Bipolar

 

Clinical Psychology Review 25 (2005) 1101–1122

Psychological approaches to bipolar disorders: A theoretical critique M.J. Power   T Section of Clinical and Health Psychology, University of Edinburgh, Medical School, Teviot Place,  Edinburgh EH8 9AG, UK 

Received 21 September 2004; received in revised form 6 May 2005; accepted 13 June 2005

Abstract

An outline is presented of five main psychological models of the bipolar disorders. These approaches include the Behavioural Activation/Inhibition Systems model, the Cognitive Therapy model, the Interpersonal and Social Rhythm Therapy model, the Interacting Cognitive Subsystems model, and the SPAARS model. Strengths and weaknesses are highlighted for each approach. It is concluded that although there is no model that can adequately account for even the key features of the bipolar disorders (such as periodicity, shifts in the valence of the selfconcept, mixed affective states, and patterns of recovery and relapse), nevertheless, more recently developed multilevel approaches to emotion offer more sophisticated possibilities for modeling these complex disorders. D  2005 Elsevier Ltd. All rights reserved.

1. Introduction

Bipolar disorders occur in approximately 1% of the population (Cavanagh, ( Cavanagh, 2004; Weissman & Myers, 1978). 1978 ). They are characterized by a series of affective highs and lows with some states combining feelings of mania, depression and other moods or emotions concurrently. These disturbances are thought  to recur throughout the lifetime of 80–95% of those affected (Goodwin (Goodwin & Jamison, 1990) 1990). Bipolar  disorders can have devastating consequences for individual sufferers and their families. An estimated 9 years of life, 12 years of normal health and 14 years of major activities such as schooling, work and child rearing may be lost to the average 25-year-old woman diagnosed with the disorder (Prien ( Prien & Potter, T

 Tel.: +44 131 651 3943.  E-mail address:   mjpower@staffm [email protected] ail.ed.ac.uk..

0272-7358/$ - see front matter   D   2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.cpr.2005.06.008

 

1102

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

1990). ). A high mortality risk is associated with the th e affective episodes with episodes  with approximately one quarter of  1990  bipolar disorder individuals attempting suicide (Prien & Potter, 1990). 1990). Different theories exist as to the aetiology of bipolar disorders. Psychobiological theories propose a diath di athes esisis-str stress ess mo model del,, in whi which ch str stres esss fro from m li life fe ev event entss in inte terac racts ts wit with h pre predi dispo spose sed d bi biolo ologi gical cal,,  biochemical and neurological instabilities to induce the illness in vulnerable individuals. Whilst many such models are a useful reminder that increased stress levels are linked to the onset of a variety of  disord dis orders ers,, th thee mod models els the themse mselve lvess rar rarely ely mov movee bey beyon ond d a sim simple ple lev level el of des descri cri ptiveness.  ptiveness. Additional research has linked factors such as lack of social support, family environment environment (Miklowitz, (Miklowitz, Goldstein, Goldstein,  Neuchterlein, Snyder, & Mintz, 1988 1988), ), lifest yle yle and sleep sleep irregularity (Wehr, (Wehr, Sack, & Rosenthal, 1987) 1987), and increased sensitivity with each episode (Post, (Post, 1992) 1992)  to illness instigation and relapse. However, with onee or two exc on except eption ions, s, the there re has bee been n lit little tle wor work k car carrie ried d out on psy psycho cholog logica icall asp aspect ectss of bip bipola olar  r  disorders diso rders such as resea research rch that addresses addresses the self-concept self-concept or emoti emotion on in bipo bipolar lar diso disorders. rders. Furthermore, Furthermore, many of the models of the bipolar disorders simplify the clinical characteristics of the disorders almost   beyond recognition; for example, many accounts of hypomania/ma hypomania/mania nia would imply that the goaldirect dir ected ed eng engage agemen mentt and act activi ivity ty lea leads ds onl only y to po posit sitive ive emo emotio tions, ns, whe wherea reass the act actual ual emo emotio tional nal experience of mania typically includes  considerable dysphoria, anxiety, anxiety, and irritability, with emotional lability being a characteristic feature.  Goodwin and   Jamison (1990)   reported that 70–80% of patients patients with mania presented with this mixed state picture;  Cassidy, Forest, Murry, and Carroll’s (1998)  largescale exploratory factor analysis of manic symptoms showed that  dysphoric mood  was the first major  factor in their data; and bipolar bipolar disorders show show considerable co-morbidity with anxiety disorders, drug and alcohol abuse, etc (see  Papolos, 2003, 2003,  for a recent summary). Despite the surprising lack of psychological research on bipolar disorders, there are now several  psychologicall models that have been offered in order to provide an account of at least part of the  psychologica  phenomena of the bipolar disorders. These models are either adaptations of existing approaches to  psychologicall disorders, as in the case of the adaptations of the Cognitive Therapy and Interpersonal  psychologica Psychotherapy Models, or they are based on more general frameworks as in the adaptations of the Interacti Inte racting ng Cogn Cognitiv itivee Sub-s Sub-system ystemss (ICS) and the Schem Schematicatic-Propo Propositi sitional onal-Anal -Analogic ogical-Ass al-Associat ociationi ioniststRepresentation-Systems (SPAARS) approaches. However, before these approaches are reviewed, the adaptation of Jeffrey Gray’s Behavioural Activation/Behavioural Inhibition Systems (BAS/BIS) model will be considered. For each of the models, three simple tests can be run:   b

 Q 

(1) Does the model account for some or all of the unique clinical features of the disorders such as  periodicity,, mixed affective states, extreme shifts in the valence of mood and self-concept, and  periodicity  patterns of recovery and relapse? (2) What is the explanatory power of the theory? Does it focus on one category of symptom at the expens exp ensee of oth others ers?? Doe Doess the the theory ory mak makee tes testab table le pre predic dictio tions? ns? Doe Doess the the theory ory mak makee uni unique que  predictions? (3) Is the theory relevant to normal functioning or is it a stand-alone and disorder-specific approaches? Can it account for other types of psychopathology and place the bipolar disorders in a general framework of psychopathology? These three sets of tests will not be applied in a formulaic way as we proceed through each of the models, but the models will be described in their own appropriate ways first, before returning to these questions so that they will be used to structure the final summary and conclusions.

 

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

1103

2. The behavioural activation/behavioural inhibition systems (BAS/BIS) model

The BAS/BIS model can be considered in the context of dimensional appr oaches oaches to personality, personality, motivation, and emotion. Although in many ways derived from and related to  Eysenck’s (1967)   two dimensions of extraversion–introversion and normality-neuroticism (also called emotionality ), ), there is also a also  a relevant tradition within mood and affect and  affect from at least Osgood’s semantic differential differential approach (e.g. Osgood, (e.g.  Osgood, Suci, & Tannenbaum, 1957) 1957) of examining a two-dimensional approach (e.g. Rolls, (e.g.  Rolls, 1999; Russell & Carroll, 1999; Watson, Clark, & Tellegen, 1988). 1988 ). First, however, the BAS/BIS approach will  be outlined before returning to some of the generic problems shared by these two-dimensional approaches. In hi hiss th theo eory ry of an anxi xiet ety y,   Gray Gray (19 (1976 76,, 198 1982) 2)   identi identifie fied d two key beh behavi aviour oural al sys system tems. s. Fir First, st, a Behavioural Inhibition System (BIS), a Pavlovian conditioning based system, which focuses on stimuli that have come to be conditioned to punishment or frustrative non-reward, or novel stimuli which may come to be associated with unconditioned responses (UCRs). In the presence of such stimuli, the BIS leads to behavioural inhibition, that is, to the interruption of ongoing activity, which is accompanied by increa inc reased sed aro arousa usall and att attent ention ion to the env enviro ironme nment. nt. Anx Anxiet iety y the theref refore ore   was   cons consid ider ered ed to be a consequence of activity in the BIS, in which a postulated comparator (see  Fig. 1) 1)  detects mismatches   b

 Q 

 between predicted events that do not occur, unpredicted events that do occur, and predicted punishments and non-rewards that occur. The application of Gray’s BAS/BIS model to mood disorders has tended to focus on the role of the BAS rather than the BIS (see later for some exceptions to this generalisation), though there is a clear  overlap between Gray’s Gray’s account of   of   the BIS, with its focus on behavioural inhibition and predicted aversiveness, and   Seligman’s (1975)  Learned Helplessness theory and its subsequent reformulations.  Nevertheless, the Behavioural Activation System (BAS) focuses on approach behaviour in which  positive stimuli, or reward, lead to approach and engagement. The BAS is therefore closel closely y associated with incentive incentive motiv motivatio ation n and with motor progr programmes ammes related to appro approach. ach. Although  Although   Gray (1990 (1990)) STORED REGULARITIES

THE WORLD 

COMPARATOR

  PREDICTIONS

PLANS

Fig. 1. Gray’s comparator model.

 

1104

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

thought that only the BIS was clearly identified with the emotion of anxiety, anx iety, others have been bolder and developed emotion models based on the BAS. For example, Depue ( Depue, Krauss, & Spoont, 1987; Goplerud & Depue, 1985) 1985) has proposed that BAS-type systems are poorly regulated both in bipolar  disorder individuals and in cyclothymic individuals, who can be considered to have a milder variant of a  bipolar disorder disorder.. Depue et al. focussed in particular on normal levels of cortisol product ion ion and the change in cortisol production under stress in both cyclothymic and bipolar individuals (e.g. Goplerud (e.g.  Goplerud & Depue, 1985); 1985); they found hypersecretion of cortisol and slower recovery to normal cortisol levels. On the basis of these and related findings, Depue suggested that the high arousal, high goal-directed activity, and high positive emotions are characteristic of high levels of BAS activity and of hypomania, whereas low levels of BAS activity, including disengagement   disengagement   from rewarding activities, retardation, etc., are charac cha racter terist istic ic of the dep depres ressed sed pha phase. se. In thi thiss vei vein, n,   Johnson, Johnson, Sandr Sandrow ow,, Meye Meyer, r, and Win Winters ters (200 (2000) 0) reported an increase in manic symptoms after the attainment of an important goal, with the additional feature that bipolar individuals continue to attempt to increase positive affect after goal attainment with increased goal pursuit, whereas normal individuals tend to aim for some moderation of the positive affect  for exa exampl mplee by coasting . In an an anal alog ogue ue st stud udy y with stud students ents repor reporting ting hypomanic hypomanic or depre depressed ssed symptomatology, Meyer, symptomatology,  Meyer, Beevers, and Johnson (2004) found (2004)  found that hypomania was linked with an overlyoptimistic future goal-oriented pattern, and that such  individuals reported future   goals as being less   b

 Q 

stressful and less difficult than did control individuals. Stern individuals.  Stern and Berrenberg (1979) reported (1979)  reported that on the illusion of control laboratory task individuals with a history of hypomania predicted higher rates of  success after success feedback in comparison to controls. There is some evidence therefore in support of  the claim that goal attainment and goal loss or interruption are hyper-valenced in the bipolar disorders, though it should be cautioned that patterns of over-investment over-investment in some goals and a nd under-investment in others are characteristic characteristic of unipolar depr ession ession also (Champion (Champion & Power, 1995) 1995). The work of  Carver  Carver and White (1994) in (1994)  in their operationalisation of Gray’s approach has offered a selfrepo re port rt BI BIS/ S/BA BAS S sc scal alee fo forr th thee as asse sess ssme ment nt of BI BIS S an and d BA BAS S se sens nsit itiv ivit itie ies, s, th thou ough gh qu ques esti tion onss ha have ve necessarily been asked about whether this self-report measure does measure the same construct as the  psychophysiological  psychophysio logical and behavioural measures of BIS/BAS that have been used in the laboratory (cf. Johnso Joh nson, n, Tu Turne rner, r, & Iwa Iwata, ta, 200 2003 3). Al Alth thou ough gh in th thee Ca Carv rver er an and d Wh Whit itee ap appr proa oach ch th thee BI BIS S sc scal alee ha hass remained much as conceptualised by Gray in its focus on punishment sensitivity, they have further  developed and modified the BAS scale to include subscales that assess Drive (i.e. the pursuit of desired goals), Fun Seeking, and Reward Responsiveness. Carver and White also considered the BAS to be linked lin ked to pos positi itive ve aff affect ect and the BIS to neg negati ative ve aff affect ect,, the thereb reby y rai raisin sing g the pos possib sibili ility ty tha that  t   both dimensions could be involved in mood disorders rather than only the BAS and therefore contrary to authors such as Depue and others (e.g.  Depue et al., 1987; Wright & Lam, 2004) 2004)  who have focussed only on the BAS in bipolar disorders. In an analogue study with students that included the Carver and White BIS/BAS scales, Meyer, scales,  Meyer, Johnson, and Carver (1999) found (1999)  found that in mood-disorder prone students, higher scores on the BAS Fun Seeking scale were related to higher mania symptom scores, whereas both high hig h BIS and low BAS Rew Reward ard Res Respon ponsiv sivene eness ss wer weree rel relate ated d to dep depres ressio sion n sym sympto ptom m sco scores res.. Thi Thiss  preliminary study therefore offered some support for the Carver and White modification of the BIS/BAS model, in particular in the combination of both BAS and BIS in the correlation with depression scores. A subsequent study from the same research group (Meyer, (Meyer, Johnson, & Winters, 2001) 2001)  followed up a group of 59 Bipolar 1 individuals over an average of 20 months. They found that the BAS (especially Reward Responsiveness) was predictive of increases in manic symptoms over time but was not predictive of  changes in depression. In contrast, the BIS fluctuated with levels of depression but was not predictive of 

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

 

1105

late terr le leve vels ls of de depr pres essi sion on,, wh whic ich h wo woul uld d in indi dica cate te th that at th thee BI BIS S is si simp mply ly a st stat atee ma mark rker er bu butt no nott a la vulnerability factor. factor. A further inconsist ent ent finding was reported in an epidemiological study of over 1800 19–21 year olds (Johnson (Johnson et al., 2003), 2003), which found significantly higher levels of BIS in respondents with a lifetime-ever diagnosis of depression, but none of the expected differences were found on the BAS scales. scales. The Meyer The  Meyer et al (2001) clinical (2001)  clinical study needs of course to be replicated in a larger sample, but, together  with the other inconsistent findings, it does question the relevance of the BIS/BAS approach to bipolar  disorders, in addition to the question of whether or not the self-report measure correctly operationalises the construct. construct. However, an alternative option option might be to argue as Johnson et al. have done subsequently (e.g. Cuellar, (e.g.  Cuellar, Johnson, & Winters, 2005) 2005)  that, contrary to the current classification of bipolar disorders, mania and depression are not inherently linked but simply show high levels co-morbidity; in this case, the linking of BAS to mania but not depression would be consistent, but there still remains the problem for the BIS/BAS approach of why neither BIS nor BAS were predictive of depression. There are a number of major limitations of the BAS/BIS approach in addition to some of the issues already raised. Perhaps the major limitation can be highlighted when the theory is placed in the context  of oth ther er tw two-dim o-dimensi ensional onal appr approache oachess   to mo moti tiva vati tion on an and d af affe fect ct,, wh whic ich h sh shar aree so some me of th thee sa same me limitations. Carver limitations.  Carver and White’s (1994) analysis (1994)  analysis argues that BAS is related to positive affect and BIS to negative affect and that  these   these two dimensions dimensions are orthogonal to each other. A similar analysis of affect  space is proposed by   Wats Watson on et al. (19 (1988) 88),,   in whi which ch the two ort orthog hogona onall dim dimens ension ionss are exp explic licitl itly y labeled Positive Affect and Negative Affect, with each varying from High to Low. However, a contrary approach appro ach is that Positive–Negat Positive–Negative ive (or Pleas Pleasant– ant–Unple Unpleasant asant)) forms a sin single gle bipolar dimension dimension wit h Arous Aro usal al or Ac Acti tiva vati tion on as th thee se seco cond nd or orth thog ogon onal al di dime mens nsio ion n (e (e.g .g..   Russ Russel elll & Ca Carr rrol oll, l, 19 1999 99)).  Notwithstanding these contradiction contradictionss within the two-dimensional approach is the problem that  althou alt hough gh the sel self-re f-repor portt of con consci sciou ouss af affec fectt mig might ht be cap captur tured ed by dim dimens ension ionss suc such h as Vale alence nce and Arousal, the underlying functional and neuroanatomical systems may be far more complex and not best  described by such systems. The evidence for this complexity is wide-ranging in the areas of emotion and mot motiva ivatio tion, n, bu butt inc includ ludes es wor work k in fav favour our of bas basic ic emo emotio tions ns (e. (e.g. g.   Ekman, Ekman, 1992 1992))   and incre increasing asing evidence of specific-emotion linked neuroanatomical circuits that underlie different emotions but which are not capturable simply by a positive-negative distinction (e.g. Lane (e.g.  Lane & Nadel, 2000) 2000). Even within the study stu dy of sel self-re f-repo porte rted d aff affect ect,, the evi eviden dence ce is far fro from m cle clearar-cut cut tha thatt any of the two two-di -dimen mensio sional nal approaches are sufficient, contrary to many of the assumptions to date (Power, ( Power, in press) press). The question therefore for the BIS/BAS approach, like with the other two-dimensional approaches, is what exactly are the two dimensions, how do they map onto emotion and mood, and how do they deal with the contradictory evidence base? The limitations of the BIS/BAS approach are also paralleled in the shift from the all-embracing concept of stress, dominant in the 1950s and 1960s, to more sophisticated emotion-based theories which will be returned to later (see e.g.   Lazarus, 1966, 1991, 1991,   for a dramatic example of such a change). Although there are still recent equivalents of the BAS/BIS such as in Rolls in  Rolls (1999) reinforcement-based (1999)  reinforcement-based theory of emotion, these approaches simplify the complex phenomena which they attempt to explain. For example, Gray developed his comparator model illustrated above (see  Fig. 1) 1)  to include a theory of  consciousness and the self, in which the function of consciousness is that of an error-detector for the interruption of automatic processes (e.g. Gray, (e.g.  Gray, 1999) 1999). Again, such an approach simplifies the concept of  the self and of consciousness, for example, with the assumption that both are unitary constructions, whereas the phenomena of the emotional disorders suggest otherwise (e.g.  Dalgleish & Power, 2004a) 2004a).

 

1106

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

3. The cognitive therapy model

An outline of the classic Cognitive Therapy T herapy model of unipolar depression is presented in Fig. in  Fig. 2 (based 2 (based on Beck, on  Beck, Rush, Shaw, & Emery, 1979). 1979 ). In summary, childhood experiences lead to the development of  dysfunctional schemas that centre around themes such as the need to be loved or the need to achieve. Disconfirmi Disco nfirming ng expe experienc riences es or life events later in life, for exam example, ple, in the transition transition from adol adolescen escence ce to adulthood, may lead to activation  of  of the dysfunctional schema in that the person no longer believes that  he or she is lovable, or believes that he or she has been a failure throughout life. The activation of the dysfunctional beliefs causes the production of negative automatic thoughts (e.g. I am unlovable , I am a failure , etc.), which in turn cause the onset of the relevant mood such as the depressed state in the vulnerable individual. This classic Cognitive Therapy model of unipolar depression has now been adapted to a number of  different disorders that include anxiety disorders, schizophrenia, personality disorders, and, more recently, more  recently,  bipolar  bipol ar disor disorders. ders. The two main presen presentatio tations ns of the adapt adaptation ation to bip bipo ola larr di diso sord rder erss by Be Beck ck et al al.. ( Ne  Newman, wman, Leahy, Beck, Reilly-Harrington, & Gyulai, 2002) 2002) and by Lam et al. (Lam, (Lam, Jones, Hayward, & Bright, 1999; Wright & Lam, 2004 2004)) ar e both relatively informal, in the sense that they assume the basic cognitive therapy model outli outlined ned in Fi Fig. g. 2, wi with th th thee ma main in di diff ffer eren ence cess be bein ing g th thee ty type pess of dy dysfu sfunc ncti tion onal al at atti titu tude dess or sc sche hemas mas th that  at   b

 Q 

 b

 Q 

 b

 Q 

 b

 Q 

are implicated. Newman implicated. Newman et al. (2002) have (2002)  have argued in bipolar disorders that schemas act in a  bidir  bidirectio ectional nal way, that is, even though a schema might focus on lovability, the polarity of the schema shifts with mood state and life events from one extreme to the other, from at one time representing that they are totally unlovable to another time representing that everybody loves them. The Newman et al. model is therefore also a diathesis-stress model in which different types of life events may be related to different manic or  depressive states; for example, events that lead to sleep disruption or to successful goal pursuits may m ay be more likely to trigger mania. In a study of students with lifetime-ever diagnosis of bipolar disorders, Reillydisorders,  ReillyHarrington, Alloy, Fresco, and Whitehouse (1999) found (1999)  found that the interaction between negative life events and cognitive style as measured by the Attributional Style Questionnaire or the Dysfunctional Attitudes Scale (DAS) was predictive of both depressive symptoms and manic symptoms at 1-month follow-up. A EARLY EXPERIENCE (e.g. criticism and rejection from parents)

FORMULATION OF DYSFUNCTIONAL ASSUMPTIONS (e.g. unless I am loved I am a m worthless)

CRITICAL INCIDENTS (e.g. loss events)

ACTIVATION OF ASSUMPTION

NEGATIVE AUTOMATIC THOUGHTS

DEPRESSION

Fig. 2. Beck’s Cognitive Therapy model.

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

 

1107

further study by Johnson by  Johnson and Fingerhut (2004)  of 60 Bipolar 1 individuals reported support for the DAS  being predi predictive ctive of depre depressive ssive sympt symptomato omatology logy at 6 month follow follow-up, -up, but the autho authors rs report reported ed that  cognitive factors such as the DAS are not predictive of changes in manic symptoms. However, their results for mania are inconclusive because they did not examine subscales of the DAS, nor do they report the relevant analyses that would have fully tested the predictor models. The second main cognitive therapy model has been presented by Lam by Lam et al. (1999), (1999) , which has focussed on specific subsets of dysfunctional attitudes. In order to test the model, Lam, model,  Lam, Wright, and Smith (2004) used the Dysfunctional Attitude Scale with a sample of 143 Bipolar Type 1 patients; exploratory factor  analysis suggested factors that included goal-attainment   (e.g. If I try hard enough, I should be able to excel at anything I attempt  ) and antidependency  (e.g.  (e.g. I do not need the approval of other people to be happy ), ), two factors that distinguished the bipolar patients from a comparison sample of unipolar patients. Consistent therefore with the proposals for the BAS approach discussed above, the dysfunctional attitudes that are characteristi characteristicc in bipolar disorder include goal-striving, success in which is postulated to lead to euphoria (see Fig. (see  Fig. 3). 3). A positive feedback loop in vulnerable individuals leads to attempts to enhance  positive mood state with increasingly driven goal-attainment behaviour behaviour,, disregard of feedback from others (exacer  bated by activation of the antidependency  beliefs),   beliefs), and disruption of normal routines. As Leahy (1999) has (1999)  has proposed, manic individuals tend to be risk-lovers   in which their decision-making   b

 Q   Q 

 Q 

 b

 Q 

  b

 b

 Q 

 b

 Q 

  b

 Q 

 processes assume unlimited  processes unlimited resources, infallibility of predictions, and control over outcomes. Johnson et al. (2000) offer (2000)  offer support for the proposal that specifically goal-attainment goal-attainment life events rather  than positive events in general are related to increases in manic symptoms; and  Scott, Stanton, Garland, and Fer Ferrie rierr (20 (2000 00))   found found tha thatt eut euthym hymic ic bip bipola olarr ind indivi ividua duals ls con contin tinued ued to sho show w ele elevat vated ed sco scores res on dysfunctional attitudes related to Perfectionism in addition to continuing to display social problemsolving difficulties (as measured measured by the Means-Ends Problem-Solving Means-Ends  Problem-Solving procedure). A fu furt rthe herr st stud udy y by Sc Scot ottt (Sc Scot ottt & Po Pope pe,, 20 2003 03))   compare compared d bipo bipolar lar parti participan cipants ts who were eith either  er  remitted, hypomanic or depressed at the time of the assessment. The hypomanic participants were found to score at intermediate levels between the remitted and the depressed groups on dysfunctional attitudes, including those attitudes related to Perfectionism or goal-attainment. Moreover, on the Rosenberg SelfEsteem Scale hypomanic participants were found to score somewhat higher on both the Positive Esteem and the Negative Esteem subscales in comparison to the remitted and depressed bipolar participants. These findings again, as noted in the BAS comments above, demonstrate that the hypomanic/manic  phases of bipolar disorders are not simply the positive opposite of the negative depressed phases. Indeed, findings such as these suggest that the name  bipolar disorder   is a misnomer in itself. Criticisms of the Cognitive Therapy model, like with the BAS/BIS approach, have to begin with a focus on the simplicity of the model. One can only feel sorry for the sheer amount of work that poor  dysfunctional attitudes (and this is not to deny the role of other cognitive processes such as memory, attention, and reasoning) are having to do as the Cognitive Therapy model is extended to more and more disorders (and I make this criticism as a one-time fan of dysfunctional attitudes—see e.g. Power e.g.  Power et al., 1994); 1994 ); eventually there will be a disorder for each dysfunctional attitude! The serious points here are summarised in the following points.   b

 Q   Q 

(1) The classic Cognitive Therapy framework has too simple a monolithic view of the self-concept; this problem is highlighted in the need for   bidirectional schemas  in the adaptation for bipolar  disorders, which can either be overly positive or overly negative according to context and mood. However, it is unclear how a single schema would change its content and processing features so   b

 Q 

 

1108

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

DYSFUNCTIONAL ATTITUDES Goal Attainment: 1. I should be happy all the time 2. A person should do well at everything he undertakes

BIOLOGICAL

MOOD Euphoric or dysphoric Mood

Genetic predisposition to manic depressive illness

BEHAVIOUR Relating to Mania: Highly driven behaviour Lack of routine etc Relating to Depression:  Depression:  Self-blame for failure to meet standards Rumination about negative implications of depressive symptoms etc

Fig. 3. Lam’s Dysfunctional Attitude model.

dramatically in order at one time to be excessively positive and at a later time to be excessively negative, nor is this problem tackled by Newman by  Newman et al. (2002). (2002). A more elegant solution would be to consider the more complex self-concept structures that have been considered in social cognitive  psychology (see later). (2) The classic Cognitive Therapy framework has a single level of information processing (as do many other oth er cla classi ssicc mod models els of str stress ess and emotion; emotion; see   Power Power & Dal Dalgle gleish ish,, 199 1997 7), wh whic ich h li limi mits ts th thee explan exp lanato atory ry pow power er of the mod model el and whi which ch wea weaken kenss it in com compar pariso ison n to mod modern ern mul multiti-lev level el information processing approaches such as ICS and SPAARS, which will be presented later. (3) The Cognitive Therapy approach has an overly cognitive and inadequate theory of emotion. The main tenet of the original model was that cognition (e.g. Negative Automatic Thoughts) causes emotion, though of necessity feedback loops have been added into cognition-emotion cycles in subsequent cognitive therapy models (e.g. in   Clark’s, 1986, 1986,   model of panic). Nevertheless, the approach is neutral on what emotions are, whether or not they relate to each other or stand alone, how they develop, whether they can combine or be in conflict with each other, and so on. In other 

 

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

1109

words, there is no theory of emotion within Cognitive Therapy, but  emotions  are givens within the model that do not need further theoretical deconstruction. A theory of emotional disorder that  does not provide a theory of emotion is sadly lacking in its scope. (4) Ada Adapta ptatio tions ns on only ly to the con conten tentt of sch schema emas/a s/atti ttitud tudes es are   insuf insuffic ficien ientt to cap captur turee the act actual ual differences between the full range of psychological disorders (Power ( Power & Champion, 1986; Power & Dalgleish, 1997). 1997). The complex phenomenology of manic episodes cannot simply be accounted for   by successful goal-attainmen goal-attainmentt leading to the activation of positive goal-attainment schema; the elevation of negative emotions as well as positive, the elevation of negative beliefs and negative self-esteem as well as positive beliefs and positive self-esteem, all point to the need for more sophisticated models of the type to be considered later. The findings also add weight to the view that mania is not the polar opposite of depression.   b

 Q 

4. The interpersonal and social rhythm therapy (IPSRT) approach

The IPSR IPSRT T approa approach ch to bip bipola olarr disord disorders ers   combine combiness the Inter Interperso personal nal Psych Psychothe otherapy rapy appro approach ach deve de velo lope ped d by Wei eiss ssma man, n, Kl Kler erma man, n, et al. al. (Kle Klerma rman, n, Weis eisman mann, n, Rou Rounsa nsavil ville, le, & Che Chevro vron, n, 198 1984; 4; Weissm eissman, an, Marko Markowitz, witz, & Klerma Klerman, n, 2000 2000)) for the treatment of unipolar depression,   together with a circadian rhythm model developed by the IPT group at the University of Pittsburgh (Ehlers, ( Ehlers, Frank, & Kupfer, 1988; Swartz, Frank, Spielvogle, & Kupfer, 2004). 2004). The Pittsburgh group have of course not   been the only researchers and clinicians and  clinicians to have identified the significance of circadian rhythm rhy thm disruption in bipolar disorders (see e.g. Goodwin e.g.  Goodwin & Jamison, 1990; Healy & Williams, 1988, 1989) 1989), but they have  been the only group to have developed it into a framework in which to understand bipolar disorders combined with a clinical intervention. The focus therefore will be on the IPSRT model rather than the earlier circadian dysrhythmia models. Thee IP Th IPT T ap appr proa oach ch to de depr pres essi sion on wa wass or orig igin inal ally ly de deve velo lope ped d as a co cont ntro roll in inte terv rven enti tion on in a  pharmacotherapy trial for unipolar depression. It is deliberately pragmatic rather than theoretical in its approach, in that Weissman and Klerman interviewed expert colleagues about how they worked in  practice with their depressed clients. The presentation of the model is therefore necessarily different to the othe ot herr mo mode dels ls in th this is pa pape perr be beca caus usee of th thee or orig igin inat ator orss ex expl plic icit it di disa savo vowa wall of th theo eory ry in fa favo vour ur of   practitioners’ anecdotal accounts of what worked for them in therapy. Nevertheless, it is possible to begin to locate the approach theoretically in a post-hoc fashion as it has developed into an evidence-based approach for unipolar depression. The approach basically draws on social-psychodynamic models that  reflect for example the influence of the social psychoanalytic approach of Harry Stack  Sullivan  Sullivan (1953) in (1953)  in the US, combined with the equal influence in the US of the social psychiatrist Adolf  Meyer Adolf  Meyer (1957). (1957). IPT works with three phases in therapy. The first phase focuses on assessment and formulation. In addition addi tion to histo history ry and diag diagnost nostic ic inte interview rviewing, ing, IPT incl includes udes the so-ca so-called lled   Interpersonal Interpersonal Inventory which is an interview assessment of the social network and social support of the client. The second phase of therapy selects one of four focus areas on which to base the intervention: 1) Interpersonal Role Disputes—for example, marital conflict, problems at work; 2) Role Transitions—for example, the transition from adolescence to adulthood, retirement; 3) Grief—the loss of a significant other; 4) Interpersonal Deficits problems in establishing and maintaining relationships.

 

1110

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

The third and final phase of the therapy is the   Termination   phase in which the therapist works to establ est ablish ish the therap rapeut eutic ic gai gains ns and to dea deall wit with h the issues issues ari arisin sing g fro from m the end ending ing of the therap rapy y. Oth Other  er  characteri chara cteristics stics of the thera therapy py incl include ude its short short-term -term here-and-now here-and-now focus (see   Klerman Klerman et al. al.,, 198 1984; 4; Weissman et al., 2000). 2000). The second second main feature of IPSRT is its focus  on social circadian rhythms. The  Social Rhythm Metric (SRM) (Monk, (Monk, Kupfer, Frank, & Ritenour, 1991) 1991)  is completed as part of the assessment. The SRM seeks to identify specific triggers that are likely to disrupt normal social rhythms for the individual, to monitor  mood, and to monitor social interactions. The plan then is to regularise social rhythms, especially under  circumstances of vulnerability for the individual, or when certain prodromes or early warning signs of a manic man ic or dep depres ressiv sivee epi episod sodee are ide identi ntifie fied d (a fea featu ture re tha thatt is now als also o cha charac racter terist istic ic of mos mostt CBT approaches to bipolar disorders and is of proven efficacy; see   Lam et al., 1999; Perry et al., 1999; Schwannauer, 2004; SIGN, 2005). 2005). Psychoeducation about about bipolar disorders disorders is also a key component, but  again is also a feature of other bipolar interventions (SIGN, ( SIGN, 2005) 2005). Similar to Cognitive Therapy therefore, IPSRT is primarily a treatment-driven approach to bipolar  disorders, so in many ways the theoretical base of the approach is seen as secondary to the question of  whether or not the intervention works. Because of the focus on application and intervention, there is a loose mapping between theory and practice; thus, thus, it would be possible to map IPSRT onto onto one  one of several several viable theories, just as it would be possible to do this with Cognitive Behaviour Therapy (Power, ( Power, 2002) 2002). The lack of an explicit theoretical base for IPT and for IPSRT has to be therefore its major weakness; good science requires good theory for the development and testing of hypotheses. Other weaknesses of  the IPSRT approach follow from this lack of theory; thus, it is unclear how important circadian and other  dysrhythmias are for both mania and depression and to what extent they can account for all features of  symptomatology; it is unclear how the four focus areas of IPT derived for unipolar depressions map onto  bipolar disorders given evidence from fr om elsewhere pointing to for to  for example the the importance of goal-related events rather than positive or negative events in general (Johnson, ( Johnson, 2005) 2005). However, the strengths of  IPSRT and CBT approaches are in the boundaries boundaries they set for other more theoretically-driven theoretically-driven approaches to bipolar disorders; and for the clinical insights that they offer into factors associated with onset, maintenance, relapse, recovery, and the periodicity of the disorders. We will turn therefore to two multilevel cognitive approaches to emotion, both of which have been derived as more general frameworks but   both of which have recently been adapted specifically for bipolar disorders.

5. The interacting cognitive subsystems (ICS) approach

The Interacting Cognitive Subsystems (ICS) approach (Barnard, ( Barnard, 1985, 2004; Barnard & Teasdale, 1991; Teasdale & Barnard, 1993) 1993) is a recent exemplar of one of a class of multi-level, multi-system approaches, which, in addition to their potential application to emotion, can provide accounts of a wide variety of  cognitive skills and processes. The link between cognition and emotion is not easily pinned down in such models, because the relationship is seen as complex and interactive and so, in contrast to some of the earlier  models discussed above, one cannot simply point at the model and say this is the emotion box : emotion is a process distributed over many subsystems and represents a high-level integration of a variety of such  processes,  proce sses, which vary from occasi occasion on to occas occasion ion and from emoti emotion on to emotio emotion. n. There are nine main cognitive subsystems in Teasdale in  Teasdale and Barnard’s (1993) ICS (1993)  ICS approach: the first set  of subsystems is sensory-related and includes the Acoustic and Visual subsystems; the second set is the  b

 Q 

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

 

1111

Central ral subs subsystem ystemss and incl includes udes the Morp Morphono honolexi lexical, cal, the Propo Propositio sitional, nal, the Impli Implicatio cational, nal, and the Cent Object subsystems; and the third set is Affector subsystems and includes the Articulatory, the Body State, and the Limb subsystems. These subsystems process information partly in parallel and partly sequentially according to the type of task and other requirements acting on the overall system. Rather  than provide a detailed description ofofallemotion. subsystems however, the focus will be one those that are especially important in the occurrence In relat relation ion   to emot motion ion,, the key sub subsys system temss are th thee soso-cal called led Pro Propos positi itiona onall and Imp Implic licati ationa onal, l, as illustrated in Fig. in Fig. 4. 4. These two systems systems represent a common distinction made in psycholinguistics psycholinguistics because because of the need to have both mult iple iple levels and and multiple representations in models of the comprehension and  production of language (e.g. Power, (e.g.  Power, 1986) 1986). The units of representation in the Propositional subsystem are  propositions, which are the smallest semantic units that can have a truth value; thus, the phrases Tony Blair   or  George Bush  do not have truth values in themselves but are merely names about which nothing is asserted. Only when they are included in larger units such as Tony Blair eats British beef   or  George Bush is a fine politician  do the units become propositional because they are either true or false. In contrast, the higher level semantic representations at the Implicational level in ICS ar e referred to as schematic schem atic mode models ls . Lik Likee the more com common monly ly use used d ment mental al mode models ls   approach approach (e.g. (e.g.   Johnson-Laird, 1983), 1983 ), schematic models combine information from a variety of sources and are more generic and   d

T

 d

T

 d

T

 d

T

d

T

  d

T

holistic and integrate information from other subsystems; so, within ICS, schematic models draw upon thee wh th whol olee ra rang ngee of ot othe herr su sub bsystem systemss fee feedin ding g inf inform ormati ation on int into o the Imp Implic licati ationa onall sub subsys system tem.. In the simplified example shown in   Fig. 4, 4,   we have highlighted three possible inputs to the Implicational subsystem from the Visual, the Body State, and the Propositional subsystems. These four subsystems together are the most important ones in the production of emotion in the ICS approach. In the approach therefore emotion is treated as a distributed phenomenon that is the result of the combination within the Implicational subsystem of outputs from a number of cognitive subsystems rather than simply being the output from a specific cognitive appraisal. Although ICS is not alone in i n considering emotion to be the result of processing in multiple cognitive systems (see for example   Leventhal and Scherer’s, 1987, 1987, model), it provides one of the most detailed and elegant multi-system approaches to the understanding of  emotion.   Teasdale and Barnard (1993) also emotion. (1993)  also extend an earlier criticism of Cognitive Therapy made by Power and Champion (1986); (1986);   namely, that Cognitive Therapy focuses on a single level of meaning, whereas profit can be had from considering two levels of meaning such as propositions and mental models. Teasdale and Barnard argue that much of the challenging of negative thoughts and beliefs in standard Cognitive Therapy occurs at the propositional level and may often ignore the higher level PROPOSITIONAL SUB-SYSTEM

VISUAL SUB-SYSTEM

IMPLICATIONAL SUB-SYSTEM

BODY STATE SUB-SYSTEM

Fig. 4. Teasdale and Barnard’s ICS model.

 

1112

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

Implicational meaning; the net outcome of such a process can be that the individual is browbeaten into rejecting the negative proposition, but becomes more depressed rather than less depressed because for  example a higher level model is confirmed in which the individual is always wrong. The adaptation of the ICS approach to bipolar disorders has recently been outlined by  Barnard (2003, 2004),, in which 2004) also considers its application schizophrenia. In the extension of ICS to a range of   psychopatho  psychopathology logy,he , Barnard considers variation in to four sets of processes: (1) the content of semantic representations; (2) the rate of change in mental images; (3) the mode in which processes operate; (4) the synchronization of the processes that generate meaning. In depression, the Implicational and Propositional systems are considered to enter a state of  interlock , in which schematic models continually generate negative propositions which in turn feedback to the Implicational system thereby regenerating the original schematic model. By contrast to this low rate of  change in depression, schematic models are seen to have a high rate of change in mania and involve the  processing of positive or mixed schematic models but with largely unevaluated propositional representation representa tionss there thereby by remai remaining ning outside of focal awareness. awareness. Schiz Schizophre ophrenia nia is cons considere idered d to arise from desynchrony between schematic and propositional levels of meaning, which then has consequences for the int integr egrati ation on of inf inform ormati ation on ari arisin sing g fro from m oth other er sub subsys system tems; s; Bar Barnar nard d pro propos poses es tha thatt the sam samee mechanism may underlie the production of delusions in mania when combined with the fast rate of   processing. Partial products from fast changing schematic models might then combine to form new schematic models, which under optimal conditions can lead to imaginative thinking, but under more extreme circumstances can lead to delusional models. In contrast therefore to the depressed state in which propositional and implicational meanings become interlocked , in mania the opposite occurs and  propositional and implicational meanings are said to enter a runaway state ; in such states, there is little or no refle reflection ction on spec specific ific propositiona propositionall meani meanings, ngs, which permi permits ts the development development of model modelss and meanings that are unscrutinised, inconsistent with each other, and rapidly changing. Overall, it is too soon to judge the ICS approach to bipolar disorders because of its recent and only sketchy development. ICS is of course a general cognitive model that will stand or fall on how useful an  b

 Q 

 b

 Q 

account it provides of a wide range of cognitive processes, not just its account of bipolar disorders. Because it is in part a framework and in part a theory, there is a degree to which it is not falsifiable; thus, there is some flexibility about which particular theory a framework can instantiate. For example, the low level cognitive architecture for ICS could be based on connectionist networks, but it could be based on a viable alternative; the subsystems could be treated as functional modules within a modular processing system, or, alternatively, it could be argued that the cognitive subsystems were simply a functional description of a set of processes that were instantiated in the brain in a radically different fashion (see Gazzaniga, 1988, 1988,   for an exc excell ellent ent dis discus cussio sion n of thi thiss iss issue ue in rel relati ation on to mod modula ularit rity). y). In rel relati ation on to emotion, it remains to be seen how useful the ICS approach is from both the empirical point of view and the clinical point of view. However, some interesting preliminary supporting data for the ICS account of  unipolar depression have been presented by Teasdale, by  Teasdale, Taylor, Cooper, Hayhurst, and Paykel (1995) who (1995)  who found that depressed patients completed sentence stems with  positive  words or phrases consistent with high-level schematic models rather than simply offering negative sentence completions which might be  predicted by theories such as Beck’ Beck’ss or Bower’s. Bower’s. Sheppard  Sheppard and Teasdale (2000)  have replicated this

 

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

1113

finding and have also shown that at 2 months follow-up, mood-improved patients presented completed senten sen tence ce ste stems ms in a mor moree fun functi ctiona onal, l, les lesss pos positi itive ve way way.. We mig might ht not notee how howeve everr tha thatt the central central distinction between the Propositional and the Implicational levels of meaning may not be as clear-cut in  practice as it appears, while not disputing that it is advantageous to make distinctions between different  levels meaning. Ultimately, of course, generated in ICS a pattern-matching process, albeit albe it aofsoph sophistic isticated ated set of such processemotion processes; es; thus thus,,isICS incorpora inco rporates tes avia highly high ly soph sophistic isticated ated cognitive cognit ive theory, but it lacks a sophisticated theory of emotion, nor is it clear what unique implications, if any, the approach has for therapeutic practice with the bipolar disorders. In this manner therefore, ICS appears to swim against against the tide of current goal-ba goal-based sed appraisal theories theories of emoti emotion on that we have highlighted highlighted elsewhere (Power (Power & Dalgleish, 1997) 1997); although it might be argued that goal-based discrepancies can be modeled within ICS, such a process is not an explicit feature of the ICS approach in contrast to the SPAARS model to be considered next. Swimming against the tide is obviously not a criticism in itself  and may, indeed, prove to be remarkably percipient. For the present however we believe that the tide  is running in the right direction and will provide the direction that will be taken in the next section.

6. The sche schemat matic, ic, prop proposit osition ional, al, anal analogic ogical, al, assoc associati iative ve rep represe resentat ntation ion syst systems ems (SP (SPAARS AARS)) approach

On the basis of more recent philosophical and psychological models (Power (Power & Dalgleish, 1997) 1997), a number num ber of com compo ponen nents ts of emo emotio tion n can be ide identi ntifie fied; d; an ini initia tiatin ting g eve event nt (ex (exter ternal nal or int intern ernal) al),, an interpretation, an appraisal of the interpretation especially in relation to goal relevance, a physiological reaction, an action potential or tendency to action, conscious awareness, and overt behaviour. All of  these components are present normally in an emotion episode, with the possible exceptions of conscious awareness and overt behaviour; we have suggested that the concept of  emotion  is a holistic one that  typically includes all of these components, but that emotion is not identifiable with any one component. This approach is contrary to prior theories that have equated emotion for example with the conscious feeling  (as in the so-called feeling theories ), ), nor can emotion be equated with the physiology and overt behaviour. As with ICS, emotion is considered to be integrative across multiple processes and systems and not identifiable with any one of them.   d

b

 Q 

  b

T

 Q 

The SPAARS cognitive model of emotion is summarised in Fig. in  Fig. 5 (the 5  (the letters are merely a mnemonic for the dif differ ferent ent typ types es of rep repres resent entati ation on sys system tems—t s—the he Sch Schema ematic tic Mod Model, el, the Pro Propos positi itiona onal, l, the Associ Ass ociati ative ve and the Ana Analog logica ical). l). The app approa roach ch is mul multiti-lev level el and inc includ ludes es fou fourr dif differ ferent ent lev levels els of  representation. It would of course be possible for these representation systems to be ordered sequentially thereby forming a single level along the lines of the original cognitive therapy model (Beck, ( Beck, 1976) 1976) in which, for example, schemas produce negative automatic thoughts (propositional representations) which then the n cau cause se the emo emotio tion. n. How Howeve ever, r, in SP SPAAR AARS S the pro proces cessin sing g of th thee sch schema ematic tic,, pro propos positi itiona onal, l, and associ ass ociati ative ve lev levels els may occ occur ur in par parall allel el in a man manner ner com compar parabl ablee to   Leventhal Leventhal’s ’s (e.g. 1980 1980))   early influential multi-level theory. The initial processing of stimuli occurs through a number of mode-specific or senso sensory-sp ry-specifi ecificc syste systems ms such as the visual, the audi auditory tory,, the tactile, the propr proprioce ioceptiv ptive, e, and the olfactory which we have grouped together and termed the Analogical representation system, but which in practice also constitute a set of parallel processing modules. The importance of such systems in emotio emo tions ns and emo emotio tional nal dis disord orders ers is cle clearl arly y evi eviden dentt for exa exampl mplee in Pos Postt Tr Traum aumati aticc Str Stress ess Dis Disord order  er  (Dalgleish & Power, 2004b) 2004b) in which certain sights, sounds or other bodily sensations may become

 

1114

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

EVENT

ANALOGICAL

SCHEMATIC MODEL LEVEL

Route 1

ASSOCIATIVE LEVEL

Route 2

SYSTEM

PROPOSITIONAL LEVEL

Fig. 5. Power and Dalgleish’s SPAARS model.

inherent parts of the memory and experience of a traumatic event. The output from analogical processing then feeds into the three representation systems that operate in parallel. At the lower level there is an associ ass ociati ative ve sys system tem whi which, ch, in ter terms ms of pos posssible archi architect tectures, ures, coul could d   tak e the form of a number of  modularised connectionist networks (see e.g. Power e.g.  Power & Dalgleish, 1997, 1997, or  Williams,  Williams, Watts, MacLeod, & Mathews, 1998, 1998, for further discussion of  cognitive architectures ). ). The intermediate level of semantic representation within SPAARS is the Propositional level. This is the most language-like level of representation. Although such propositional representations have played a key role in the generation of emotion in a number of theories, such as the role of propositional level automatic thoughts in Beck’s Cognitive Therapy discussed earlier, we propose that there is no direct  route from propositions to emotion (in agreement with Teasdale with  Teasdale & Barnard, 1993) 1993), but instead argue that  they feed either through appraisals at the schematic model level or directly through the associative route (in contrast to Teasdale to Teasdale & Barnard, 1993) 1993). For example, particular words or phrases may become directly linked to emotion for certain individuals; thus, swear words come in a whole range of culture-specific forms. These words and phrases are normally designed to elicit an emotional reaction in the recipient, which is typically through the direct access associative route. Each individual accumulates a set of  unique personal words and phrases which may also directly access emotion through the associative route: significant names and significant places provide two such examples (cf. the classic cocktail party  phenomenon  of hearing one’s name spoken whilst engaged in another conversation;  Cherry, 1953) 1953), the emotion-laden nature of which may become acutely apparent to the individual following for example  bereavement when names and places associated with the loved one can trigger overwhelming feelings of  sadness, anger, and other emotions (Power, ( Power, 1999) 1999). The highest level of semantic representation, illustrated above in   Fig. 5, 5,   is labeled the Schematic Mode Mo dell le leve vel. l. Th Thee te term rm is ta take ken n fr from om   Teasda Teasdale le and Barna Barnard rd (199 (1993) 3);;   it is de desi sign gned ed to ca capt ptur uree th thee advantages of a mental model level of representation (Johnson-Laird, ( Johnson-Laird, 1983) 1983), a level that is designed to integrate information in a flexible and dynamic fashion in combination with the advantages of the more traditional schema approach, which provides a good account of repetitive and invariant relationships   b

 Q 

 d

T

 

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

1115

 between concepts but whic which h is weakest therefore where more flexible representations are needed (cf. Kahneman & Miller, 1986). 1986). In relation to emotion, the Schematic Model level is extremely important   because it is at this lev level el that  the   the generation of emotion occurs through the process of effortful appraisal (shown as Route 1 in Fig. in  Fig. 5). 5). The key processes through this route include therefore the interpretation and apprai app raisal sal of any releva rel evant nt elsewhere input, inp ut, whe whethe therr of &ext extern ernal al or int intern ernal al ori origin gin, , acc accord ing to the bas)ic basic appraisal processes considered ( Oatley (Oatley Johnson-Laird, 1987; Power &ording Dalgleish, 1997 1997) . An important feature of emotional disorder follows from the proposal in SPAARS that some of the disorders may be derived from the coupling of two or more basic emotions, or may involve the coupling of different semantic levels within an emotion module. Basic emotions are also considered to be the  building blocks from which more complex emotions are derived; they are typically considered to originate in innate   systems, be universal universal in their expression, and to appear   appear   early during the infant’s development (e.g.   Ekman, Ekman, 1992 1992)). SPAARS follows the proposal made by   Oatley Oatley & Johns Johnson-La on-Laird ird (1987) that (1987)  that there are five basic emotions of Sadness, Happiness, Anger, Fear and Disgust and that all other emotions can be derived from this basic set. The proposal in relation to a number of emotional disorders is that in many cases the coupling of two or more of these basic emotions provides the basis of  the disorder. For example, some forms of unipolar depression may occur from the coupling of Sadness and Disgust in which the individual feels both sad because of some actual or imagined loss, but, in addition, turns disgust against the self because of perceived inadequacy or culpabi lity. Although previous Although previous theori the orists sts hav havee der derive ived d dep depres ressio sion n fro from m oth other er com combin binati ations ons,, for exa exampl mple, e,   Freud Freud (19 (1917) 17)   derived Melancholia from Sadness and Anger, and more recent theorists the orists have proposed that that the comorbidity of  depression and anxiety has theoretical implications (e.g.   Watson & Clark, 1992) 1992), it is suggested that  disgust’s crucial role, especially in the form of self-disgust, has gone largely  unrecognised in relation t o  both the emotional disorders and a number of other drive-related disorders (Power & Dalgleish, 1997) 1997). The advantage of a multi-level system such as SPAARS can be illustrated by reference to particular  empirical and clinical findings that originate from work with the bipolar disorders. For example, there have been a number of studies following in the psychoanalytic tradition of the manic defense  (e.g. Lewin, 1951), 1951), in which mania is seen to be a defense against an underlying state of depression (see studies by   Winters & Neale, 1985, 1985,   and the summary in   Bentall & Kinderman, 1999, 1999, or in  Bentall, 2003). 2003 ). Lyon,  Lyon, Startup, and Bentall (1999) found (1999)  found that on an explicit test of attributional style, currently depressed depre ssed bipo bipolar lar pati patients ents gave chara characteri cteristic stic selfself-relat related ed attri attributi butions ons for nega negative tive even events, ts, wherea whereass curren cur rently tly man manic ic pat patien ients ts gav gavee nor normal mal att attrib ributi utions ons.. In con contra trast, st, on imp implic licit it tes tests ts suc such h as nam naming ing latencies and recall, the currently manic patients showed the same negative biases that the depressed  patients showed. Although this line of work   might   be int interp erpret retabl ablee in the lig light ht of the pur purpo porte rted d  psychodynamic defense mechanism, it could equally be argued that, as in SP SPAARS, AARS, that two or more emotions can be produced in parallel by the different Schematic and Associative routes, or that one emotion or mood state may replace another that is experienced more aversively (Power (Power & Dalgleish, 1997), 1997 ), as will be discussed in detail later. In relation to the self-concept, there now exist a number of models that have been applied to the normal nor mal sel self-co f-conce ncept pt and to the sel self-c f-conc oncept ept in uni unipol polar ar dep depres ressio sion n tha thatt can be app applie lied d fru fruitf itfull ully y to  bipolar disorders. For example, example, Power  Power (1987, 1991; Power & Dalgleish, 1997) has 1997)  has argued that the selfconcept can become modularised  (i.e.   (i.e. the processes become autonomous and encapsulated and are not  easily interruptable) around particular self-aspects such as certain emotions, roles, or goals, a process that  leads to a state referred to as the Ambivalent Self  . A similar proposal has been made by   Showers (1992) in (1992)  in her notion of  compartmentalization   which shares some similarities with the modularisation   b

 b

 Q 

  b

  b

 Q   Q 

 Q 

 Q 

 

1116

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

 proposal. Two studies reported in in Power,  Power, de Jong, and Lloyd (2002)   used Shower’s card sort sort task to explore the self-concept in individuals with bipolar disorders. The results suggested that in  Showers’ (1992) 1992) terms  terms of  compartme compartmentalisation ntalisation  and in our two terms of  ambivalent self   and modularisation (Pow Power er & Dal Dalgle gleish ish,, 199 1997 7), the there re is goo good d rep replic licabl ablee evi eviden dence ce tha thatt the sel self-co f-conce ncept pt is org organi anised sed   b

 Q 

  b

 Q   Q 

  b

 Q 

differently in bipolar disorders. is,being in bipolar disorders key self-aspects being a student, beingor  a mother, being a lawyer, being aThat lover, a tennis player, etc.) are almost(e.g. always entirely positive entirely negative in their content, whereas in non-bipolar controls the equivalent self-aspects are more often described with mixed positive and negative content (e.g. I’m a pretty good tennis player, but my  backhand is a bit weak, and my serve could do with improving. . .  as opposed to I’m such a fantastic tennis player who should have been at Wimbledon . . . ). ). With such modularised extreme self-aspects, it  is easy to see how the currently dominant self-aspect would be part of a feedback system that maintained overly positive or overly negative mood states; furthermore, the dominance of a positive self-aspect (i.e. the dominant active Schematic Model in SPAARS) in mania would not prevent automatic processes occurring in currently non-dominant negative self-aspects (i.e. the Associative level in SPAARS). Such  processes would appear to operate for example  as if   there was a manic defense , but SPAARS would equally predict the opposite defense , the depression defense  against mania if the same logic was followed. Rather than go down this line however, the SPAARS proposal is that so-called phenomena   b

 Q 

  b

 Q 

  b

  b

 Q 

  b

 Q 

 Q 

  b

 Q 

such as the manic defense  merely reflect multi-level processes in which some processes are conscious effortful and explicit, whereas in parallel in  parallel other processes are processes  are automatic, effortless, and implicit. Even in individu indi viduals als with hypo hypomania mania,, as   Scott and Pope (2003)   have shown, there are raised levels of   both  positive and negative self-esteem on a straightforward self-report measure, so an account based on schematic and automatic processes seems preferable to one based solely on a defensive process. In the meantime, it is clear that in bipolar individuals this dysfunctional self-organization represents part of the recurring vulnerability to the disorder. The clinical and psychotherapeutic implications are likely to be considerable and well worth further exploration. One explicit proposed application of the SPAARS model to bipolar disorders has been made by Steven   Jones (200 (2001) 1),, a summary of which is presented in   Fig. 6. 6.   Jones Jones has sug sugges gested ted tha thatt wit within hin SPAARS, the Analogical system becomes disrupted in a number of ways such as through circadian rhythms changing, an increase in energy, the experience of a positive event, and so on. This change at the Analog Ana logica icall lev level el the then n fee feeds ds int into o the Sch Schema ematic tic Mod Model, el, Ass Associ ociati ation, on, and Pro Propos positi itiona onall Lev Levels els.. For  example, at the Schematic Model level a positive model of hypomania, models of feeling superior, or  simply sim ply a rel relief ief tha thatt a dep depres ressed sed episode episode has fin finish ished, ed, can beg begin in to exa exacer cerbat batee and set up pos positi itive ve feedba fee dback ck loo loops ps aro around und all of the sys system tems. s. At the Ass Associ ociati ative ve lev level el aut automa omatic tic app apprai raisal salss and bia biases ses learned from previous experiences feed into the creation and perception of further experiences (e.g. further enforced sleep deprivation and drug and alcohol abuse in an attempt to maintain the positive high whilst whi lst red reduci ucing ng neg negati ative ve asp aspect ectss ari arisin sing g fro from m exh exhaus austio tion, n, etc etc.); .); whi whilst lst at the Pro Propo posit sition ional al lev level el  propositions of the form I feel good, creative, attractive. . .  abound and again form part of the positive feedback loops that serve to maintain and exacerbate the initial state of change at the Analogical level. Jones (2001) adaptation (2001)  adaptation of SPAARS for bipolar disorders provides an excellent starting point within which to model some of the phenomena of the bipolar disorders. However, it is clearly not the whole story, as we have suggested recently. Power recently.  Power and Schmidt (2004) have (2004)  have also tried to develop the emotion theory side of SPAARS in order to cover additional aspects of bipolar disorders. Part of the original SPAARS formulation (Power (Power & Dalgleish, 1997) 1997)  is that not only do positive feedback loops arise  within emotion modules but that they can also arise   between   modules modules such that emoti emotions ons them themselve selvess can   b

 Q 

 

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

1117

SCHEMATIC MODEL MANIA-LINKED EMOTIONS

 Positive view of mania  Feelings of superiority

ANALOGICAL SYSTEM EVENT

ASSOCIATIVE LEVEL  Automatic appraisals, appraisals,  biases

Circadian System  Disruption;  Increased Energy

 

MANIA-LINKED EMOTIONS

PROPOSITIONAL “I feel good, creative,  attractive ...” 

Fig. 6. Jones’ adaptation of the SPAARS model for bipolar disorders.

 become coupled and set up positive feedback loops between each other other.. Our original analysis, noted above, abov e, incl included uded the propo proposal sal that   unipolar   depressi depression on may con consis sistt of the cou coupli pling ng of sad sadnes nesss and disgust. In the case of mania, then happiness (elation) can become coupled with one or more other basic emotions, such as sadness in mixed states, anxiety in dysphoric mania, and anger in irritable/aggressive episodes. One of the possible implications of the coupling proposal in SPAARS is that, as Johnson et al. have recently suggested (e.g.   Cuellar et al., 2005), 2005), the correlation of mania and depression may not  represent the nosological entity that the terms  bipolar disorder   or   or  manic-depression  imply,   imply, but, rather, may simply represent co-morbidity. In other words both mania and depression clearly can occur without  each other (e.g. approximately 25% of individuals diagnosed with bipolar disorders never experience significant depression); the proposal in SPAARS is that the correct level of description of emotional disorders should begin with the basic emotions and their tendency to couple with each other; just as episodes of depression may be analysed within SPAARS as couplings of sadness and disgust or sadness and anxiety, so, equivalently, episodes of mania may be better analysed as couplings of happiness and anger, happiness and anxiety, or whatever, thereby reflecting the fact noted in the Introduction that most  so-called manic states are actually mixed states  when symptoms are carefully assessed (Cassidy (Cassidy et al., 1998). 1998 ). We are currently engaged in an assessment of basic emotions in bipolar disorders that should address these predictions. One further point that can be made about the SPAARS approach to mania is that the fast-changing appraisals that occur because of high risk taking and self-created events may lead to fast-changing emotion couplings, but these emotions tend to be experienced in an immersed  way   way rather than in a selfreflective way (Dalgleish (Dalgleish & Power, 2004a) 2004a). We have recently argued that the conscious experience of  emotio emo tion n can var vary y acc accord ording ing to whe whethe therr the ind indivi ividua duall ret retain ainss a sel self-re f-refle flecti ctive ve cap capaci acity ty du durin ring g the emotion experience (e.g. a high level schematic model of  I am now having a panic attack  ) versus  becomes completely immersed in the emotion and does not maintain a reflective capacity (e.g. death by  panic   in whi which ch aut automa omatic tic seq sequen uences ces are dom domina inant) nt).. Par Partt of the the therap rapeut eutic ic end endeav eavour our in wor workin king g  b

  b

 Q   Q 

 b

 Q 

 Q 

 b

  b

 Q 

 Q   Q 

 b

 Q 

 

1118

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

 psychologically schematic models  models of the  psychologically with mania must be therefore to help reinstate self-reflective schematic self in order to interrupt the automatic positive appraisals and action sequences that  Barnard that  Barnard (2003, 2004) has 2004)  has also recently commented on in the ICS approach. In terms of the limitations of the SPAARS approach, similar to the ICS model, the application of  SPAARSatovery the flexible bipolar framework disorders isinrecent; to bethe seen therefore of whether not it merely  provides whichittoremains re-describe phenomena bipolarordisorders. At a theoretical level, the fact that it has so much explanatory power in that it can be applied to all normal emotions and to all emotional disorders might make it unfalsifiable as a framework because different  specific theories can be incorporated; however, we are not of the view that it is unfalsifiable in that key components and putative processes lead to very clear predictions about emotional conflict, the coupling of emotions, the modularisation of emotions, the implications for the self-concept, and the nosological deriva der ivatio tion n of emo emotio tional nal dis disord orders ers.. The SP SPAAR AARS S app approa roach ch do does es at lea least st beg begin in to off offer er the lev level el of  intricacy needed for complex disorders such as the bipolar disorders, but at the same time the unique clinical clin ical features of features  of the bipolar disorders have not yet been clearly and explicitly modeled, though work   by Jones  by  Jones (2001) has (2001)  has presented an important first step in this direction. Nevertheless, the real test of the useful use fulnes nesss of SP SPAAR AARS S as a the theory ory wil willl be whe whethe therr or not it mak makes es tes testab table le emp empiri irical cal pre predic dictio tions, ns, whether it adds to our our understanding of these disorders, dis orders, and, ultimately, whether it has implications for  clinical practice. As   Power and Schmidt (2004)   have suggested, we believe that there are important  empirical and clinical implications of the SPAARS model for both unipolar and bipolar disorders, but  these implications remain to be fully tested.

7. Final points and conclusions

A very cr ude ude attempt attempt to summarise some of the strengths and weaknesses of each model has been  presented in Table in  Table 1, 1, with each of the 5 main models assessed against the three overarching criteria listed in the Introduction. It is clear that different models have different strengths and weaknesses and, equally, there are many other criteria, or more detailed criteria against which the models can be tested. As apparent from   Table 1, 1, our conclusion is that current multi-level models of cognition and emotion, specifically, the ICS and the SPAARS models, provide the best ways forward for the foreseeable future. Although neither ICS nor SPAARS were specifically developed to account for the bipolar disorders, their th eir rec recent ent app applic licati ations ons to the these se and to oth other er dis disord orders ers ill illust ustrat ratee the pot potent ential ial str streng engths ths,, emp empiri irical cal  predictions, and therapeutic implications that these models may have. Moreover Moreover,, these multi-level

Table 1 A summary of the adequacy of the 5 main theories evaluated in terms of three summary criteria

BIS/BAS Cognitive Therapy IPSRT ICS SPAARS

Clinical features of BD

Theoretical adequacy

Applicability to normal and abnormal

Low Medium High Medium Medium

Low Medium Low High High

High Medium Low High High

 

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

1119

approaches provide explanations or readily incorporate other phenomena such as the manic-defense , circadian dysrhythmias, mixed affective states, changes in the self-concept, and so on. In sum summar mary y, psy psycho cholog logica icall app approa roache chess to bip bipola olarr dis disord orders ers hav havee rea reache ched d the sta starti rting ng lin line; e; it is hoped that something of both the potential and the excitement for developments in this area have been   b

 Q 

conv co nvey eyed ed approaches. in th this is pa pape per r. place Forr to Fo too long lo ng, , th thee bi bipo pola larr di diso rder ersstheha have ve be been en ab aban ando done ned dat to simp mpli list stic ic  biological In ofo this simplicity, simplicity , there issord now opportunity to look thesi complex intera int eracti ctions ons bet betwee ween n bio biolog logica ical, l, psy psycho cholog logica ical, l, and soc social ial pro proces cesses ses,, for ind indeed eed all thr three ee typ types es of   processes are closely involved in the vulnerability vulnerability,, onset, recovery recovery,, relapse, and periodicity in these disorders. There is one area of research that integrates the biological, the psychological, and social  better than any other, and that is in the study of emotion. At one and the same time emotions are esse es sent ntia iall lly y bi biol olog ogic ical al fo forr th they ey ar aree ro root oted ed in th thee bi biol olog ogy y of ou ourr ow own n an and d ot othe herr sp spec ecie ies, s, th they ey ar aree  psychological and can preoccupy consciousness like no other experience, and they are social in that  they th ey ar aris isee in so soci cial al si situ tuat atio ions ns mo more re th than an in an any y ot othe herr. It is ho hope ped d th that at th thee su summ mmar ary y of th thes esee fi five ve different psychological approaches to the bipolar disorders has conveyed something of the strengths and wea weakne knesse ssess of eac each, h, whi while le pro provid viding ing poi pointe nters rs to fut future ure dir direct ection ionss th that at thi thiss bur burge geoni oning ng are areaa of  research and clinical need should take.

References Barnard, P. (1985). Interacting cognitive subsystems: A psycholinguistic approach to short-term memory. In A. Ellis (Ed.),  Progresss in the Psychology of Language,  vol. 2. Hove7  Lawrence Erlbaum.  Progres Barnard, P. J. (2003). Asynchrony, implicational meaning and the experience of self in schizophrenia. In T. Kircher, & A. David (Eds.),  The self in neuroscience and psychiatry. Cambridge7  Cambridge University Press. Barnard, P. J. (2004). Bridging between basic theory and clinical practice.   Behaviour Research and Therapy,  42 , 977–1000. Barnard, Barn ard, P. J., & Teasd easdale, ale, J. D. (199 (1991). 1). Inte Interacti racting ng cogn cognitive itive subsystems: subsystems: A syste systemic mic appr approach oach to cogn cognitive itive-affe -affective ctive interaction and change.  Cognition and Emotion,  5 , 1 –39 –39.. Beck, A. T. (1976).  Cognitive therapy and the emotional disorders . New York 7  Meridian. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979).   Cognitive therapy of depression: A treatment manual . New York 7 Guilford Press. Bentall, R. (2003).  Madness explained: Psychosis and human nature . London7   Penguin.  Self-regulation, and psychosis: The role of social7cognition Bentall, & Kinderman, (1999).(Eds.),  Handbookaffect of cognition and emotion In T. R., Dalgleish, & M. J.P.Power . Chichester    Wiley. in paranoia and mania . Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: The BIS/BAS scales.  Journal of Personality and Social Psychology,  67 , 319–333. Cassidy, F., Forest, K., Murry, E., & Carroll, B. J. (1998). A factor analysis of the signs and symptoms of mania.  Archives of   General Psychiatry,   55, 27–32. Cavanagh, J. (2004). Epidemiology and classification of bipolar disorder. In M. J. Power (Ed.),  Mood disorders: A handbook of    science and practice. Chichester 7   Wiley. Champion, L. A., & Power, M. J. (1995). Social and cognitive approaches to depression: Towards a new synthesis.  British  Journal of Clinical Psychology,  34 , 485–503. Cherry, E. C. (1953). Some experiments on the recognition of speech, with one and with two ears.  Journal of the Acoustical  Society of America,  25 , 975–979. Clark, D. M. (1986). A cognitive approach to panic.   Behaviour Research and Therapy,  24 , 461–470. Cuellar, A. K., Johnson, S. L., & Winters, R. (2005). Distinctions between bipolar and unipolar depression.  Clinical Psychology  Review,  25 , 307–339.

Dalgleish, T., & Power, M. 2002]. J. (2004a). The I of theReview storm:,  11 Relations between self and conscious emotion experience [A critique  Psychological  111 1, 812–881. of Lambie and Marcel,

 

1120

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1122

Dalgleish, T., & Power, M. J. (2004b). Emotion specific and emotion-non-specific components of posttraumatic stress disorder  Research and Thera Therapy py,  42 , 1069–1088. (PTSD): Implications for a taxonomy of related psychopathology.  Behaviour Research Depue, R. A., Krauss, S. P., & Spoont, M. R. (1987). A two-dimensional threshold model of seasonal bipolar affective disorder. In D. Magnusson, & A. Ohman (Eds.),   Psychopathology: An interactionist perspective . New York 7  Academic Press. Ehlers, C. L., Frank, E., & Kupfer, D. J. (1988). Social zeitgebers and biological rhythms.  Archives of General Psychiatry,  45 , 948–952. Ekman, P. (1992). An argument for basic emotions.  Cognition and Emotion,  6 , 169–200. Eysenck, H. J. (1967).  The biological basis of personality . Springfield7  Charles C. Thomas. Freud, S. (1917/1984). Mourning and melancholia.   Pelican Freud Library,  vol. 11. Harmondsworth7   Penguin. Gazzaniga, M. S. (1988). Brain modularity: Towards a philosophy of conscious experience. In A. Marcel, & E. Bisiach (Eds.), Consciousness in contemporary science. Oxford7  Oxford University Press. Goodwin, F. K., & Jamison, K. R. (1990).   Manic depressive illness . Oxford7  Oxford University Press. Goplerud, E., & Depue, R. A. (1985). Behavioral response to naturally occurring stress in cyclothymia and dysthymia.  Journal  of Abnorm Abnormal al Psych Psychology ology,  94 , 128–139. Gray, J. A. (1976). The behavioural inhibition system: A possible substrate for anxiety. In M. P. Feldman, & A. M. Broadhurst  (Eds.),   Theoretical and experimental bases of behaviour modification. Chichester 7   Wiley. Gray, J. A. (1982).   The neuropsychology of anxiety. Oxford7  Oxford University Press. Gray, J. A. (1990). Brain systems that mediate both emotion and cognition.  Cognition and Emotion,  4 , 269–288. Gray, J. A. (1999). Cognition, emotion, conscious experience and the brain. In T. Dalgleish, & M. J. Power (Eds.),  Handbook of   cognition cognit ion and emoti emotion on. Chichester 7   Wiley. Healy, D., & dysrhythmia Williams, J. M. G. pathogenesis (1988). Dysrhythmia, dysphoria, and depression: interaction of learned helplessness and   Psychological BulletinThe circadian in the of depression. ,   103 , 163–178. Healy, D., & Williams, J. M. G. (1989). Moods, misattributions and mania: An interaction of biological and psychological factors in the pathogenesis of mania.  Psychiatric Developments,  7 , 49–70. Johnson, S. L. (2005). Mania and dysregulation in goal pursuit: A review.   Clinical Psychology Review,  25 , 241–262. Johnson, S. L., & Fingerhut, R. (2004). Negative cognitions predict the course of bipolar depression, not mania.  Journal of   Cognitive Psychotherapy: An International Quarterly,  18 , 149–162. Johnson, S. L., Sandrow, D., Meyer, B., & Winters, R. (2000). Life events involving goal-attainment and the emergence of  manic symptoms.  Journal of Abnormal Psychology,  1  109 09, 721–727. Johnson, S. L., Turner, R. J., & Iwata, N. (2003). BIS/BAS levels and psychiatric disorder: An epidemiological study.  Journal  of Psychopathology and Behavioral Assessment ,  25 , 25–36. Johnson Joh nson-Lair -Laird, d, P. N. (198 (1983). 3).   Mental Mental mod models els:: Towa oward rdss a cog cognit nitive ive sci scienc encee of lan languag guage, e, inf infer erenc encee and con consci scious ousnes nesss . Cambridge 7  Cambridge University Press. Jones, S. H. (2001). Circadian rhythms, multilevel models of emotion and bipolar disorder—an initial step towards integration. Clinical Psychology Review,  21 , 1193–1209.  Psychological Review,  93 Kahneman, Miller, D. T. Norm theory: Comparing reality to (1984). its alternatives. , 136– 153 153...   Interpersonal psychotherapy of depression Klerman, G.D., L.,&Weissman, M.(1986). M., Rounsaville, B. J., & Chevron, E. S.  New York 7  Basic Books. Lam, D., Wright, K., & Smith, N. (2004). Dysfunctional assumptions in bipolar disorder.   Journal of Affective Disorders,  79 , 193–199. Lam, D. H., Jones, S. H., Hayward, P., & Bright, J. A. (1999).  Cognitive therapy for bipola bipolarr disor disorder  der . Chichester 7   Wiley. Lane, R. D., & Nadel, L. (Eds.) (2000).   Cognitive neuroscience of emotion. New York 7   Oxford. Lazarus, R. S. (1966).  Psychological stress and the coping process . New York 7  McGraw-Hill. Lazarus, R. S. (1991).  Emotion and adaptation . New York 7  Oxford University Press. Leahy Lea hy,, R. L. (19 (1999) 99).. De Decis cision ion mak making ing and ma mania nia..  Journal of Cognitive Psychotherapy: An International Quarterly,  13, 83–105. Leventhal, H. (1980). Toward a comprehensive theory of emotion. In L. Berkowitz (Ed.),  Advances in Experimental Social   Psychology,  vol. 13. New York 7  Academic Press. Leventhal, H., & Scherer, K. (1987). The relationship of emotion to cognition: A functional approach to a semantic controversy. Cognition and Emotion,   1, 3 –28 –28.. Lewin, B. D. (1951).  The psychoanalysis of elation . London7  Hogarth Press.

Lyon, H. M., Startup, M., & Bentall,disorder. R. P. (1999). Social cognition and the manic defense: Attributions, selective attention, and  Journal of Abnormal Psychology self-schema in bipolar affective ,  108 , 273–282.

 

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

 

1121

Meyer, A. (1957).  Psychobiology: A science of man . Springfield7  Charles Thomas. Meyer, B., Beevers, C. G., & Johnson, S. L. (2004). Goal appraisals and vulnerability to bipolar disorder: A personal projects analysis.   Cognitive Therapy and Research,  28 , 173–182. Meyer Mey er,, B., Joh Johnso nson, n, S. L., & Car Carver ver,, C. S. (19 (1999 99). ). Exp Explor loring ing beh behavi aviora orall act activa ivatio tion n and inh inhibi ibitio tion n sen sensit sitivi ivitie tiess amo among ng college colle ge stud students ents at risk for bipo bipolar lar spec spectrum trum symp symptoma tomatolo tology gy..  Journal of Psychopathology and Behavioral Assessment , 21, 275–292. Meyer, B., Johnson, S. L., & Winters, R. (2001). Responsiveness to threat and incentive in bipolar disorder: Relations of the Psychopatho hopathology logy and Behav Behavioral ioral Assessment  Assessment ,  23 , 133–143. BIS/BAS scales with symptoms.  Journal of Psyc Miklowitz, D. J., Goldstein, M., Neuchterlein, K., Snyder, M., & Mintz, J. (1988). Family factors and the course of bipolar  affective disorder.   Archives of General Psychiatry,  45 , 225–230. Monk, T. H., Kupfer, D. J., Frank, E., & Ritenour, A. M. (1991). The Social Rhythm Metric (SRM): Measuring daily social rhythms over 12 weeks.  Psychiatry Research,  36 , 195–207.  Newman, C. F. F.,, Leahy Leahy,, R. L., Beck, A. T. T.,, Reilly-Harringto Reilly-Harrington, n, N., & Gyulai, L. (2002).  Bipolar disorder: A cognitive therapy approach. Wa Washington shington7  American Psychological Association. Oatley, K., & Johnson-Laird, P. N. (1987). Towards a cognitive theory of emotions.  Cognition and Emotion,  1 , 29–50. Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957).  The measurement of meaning . Urbana7  University of Illinois. Papolos, D. F. (2003). Bipolar disorder and comorbid disorder: The case for a dimensional nosology. In B. Geller, & M. P. DelBello (Eds.),  Bipolar disorder in childhood and early adolescence . New York 7  Guilford Press. Perry, A., Tarrier, N., Morriss, R., McCarthy, E., & Limb, K. (1999). Randomised controlled trial of efficacy of teaching  patients with bipolar disorder to identify early symptoms of relapse and obtain treatment.  British Medical Journal ,   318,

149–153. Post, R. M. (1992). Transduction of psychosocial stress into the neurobiology of recurrent affective disorder.  American Journal  of Psychiatry,  1  149 49, 99 999 9 –1010 –1010.. Power, Powe r, M. J. (198 (1986). 6). A techn technique ique for meas measuring uring processing processing load duri during ng spee speech ch prod productio uction. n.   Journal of Psych Psycholingui olinguistic stic  Research,  15 , 371–382. Power, M. J. (1987). Cognitive theories of depression. In H. J. Eysenck, & I. Martin (Eds.),   Theoretica Theoreticall foundat foundations ions of    behaviour therapy. New York 7  Plenum Publishing Corporation. Power, M. J. (1991). Depression: The role of the self-concept.   Analise Psicologica,  9 , 33–41. Power, M. J. (1999). Sadness and its disorders. In T. Dalgleish, & M. J. Power (Eds.),  Handbook of cognition and emotion . Chichester 7   Wiley. Power, Powe r, M. J. (200 (2002). 2). Integrative Integrative thera therapy py from a cogn cognitive itive-beh -behavio avioural ural perspective. perspective. In J. Holm Holmes, es, & A. Bate Bateman man (Eds.),  Integration in psychotherapy: Models and methods. Oxford7  Oxford University Press. Power, M. J. (in press).  The structure of emotion: An empirical comparison of six models.  Cognition and Emotion. Power, M. J., & Champion, L. A. (1986). Cognitive approaches to depression: A theoretical critique.  British Journal of Clinical   Psychology,   25, 201–212. 7

 Cognition and emotion: From order to disorder . Hove  Psychology Press. Power, M. J., J., de & Jong, Dalgleish, (1997). Power, M. F., &T. Lloyd, A. (2002). The organisation of the self-concept in bipolar disorders: An empirical study and replication.   Cognitive Therapy and Research,  2  26  6 , 553–561. Power, M. J., Katz, R., McGuffin, P., Duggan, C. F., Lam, D., & Beck, A. T. (1994). The Dysfunctional Attitude Scale (DAS): A comparison of forms A and B and proposals for a new subscaled version.  Journal of Research in Personality,  28 , 263– 276 276.. Power, M. J., & Schmidt, S. (2004). Emotion-focused treatment of unipolar and bipolar mood disorders.   Clinical Psychology and Psychotherapy,  11, 44–57. Prien, R., & Potter, W. W. (1990). NIMH Wo Workshop rkshop on treatment of bipolar disorders. Psychopharmacology Bulletin,  26 , 409– 427 427.. Reilly-Harrington, N. A., Alloy, L. B., Fresco, D. M., & Whitehouse, W. G. (1999). Cognitive styles and life events interact to  predict bipolar and unipolar symptomato symptomatology logy..  Journal of Abnormal Psychology,  10  108 8, 567–578. Rolls, E. T. (1999).  The brain and emotion . Oxford7  Oxford University Press. Russell, J. A., & Carroll, J. M. (1999). On the bipolarity of positive and negative affect.   Psychological Bulletin,  1  125 25, 3 –30 –30.. Schwannauer, M. (2004). Cognitive behavioural therapy for bipolar affective disorder. In M. J. Power (Ed.),  Mood disorders: A handbook of science and practice . Chichester 7   Wiley. Scott, J., & Pope, M. (2003). Cognitive styles in individuals with bipolar disorder.   Psychological Medicine,  30 , 467–472.

Sco Psychological Scott, tt, J., Sta Stanto nton, n,Medicine B., Gar Garlan land, d, A., & Fer Ferrie rier, r, I. N. (20 (2000 00). ). Cog Cognit nitive ive vu vulne lnerab rabili ility ty in pa patie tients nts wit with h bip bipola olarr dis disord order er.. ,  30 , 467–472.

 

1122

 M.J. Power / Clinical Psychology Review 25 (2005) 1101–1 1101–1122 122

Seligman, M. E. P. (1975).   Helplessness: On depression, development and death . San Francisco7  Freeman. Sheppard, L. C., & Teasdale, J. D. (2000). Dysfunctional thinking in major depressive disorder: A deficit in metacognitive monitoring? Journal of Abnormal Psychology,  109 , 768–776. Showers, C. J. (1992). Compartmentalization of positive and negative self-knowledge: Keeping bad apples out of the bunch.  Journal of Personality and Social Psychology,  62 , 1036–1049. Bipolar Affec Affective tive Disor Disorder: der: A Nation National al Clinic Clinical al Guide Guideline line. Edin SIGN (200 (2005). 5).   Bipolar Edinburg burgh h7   Scottish Scottish Intercollegiate Guidelines  Network. Stern, G. S., & Berrenberg, J. L. (1979). Skill-set, success outcome, and mania as determinants of the illusion of control.  Journal of Research in Personality,  1  13 3, 206–220. Sullivan, H. S. (1953).  The interpersonal theory of psychiatry . New York 7  Norton. Swartz, H. A., Frank, E., Spielvogle, H. N., & Kupfer, D. J. (2004). Interpersonal and social rhythm therapy. In M. J. Power  (Ed.),  Mood disorders: A handbook of science and practice. Chichester 7   Wiley. Teasdale, J., & Barnard, P. (1993).  Affect, cognition and change . Hove7  Lawrence Erlbaum Associates. Teasd easdale, ale, J. D., Taylor, Taylor, M. J., Cooper, Cooper, Z., Hayhurst, Hayhurst, H., & Payk Paykel, el, E. S. (199 (1995). 5). Depressive Depressive thin thinking king shifts in cons construc truct t   104 4, 500–507. accessibility or in schematic mental models.  Journal of Abnormal Psychology,  10 Watson, D., & Clark, L. A. (1992). Affects separable and inseparable: On the hierarchical arrangement of the negative affects.  Journal of Personality and Social Psychology,  62 , 489–505. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales.  Journal of Personality and Social Psychology,  5  54 4, 1063–1070. Wehr, T. A., Sack, D. A., & Rosenthal, N. E. (1987). Sleep reduction as a final common pathway in the genesis of mania.  American Journal of Psychiatry,  1  144 44, 201–204. Weissman, M. M., Markowitz, J. C., & Klerman, G. L. (2000).   Comprehensive guide to interpersonal psychotherapy . New York 7  Basic Books. Weissman, M. M., & Myers, J. K. (1978). Affective disorders in a US urban community: The use of Research Diagnostic Criteria in an epidemiological survey.   Archives of General Psychiatry,   35, 1304–1311. Williams Wil liams,, J. M. G., Watts, Watts, F. N., MacLeod, C., & Math Mathews, ews, A. (199 (1998). 8). Cognitive Cognitive psyc psycholo hology gy and emotional emotional diso disorders rders (2nd ed.) ed.).. Chic Chicheste hester  r 7   Wiley. Winters, K. C., & Neale, J. M. (1985). Mania and low self-esteem.  Journal of Abnormal Psychology,  94 , 282–290. Wright, K., & Lam, D. (2004). Bipolar affective disorder: Current perspectives on psychological theory and treatment. In M. J. Power (Ed.),  Mood disorders: A handbook of science and practice. Chichester 7   Wiley.