychiatry 4:561–571. Beyer JL, Krishnan KRR (2002): Volumetric brain imaging findings in mood disorders. Bipolar Disorders 4:89 –104. Blood AJ, Zatorre RJ (2001): Intensely pleasurable responses to music correlate with activity in brain regions implicated in reward and emotion. Proceedings of the National Academy of Sciences of the United States of America 98:11818 –11823. Brammer M, Bullmore ET, Simmons A, Williams SCR, Grasby PM, Howard RJ, et al (1997): Generic brain activation mapping in functional magnetic resonance imaging: a nonparametric approach. Magn Reson Imaging 15:763–770. Breakspear M, Brammer MJ, Robinson P (2003): Construction of multivariate surrogate sets from nonlinear data using the wavelet transform. Physica D 182:1–22. Breiter HC, Gollub RL, Weisskoff RM (1997): Acute effects of cocaine on human brain activity and emotion. Neuron 19:591– 611. Brown GW, Harris TO (1978): The Social Origins of Depression. London: Free Press. Bullmore ET, Long C, Suckling J, Fadili J, Calvert GA, Zelaya F, et al (2001): Coloured noise and computational inference in neurophysiological (fMRI) time series analysis: resampling methods in time and wavelet domains. Hum Brain Mapp 12:61–78. Bullmore ET, Suckling J, Overmeyer S, Rabe-Hesketh S, Taylor E, Brammer M (1999): Global, voxel and cluster tests, by theory and permutation, for a difference between two groups of structural MR images of the brain. IEEE Trans Med Imaging 18:32–34. Clark DM, Teasdale JD (1982): Diurnal variation in clinical depression and accessibility of memories of positive and negative experiences. J Abnorm Psychol 91(2):87–95. Critchley HD, Elliott R, Mathias CJ, Dolan RJ (2000): Neural activity relating to generation and representation of galvanic skin conductance responses:
    
    P.A. Keedwell et al
    A functional magnetic resonance imaging study. J Neuroscience 20:3033–3040. Drevets WC, Gautier C, Price JC, Kupfer DJ, Kinahan PE, Grace AA, et al (2001): Amphetamine-induced dopamine release in human ventral striatum correlates with euphoria. Biol Psychiatry 49:81–96. Dunn RT, Kimbrell TA, Ketter TA, Frye MA, Willis MW, Luckenbaugh DA, et al (2002): Principal components of the Beck Depression Inventory and regional cerebral metabolism in unipolar and bipolar depression. Biol Psychiatry 51:387–399. Ekman P, Friesen WV (1976): Pictures of Facial Affect. Palo Alto: Consulting Psychologists. Elliott R, Friston KJ, Dolan RJ (2000): Dissociable neural responses in human reward systems. J Neuroscience 20:6159 – 6165. Elliott R, Newman JL, Longe OA, William Deakin JF (2004): Instrumental responding for rewards is associated with enhanced neuronal response in subcortical reward systems. Neuroimage 21:984 –990. Elliott R, Rubinsztein JS, Sahakian BJ, Dolan RJ (2004): The neural basis of mood-congruent processing biases in depression. Arch Gen Psychiatry 59:597– 604. Fawcett J, Clark DC, Schnefter WA, Gibson RD (1983): Assessing anhedonia in psychiatric patients: The Pleasure Scale. Arch Gen Psychiatry 40:79 – 84. Garrett AS, Maddock RJ (2001): Time course of the subjective emotional response to aversive pictures: Relevance to fMRI studies. Psychiatry Res: Neuroimaging 108(1):39 – 48. Haxby JV, Hoffman EA, Gobbini M (2002): Human neural systems for face recognition and social communication. Biol Psychiatry 51(1):59 – 67. Henriques JB, Davidson RJ (2000): Decreased responsiveness to reward in depression. Cog and Emotion 14:711–724. Keedwell PA, Andrew C, Williams SCR, Brammer MJ, Zelaya F, Phillips ML (2003): The neural Correlates of Depression (Abstract). Biol Psychiatry 53:171S. Keedwell PA, Andrew C, Williams SCR, Brammer MJ, Zelaya F, Phillips ML (in press): A double dissociation of ventromedial prefrontal cortical responses to sad and happy stimuli in depressed and healthy individuals. Biol Psychiatry. Klein DF (1974): Endogenomorphic depression: a conceptual and terminological revision. Arch Gen Psychiatry 31:447– 454. Knutson B, Fong GW, Adams CM, Varner JL, Hommer D (2001): Dissociation of reward anticipation and outcome with event-related fMRI. Neuroreport 12 (17):3683–3687. 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.
    
    BIOL PSYCHIATRY 2005;58:843– 853 853
    Lang PJ, Bradley MM, Cuthbert BN (1997): International Affective Picture System (IAPS): Technical Manual and Affective Ratings. Gainesville, FL: NIMH Center for the Study of Emotion and Attention. 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 [comment]. Am J Psychiatry 159:1830 –1840. Lloyd GG, Lishman WA (1975): Effect of depression on the speed of recall of pleasant and unpleasant experiences. Psychol Med 5(2):173–180. 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. Naranjo CA, Tremblay LK, Busto UE (2001): The role of the brain reward system in depression. Prog Neuro-Psychopharm & Biol Psychiatry 25:781– 823. O’Doherty J, Kringelbach ML, Rolls ET, Hornak J, Andrews C (2001): Abstract reward and punishment representations in the human orbitofrontal cortex. Nat Neurosci 4 (1):95–102. Phan KL, Wager T, Taylor SF, Liberzon I (2001): Functional Neuroanatomy of Emotion: A meta-analysis of Emotion Activation Studies in PET and fMRI. Neuroimage 16:331–348. Rampello L, Ferdinando N, Francesco N (2000): Dopamine and depression: Therapeutic implications. CNS Drugs 13 (1): 33– 45. Sackheim HA (2001): The definition and meaning of treatment resistant depression. J Clin Psychiatry 62 (Suppl 16):1–17. Schneider F, Grodd W, Weiss U, Klose U, Mayer KR, Nagele T, et al (1997): Functional MRI reveals left amygdala activation during emotion. Psychiatry Res 76 (2–3):75– 82. Senior C (2004): Beauty in the brain of the beholder. Neuron 38:525–528. Small DM, Gregory MD, Mak YE, Gitelman D, Mesulam MM, Parrish T (2003): Dissociation of neural representation of intensity and affective valuation in human gustation. Neuron 39 (4):701–711. Smith AP, Henson RN, Dolan RJ, Rugg MD (2004): fMRI correlates of the episodic retrieval of emotional contexts. Neuroimage 22:868 – 878. Snaith P (1993): Anhedonia: a neglected symptom of psychopathology. Psychol Medicine 23:957–966. Spielberger CD (1983): State-Trait Anxiety Inventory for Adults. California: Mind Garden. Surguladze S, Keedwell PA, Brammer M, Young A, Andrew C, Travis M, et al (2005): A differential pattern of neural response and identification bias towards negative versus positive facial expressions in Major Depressive Disorder. Biol Psychiatry 57:201–209. Talairach J, Tournoux P (1998): Co-planar Stereotactic Atlas of the Human Brain. Stuttgart, Germany: Georg Thieme Verlag. World Health Organization (1992): ICD-10 Classification of Mental and Behavioural Disorders. Geneva: World Health Organization.
    
    www.sobp.org/journal

Add Comment

To add a comment, you need to login or signup (you can signup without leaving this page)