Neural correlates of happy and sad mood in ... - Semantic Scholar

www.elsevier.com/locate/ynimg NeuroImage 26 (2005) 206 – 214

Same or different? Neural correlates of happy and sad mood in healthy males Ute Habel,a,T Martina Klein,a Thilo Kellermann,a N. Jon Shah,b and Frank Schneider a a

Department of Psychiatry and Psychotherapy, RWTH Aachen University, Pauwelsstr. 30, 52074 Aachen, Germany Institute of Medicine, Research Center Ju¨lich, Germany

b

Received 2 September 2004; revised 9 December 2004; accepted 12 January 2005 Available online 10 March 2005

Emotional experience in healthy men has been shown to rely on a brain network including subcortical as well as cortical areas in a complex interaction, which may be substantially influenced by many internal personal and external factors such as individuality, gender, stimulus material and task instructions. The divergent results may be interpreted by taking these considerations into account. Hence, many aspects remain to be clarified in characterizing the neural correlates underlying the subjective experience of emotion. One unresolved question refers to the influence of emotion quality on the cerebral substrates. Hence, 26 male healthy subjects were investigated with functional magnetic resonance imaging during standardized sad and happy mood induction as well as a cognitive control task to explore brain responses differentially involved in positive and negative emotional experience. Sad and happy mood in contrast to the control task produced similarly significant activations in the amygdala–hippocampal area extending into the parahippocampal gyrus as well as in the prefrontal and temporal cortex, the anterior cingulate, and the precuneus. Significant valence differences emerged when comparing both tasks directly. More activation has been demonstrated in the ventrolateral prefrontal cortex (VLPFC), the anterior cingulate cortex (ACC), the transverse temporal gyrus, and the superior temporal gyrus during sadness. Happiness, on the other hand, produced stronger activations in the dorsolateral prefrontal cortex (DLPFC), the cingulate gyrus, the inferior temporal gyrus, and the cerebellum. Hence, negative and positive moods reveal distinct cortical activation foci within a common neural network, probably making the difference between qualitatively different emotional feelings. D 2005 Elsevier Inc. All rights reserved. Keywords: Happy; Sad; Neural correlates

Introduction The investigation of emotional experience in experimental settings has to rely on the application of mood induction T Corresponding author. Fax: +49 241 8082401. E-mail address: [email protected] (U. Habel). Available online on ScienceDirect (www.sciencedirect.com). 1053-8119/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2005.01.014

procedures. Due to the existence of a variety of such mood induction procedures, which are not all sufficiently reliable, valid, and comparable, comprising controlled stimulus material, comparisons of different study results are limited. In addition, only a handful of studies have specifically investigated emotional experience applying neuroimaging methods with a direct mood induction instruction. Among the first, Pardo et al. (1993) reported an rCBF increase in inferior and orbitofrontal regions with positron emission tomograpy (PET-15O) during self-induced dysphoria in seven healthy subjects. George et al. (1995) applied the PENN Facial Emotion Discrimination stimuli in a PET 15O study for mood induction in a small number of healthy women. During sadness, they described rCBF increases in the right anterior cingulate and the bilateral inferior frontal gyri. Accordingly, we were led to develop a mood induction procedure that includes ecologically valid and socially relevant emotional stimuli, which are controlled, comparable, and simultaneously applicable during neuroimaging. Straight angle monochromatic photographs of happy and sad facial expressions varying in intensity made up the stimuli. The procedure for its construction has been detailed before (Schneider et al., 1994a). The final procedure consists of two components and is based on the presentation of 40 slides of male and female facial expressions all according to the required emotional quality (happy or sad). Evidence for the validity and objectivity of this standardized mood induction procedure was shown on a behavioral level in normal subjects (Weiss et al., 1999) and characteristic valence-specific regional cerebral and autonomic effects have been demonstrated measuring regional brain activity with 133Xenon clearance method (Schneider et al., 1994b), H2O15-PET (Schneider et al., 1995), and functional magnetic resonance imaging (fMRI; Habel et al., 2004; Schneider et al., 1997, 1998) in healthy subjects and schizophrenia patients. The main finding was an amygdala activation as a correlate of sad mood which was demonstrated with a ROI approach with PET (Schneider et al., 1995) and fMRI (Habel et al., 2004; Schneider et al., 1997, 1998). Correlations of subjective mood experience with activation changes in the amygdala (Habel et al., 2004; Schneider et al., 1995, 1997, 1998) supported a special

U. Habel et al. / NeuroImage 26 (2005) 206–214

role of the amygdala in regulating emotional states. However, other authors failed to report amygdala activation during mood induction and reported evidence primarily for the participation of the ACC and prefrontal cortex instead (Mayberg et al., 1999; Teasdale et al., 1999). The differing outcomes may at least in part be due to the divergent mood induction procedures. Some studies used visual material (Lane et al., 1997; Teasdale et al., 1999), others the recall of personal events of different emotional valence (Mayberg et al., 1999). However, as Lane et al. (1997) demonstrated, different regional rCBF changes accompanied film (amygdala) or recall induced (anterior insula) sad mood. Hence, many aspects remain to be clarified concerning the neural substrates of different basic emotions. One open question refers to the influence of emotion quality on the cerebral substrates. Phan et al. (2003) found indications for common as well as emotion specific activations in their meta-analysis. Common activations were found in the prefrontal area, emotion specific activations, for example, in the subgenual ACC for sadness and in the basal ganglia for happiness, while another recent meta-analysis of 106 PET and fMRI studies (Murphy et al., 2003) reported a similar activated network for happiness and sadness. However, apart from methodological considerations of these meta-analyses such as study inclusion criteria, the authors encounter the difficulty of combining studies involving a general processing of emotional material (such as emotion discrimination tasks) and those addressing directly the subjective emotional feeling applying different mood induction procedures. We suggest, however, that nonemotional task instructions (such as gender discrimination) or passive processing of emotional material is quite different and not comparable to the active subjective feeling of an emotion while being emotionally engaged, even when taking the fact into account that all mood induction procedures take place in an experimental setting which is quite artificial compared to natural emotional experiences. The objective of the present study was therefore to investigate the common and specific neural correlates of happy and sad mood applying our standardized mood induction procedure, while controlling for the influence of gender. Furthermore, the mood induction conditions were compared to a non-emotional control condition based on the same stimulus material as well as compared directly with each other. We hypothesized a greater subcortical involvement during sad compared to happy mood. Hence, a special focus of interest has been placed on the valence-specific involvement of the amygdala, since this region has been found to play an important role during the subjective experience of emotion, predominantly negative emotion.

Materials and methods Subjects Healthy male subjects (n = 26) with a mean age of 33.4 F 8.1 years (range 21–46) and a mean education of 12.1 F 3.8 years (range 9–18) participated in the study. They were recruited and used as comparison subjects for schizophrenia patients and the group differences have been reported previously (Habel et al., 2004; Schneider et al., submitted for publication). Here, healthy subjects were analyzed with the aim of characterizing cerebral differences between happy and sad mood induction. Screenings by comprehensive assessment procedures guaranteed that subjects

207

were healthy (no lifetime DSM IV diagnosis, no first-degree relatives with psychiatric diseases). For all subjects, the usual exclusion criteria for fMRI were applied (neurological diseases, disorders which affect cerebral metabolism, age under 18 and over 46). After a complete description of the study, written informed consent was obtained. The local IRB approved the protocol. Procedure Subjects participated in the standardized mood induction procedure, described before and in greater detail previously (Schneider et al., 1994a). Briefly, 117 Caucasian professional actors were instructed to display the emotions while their pictures were taken. The models were draped in black fabric and photographed against a black backdrop to eliminate all clothing and ambient distractors. This set of photographs was reviewed by six raters for asymmetry and for ambiguity of expressed emotion. Only unitary and genuine facial expressions were retained. The method demonstrated small intra-individual variability and high retest reliability behaviorally. Presenting happy and sad facial expressions, the task instructions were as follows: bDuring this task, I would like you to try to become happy [sad]. To help you do that, I will be showing you slides with faces expressing happiness [sadness]. Look at each face and use it to help you to feel happy [sad].Q Subjects viewed the stimuli at their own pace and moved on to the next face with the aid of a response device by pressing buttons using both thumbs simultaneously. An intermediate gender discrimination task requiring subjects to determine the gender of the character portrayed on each slide (by pressing the left/right response button) served as a cognitive non-emotional control condition using the same emotional (happy and sad) faces that were presented during the mood induction conditions. During resting baseline conditions, subjects had to keep their eyes open and lie calmly without further instruction. A single slide preceding the activation phases provided subjects with a brief instruction on the task (bsadnessQ, bhappinessQ, bgender discriminationQ). Hence, the number of stimuli seen during each block varied between subjects with more stimuli expected during the cognitive control condition. Correspondingly, a one-way ANOVA with emotion as within subject factor revealed significantly different progression rates, i.e., number of stimuli seen by the subjects in the different conditions (F = 138.7, df = 2,44; P = 0.0001). More stimuli were seen during the cognitive control condition (105.1 F 34.7) than during the emotional conditions (sadness: 15.6 F 16.5, happiness: 18.2 F 19.4). Hence, subjects participated in 3 conditions (happy and sad mood induction, gender discrimination) while fMRI data were acquired (Fig. 1). Conditions consisted of 4 baseline and 3 activation blocks each. One additional fMRI paradigm was also performed afterwards by the subjects, reported elsewhere (Schneider et al., submitted for publication). The three conditions were administered in counterbalanced order (Latin square design) with the non-emotional cognitive condition amid the two emotional conditions. Data acquisition Subjective responses Dependent measure for quantifying the mood induction effect was the Positive and Negative Affect Scale (PANAS, Watson et al.,

208


Fig. 1. Block design of the mood induction paradigm. Experimental design consisting of 3 runs, one for sad mood induction, one for happy mood induction, and one for the gender discrimination cognitive control condition. Each run consisted of 15 scans per baseline and 15 per activation, with 4 baselines and 3 activation phases in each run and a total duration of 7 min 12 s.

1988), a 5-point unipolar intensity scale, which includes 20 items for factor-referenced emotional descriptors for orthogonal positive and negative dimensions. fMRI Scanning was performed with a 1.5-T Siemens scanner based on echo planar imaging (EPI) using BOLD contrast. Transaxial functional images (EPI, 64 64, 30 slices, voxel size: 3 3 4 mm3, TE = 66 ms, TR = 4 s, a = 908) positioned parallel to the intercommissural line (AC–PC) were acquired covering the whole brain including subcortical brain regions. Relying on a blocked design, each mood induction condition and the cognition condition consisted of 108 measurements each. The fMRI measurements were triggered by a laptop for stimulus presentation. The first three volumes were discarded to allow for equilibration of the scanner. The remaining volumes were equally distributed (n = 15 scans per block for baseline and activation phases during emotion as well as cognition) during each run: each run (happiness, sadness, cognition) entailed 4 baseline blocks and 3 activation phases, 60 s each, and lasted for 7 min 12 s (Fig. 1). Subjective ratings were assessed after each condition.

Inspection of the brain volumes yielded no spatial offset of more than one voxel in any of the subjects. After realignment and coregistration, spatial normalization (2 2 2 mm) and spatial smoothing (full width at half maximum, 10 mm) were performed. Statistical parametric maps were calculated independently for each subject by using a delayed boxcar convolved with a hemodynamic response function. For each individual, contrast images were created contrasting the emotional (sadness, happiness) and the cognitive condition as well as contrasting the two emotional conditions directly. The group analysis for comparisons with the non-emotional condition was performed according to a random effect model, applying one-sample t tests. The resulting maps revealed strong activation and were hence thresholded at P = 0.05, corrected for multiple comparisons (spatial extent z 10 contiguous voxels), based on the theory of Gaussian fields as used in SPM99. For the direct comparison of both emotional conditions, the threshold was lowered to P = 0.001, uncorrected (spatial extent z 10 contiguous voxels), to increase sensitivity for valence-dependent differences according to our hypotheses, since correcting for multiple comparisons left no suprathreshold clusters.

Data analysis

Results

Subjective ratings The mood induction effect was quantified with the positive and negative score of the PANAS. Since the PANAS is an indirect measurement of positive and negative affect, including for example items assessing attention, alertness, and activity for positive affect, the cognitive condition also scores on these scales. Hence, for assessing the pure affective changes during mood induction, the affective ratings of the cognitive control condition were subtracted from the affective ratings of the mood induction condition (parallel to the fMRI data analysis). These scores were analyzed in a repeated measures analysis of variance (ANOVA) with 2 repeated factors, Mood Induction (happy, sad) and Rating Scale (positive, negative score of the PANAS), to check the mood induction effect. The Scale Mood induction interaction served as a test for the hypothesis.

Subjective ratings

fMRI Data analysis was performed with SPM99 (http://www.fil.ion. ucl.ac.uk/spm). Functional data were realigned to the tenth image.

ANOVA of the PANAS data yielded a Scale Mood induction interaction (F = 26.7, df = 1, 25; P = 0.0001) and a significant main effect for scale (F = 12.40, df = 1, 25, P = 0.002), indicating successful mood induction. During happy mood induction, there was more positive affect (28.3 F 6.9) than negative affect (12.0 F 3.6). For sad mood, negative affect increased (16.2 F 5.2) while positive affect was reduced (22.9 F 7.4). Cognition revealed more positive (27.5 F 7.0) than negative affect (12.8 F 4.0). fMRI: sadness–cognition Main activation foci were detected mainly left lateralized in subcortical as well as cortical regions, namely in the amygdala– hippocampal area extending into the parahippocampal gyrus as well as in the putamen, the insula, the prefrontal cortex (dorsolateral prefrontal (DLPFC), orbitofrontal (OFC), and superior frontal gyrus) extending into the ACC, the middle and superior


temporal gyri, the precuneus extending into the posterior cingulate, and the right fasciculus occipito-frontalis (see Table 1; Figs. 2 and 3). fMRI: happiness–cognition The whole brain analysis demonstrated significant activation in the DLPFC, the anterior and posterior cingulate cortices including the cingulate gyrus and extending into the precuneus, the angular gyrus, the paracentral lobule, the middle temporal gyrus and the parahippocampal area including the amygdala (Table 1, Figs. 2 and 3).

209

amygdala–hippocampal area focusing on a ROI of 6 6 6 mm3 centered around the maximum of activation. For the sphere of 6 mm centered on the maximum of activation in the amygdala–hippocampal area (x = 34, y = 10, z = 16), a moderate but significant correlation emerged with the subjective experience of sadness (r = 0.36, P = 0.04; with the negative scores of the PANAS). Hence, amygdala activation increased with an intensified subjective experience of negative affect. No significant correlation emerged for happy mood and a comparable ROI around the activated cluster in the parahippocampal gyrus including the amygdala (x = 30, y = 30, z = 20).

Sadness vs. happiness

Discussion

The direct comparison of both mood induction conditions revealed higher activation in sadness in the bilateral ventrolateral prefrontal cortex (BA 47) and the left ACC (BA 32), as well as the bilateral superior temporal gyrus and left transverse temporal gyrus (Table 2, Fig. 4).

Controlling for the influence of gender on emotional experience by investigating a sample of healthy males, mood induction of happiness and sadness using standardized emotional facial expressions was associated with a distributed network including subcortical as well as cortical regions. Compared to a nonemotional condition, the activation pattern of the two emotional conditions mainly revealed similarities, namely activations in the subcortical amygdala–hippocampal–parahippocampal area as well as in the prefrontal and temporal cortex, the anterior cingulate, and the precuneus. Contrasting both emotional conditions directly, valence-specific characteristics emerged and support the conclusion, that despite a general convergence, sad and happy mood states rely on the dominance of distinct key components within these neural networks.

Happiness vs. sadness Happiness was associated with higher activation in the right DLPFC, the left cingulate gyrus, the right inferior temporal gyrus, and bilaterally in the cerebellum (Table 2, Fig. 4). Covariation of BOLD effect with subjective ratings A correlation analysis was performed for the amygdala as a region of interest (ROI) and in view of our preceding findings concerning the role of the amygdala during mood induction. Correlations were calculated between positive and negative scores of the PANAS and the activation during sadness/happiness in the

Subcortical involvement: the amygdala The amygdala has been assigned a major role in emotion processing by combining external information with internal

Table 1 SPM ANOVA results for sadness and happiness compared to the non-emotional control condition Sadness–Cognition

Happiness–Cognition

Region

Side

x

y

z

n

Max. SPM{T}

Dorsolateral prefrontal cortex Orbitofrontal cortex Orbitofrontal cortex Superior frontal gyrus Middle temporal gyrus Middle temporal gyrus Superior temporal gyrus Superior temporal gyrus Precuneus Parahippocampal gyrus Amygdala–hippocampal area Putamen Fasciculus occipito-frontalis Insula Dorsolateral prefrontal cortex Anterior cingulate gyrus Cingulate gyrus Posterior cingulate gyrus Middle temporal gyrus Parietal cortex, angular gyrus Paracentral lobule Parahippocampal gyrus including amygdala

L L L L L L L L L L L L R L L L R L L L R L

22 40 12 10 54 54 44 30 6 26 34 26 24 38 24 4 14 4 54 44 16 30

22 38 56 56 6 32 58 12 64 18 10 8 42 28 28 34 40 50 10 70 44 30

44 16 8 28 10 4 20 22 20 20 16 2 16 2 44 8 42 22 22 34 56 20

484 73 1586 15 497 80 632 16 1562 133 10 14 29 10 207 1039 67 1272 171 606 12 212

8.37 7.76 9.45 6.37 10.08 7.52 8.35 6.80 9.04 7.33 6.46 6.75 6.48 6.42 6.86 8.52 7.50 7.98 7.01 8.68 5.90 7.50

Significance threshold was P = 0.05 corrected for multiple comparisons (spatial extent z 10 contiguous voxels). x, y, and z are MNI coordinates and converted to Talairach coordinates (Talairach and Tournoux, 1988) for anatomical localization; they refer to the center of gravity of the cluster.

210


Fig. 2. Amygdala–hippocampal activation during sad and happy mood compared to a non-emotional control condition. Group activation (one-sample t test) overlaid on an EPI template for the contrast sadness–cognitive condition and happy–cognitive condition (P = 0.05 corrected for multiple comparisons, extent threshold 10 voxels).

reactions and hence constituting the basis for emotional responses. Based on earlier findings, we hypothesized amygdala involvement predominantly during sad mood induction due to its often suggested greater role in negative affect. However, amygdala– hippocampal participation extending into the parahippocampal area was observed similarly during both mood states and direct comparisons of both mood conditions failed to show stronger involvement during sad mood, yet our findings do not suggest a differential valence-specific involvement. Findings hence support growing evidence reporting amygdala responses to positive as well as negative emotions for different emotional tasks (Breiter et al., 1996; Hamann et al., 2002; Liberzon et al., 2003; Williams et al., 2004). Similarly, in our own preceding mood induction studies, amygdala activation was not only found during negative (Schneider et al., 1995, 2000) but also during positive affect (Schneider et al., 1997). A dominance in negative affect could sometimes be misleadingly suggested through differential findings in controls and schizophrenia patients for negative affect only (Habel et al., 2004).

However, while during sad mood, the activation was more focused in the amygdala–hippocampal area, the maximum of activation during happy mood lay more in the parahippocampal area. Furthermore, a correlation with the subjective emotional ratings emerged for sad mood only, indicating that amygdala responses appear thus not to be valence specific, but show a tendency for a stronger involvement in negative emotional experience. This is in accordance with our preceding region-ofinterest approach based on extracted signal courses in these subjects and used for comparisons with schizophrenia patients (Habel et al., 2004). Hence, activation in the amygdala–hippocampal area in this group of healthy subjects was demonstrated using two different analysis approaches, whole brain analysis (corrected for multiple comparisons) as well as an additional regional analysis reported before (Habel et al., 2004). This attempt of balancing sensitivity and specificity might help validating results in this critical region prone to susceptibility, movement, and other artefacts. Yet, we would like to emphasize that these conclusions on the role of the amygdala can be drawn for

Fig. 3. Rendered whole brain activation pattern during sad and happy mood compared to a non-emotional control condition. Group activation (one-sample t test) for the contrast sadness–cognitive condition and happiness–cognitive condition (P = 0.05 corrected for multiple comparisons, extent threshold 10 voxels).


211

Table 2 SPM ANOVA results for the comparison of both emotional conditions Sadness–Happiness

Happiness–Sadness

Region

Side

x

y

z

n

Max. SPM{T}

Ventrolateral prefrontal cortex Ventrolateral prefrontal cortex Anterior cingulate Transverse temporal gyrus Superior temporal gyrus Superior temporal gyrus Dorsolateral prefrontal cortex Cingulate Gyrus Inferior temporal gyrus Cerebellum Cerebellum

R L L L R L R L R R L

30 30 6 36 46 64 22 14 40 14 22

18 16 36 28 24 22 22 20 20 42 36

22 20 16 10 8 12 36 44 24 30 30

59 24 21 36 46 18 305 15 18 70 29

4.43 4.23 4.32 4.04 4.03 3.56 4.60 3.66 3.77 3.96 4.16

Significance threshold was P = 0.001 uncorrected (spatial extent z 10 contiguous voxels). x, y, and z are MNI coordinates and converted to Talairach coordinates (Talairach and Tournoux, 1988) for anatomical localization; they refer to the center of gravity of the cluster.

emotional experience only, with probably quite different results for other emotional processes. Sadness and happiness: same or different? The cerebral correlates of sadness and happiness seem to be part of the often reported brain network implicated in emotional

experience, including the DLPFC and OFC (Beauregard et al., 1998, Lane et al., 1997), the ACC (see Phan et al., 2003), the temporal cortex (Damasio et al., 2000; Lane et al., 1997; Lee et al., 2004), the precuneus, and the amygdala extending into the hippocampus and parahippocampal gyrus. Despite the varying methodologies of the different studies, our results are generally in accordance with the meta-analysis of

Fig. 4. Valence-specific activation. Group activation for the contrast sadness vs. happiness revealed more anterior cingulate activation while the contrast of happiness vs. sadness demonstrated a greater DLPFC participation (P = 0.001, extent threshold 10 voxels).

212


Murphy et al. (2003), but also with specific findings, for example, in professional actors (Pelletier et al., 2003), confirming similar cerebral networks for the emotional experience of sadness and happiness. When compared to a cognitive non-emotional task, prefrontal activation emerged in the DLPFC and OFC, extending into the ACC up to the subgenual division (subgenual cingulate cortex, SCC) with both emotions. This also supports the metaanalysis of Phan et al. (2003) consisting of 55 PET and fMRI studies in which prefrontal activation in response to different individual emotions and mood induction procedures has been found within the ventral–rostral (BA 9 and 10) area of the medial prefrontal cortex extending into the affective subdivision of the ACC (BA 24, 32, 33). The authors suggest that prefrontal activation may reflect the cognitive evaluative aspects (attention, awareness, appraisal) characteristic for all emotions and mood induction procedures. The ACC activation on the other hand was found to be typical for mood induction elicited by recall or imagery as well as during emotional tasks with cognitive demand. This applies to our mood induction procedure involving externally triggered but also internally generated mood, in which cognitive processes and emotional memories play a substantial role. Correspondingly, hippocampal, parahippocampal, and precuneus activations indicate such memory processes. Debriefing of our subjects after each mood induction condition revealed that besides the visual stimulation with emotional faces, retrieval of emotional autobiographical memories has often been used for mood induction (in 15 subjects during happiness and 17 during sadness). However, within this emotional cerebral network, sadness and happiness seem to rely on distinct main activation foci and can hence be differentiated by valence-specific characteristics in their neural correlates. While sadness was specifically associated with more ACC activation, happiness demonstrated a greater activation cluster in a more posterior and dorsal part of the cingulate gyrus. Although the activated cluster in sadness was part of the affective subdivision of the ACC, it is not part of SCC. This supports the notion of Reiman et al. (1997) that it is the internally generated mood that activates the subgenual part rather than sadness itself. Furthermore, Markowitsch et al. (2003) found SCC participation also during the retrieval of happy compared to sad autobiographical memories. Probably a more dorsal part of the ACC (BA 32) may play a specific role in sadness, since similar activation has been reported during visually elicited grief in bereaved women (Gundel et al., 2003) and externally triggered sad mood via film clips, which was correlated to the subjective experience of sadness (Levesque et al., 2003). As not only conscious (Blair et al., 1999) but also unconscious perception (Killgore and Yurgelun-Todd, 2001) of sadness in facial expressions is also associated with ACC activation, this points to an emotion-specific regulatory function of this region, besides the valence independent role of the ACC in emotional evaluation and control (Hariri et al., 2003; Phan et al., 2003) together with its involvement in cognitive control (Badre and Wagner, 2004) and attention (Fan et al., 2003). Further specific correlates of sadness have been found in the bilateral VLPFC and temporal cortex. VLPFC activation has been demonstrated during the retrieval of sad memories (Markowitsch et al., 2003) and in response to sad films (Beauregard et al., 1998; Levesque et al., 2003). Structural equation modeling revealed a special influence of the amygdala to the VLPFC during emotional conditions (Kilpatrick and Cahill, 2003). Hence, this area seems to be involved in the regulation of emotional experience probably by modulating the negative emotion.

Since the superior temporal cortex is an area sensitive to facial stimuli, this activation may be specific to our mood induction procedure and may correspond to the higher attention paid to sad vs. happy stimuli due to a greater social relevance of sad faces. Selective attention to facial emotion has been demonstrated to modulate the activity of the right superior temporal cortex (Narumoto et al., 2001), suggesting a special role of this area in social perception. On the other hand, besides the cingulate gyrus, the right DLPFC, right inferior temporal gyrus, and bilateral cerebellar regions responded stronger to positive mood. Specific left DLPFC activation has been reported during the retrieval of happy mood (Markowitsch et al., 2003) and has been associated with reencoding during the retrieval of autobiographic episodes. Although part of this activation seems to be located within the white matter (Fig. 4), the maximum of activation is well within the gray matter (following the Talairach coordinates). Hence, the extent of activation as well as normal deviations when overlaid on the SPM template may account for why part of the activation cluster extended into white matter. The role of the cerebellum in emotion processing has been acknowledged (Schmahmann, 2000) although not specifically linked to positive emotions. It has been suggested from clinical data that cerebellar structures may adjust the vocal expression (crying, laughing) of emotion to the cognitive and situational context (Parvizi et al., 2001). The inferior temporal region has been found to be characteristically activated during visually elicited grief (Gundel et al., 2003); however, no positive emotion has been included in this study. Yet, valence-specific functions of these areas have also been indicated. The perception of humor has been associated with cerebellar as well as inferior temporal structures (Wild et al., 2003), and ictal pleasant feelings during focal epilepsies have been traced back to the inferior basal temporal lobe (Stefan et al., 2004). In contrast to other results (Lee et al., 2004), our findings as well as comparable studies (Markowitsch et al., 2003; Pelletier et al., 2003) do not confirm a valence-specific hemispheric specialization in emotion processing. The discussion concerning a greater right-sided involvement during negative mood and a prevailing left hemispheric dominance for positive emotions is still unresolved. Hemispheric lateralization in emotion may be less obvious than previously thought and influenced by many external as well as internal factors. The contrast between our emotional and non-emotional control condition left mainly a left lateralized cerebral activation in both mood conditions, which may be caused by the comparison of our specific mood induction procedure based on visual material and recalled emotional autobiographical memories with the gender discrimination task. Finally, some methodological constraints have to be considered. While controlling for gender influences by investigating a large sample of healthy males, gender differences could not be addressed. Further factors of influence such as individuality, interindividual variability (Canli et al., 2001), and task instruction could not be taken into account since we focused on group results using a standardized mood induction technique. Furthermore, the use of uncorrected P values (i.e., 0.001 in this case) poses the problem of false positives (i.e., type I errors) due to multiple comparisons. However, since the activated clusters at this level agree with the literature, it might be concluded that lowering the threshold yielded less false negatives (or type II errors), which would be equipollent to increased sensitivity. Results are also supported by the fact that mainly the same regions survive a more


conservative threshold (corrected for multiple comparisons) in the comparisons between the emotional and control conditions. Nevertheless, results also suggest that differences between various emotions may be more subtle than previously thought. The cognitive non-emotional gender discrimination task imposes a differential effort compared to the mood induction and points once again to the problem of including adequate control tasks in experiments that apply emotional tasks. While the effect of such a differential effort on the activation pattern cannot be excluded when comparing the emotional and the non-emotional control condition, this point is negligible when analyzing the valencespecific cerebral correlates of sad vs. happy mood, which was the major focus of interest in this study. Finally, blocked designs are prone to habituation effects. However, testing for such effects exemplarily in a few regions of interest, we could not find any indications of habituation in the amygdala (see Habel et al., 2004), the ACC, and DLPFC. Hence, we conclude that basic emotions such as sadness and happiness rely on the activation of a common neural network including subcortical and cortical areas in which specific components are involved differentially and specifically for single emotions, whereby these complex interactions may substantially be influenced by external environmental and internal personal factors. Acknowledgments This work was supported by the German Research Foundation DFG, Schn 362/10-1. We thank B. Elghawaghi, H. Greenland, M. Grosse-Ruykeny, U. Felger, and B. Bewernick for assistance and support. References Badre, D., Wagner, A.D., 2004. Selection, integration, and conflict monitoring; assessing the nature and generality of prefrontal cognitive control mechanisms. Neuron 41, 473 – 487. Beauregard, M., Leroux, J.M., Bergman, S., Arzoumanian, Y., Beaudoin, G., Bourgouin, P., Stip, E., 1998. The functional neuroanatomy of major depression: an fMRI study using an emotional activation paradigm. NeuroReport 9, 3253 – 3258. Blair, R.J., Morris, J.S., Frith, C.D., Perrett, D.I., Dolan, R.J., 1999. Dissociable neural responses to facial expressions of sadness and anger. Brain 122, 883 – 893. Breiter, H.C., Etcoff, N.L., Whalen, P.J., Kennedy, W.A., Rauch, S.L., Buckner, R.L., Strauss, M.M., Hyman, S.E., Rosen, B.R., 1996. Response and habituation of the human amygdala during visual processing of facial expression. Neuron 17, 875 – 887. Canli, T., Zhao, Z., Desmond, J.E., Kang, E., Gross, J., Gabrieli, J.D., 2001. An fMRI study of personality influences on brain reactivity to emotional stimuli. Behav. Neurosci. 115, 33 – 42. Damasio, A.R., Grabowski, T.J., Bechara, A., Damasio, H., Ponto, L.L., Parvizi, J., Hichwa, R.D., 2000. Subcortical and cortical brain activity during the feeling of self-generated emotions. Nat. Neurosci. 3, 1049 – 1056. Fan, J., Flombaum, J.I., McCandliss, B.D., Thomas, K.M., Posner, M.I., 2003. Cognitive and brain consequences of conflict. NeuroImage 18, 42 – 57. George, M.S., Ketter, T.A., Parekh, P.I., Horwitz, B., Herscovitch, P., Post, R.M., 1995. Brain activity during transient sadness and happiness in healthy women. Am. J. Psychiatry 152, 341 – 351.

213

Gundel, H., O’Connor, M.F., Littrell, L., Fort, C., Lane, R.D., 2003. Functional neuroanatomy of grief: an FMRI study. Am. J. Psychiatry 160, 1946 – 1953. Habel, U., Klein, M., Shah, N.J., Toni, I., Zilles, K., Falkai, P., Schneider, F., 2004. Genetic load on amygdala hypofunction during sadness in non-affected brothers of schizophrenia patients. Am. J. Psychiatry 161, 1806 – 1813. Hamann, S.B., Ely, T.D., Hoffman, J.M., Kilts, C.D., 2002. Ecstasy and agony: activation of the human amygdala in positive and negative emotion. Psychol. Sci. 13, 135 – 141. Hariri, A.R., Mattay, V.S., Tessitore, A., Fera, F., Weinberger, D.R., 2003. Neocortical modulation of the amygdala response to fearful stimuli. Biol. Psychiatry 53, 494 – 501. Killgore, W.D., Yurgelun-Todd, D.A., 2001. Sex differences in amygdala activation during the perception of facial affect. NeuroReport 12, 2543 – 2547. Kilpatrick, L., Cahill, L., 2003. Amygdala modulation of parahippocampal and frontal regions during emotionally influenced memory storage. NeuroImage 20, 2091 – 2099. Lane, R.D., Reiman, R.M., Ahern, G.I., Schwartz, G.E., Davidson, R.J., 1997. Neuroanatomical correlates of happiness, sadness and disgust. Am. J. Psychiatry 154, 926 – 933. Lee, G.P., Meador, K.J., Loring, D.W., Allison, J.D., Brown, W.S., Paul, L.K., Pillai, J.J., Lavin, T.B., 2004. Neural substrates of emotion as revealed by functional magnetic resonance imaging. Cogn. Behav. Neurol. 17, 9 – 17. Levesque, J., Eugene, F., Joanette, Y., Paquette, V., Mensour, B., Beaudoin, G., Leroux, J.M., Bourgouin, P., Beauregard, M., 2003. Neural circuitry underlying voluntary suppression of sadness. Biol. Psychiatry 53, 502 – 510. Liberzon, I., Phan, K.L., Decker, L.R., Taylor, S.F., 2003. Extended amygdala and emotional salience: a PET activation study of positive and negative affect. Neuropsychopharmacology 28, 726 – 733. Markowitsch, H.J., Vandekerckhovel, M.M., Lanfermann, H., Russ, M.O., 2003. Engagement of lateral and medial prefrontal areas in the ecophory of sad and happy autobiographical memories. Cortex 39, 643 – 665. Mayberg, H.S., Liotti, M., Brannan, S.K., McGinnis, S., Mahurin, R.K., Jerabek, P.A., Silva, J.A., Tekell, J.L., Martin, C.C., Lancaster, J.L., Fox, P.T., 1999. Reciprocal limbic–cortical function and negative mood: converging PET findings in depression and normal sadness. Am. J. Psychiatry 156, 675 – 682. Murphy, F.C., Nimmo-Smith, I., Lawrence, A.D., 2003. Functional neuroanatomy of emotions: a meta-analysis. Cogn. Affect. Behav. Neurosci. 3, 207 – 233. Narumoto, J., Okada, T., Sadato, N., Fukui, K., Yonekura, Y., 2001. Attention to emotion modulates fMRI activity in human right superior temporal sulcus. Brain Res. Cogn. Brain Res. 12, 225 – 231. Pardo, J.V., Pardo, P.J., Raichle, M.E., 1993. Neural correlates of selfinduced dysphoria. Am. J. Psychiatry 150, 713 – 719. Parvizi, J., Anderson, S.W., Martin, C.O., Damasio, H., Damasio, A.R., 2001. Pathological laughter and crying: a link to the cerebellum. Brain 124, 1708 – 1719. Pelletier, M., Bouthillier, A., Levesque, J., Carrier, S., Breault, C., Paquette, V., Mensour, B., Leroux, J.M., Beaudoin, G., Bourgouin, P., Beauregard, M., 2003. Separate neural circuits for primary emotions? Brain activity during self-induced sadness and happiness in professional actors. NeuroReport 14, 1111 – 1116. Phan, K.L., Wager, T., Taylor, S.F., Liberzon, I., 2003. Functional neuroanatomy of emotion: a meta-analysis of emotion activation studies in PET and fMRI. NeuroImage 16, 331 – 348. Reiman, E.M., Lane, R.D., Ahern, G.L., Schwartz, G.E., Davidson, R.J., Friston, K.J., Yun, L.S., Chen, K., 1997. Neuroanatomical correlates of externally and internally generated human emotion. Am. J. Psychiatry 154, 918 – 925. Schmahmann, J.D., 2000. The role of the cerebellum in affect and psychosis. J. Neurolinguist. 13, 189 – 214.

214


Schneider, F., Gur, R.C., Gur, R.E., Muenz, L.R., 1994a. Standardized mood induction with happy and sad facial expressions. Psychiatry Res. 51, 19 – 31. Schneider, F., Gur, R.C., Jaggi, J.L., Gur, R.E., 1994b. Differential effects of mood on cortical cerebral blood flow: a 133xenon clearance study. Psychiatry Res. 52, 215 – 236. Schneider, F., Gur, R.E., Harper Mozley, L., Smith, R.J., Mozley, P.D., Censits, D.M., Alavi, A., Gur, R.C., 1995. Mood effects on limbic blood flow correlate with emotional self-rating: a PET study with oxygen-15 labeled water. Psychiatry Res.: NeuroImaging 61, 265 – 283. Schneider, F., Grodd, W., Weiss, U., Klose, U., Mayer, K.R., N7gele, T., Gur, R.C., 1997. Functional MRI reveals left amygdala activation during emotion. Psychiatry Res. 76, 75 – 82. Schneider, F., Weiss, U., Kessler, C., Salloum, J.B., Posse, S., Grodd, W., Mqller-G7rtner, H.-W., 1998. Differential amygdala activation in schizophrenia during sadness. Schizophr. Res. 34, 133 – 142. Schneider, F., Habel, U., Kessler, C., Salloum, J.B., Posse, S., 2000. Gender differences in regional cerebral activity during sadness. Hum. Brain Mapp. 9, 226 – 238 (Erratum: 2001, 13, 124). Schneider, F., Habel, U., Klein, M., Toni, I., Zilles, K., Falkai, P., Shah, N.J., submitted for publication. Behavioral correlates of olfactory

stimulation in non-affected brothers of schizophrenia patients. Biol. Psychiatry. Stefan, H., Schulze-Bonhage, A., Pauli, E., Platsch, G., Quiske, A., Buchfelder, M., Romstock, J., 2004. Ictal pleasant sensations: cerebral localization and lateralization. Epilepsia 45, 35 – 40. Talairach, J., Tournoux, P., 1988. Co-planar stereotaxic atlas of the human brain. Thieme, Stuttgart. Teasdale, J.D., Howard, R.J., Cox, S.G., Ha, Y., Brammer, M.J., Williams, S.C., Checkley, S.A., 1999. Functional MRI study of the cognitive generation of affect. Am. J. Psychiatry 156, 209 – 215. Watson, D., Clark, L.A., Tellegen, A., 1988. Development and validation of brief measures of positive and negative affect: the PANAS scales. J. Pers. Soc. Psychol. 54, 1063 – 1070. Weiss, U., Salloum, J.B., Schneider, F., 1999. Correspondence of emotional self-rating with facial expression. Psychiatry Res. 86, 175 – 184. Wild, B., Rodden, F.A., Grodd, W., Ruch, W., 2003. Neural correlates of laughter and humour. Brain 126, 2121 – 2138. Williams, M.A., Morris, A.P., McGlone, F., Abbott, D.F., Mattingley, J.B., 2004. Amygdala responses to fearful and happy facial expressions under conditions of binocular suppression. J. Neurosci. 24, 2898 – 2904.