emotion processing and regulation in bipolar disorder: a review

Review Article

Emotion processing and regulation in bipolardisorder: a review

Emotion processing involves the detection andevaluation of salient stimuli as well as the regula-tion of one�s emotional (affective) response to thesestimuli (1). Models of emotional processes includea number of components, such as: (i) attention toand perception of information that could poten-tially elicit emotional responses, (ii) subsequentemotional arousal to such stimuli, and (iii) theregulation of that arousal via several potentialmechanisms (i.e., altering attention or perceptionof stimuli and arousal responses and ⁄or usingcognitive reappraisal or other cognitive or

behavioral strategies to modulate arousal level).Dysregulated emotional responses can lead topathological mood states, including those seen inbipolar disorder (BP) (2, 3). BP is distinguishedprimarily by acute dysfunctional mood states,alternating between mania [BP type I (BP-I)] orhypomania [BP type II (BP-II)], and depression.This characteristic mood lability suggests a possi-ble dysfunction in neural networks involved inemotion processing and regulation. BP exacts atremendous cost not only to the patient, family,and friends, but also to society. The World Health

Townsend J, Altshuler LL. Emotion processing and regulation in bipolardisorder: a review.Bipolar Disord 2012: 14: 326–339. � 2012 The Authors.Journal compilation � 2012 John Wiley & Sons A ⁄S.

Objectives: Bipolar disorder (BP) is characterized by a dysfunction ofmood, alternating between states of mania ⁄ hypomania and depression.Thus, the primary abnormality appears to be an inability to regulateemotion, the result of which is emotional extremes. The purpose of thispaper is to review the current functional magnetic resonance imaging(fMRI) literature on adult patients with BP using emotion processing orregulation paradigms.

Methods: A search was conducted on PubMed using the keywords:bipolar disorder, fMRI, mania, bipolar depression, bipolar euthymia,emotion, and amygdala. Only those studies that were conducted in adultpatients using an emotion activation task were included in the final review.

Results: Using tasks that assess neural functioning during emotionprocessing and emotion regulation, many fMRI studies have examinedBP subjects during mania and euthymia. Fewer fMRI studies have beenconducted during depression, and fewer still have included the samesubjects in multiple mood states. Despite these limitations, these studieshave demonstrated specific abnormalities in frontal–limbic regions.Using a variety of paradigms, investigators have specifically evaluatedthe amygdala (a structure within the limbic system known to be criticalfor emotion) and the prefrontal cortex (PFC) (a region known to have aregulatory function over the limbic system).

Conclusions: These investigations reveal that amygdala activationvaries as a function of mood state, while the PFC remains persistentlyhypoactivated across mood states. Emotional dysregulation and labilityin mania and depression may reflect disruption of a frontal–limbicfunctional neuroanatomical network.

Jennifer Townsenda,b and Lori LAltshulera,b,c

aDepartment of Psychiatry and Biobehavioral

Sciences, David Geffen School of Medicine,bSemel Institute for Neuroscience and Human

Behavior, University of California, Los Angeles,cDepartment of Psychiatry, VA Greater Los Angeles

Healthcare System, West Los Angeles,

Los Angeles, CA, USA

doi: 10.1111/j.1399-5618.2012.01021.x

Key words: amygdala – bipolar disorder –

emotion processing – emotion regulation –

functional neuroimaging – vlPFC

Received 27 September 2011, revised and

accepted for publication 22 March 2012

Corresponding author:

Lori L. Altshuler, M.D.

Department of Psychiatry and Biobehavioral

Sciences

David Geffen School of Medicine

University of California, Los Angeles

300 UCLA Medical Plaza, Suite 1544

Box 957057

Los Angeles, CA 90095-7057

USA

Fax: 310-794-9915

E-mail: [email protected]

Bipolar Disorders 2012: 14: 326–339 � 2012 John Wiley and Sons A/S

BIPOLAR DISORDERS

326

Organization regards BP as one of the leadingcauses of disability throughout the world (4). Withthe high direct and indirect financial (5) andemotional costs of BP, it is imperative to betterunderstand and delineate its underlying causes forthe ultimate purpose of improved diagnosis andtreatment.As BP is primarily a disorder of mood, many BP

functional magnetic resonance imaging (fMRI)studies utilize tasks involving emotion processingand regulation. The inability to regulate emotion isthe primary abnormality of BP, the result of whichis the characteristic acute mood states of thedisorder. Using a variety of paradigms, manyneuroimaging studies specifically investigated theamygdala, a structure within the limbic systemknown to be critical for emotion, and theprefrontal cortex (PFC), a region known to havea regulatory function over the limbic system. MostfMRI studies examine BP subjects while euthymic,while fewer fMRI studies have been conducted onsubjects during mania and depression, and fewerstill include the same subjects in multiple moodstates. Despite these limitations, structural andfunctional imaging studies have demonstratedspecific abnormalities in frontal–limbic regionsand suggest that abnormalities in these circuitsmay result in the emergence of the manic anddepressive symptoms seen in BP (6, 7). This reviewwill focus only on the fMRI literature on emotionand emotion regulation in adults with BP. Thereexists considerable controversy on the childhoodand adolescent diagnosis of BP, and such discus-sion is beyond the scope of this review. Emotionaldysregulation in mania and depression may reflectdisruption of a functional neuroanatomical net-work with potential structural correlates, and thisnetwork will be the main focus of this review.

Healthy functioning

The amygdala, insula, anterior cingulate cortex(ACC), medial PFC (mPFC), and ventral lateralPFC (vlPFC) are considered key neural substratesof an emotion processing and regulation circuit (1).Neuroimaging studies have demonstrated a role forthe amygdala and insula in normal emotionprocessing, and for the medial and lateral regionsof the vlPFC in mood regulation (8, 9) andassociative emotional memory functions (10, 11).The amygdala plays a key role in signaling

danger and, more generally, signaling the emo-tional salience of sensory information to the rest ofthe brain in preparation for an appropriateresponse (12). Functional neuroimaging studiesin healthy subjects, using neuropsychological

paradigms that involve simple viewing of emotionstimuli, reliably activate the amygdala (13–15).Most emotion studies of healthy subjects useemotional faces as stimuli and demonstrate robustbilateral amygdala activation in response to faceswith both positive and negative valence (for areview, see 16), with the strongest amygdalaactivation occurring during the viewing of fearfulfaces. The amygdala is activated by valencedstimuli and may serve to strengthen the portrayalof emotional stimuli in sensory representationareas of the occipital and temporal cortex (17,18). Through neuronal projections to the brain-stem and hypothalamus, amygdala activation stim-ulates the autonomic nervous system (ANS) andthe hypothalamic–pituitary–adrenal (HPA) axis.This neural activity promotes physical and behav-ioral responses to the stimuli. The actions of thefast-acting ANS and slower-acting HPA systemenable an adaptive modulation of the stressresponse (19). Thus, the amygdala plays an essen-tial role in both the perception of emotional stimuliand the subsequent arousal reactions to suchincoming information. As such, it has been studiedextensively in healthy populations and in patientswith mood disorders.Emotion regulation involves the initial steps of

emotion processing (i.e., attention and perceptionof information eliciting emotional arousal) and thesubsequent mechanisms of modulating emotionalarousal. fMRI studies that involve modifyingemotional arousal, regardless of the mechanismemployed, typically reduced amygdala activationand activated the vlPFC and other frontal regions(13–15, 20). The vlPFC has been implicated inprocessing emotional salience (21) and motivation(22), and plays a role in integrating emotionalinformation and regulating the intensity of emo-tional responses (21, 23). Emotion regulation mayoccur via pathways between the vlPFC and auto-nomic systems that govern visceral responsesassociated with affective stimuli (24). The vlPFChas been implicated in other behavioral (response)inhibition studies and may serve a broader role ininhibiting undesirable or inappropriate states oractions (25). Dysfunction in this region mayprovide a mechanism for understanding the failureto modulate activation of a region involved inemotion, such as the amygdala.Tasks that require simple emotion processing (as

in the case of passive viewing) show greateramygdala activation than tasks that require emo-tion regulation (15, 26, 27). Emotion regulationcan generally be seen as the alteration of ongoingemotional responses through processes like alteredattention via the ACC or through implicit or

Emotion processing and regulation in bipolar disorder

327

explicit emotion regulation via frontal lobe activa-tion (28, 29). The frontal lobe connections toamygdala and other limbic structures may beinhibitory in nature, allowing for a dampening orreduction of amygdala activation and, in turn,perhaps a dampening of emotional arousal. fMRIstudies of emotion regulation in healthy subjectsdemonstrate engagement of frontal regions, includ-ing the vlPFC and ACC (27, 29, 30), and down-regulation of limbic regions by way of vlPFCactivation (15, 28).One commonly used technique to regulate emo-

tion is cognitive reappraisal, wherein subjectsrecruit cognitive resources to consciously reframethe context of salient emotional stimuli to modu-late their emotional response. In behavioral stud-ies, downregulation of emotion via cognitivereappraisal has been shown to decrease physiolo-gical arousal (31–33) and subjective reports ofdistress (34). Emotion regulation studies usingcognitive reappraisal in healthy subjects demon-strate reliable activation in the vlPFC, pre-supple-mentary motor area (pre-SMA), and lessfrequently in the ACC (34–39). Specifically, duringemotion downregulation using cognitive reap-praisal, healthy subjects show increased activationin the vlPFC, mPFC, and ACC (34, 40). Theseregulation studies also demonstrate reduced activ-ity in the amygdala during cognitive reappraisal,consistent with the vlPFC�s putative role in inhib-iting limbic activity.The regions implicated in emotion regulation

studies have extensive anatomical connections withthe amygdala (23, 24, 41). The amygdala hasextensive reciprocal connections with the frontallobe, with direct connections to the mPFC andvlPFC (42). Modulatory input from cortical brainregions, including the mPFC and vlPFC, dampenamygdala activation in the wake of emotionallyarousing events (12). Human (43) and non-humanprimate (44) anatomical studies show extensivereciprocal connections between the amygdala andPFC. Neurochemical studies in animals suggestthat this inhibitory connection between the amy-gdala and PFC is modulated via GABAergicinterneurons (45, 46). Prior studies of non-humanprimates have demonstrated a range of corticalregions, including the vlPFC, mPFC, insula, andtemporal lobe, with reciprocal connections to theamygdala (47, 48). The mPFC, with dense recipro-cal connections to the amygdala (49), may serve tomodulate amygdala activity or, more specifically,to regulate emotional response to salient stimuli.The insula is another region commonly implicatedin emotion processing and also has extensiveanatomical connections and shows significant

effective connectivity with the amygdala, ACC,and vlPFC (50–52). Sad and disgusted faces areknown to preferentially activate the insula inhealthy subjects (16), and this region may berecruited during the emotional response to poten-tially distressing interoceptive stimuli (53).In addition to studies investigating anatomical

connections between the amygdala and frontalregions, functional and effective connectivity stud-ies have demonstrated specific changes in connec-tivity as a response to emotional stimuli. Studies ofhealthy subjects report significant effective connec-tivity between the amygdala, vlPFC, insula, andACC (52), and significant negative functionalconnectivity between the amygdala and frontalregions (15, 27). Researchers have found increasedfunctional connectivity between the amygdala andregions mediating autonomic activity, such as theACC and brainstem, during stress induction (54).These studies suggest an antagonistic relationship(i.e., negative connectivity) between the amygdala –which attributes emotion salience to stimuli – andregions of the PFC and ACC, which are involvedin attention and cognition regulation. Thus, theamygdala can be seen as part of an adaptive systemthat increases the salience of emotional informa-tion (12). A counteractive system involving thePFC may regulate the impact of emotional stimuli,especially when such limbic activity interferes withcurrent task demands (18). As such, the effects of astressor may relate to activation of the autonomicstress response by the amygdala (54). Prolongedamygdala activation (or diminished frontal modu-lation) during the recovery period following expo-sure to emotion stimuli may relate to thedevelopment of psychopathology. An overload instress may impact the feedback circuit between theamygdala and vlPFC and mPFC, and contribute tothe pathogenesis of mood and anxiety disorders.This amygdala–ACC–mPFC and amygdala–vlPFC connectivity in healthy subjects is balancedto reflect emotion regulation through cognitivecontrol, and such a balance may be disrupted inpsychiatric populations (55). The decoupling ofthese regions has been demonstrated in relation toemotion dysregulation (2) and remains to be fullyexplored, specifically in BP.

Mania

Functional imaging studies during mania havebeen limited, no doubt because of difficulties inhaving manic patients remain still during scanning.Activation studies published to date have primarilyassessed frontal–amygdala activity using para-digms known to activate the amygdala in healthy

Townsend and Altshuler

328

subjects (15, 56–58). As an enlarged amygdala (59–61) and heightened activation (62) have beenreported in some subjects with BP, many fMRIstudies utilized emotion paradigms known toactivate the amygdala, such as those involvingemotional faces or scenes (15), in order to assessemotion reactivity and frontal–amygdala circuitryduring mania. Table 1 reviews the current fMRIliterature on adult subjects with mania whileperforming a variety of task involving emotion.One of the most consistent findings in mania is

heightened limbic activation in response to emo-tional stimuli, primarily emotional faces. A num-ber of studies specifically found increased leftamygdala activation compared to healthy subjectsduring simple emotion processing (i.e., viewing) ofemotional faces (63), emotional scenes (64), posi-tive pictures (65), and during implicit, but notexplicit, emotion processing (66). Another study,using patients in a variety of mood states, alsofound increased left amygdala activation inresponse to fearful faces (62), stimuli known toactivate the amygdala most reliably (16). Thisincreased left amygdala response during mania

correlated positively with Young Mania RatingScale (YMRS) scores (65). These studies areinteresting in light of several structural MRIstudies that reported an enlarged amygdala insubjects with BP (59, 60), although not all struc-tural studies in BP have found amygdala enlarge-ment (67). Although there is considerableconvergence in amygdala findings, particularly inthe left amygdala, at least one study found nosignificant differences in overall amygdala activa-tion between manic and healthy subjects (68). Thismay have been due to the relatively small samplesize (n = 10) or the use of only ‘‘sad’’ for thenegative faces in this study. However, manicsubjects did exhibit a decreased load response inthe bilateral amygdala when examining sad faces ofincreasing intensity (68), suggesting that the inten-sity of the emotional stimuli may contribute to whynot all studies in mania found amygdala hyperac-tivation. Another study found hyperactivation ofthe left amygdala in manic subjects irrespective ofthe valence of the stimuli, but found hypoactiva-tion of the right amygdala only during negativeemotion scenes (but not other stimuli) (64). Of the

Table 1. Amygdala and prefrontal cortex results of functional magnetic resonance imaging emotion studies in mania

Study N Paradigm Method Amygdala vlPFC

Elliot et al. 2004 (71) M = 8HS = 11

Emotional Go ⁄ NoGo (happy andsad words)

3TWhole-brain

NS fl L ⁄ R

Lennox et al. 2004 (68) M = 10HS = 12

Faces (happy, sad) 3TWhole-brain

NSa NS

Altshuler et al. 2005 (63) M = 9HS = 9

Faces (negative): match 3TROI: amygdala

› L fl L ⁄ R

Chen et al. 2006 (66) M = 8D = 8HS = 8

Faces (fear, sad, happy): implicitand explicit processing

3TWhole-brain

› Implicitfl Explicit

NS

Foland et al. 2008 (27) M = 9HS = 9

Faces (negative): match andlabel

3TWhole-brainPPI

› L fl L ⁄ R

Killgore et al. 2008 (72) M = 13HS = 13

Faces (fear): passive view 1.5TWhole-brainROI: amygdala

NS fl R

Bermpohl et al. 2009 (65) M = 10HS = 10

IAPS 1.5TWhole-brain

› L NS

Chen et al. 2010 (74) M = 12E = 9HS = 12

Faces (happy, sad, disgust, fear,surprise, anger): rate intensity

1.5TROI: amygdala, OFC

— › R

Van der Schot et al. 2010 (69) M = 12E = 18D = 12HS = 18

Faces (neutral, happy, fearful):masked and unmasked


fl L ⁄ R fl L ⁄ R

Strakowski et al. 2011 (64) M = 40HS = 36

CPT-END: IAPS 4TROI: amygdala, vlPFC

› Lfl Rb

fl L ⁄ R

vlPFC = ventral lateral prefrontal cortex; M = manic subjects; HS = healthy subjects; D = depressed subjects; E = euthymic subjects;IAPS = International Affective Picture Set; CPT-END = Continuous Performance Task with Emotional and Neutral Distracters;ROI = region of interest; PPI = prepulse inhibition; OFC = orbitofrontal cortex; NS = not significant; L = left; R = right.aAmygdala increases as load response increases in healthy subjects, but not in manic subjects.bNo main group effect. Interaction effects with healthy subjects emotion cues > manic emotional cues.


329

ten emotion studies in mania, only two (66, 69)found decreased amygdala activation in manicsubjects compared to healthy subjects. In one ofthese studies, Chen et al. (66) did report increasedamygdala activation in manic compared to euthy-mic BP subjects. Both of these studies hadrelatively modest sample sizes (n = 8 andn = 12, respectively) and utilized both positiveand negative valenced faces. These varied resultssuggest that the nature of amygdala abnormalitiesobserved during mania may depend, in part, on thevalence and salience of the stimuli used. Futurestudies that utilize more standardized paradigmsmay help to alleviate some of the ambiguity in theliterature.Several neuroimaging studies demonstrated

attenuated vlPFC function and ⁄or heightenedamygdala activation in manic compared to healthysubjects (58, 63, 66, 70, 71). fMRI studies of maniaspecifically probing emotion processing demon-strate hypoactivation of the vlPFC during theprocessing of negative faces (63), fear perception(72), masked positive and negative faces (69),emotional and neutral cues (64), and negativelycaptioned pictures (73). In fact, of the ten studies ofemotion in mania, six found evidence of hypo-activation of the vlPFC (27, 63, 64, 69, 71, 72),three found no group differences (65, 66, 68), andonly one found increased vlPFC activation inmanic subjects (74). While the previous studiesutilized negative emotional stimuli, it is importantto note that the one study of mania that foundincreased vlPFC activation used both positive andnegative emotional faces (74). This result is ofinterest because studies of healthy comparisonsubjects found engagement of the vlPFC duringthe processing of negative (i.e., fearful and angry)but not positive (happy) faces (16). As for the threestudies that failed to report any group differencesin the vlPFC, it is difficult to draw any specificconclusions since all three studies had relativelymodest sample sizes (n £ 10) and may have beenunderpowered to detect such differences.Increased amygdala activation and decreased

negative connectivity between the vlPFC andamygdala (27, 75) and between the amygdala andACC (76) have also been reported during mania.The amygdala has extensive reciprocal connectionsto the frontal lobe, including direct connections tothe mPFC, including the ACC and the orbitofron-tal cortex (OFC) (42). There is a striking conver-gence of studies using not only emotion paradigms,but also response inhibition paradigms, whichsuggests an attenuation of vlPFC function duringmania (58, 71). Furthermore, hypoactivation ofother frontal regions, including the dorsolateral

PFC (dlPFC) and mPFC, have also been reportedduring mania (62, 72). Extensive anatomical con-nectivity exists between the vlPFC, amygdala,cingulate, and other frontal regions (23, 24, 41).Ventral prefrontal areas are also connected withthe temporal polar and entorhinal temporal cortexand so have connections to the limbic cortex (42,77). Thus, it is not surprising to see abnormalitiesin these same regions during mania.Finally, one study of manic patients found

increased left insular activation in response to sadfaces (68). The presence of an attenuated vlPFCresponse and a heightened amygdala and insularesponse suggests another aspect to the alterationsin prefrontal–limbic circuits seen in mania. Studiesof healthy subjects reported significant effectiveconnectivity between the amygdala, insula, ACC,and vlPFC (52). Dysfunction in the vlPFC couldhelp to explain a failure to appropriately modulateother limbic brain regions and perhaps result in arange of intensity of mood shifts to hypomania andmania in patients with BP.

Depression

Patients with BP spend the majority of their time inepisodes of depression, not mania (78–80). Asbipolar depression is associated with high levels ofmorbidity and mortality across the life span, effortsaimed at identifying biological mechanisms thatcontribute to this phase of the illness are impera-tive. Functional neuroimaging studies of patientswith unipolar depressive disorder have begun toreveal dysregulated neuroanatomical circuits asso-ciated with this syndrome (2, 81). Similar studies inpatients with bipolar depression are more limited.Relatively few fMRI studies have been per-

formed in patients with BP during the depressedphase using emotion processing paradigms knownto activate limbic structures (66, 70, 73, 76, 82–84), with the latter two studies using subjects indifferent mood states (see Table 2 for a summaryof the amygdala and vlPFC fMRI findings inpatients with BP while depressed). Unlike studiesin mania, there was greater variability in theactivation patterns reported in bipolar depressedversus healthy comparison samples. In a studyinvolving positive- versus neutral-captioned pic-tures, distinct patterns of regional activation werefound with bipolar depressed patients, showingincreased right-sided subcortical activation (basalganglia, thalamus, hypothalamus, and left amyg-dala) compared to healthy subjects (73). A trendof left amygdala hyperactivation was observed inanother study using sad faces as stimuli, but thisamygdala hyperactivation was not statistically


330

significant or was not seen for other emotionstimuli (84). The remaining studies of emotionprocessing in bipolar depression did not consis-tently find limbic hyperactivation. Using a para-digm to assess implicit versus explicit facialemotion recognition, one study found that bipolardepressed subjects tended to overactivate fronto–striato–thalamic regions, but not the amygdala, inresponse to fearful faces (66). A more recentstudy using masked fearful and happy faces foundno significant differences between bipolar de-pressed and healthy subjects in the amygdala,but did find frontal hypoactivation, including ofthe vlPFC, in bipolar depressed subjects (69). Thisdecreased frontal lobe activation in bipolardepressed subjects was found in other studiesusing a variety of emotion paradigms (69, 73, 82)and represents a more consistent finding inbipolar depression.A few studies suggest reduced limbic activation

in bipolar depression. Unlike the previous studies,hyperactivation of amygdala was not seen inbipolar depressed subjects while performing aface-matching task. In fact, there was no significantamygdala activation seen in bipolar depressedsubjects (82), while reduced activations were seenin bilateral vlPFC [Brodmann�s area (BA) 47] andright dlPFC, and increased activation was reportedin mPFC (BA10) (82) in response to faces withnegative emotions. Using functional connectivitytechniques, Versace et al. (83) found increasedright and left amygdala–vlPFC connectivity inbipolar depressed subjects compared to healthysubjects while viewing sad faces, but decreasedamygdala–OFC connectivity in response to happyfaces. These results are consistent with findings in

mania that also show reduced amygdala–vlPFCconnectivity (27, 75) and suggest alterations inhealthy functional connectivity in both mania anddepression. The latter study (75) followed BPsubjects from initial manic mood state to thepresentation of a depressed mood and found thatfunctional connectivity between the amygdala andfrontal region changed as subjects changed moodstates. Another study, examining subjects in allthree mood states, also found decreased frontal–limbic connectivity while viewing fearful andhappy faces (76).These studies suggest hypoactivation of the

frontal lobe, especially the vlPFC, and abnormalconnectivity between frontal and limbic structuresthat may vary as a function of mood state and of thespecific emotion stimuli used. Abnormal amygdalaactivation is reported across many of the studies,but the direction of the finding may be dependenton the specific stimuli used. Significant increasedamygdala activation was reported in only one studythat utilized negative-captioned pictures (with thepictures themselves designed to evoke no emotionalresponse) (73). Studies using negative emotionalfaces as stimuli found no significant amygdaladifferences (66, 69), with one finding a trend level ofdecreased amygdala activation (82) between bipolardepressed and healthy groups. Given the multiplic-ity of paradigms used, the small sample sizes used(all studies had n £ 15), and the heterogeneity ofresults, it is difficult at this time to discern a clearfunctional neuroanatomical circuit abnormalityconsistently demonstrated during bipolar depres-sion, apart from frontal hypoactivation and de-creased frontal–limbic connectivity. Clearly, morestudies are needed, with larger sample sizes.

Table 2. Amygdala and prefrontal cortex results of functional magnetic resonance imaging emotion studies in bipolar depression

Study (depressed) N Paradigm Method Amygdala vlPFC

Malhi et al. 2004 (73) D = 11HS = 10

Captioned pictures (negative, neutral,positive): passive view

1.5TWhole-brain

› L fla

Chen et al. 2006 (66) M = 8D = 8HS = 8

Faces (fear, sad, happy): implicit andexplicit processing

3TWhole-brain

NS NS

Altshuler et al. 2008 (82) D = 11HS = 17

Faces (negative): match 3TWhole-brain

NS fl L ⁄ R

Almeida et al. 2010 (84) D = 15E = 15HS = 15

Faces (fear, happy, sad): label 3TROI: amygdala

NSb N ⁄ A

Van der Schot et al. 2010 (69) M = 12E = 18D = 12HS = 18

Faces (neutral, happy, fearful): maskedand unmasked


NS fl L ⁄ R

vlPFC = ventral lateral prefrontal cortex; D = depressed subjects; HS = healthy subjects; M = manic subjects; E = euthymic subjects;ROI = region of interest; OFC = orbitofrontal cortex; L = left; R = right; NS = not significant; N ⁄ A = not applicable.a Decrease observed in another region of the inferior frontal gyrus (Brodmann’s area 9) in the negatively valenced condition.bTrend level increased amygdala in bipolar depressed group.


331

Euthymia

The enduring impairments in euthymic patientsprovide clues to the brain regions involved in theprimary pathology of BP or that are impacted bythe presence of the disorder. Studies suggest thatpatients with BP continue to display mood insta-bility or increased mood reactivity even in theabsence of an acute episode (79). An enduringregional brain abnormality in euthymia may affectthe neural circuit involved in homeostatic moodregulation. This abnormality, in turn, may pre-dispose BP patients to relapse into acute moodstates.Of the 14 studies examining emotion processing

in euthymic BP subjects, the majority – nine studies– found no significant differences between euthy-mic BP subjects and healthy subjects in amygdala

activation (85–90). Table 3 provides a summary ofthe current fMRI literature of emotion processingand regulation in bipolar euthymia. It is interestingto note that all of the studies that found nosignificant amygdala differences between euthymicBP and healthy groups employed negative emo-tional faces as stimuli. Two of these studiesexamined medication effects and neither foundsignificant effects of medication on amygdalaactivation in bipolar euthymia (88, 90). By con-trast, three studies found increased amygdalaactivation in the right hemisphere using an emo-tional Stroop task (91) and an emotional face task(74), and in the left hemisphere using happy faces(92). Finally, two studies found decreased amyg-dala activation in euthymic BP subjects comparedto healthy comparison subjects. One study, usingthe emotional Stroop task, found decreased left

Table 3. Amygdala and prefrontal cortex (PFC) results of functional magnetic resonance imaging (fMRI) emotion studies in bipolar euthymia

Study (Euthymic) N Paradigm Method Amygdala vlPFC

Malhi et al. 2005 (93) E = 12HS = 12

Emotional Stroop: (affectiveversus neutral)

3TWhole-brain, GSR

fl L fl L

Lagopoulos and Malhi2007 (91)

E = 10HS = 10

Emotional Stroop: (negativeversus neutral)

3TWhole-brain, GSR

› R fl L ⁄ R (PFC)

Malhi et al. 2007 (89) E = 10HS = 10

Faces (fear, disgust, neutral):label

3TWhole-brain

NS fl L

Wessa et al. 2007 (101) E = 17HS = 17

Emotional Go ⁄ NoGo: (fear,happy, and neutral faces)

1.5TWhole-brain

NS NS

Hassel et al. 2008 (88) E = 19HS = 24

Faces (happy, fear, neutral):label

3TWhole-brainROI: amygdala

NS fl L ⁄ R (PFC)

Robinson et al. 2008 (90) E = 15HS = 16

Faces (negative): match andlabel

3TWhole-brainROI: amygdala, vlPFC

NS › R

Almeida et al. 2009 (85) E = 21HS = 25

Faces (happy, neutral): label 3TWhole-brainDCM: parahippocampus,vmPFC, dlPFC

NS NS

Hassel et al. 2009 (87) E = 14HS = 16

Faces (happy, fear): label 3TROI: amygdala, dPFC

NS fl R

Almeida et al. 2010 (84) D = 15E = 15HS = 15

Faces (fear, happy, sad): label 3TROI: amygdala

NS N ⁄ A

Chen et al. 2010 (74) M = 12E = 9HS = 12

Faces (happy, sad, disgust, fear,surprise, anger): rate intensity


› R › R

Surguladze et al. 2010 (92) E = 20HS = 20

Faces (happy, fear): label Whole-brainROI: amygdala

› La NS

Van der Schot et al.2010 (69)

M = 12E = 18D = 12HS = 18

Faces (neutral, happy, fearful):masked and unmasked


fl L ⁄ R fl L ⁄ R

Lagopoulos and Malhi2011 (99)

E = 11HS = 11

Faces (neutral or disgust): label 3TfMRI, GSR

NS NS

vlPFC = ventral lateral prefrontal cortex; E = euthymic subjects; HS = healthy subjects; D = depressed subjects; M = manic subjects;GSR = galvanic skin response; ROI = region of interest; DCM = dynamic causal modeling; vmPFC = ventromedial prefrontal cortex;dlPFC = dorsolateral prefrontal cortex; dPFC = dorsal prefrontal cortex; OFC = orbitofrontal cortex; L = left; R = right; NS = notsignificant; N ⁄ A = not applicable.aFor intense happy faces only.


332

amygdala activation in a small sample (n = 12) ofeuthymic BP women (93); and another study, usingmasked emotional faces, found bilateral amygdalahypoactivation (69). From the literature, it appearsthat abnormalities of amygdala activation aremore consistently found during acute mood statesthan during euthymia.Similar to these fMRI studies, positron emission

tomography (PET) studies also suggest frontal–limbic circuit abnormalities in euthymic BPpatients. One PET study measured regional bloodflow while euthymic subjects completed a novelmotor sequence task (94). Patients displayed aunique pattern of widespread limbic networkactivity in response to the new sequence, whereashealthy subjects activated a spatial attentioncircuit. Patients failed to allocate attentionalresources and instead utilized limbic circuitry(i.e., suggesting arousal rather than attention) toalter their performance on a non-emotional task.This pattern of BP subjects showing significantlygreater engagement of limbic regions in response tonon-emotional tasks was seen in a number of fMRIstudies (95–97). This finding was also replicated inan fMRI study using medication-free euthymicpatients (98). These common abnormalities duringeuthymia – of increased limbic activation duringnon-emotional tasks and decreased vlPFC activa-tion during emotion processing – suggest enduringtrait deficits of BP. Thus, euthymic BP subjectsappear to engage limbic structures similarly duringemotional tasks, but differ from control subjects byactivating these same regions even in the absence ofemotional stimuli.While amygdala function appears normalized

during emotion tasks, many of the previous studiesreported abnormalities in frontal functioning dur-ing emotion processing in euthymia. Many studies,using a variety of emotional stimuli, founddecreased frontal activation in bipolar euthymia(69, 88, 91, 93, 99), most frequently in the vlPFC.Decreased activation in the vlPFC during emotionprocessing was also observed in two studies of BPsubjects in euthymic or depressed mood states(100). However, not all studies have reported thisfrontal hypoactivation. Two studies foundincreased activation in the right vlPFC (74, 90) ineuthymic compared to healthy subjects whileviewing emotional faces. It is less clear why thesetwo studies found frontal abnormalities in theopposite direction from those in the majority ofstudies; differences in task design analysis [regionof interest (ROI) in one versus whole-brain anal-ysis in the other] and subjects (small sample sizes)likely contributed to the vlPFC differences betweenthese two studies and the other ones.

Decreases in dlPFC activation were also re-ported in a number of these studies (88, 91),perhaps suggesting impairment in frontal regionsbeyond those predictably activated in studies ofemotion in healthy subjects. Furthermore, fivestudies of emotion in bipolar euthymia foundhyperactivation of striatal regions, including thebilateral caudate (99, 101), left (92) and right (87)putamen, and left caudate ⁄putamen (88). Imagingstudies using affective induction and emotionregulation paradigms that are designed to identifyregions most sensitive to provocation by externalstressors may help to elucidate the mechanismsmediating clinical relapse into mania or depression.Neural correlates of these clinical trait or diseasediathesis markers are just beginning to be charac-terized.

Comparison of mood states

Very few neuroimaging studies have directly com-pared BP groups in different mood states using thesame scanning protocol and fewer still have com-pared the same subjects while in different moodstates. To date, only five studies have directlycompared BP groups using the same emotionalparadigm (66, 69, 74, 83, 85) and only one used thesame subject in two mood states (74). Manicsubjects, when compared directly to bipolardepressed subjects, show hyperactivation of theamygdala and left fusiform gyrus, a region impor-tant in the processing of emotional faces, during animplicit emotion processing task (66). Other stud-ies, however, found less amygdala activation inmanic compared to euthymic subjects (74) andbetween manic and depressed subjects (69). Both ofthese studies included both positive and negativeemotional faces. Chen et al. (74), as the only studythat compared the same subjects during mania andeuthymia, found a significant effect of time ·group only in the left amygdala ⁄ hippocampus.From studies of healthy subjects, amygdala acti-vation is most robust during the viewing of fearfuland then sad faces (16). As such, the direction ofamygdala abnormalities seen in mania may dependon the specific valence of the stimuli used. Futurestudies that specifically explore these issues areneeded, especially using larger sample sizes ofsubjects in multiple mood states.Manic subjects, compared to depressed subjects,

show hypoactivation in frontal lobe regions,including the vlPFC (66, 69) and dlPFC (69).These same abnormalities are found in BP subjectsacross mood states, but these studies provideevidence that the most profound hypoactivationis seen during mania and that although frontal lobe


333

activation normalizes during euthymia, it remainshypoactive compared to healthy comparison sub-jects. This similar pattern of more pronouncedabnormalities emerging during the acute moodstates is seen in bipolar depression. Bipolardepressed subjects compared to bipolar euthymicsubjects showed hyperactivation of the left amyg-dala to sad faces (85) and of the right amygdala tomasked emotional faces (69). Bipolar depressedsubjects, but not bipolar euthymic subjects, alsoshowed decreased amygdala–vlPFC functionalconnectivity compared to healthy subjects whileviewing happy faces (83).Our group, using a well-validated emotional

faces paradigm in previously published groups ofmanic (63) and depressed (82) BP subjects, per-formed direct comparisons of these manic anddepressed groups, as well as a direct comparison tobipolar euthymic subjects. Prior published workshowed significant amygdala hyperactivation inmanic compared to healthy subjects (63). Manicsubjects (n = 9) also showed significant amygdalahyperactivation compared to both bipolardepressed (n = 11) and bipolar euthymic (n = 9)subjects (Fig. 1). Unfortunately, there were too fewsubjects in the same mood state to provide anymeaningful analysis, but, clearly, future longitudi-nal studies will provide valuable information in thestate versus trait issues that are raised in BPneuroimaging studies.

Limitations

Some of the limitations of the bipolar neuroimag-ing literature are general to neuroimaging, whileothers are specific to this patient population.Problems with neuroimaging studies, particularlywith psychiatric populations, include the use ofrelatively small sample sizes that may have under-powered the studies. Also, neuroimaging is arelatively new science and uses techniques not yetperfected, leading to interpretations of data thatare not yet fully understood (102). As such, carefulattention to the design of studies, with the limita-tion of each specific neuroimaging method in mind(i.e., PET studies have poor spatial resolution,while fMRI studies have poor temporal resolution)is required. For example, the use of global signalscaling in many studies has been called intoquestion, since use of such a statistical methodmay greatly distort findings in structures related toemotional processing. This distortion may contrib-ute to conflicting findings (103). In addition, theamygdala and vlPFC are regions sensitive to signalloss and image distortions from susceptibility nearthe air ⁄ tissue borders (104). As such, it is impor-tant to recognize that the regions most implicatedin BP are also the same regions most challenging toimage using fMRI. The use of sequences optimizedto minimize such artifacts, such as using a shortecho time, are essential in emotion studies.

Fig. 1. Manic bipolar disorder (BP) subjects (n = 9) show significant amygdala hyperactivation compared to (A) bipolar depressedsubjects (n = 11) (x = –24, y = –8, z = –14) and (B) euthymic BP subjects (n = 9) (x = –18, y = –6, z = –16) during anemotional face matching task.


334

Other limitations are present that are inherent inworking with this population. The first andperhaps most pertinent is the effect of medicationon both the neuroimaging data and the interpre-tation of findings. Virtually all studies using BPpopulations include patients on a variety ofpsychotropic medications, which limits the abilityto compare across groups. Studies that try to avoidthis complication by using unmedicated popula-tions (98) also limit the ability to generalize theirfindings, since these unmedicated patients may bean unrepresentative subset of BP patients. Theethics of stopping or delaying medication forresearch make it difficult to design a controlled,double-blind study that could provide more defin-itive results. Also, certain medications may affectblood flow, which is indirectly measured in fMRIstudies (105), may affect activation in select regions(7), and can change gray matter volume in BPpatients (106). Thus, medication issues must beconsidered when reviewing, interpreting, and gen-eralizing findings from studies. However, it ispossible that future studies could recruit patientsas soon as they present with acute mood systemsand carefully monitor the effects of medication onboth mood state and neural functioning.The emotion dysregulation seen in both depres-

sion and mania suggests possible dysfunction ofneural networks involved in emotion regulationand network interconnectivity. Connectivityamong neural networks involved in emotion reg-ulation remains understudied in BP, despite thefact that emotion dysregulation is a definingcriterion for the disorder. Future studies thatincorporate functional and structural connectivitywill help to deepen our understanding of theunderlying pathophysiology in BP.Another problem, not specific to neuroimaging,

is the use of heterogeneous populations by includ-ing multiple mood states in a single study withoutaccounting for state differences (70, 107–109) orfailing to report the mood state (96). A similarconfound arises from combining BP-I ⁄BP-IIpatients (110). In addition, many studies either failto report the presence ⁄absence of other psychiatriccomorbidities, including substance abuse andattention-deficit hyperactivity disorder (ADHD),or fail to take these comorbidities into account inthe analysis. BP subjects may also be heteroge-neous with regard to illness duration, number ofprior acute mood episodes, and past history ofpsychosis. These potential confounds make itdifficult to differentiate trait differences from medi-cation and illness effects on fMRI activationpatterns. Nevertheless, neuroimaging provides auseful and powerful tool to extend the existing

psychological literature and is essential for linkingunderlying anatomical and functional changes tothose seen behaviorally.

Conclusions

Despite the heterogeneity of findings, most studiesconverge with regard to the amygdala and vlPFC.In subjects with BP, the amygdala shows hyperac-tivation in mania, variable activation during depres-sion, and normal activation during euthymia whencompared to healthy subjects. The other regionsthat show state-dependent findings include theACC, which shows reduced activation duringmania (64, 68, 72), and the striatum, which showsincreased activation during euthymia (87, 88, 92,99, 101). These regions are involved in the emotionsalience network and the response inhibition net-work, respectively, and have been implicated in BP.The trait-related characteristic seen that endures

across mood states is hypoactivation of the ventralPFC. Reduced vlPFC activation has been reportedin the literature in manic, depressive, and euthymicmood states. The vlPFC may therefore participatein aspects of emotional processing and expressionthat are not exclusively related to a specific valenceof emotion. Lack of normal functioning in thisregion might result in dysregulation of mood toeither pole (111–113), and dysfunction of thevlPFC–amygdala circuit may represent a trait-dependent phenomenon of BP. Lesion studiessupport the role of the vlPFC in emotion regulation,as impairment is associated with manic and depres-sive symptoms (111, 112). Decreased vlPFC thusmay be a trait-related characteristic that enduresacross mood states in BP (114) and may participatein themodulation of emotion not exclusively relatedto a specific valence of emotion. Both structural andfunctional neuroimaging studies demonstrateabnormalities in frontolimbic circuits in both themanic and depressive symptoms seen in BP (6, 7).The vlPFC hypoactivation in subjects with BP

has been reported not only in emotion regulationtasks, but also in other tasks requiring cognitivecontrol, including response inhibition studies inmania during both emotion (71) and neutralcontexts (115). Similar hypoactivation of thevlPFC is seen during bipolar depression (57) andeuthymia (57, 91, 116, 117). Chronic vlPFC hyp-oactivation and abnormal modulatory control oflimbic structures may help to explain the continuedmood instability and reactivity that BP patientsexhibit even in the absence of an acute episode (78,79). It is possible that dysfunction of the frontal–limbic circuit increases the vulnerability of BPsubjects to lapse into mood episodes, as numerous


335

studies suggest that stressful life events predictrelapse into acute mood states in BP (see 117 for areview). This vlPFC hypoactivation may representa state-independent neural marker of BP, respon-sible for a range of cognitive and emotionaldifferences in BP subjects compared to theirhealthy counterparts. The vlPFC may act as amodulator on the experience of extreme emotion inmania or depression. Hypoactivation of this regionin BP may leave this region unable to regulatelimbic structures when it is needed. vlPFC dys-function may result in a dysregulation of emotionalreactions and the susceptibility of BP subjects torelapse into mania or depression.

Future aims

As previously noted, few existing studies involvemedication-free subjects. Thus, the issue of medi-cation is a confound in almost all of the existingliterature, and future studies are needed to clarifyspecifically the effects of medication on the variousneurocognitive domains measured. Questions alsoremain as to what extent cognitive dysfunctions arepresent before illness onset. Longitudinal studiesfollowing at-risk subjects might not only help toanswer these questions, but also address whethercognitive impairments are stable or progressive,and help us to develop insight into the processunderlying the mechanism for switches into acutemood states. Traditional fMRI methods combinedwith more nascent technologies, such as resting-state neuroimaging, will continue to shed light ontrait deficits in BP and help us to define thefunctional neuroanatomical underpinnings of thiscomplex condition.

Disclosures

LLA has received past funding from Forest Laboratories(advisory board honoraria, March 2008), Sepracor (advisoryboard honoraria, January 2010), and Eli Lilly & Co. (consul-tant, September 2010). JT does not have any commercialassociations that might pose a conflict of interest in connectionwith this manuscript.

References

1. Phillips ML, Drevets WC, Rauch SL, Lane R. Neuro-biology of emotion perception I: the neural basis ofnormal emotion perception. Biol Psychiatry 2003; 54:504–514.

2. Phillips ML, Drevets WC, Rauch SL, Lane R. Neurobi-ology of emotion perception II: implications for majorpsychiatric disorders. Biol Psychiatry 2003; 54: 515–528.

3. Critchley H. Emotion and its disorders. Br Med Bull 2003;65: 35–47.

4. Murray CJ, Lopez AD. Evidence-based health policy–lessons from the global burden of disease study. Science1996; 274: 740–743.

5. Simon GE. Social and economic burden of mood disor-ders. Biol Psychiatry 2003; 54: 208–215.

6. Altshuler LL, Townsend JD. Functional brain imaging inbipolar disorder. In: Strakowski SM ed. The BipolarBrain: Integrating Neuroimaging and Genetics. NewYork: Oxford University Press, 2012; 53–77.

7. Strakowski SM, Delbello MP, Adler CM. The functionalneuroanatomy of bipolar disorder: a review of neuroim-aging findings. Mol Psychiatry 2005; 10: 105–116.

8. Baker SC, Frith CD, Dolan RJ. The interaction betweenmood and cognitive function studied with PET. PsycholMed 1997; 27: 565–578.

9. Northoff G, Richter A, Gessner M et al. Functionaldissociation between medial and lateral prefrontal corticalspatiotemporal activation in negative and positive emo-tions: a combined fMRI ⁄MEG study. Cereb Cortex 2000;10: 93–107.

10. Price JL. Comparative aspects of amygdala connectivity.Ann NY Acad Sci 2003; 985: 50–58.

11. Bookheimer S. Functional MRI of language: newapproaches to understanding the cortical organization ofsemantic processing. Annu Rev Neurosci 2002; 25: 151–188.

12. LeDoux JE. Emotion circuits in the brain. Annu RevNeurosci 2000; 23: 155–184.

13. Adolphs R. Fear, faces, and the human amygdala. CurrOpin Neurobiol 2008; 18: 166–172.

14. Breiter HC, Etcoff NL, Whalen PJ et al. Response andhabituation of the human amygdala during visual proc-essing of facial expression. Neuron 1996; 17: 875–887.

15. Hariri AR, Bookheimer SY, Mazziotta JC. Modulatingemotional responses: effects of a neocortical network onthe limbic system. NeuroReport 2000; 11: 43–48.

16. Fusar-Poli P, Placentino A, Carletti F et al. Functionalatlas of emotional faces processing: a voxel-based meta-analysis of 105 functional magnetic resonance imagingstudies. J Psychiatry Neurosci 2009; 34: 418–432.

17. Morris JS, Friston KJ, Buchel C et al. A neuromodula-tory role for the human amygdala in processing emotionalfacial expressions. Brain 1998; 1: 47–57.

18. Pessoa L, Kastner S, Ungerleider LG. Attentional controlof the processing of neural and emotional stimuli. BrainRes Cogn Brain Res 2002; 15: 31–45.

19. Joels M, Baram TZ. The neuro-symphony of stress. NatRev Neurosci 2009; 10: 459–466.

20. Lieberman MD, Eisenberger NI, Crockett MJ, Tom SM,Pfeifer JH, Way BM. Putting feelings into words: affectlabeling disrupts amygdala activity in response to affectivestimuli. Psychol Sci 2007; 18: 421–428.

21. Cabeza R, Nyberg L. Imaging cognition II: an empiricalreview of 275 PET and fMRI studies. J Cogn Neurosci2000; 12: 1–47.

22. Tucker DM, Luu P, Pribram KH. Social and emotionalself-regulation. Ann NY Acad Sci 1995; 769: 213–239.

23. Fuster JM. The prefrontal cortex–an update: time is of theessence. Neuron 2001; 30: 319–333.

24. Morris JS, Dolan RJ. Dissociable amygdala and orbito-frontal responses during reversal fear conditioning.Neuroimage 2004; 22: 372–380.

25. Aron AR. The neural basis of inhibition in cognitivecontrol. Neuroscientist 2007; 13: 214–228.

26. Hariri AR, Tessitore A, Mattay VS, Fera F, WeinbergerDR. The amygdala response to emotional stimuli: a


336

comparison of faces and scenes. Neuroimage 2002; 17: 317–323.

27. Foland LC, Altshuler LL, Bookheimer SY, EisenbergerN, Townsend J, Thompson PM. Evidence for deficientmodulation of amygdala response by prefrontal cortex inbipolar mania. Psychiatry Res 2008; 162: 27–37.

28. Ochsner KN, Ray RR, Hughes B et al. Bottom-up andtop-down processes in emotion generation: common anddistinct neural mechanisms. Psychol Sci 2009; 20: 1322–1331.

29. Ochsner KN, Gross JJ. The cognitive control of emotion.Trends Cogn Sci 2005; 9: 242–249.

30. Goldin PR, McRae K, Ramel W, Gross JJ. The neuralbases of emotion regulation: reappraisal and suppressionof negative emotion. Biol Psychiatry 2008; 63: 577–586.

31. Gross JJ. Emotion regulation: affective, cognitive, andsocial consequences. Psychophysiol 2002; 39: 281–291.

32. Jackson DC, Malmstadt JR, Larson CL, Davidson RJ.Suppression and enhancement of emotional responses tounpleasant pictures. Psychophysiol 2000; 37: 515–522.

33. Driscoll D, Tranel D, Anderson SW. The effects ofvoluntary regulation of positive and negative emotion onpsychophysiological responsiveness. Int J Psychophysiol2009; 72: 61–66.

34. Ochsner KN, Bunge SA, Gross JJ, Gabrieli JD. Rethink-ing feelings: an FMRI study of the cognitive regulation ofemotion. J Cogn Neurosci 2002; 14: 1215–1229.

35. Levesque J, Eugene F, Joanette Y et al. Neural circuitryunderlying voluntary suppression of sadness. Biol Psychi-atry 2003; 53: 502–510.

36. Phan KL, Fitzgerald DA, Nathan PJ, Moore GJ, UhdeTW, Tancer ME. Neural substrates for voluntary sup-pression of negative affect: a functional magnetic reso-nance imaging study. Biol Psychiatry 2005; 57: 210–219.

37. McRae K, Hughes B, Chopra S, Gabrieli JD, Gross JJ,Ochsner KN. The neural bases of distraction and reap-praisal. J Cogn Neurosci 2010; 22: 248–262.

38. Koenigsberg HW, Fan J, Ochsner KN et al. Neuralcorrelates of using distancing to regulate emotionalresponses to social situations. Neuropsychologia 2010;48: 1813–1822.

39. Kanske P, Heissler J, Schonfelder S, Bongers A, Wessa M.How to regulate emotion? Neural networks for reap-praisal and distraction. Cereb Cortex 2010; 21: 1379–1388.

40. Schaefer SM, Jackson DC, Davidson RJ, Aguirre GK,Kimberg DY, Thompson-Schill SL. Modulation of amy-gdalar activity by the conscious regulation of negativeemotion. J Cogn Neurosci 2002; 14: 913–921.

41. Price JL, Carmichael ST, Drevets WC. Networks relatedto the orbital and medial prefrontal cortex; a substrate foremotional behavior? Prog Brain Res 1996; 107: 523–536.

42. Van Hoesen GW, Pandya DN, Butters N. Corticalafferents to the entorhinal cortex of the Rhesus monkey.Science 1972; 175: 1471–1473.

43. Ben-Ari Y, Institut national de la sante et de la recherchemedicale (France), Centre national de la recherche scien-tifique (France), . The Amygdaloid Complex: Proceedingsof the International Symposium on the AmygdaloidComplex. September 1-4, Senlis, France: Elsevier ⁄North-Holland Biomedical Press, 1981.

44. Roberts AC, Tomic DL, Parkinson CH et al. Forebrainconnectivity of the prefrontal cortex in the marmosetmonkey (Callithrix jacchus): an anterograde and retro-grade tract-tracing study. J Comp Neurol 2007; 502:86–112.

45. Cunningham MG, Bhattacharyya S, Benes FM. Amygd-alo-cortical sprouting continues into early adulthood:

implications for the development of normal and abnormalfunction during adolescence. J Comp Neurol 2002; 453:116–130.

46. Perez-Jaranay JM, Vives F. Electrophysiological study ofthe response of medial prefrontal cortex neurons tostimulation of the basolateral nucleus of the amygdalain the rat. Brain Res 1991; 564: 97–101.

47. Kita H, Kitai ST. Amygdaloid projections to the frontalcortex and the striatum in the rat. J Comp Neurol 1990;298: 40–49.

48. Mufson EJ, Mesulam MM, Pandya DN. Insular inter-connections with the amygdala in the rhesus monkey.Neurosci 1981; 6: 1231–1248.

49. Ghashghaei HT, Barbas H. Pathways for emotion:interactions of prefrontal and anterior temporal pathwaysin the amygdala of the rhesus monkey. Neurosci 2002;115: 1261–1279.

50. Augustine JR. Circuitry and functional aspects of theinsular lobe in primates including humans. Brain ResBrain Res Rev 1996; 22: 229–244.

51. Nagai M, Kishi K, Kato S. Insular cortex and neuropsy-chiatric disorders: a review of recent literature. EurPsychiatry 2007; 22: 387–394.

52. Stein JL, Wiedholz LM, Bassett DS et al. A validatednetwork of effective amygdala connectivity. Neuroimage2007; 36: 736–745.

53. Husted DS, Shapira NA, Goodman WK. The neurocir-cuitry of obsessive-compulsive disorder and disgust. ProgNeuropsychopharmacol Biol Psychiatry 2006; 30: 389–399.

54. van Marle HJ, Hermans EJ, Qin S, Fernandez G. Fromspecificity to sensitivity: how acute stress affects amygdalaprocessing of biologically salient stimuli. Biol Psychiatry2009; 66: 649–655.

55. Etkin A, Prater KE, Schatzberg AF, Menon V, GreiciusMD. Disrupted amygdalar subregion functional connec-tivity and evidence of a compensatory network in gener-alized anxiety disorder. Arch Gen Psychiatry 2009; 66:1361–1372.

56. Blumberg HP, Stern E, Ricketts S et al. Rostral andorbital prefrontal cortex dysfunction in the manic state ofbipolar disorder. Am J Psychiatry 1999; 156: 1986–1988.

57. Blumberg HP, Leung HC, Skudlarski P et al. A functionalmagnetic resonance imaging study of bipolar disorder:state- and trait-related dysfunction in ventral prefrontalcortices. Arch Gen Psychiatry 2003; 60: 601–609.

58. Rubinsztein JS, Fletcher PC, Rogers RD et al. Decision-making in mania: a PET study. Brain 2001; 124: 2550–2563.

59. Altshuler L, Bartzokis G, Grieder T, Curran J, Mintz J.Amygdala enlargement in bipolar disorder and hippo-campal reduction in schizophrenia: an MRI study dem-onstrating neuroanatomic specificity. Arch GenPsychiatry 1998; 55: 663–664.

60. Strakowski SM, DelBello MP, Sax KW et al. Brainmagnetic resonance imaging of structural abnormalitiesin bipolar disorder. Arch Gen Psychiatry 1999; 56: 254–260.

61. Brambilla P, Harenski K, Nicoletti M et al. MRI inves-tigation of temporal lobe structures in bipolar patients.J Psychiatr Res 2003; 37: 287–295.

62. Yurgelun-Todd DA, Gruber SA, Kanayama G, KillgoreWDS, Baird AA, Young AD. fMRI during affectdiscrimination in bipolar affective disorder. Bipolar Dis-ord 2000; 2: 237–248.

63. Altshuler L, Bookheimer S, Proenza MA et al. Increasedamygdala activation during mania: a functional magnetic


337

resonance imaging study. Am J Psychiatry 2005; 162:1211–1213.

64. Strakowski SM, Eliassen JC, Lamy M et al. Functionalmagnetic resonance imaging brain activation in bipolarmania: evidence for disruption of the ventrolateralprefrontal-amygdala emotional pathway. Biol Psychiatry2011; 69: 381–388.

65. Bermpohl F, Dalanay U, Kahnt T et al. A preliminarystudy of increased amygdala activation to positive affec-tive stimuli in mania. Bipolar Disord 2009; 11: 70–75.

66. Chen CH, Lennox B, Jacob R et al. Explicit and implicitfacial affect recognition in manic and depressed States ofbipolar disorder: a functional magnetic resonance imagingstudy. Biol Psychiatry 2006; 59: 31–39.

67. Pearlson GD, Barta PE, Powers RE et al. Medial andsuperior temporal gyral volumes and cerebral asymmetryin schizophrenia versus bipolar disorder. Biol Psychiatry1997; 41: 1–14.

68. Lennox BR, Jacob R, Calder AJ, Lupson V, Bullmore ET.Behavioural and neurocognitive responses to sad facialaffect are attenuated in patients with mania. Psychol Med2004; 34: 795–802.

69. Van der Schot A, Kahn R, Ramsey N, Nolen W, Vink M.Trait and state dependent functional impairments inbipolar disorder. Psychiatry Res 2010; 184: 135–142.

70. Lawrence NS, Williams AM, Surguladze S et al. Subcor-tical and ventral prefrontal cortical neural responses tofacial expressions distinguish patients with bipolar disor-der and major depression. Biol Psychiatry 2004; 55: 578–587.

71. Elliott R, Ogilvie A, Rubinsztein JS, Calderon G, DolanRJ, Sahakian BJ. Abnormal ventral frontal responseduring performance of an affective go ⁄ no go task inpatients with mania. Biol Psychiatry 2004; 55: 1163–1170.

72. Killgore WD, Gruber SA, Yurgelun-Todd DA. Abnormalcorticostriatal activity during fear perception in bipolardisorder. NeuroReport 2008; 19: 1523–1527.

73. Malhi GS, Lagopoulos J, Ward PB et al. Cognitivegeneration of affect in bipolar depression: an fMRI study.Eur J Neurosci 2004; 19: 741–754.

74. Chen C-H, Suckling J, Ooi C et al. A longitudinal fMRIstudy of the manic and euthymic states of bipolardisorder. Bipolar Disord 2010; 12: 344–347.

75. Cerullo MA, Fleck DE, Eliassen JC et al. A longitudinalfunctional connectivity analysis of the amygdala inbipolar I disorder across mood states. Bipolar Disord2012; 14: 175–184.

76. Wang F, Kalmar JH, He Y et al. Functional andstructural connectivity between the perigenual anteriorcingulate and amygdala in bipolar disorder. Biol Psychi-atry 2009; 66: 516–521.

77. Markowitsch HJ, Emmans D, Irle E, Streicher M,Preilowski B. Cortical and subcortical afferent connec-tions of the primate�s temporal pole: a study of rhesusmonkeys, squirrel monkeys, and marmosets. J CompNeurol 1985; 242: 425–458.

78. Kupka RW, Altshuler LL, Nolen WA et al. Three timesmore days depressed than manic or hypomanic in bothbipolar I and bipolar II disorder. Bipolar Disord 2007; 9:531–535.

79. Judd LL, Akiskal HS, Schettler PJ et al. The long-termnatural history of the weekly symptomatic status ofbipolar I disorder. Arch Gen Psychiatry 2002; 59: 530–537.

80. Keller MB, Lavori PW, Coryell W, Endicott J, MuellerTI. Bipolar I: a five-year prospective follow-up. J NervMent Dis 1993; 181: 238–245.

81. Drevets WC. Neuroimaging studies of mood disorders.Biol Psychiatry 2000; 48: 813–829.

82. Altshuler L, Bookheimer S, Townsend J et al. Regionalbrain changes in bipolar I depression: a functionalmagnetic resonance imaging study. Bipolar Disord 2008;10: 708–717.

83. Versace A, Thompson WK, Zhou D et al. Abnormal leftand right amygdala-orbitofrontal cortical functional con-nectivity to emotional faces: state versus trait vulnerabilitymarkers of depression in bipolar disorder. Biol Psychiatry2010; 67: 422–431.

84. Almeida JR, Versace A, Hassel S, Kupfer DJ, PhillipsML. Elevated amygdala activity to sad facial expressions:a state marker of bipolar but not unipolar depression. BiolPsychiatry 2010; 67: 414–421.

85. Almeida JR, Mechelli A, Hassel S, Versace A, Kupfer DJ,Phillips ML. Abnormally increased effective connectivitybetween parahippocampal gyrus and ventromedialprefrontal regions during emotion labeling in bipolardisorder. Psychiatry Res 2009; 174: 195–201.

86. Almeida JR, Versace A, Mechelli A et al. Abnormalamygdala-prefrontal effective connectivity to happy facesdifferentiates bipolar from major depression. Biol Psychi-atry 2009; 66: 451–459.

87. Hassel S, Almeida JR, Frank E et al. Prefrontal corticaland striatal activity to happy and fear faces in bipolardisorder is associated with comorbid substance abuse andeating disorder. J Affect Disord 2009; 118: 19–27.

88. Hassel S, Almeida JRC, Kerr N et al. Elevated striataland decreased dorsolateral prefrontal cortical activity inresponse to emotional stimuli in euthymic bipolar disor-der: no associations with psychotropic medication load.Bipolar Disord 2008; 10: 916–927.

89. Malhi GS, Lagopoulos J, Sachdev PS, Ivanovski B, ShnierR, Ketter T. Is a lack of disgust something to fear? Afunctional magnetic resonance imaging facial emotionrecognition study in euthymic bipolar disorder patients.Bipolar Disord 2007; 9: 345–357.

90. Robinson JL, Monkul ES, Tordesillas-Gutierrez D et al.Fronto-limbic circuitry in euthymic bipolar disorder:evidence for prefrontal hyperactivation. Psychiatry Res2008; 164: 106–113.

91. Lagopoulos J, Malhi GS. A functional magnetic reso-nance imaging study of emotional Stroop in euthymicbipolar disorder. NeuroReport 2007; 18: 1583–1587.

92. Surguladze SA, Marshall N, Schulze K et al. Exaggeratedneural response to emotional faces in patients with bipolardisorder and their first-degree relatives. Neuroimage 2010;53: 58–64.

93. Malhi GS, Lagopoulos J, Sachdev PS, Ivanovski B, ShnierR. An emotional Stroop functional MRI study of euthy-mic bipolar disorder. Bipolar Disord 2005; 7 (Suppl 5.):58–69.

94. Berns GS, Martin M, Proper SM. Limbic hyperreactivityin bipolar II disorder. Am J Psychiatry 2002; 159: 304–306.

95. Gruber O, Tost H, Henseler I et al. Pathological amyg-dala activation during working memory performance:evidence for a pathophysiological trait marker in bipolaraffective disorder. Hum Brain Mapp 2010; 31: 115–125.

96. Adler CM, Holland SK, Schmithorst V et al. Abnormalfrontal white matter tracts in bipolar disorder: a diffusiontensor imaging study. Bipolar Disord 2004; 6: 197–203.

97. Thermenos HW, Goldstein JM, Milanovic SM et al. AnfMRI study of working memory in persons with bipolardisorder or at genetic risk for bipolar disorder. Am J MedGenet B Neuropsychiatr Genet 2010; 153B: 120–131.


338

98. Strakowski SM, Adler CM, Holland SK, Mills N,DelBello MP. A preliminary FMRI study of sustainedattention in euthymic, unmedicated bipolar disorder.Neuropsychopharmacol 2004; 29: 1734–1740.

99. Lagopoulos J, Malhi G. Impairments in ‘‘top-down’’processing in bipolar disorder: a simultaneous fMRI-GSRstudy. Psychiatry Res 2011; 192: 100–108.

100. Jogia J, Haldane M, Cobb A, Kumari V, Frangou S. Pilotinvestigation of the changes in cortical activation duringfacial affect recognition with lamotrigine monotherapy inbipolar disorder. Br J Psychiatry 2008; 192: 197–201.

101. Wessa M, Houenou J, Paillere-Martinot ML et al. Fron-to-striatal overactivation in euthymic bipolar patientsduring an emotional go ⁄ nogo task. Am J Psychiatry 2007;164: 638–646.

102. Stern E, Silbersweig DA. Advances in functional neuro-imaging methodology for the study of brain systemsunderlying human neuropsychological function and dys-function. J Clin Exp Neuropsychol 2001; 23: 3–18.

103. Junghofer M, Peyk P, Flaisch T, Schupp HT. Neuroi-maging methods in affective neuroscience: selected meth-odological issues. Prog Brain Res 2006; 156: 123–143.

104. Deichmann R, Gottfried JA, Hutton C, Turner R.Optimized EPI for fMRI studies of the orbitofrontalcortex. Neuroimage 2003; 19: 430–441.

105. Loeber RT, Gruber SA, Cohen BM, Renshaw PF,Sherwood AR, Yurgelun-Todd DA. Cerebellar bloodvolume in bipolar patients correlates with medication.Biol Psychiatry 2002; 51: 370–376.

106. Gould TD, Manji HK. Signaling networks in the path-ophysiology and treatment of mood disorders. J Psycho-som Res 2002; 53: 687–697.

107. Roth RM, Koven NS, Randolph JJ et al. Functionalmagnetic resonance imaging of executive control inbipolar disorder. NeuroReport 2006; 17: 1085–1089.

108. Mitchell RL, Elliott R, Barry M, Cruttenden A, WoodruffPW. Neural response to emotional prosody in schizo-

phrenia and in bipolar affective disorder. Br J Psychiatry2004; 184: 223–230.

109. Silverstone PH, Bell EC, Willson MC, Dave S, WilmanAH. Lithium alters brain activation in bipolar disorder ina task- and state-dependent manner: an fMRI study. AnnGen Psychiatry 2005; 4: 14.

110. Caligiuri MP, Brown GG, Meloy MJ et al. An fMRIstudy of affective state and medication on cortical andsubcortical brain regions during motor performance inbipolar disorder. Psychiatry Res 2003; 123: 171–182.

111. Angrilli A, Palomba D, Cantagallo A, Maietti A, Steg-agno L. Emotional impairment after right orbitofrontallesion in a patient without cognitive deficits. NeuroReport1999; 10: 1741–1746.

112. Grafman J, Vance SC, Weingartner H, Salazar AM, AminD. The effects of lateralized frontal lesions on moodregulation. Brain 1986; 6: 1127–1148.

113. Kruger S, Seminowicz D, Goldapple K, Kennedy SH,Mayberg HS. State and trait influences onmood regulationin bipolar disorder: blood flow differences with an acutemood challenge. Biol Psychiatry 2003; 54: 1274–1283.

114. Chen C-H, Suckling J, Lennox BR, Ooi C, Bullmore ET.A quantitative meta-analysis of fMRI studies in bipolardisorder. Bipolar Disord 2011; 13: 1–15.

115. Altshuler L, Bookheimer SY, Townsend J et al. Bluntedactivation in orbitofrontal cortex during mania: a func-tional magnetic resonance imaging study. Biol Psychiatry2005; 58: 763–769.

116. Kronhaus DM, Lawrence NS, Williams AM et al. Stroopperformance in bipolar disorder: further evidence forabnormalities in the ventral prefrontal cortex. BipolarDisord 2006; 8: 28–39.

117. Altman S, Haeri S, Cohen LJ et al. Predictors of relapse inbipolar disorder: a review. J Psychiatr Pract 2006; 12: 269–282.


339

emotion processing and regulation in bipolar disorder: a review

Documents