© 2014 John Wiley & Sons A/S Published by John Wiley & Sons Ltd.

Bipolar Disorders 2015: 17: 194–204

BIPOLAR DISORDERS

Original Article

Neuropsychological performance of patients with soft bipolar spectrum disorders Lin K, Xu G, Lu W, Ouyang H, Dang Y, Guo Y, So K-F, Lee TMC. Neuropsychological performance of patients with soft bipolar spectrum disorders. Bipolar Disord 2015: 17: 194–204. © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd. Objectives: There is much evidence that shows that a substantial number of individuals with DSM-IV-defined unipolar depression (UP) manifest hypomanic sub-syndrome and bipolar diathesis. Other definitions have conceptualized the term soft bipolar spectrum (SBP) for these individuals. Little is known about the cognitive profiles of individuals with SBP. We hypothesized that they are representative of individuals with bipolar II disorder and are different from that of ‘strict’ UP. Methods: Consecutive referrals suffering major depressive episodes were categorically assigned to groups of either bipolar I disorder (n = 98), bipolar II disorder (n = 138), or UP (n = 300). Based on the SBP criteria by Akiskal and Pinto (17), patients with UP were subdivided into 81 SBP and 219 strict UP. We administered self- and clinician-administered scales to evaluate affective temperaments, and neuropsychological tests to assess seven cognitive domains. Results: Patients with SBP performed significantly better than strict UP patients in the domains of processing speed (p = 0.002), visual-spatial memory (p = 0.017), and verbal working memory (p = 0.017). Compared to patients with bipolar I disorder, patients with SBP were significantly better in set shifting (p < 0.001) and visual-spatial memory (p = 0.042). Patients with SBP performed similarly to patients with bipolar II disorder in all of the cognitive domains tested (p > 0.05). There was a group 9 cognitive domain interaction effect between bipolar I disorder, bipolar II disorder, SBP, and strict UP groups [Pillai’s F = 2.231, df = (18,1437), p = 0.002]. Conclusions: Our data suggest that patients with SBP differ from patients with UP not only in external validators (e.g., family history of bipolar disorder) and hypomanic symptoms, but also in neuropsychological performance and that the profiles of cognitive functioning were different across bipolar I disorder and ‘bipolar II spectrum’ that subsumes bipolar II disorder and SBP.

Kangguang Lina,b,c, Guiyun Xub, Weicong Lub, Huiyi Ouyangb, Yamei Dangb, Yangbo Guob, Kwok-Fai Sod,e,f and Tatia MC Leea,c,d,g a Laboratory of Neuropsychology, The University of Hong Kong, Hong Kong, bDepartment of Psychiatry, Guangzhou Psychiatric Hospital, Affiliated Hospital of Guangzhou Medical University, Guangzhou, cLaboratory of Cognitive Affective Neuroscience, The University of Hong Kong, dThe State Key Laboratory of Brain and Cognitive Sciences, The University of Hong Kong, Hong Kong, eGMH Institute of CNS Regeneration, Jinan University, Guangzhou, f Department of Ophthalmology, The University of Hong Kong, Hong Kong, gInstitute of Clinical Neuropsychology, The University of Hong Kong, Hong Kong, China

doi: 10.1111/bdi.12236 Key words: bipolar disorder – cognition – depression – soft bipolar spectrum – temperament Received 5 December 2013, revised and accepted for publication 9 May 2014 Corresponding author: Guiyun Xu, M.D. Department of Psychiatry Guangzhou Psychiatric Hospital Affiliated Hospital of Guangzhou Medical University 36 MingxinLoad, Guangzhou Guangdong Province 510370 China Fax: 86-020-81891391 E-mail: [email protected] Kangguang Lin and Tatia M.C. Lee contributed equally to this work.

The lifetime prevalence of bipolar disorder in epidemiological surveys using the Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria is between 0.3% and 1.5% for bipolar I

194

disorder and between 1.0% and 2.0% for bipolar II disorder (1–3). Yet a growing body of evidence suggests that the DSM criteria for bipolar II disorder on the requisite number (and the type) of

Cognitive deficits in bipolar spectrum disorders hypomanic symptoms and on the duration of four days are so strict that they exclude a substantial percentage of individuals who express varying manifestations of bipolar syndrome to a lesser extent (4–12). For instance, in a 20-year prospective community cohort study, Angst and colleagues (5) demonstrated that criteria for bipolar disorder by DSM-IV were unjustified in terms of clinical validators and further clinically validated a wider definition of bipolar II disorder with a prevalence of 10.9%, in which they considered a major depressive episode (MDE) plus, by their criteria, the hard definition of hypomania (over-activity, three out of seven hypomanic signs/symptoms, and brief episodes of one to three days and consequences) and the soft definition (without consequences). Zimmermann and colleagues (9), in a ten-year longitudinal study involving 2,210 participants, reported that a cumulative rate of sub-threshold bipolar disorder, in which they considered individuals who did not meet the DSM-IV criteria for hypomania either on criterion B (a minimum number of symptoms) or on criterion D (symptoms observable by others), was as high as 9.3%. Under the DSM diagnostic system, such a huge group of individuals is unrecognized and is most likely to fall into the realm of major depressive disorder or unipolar depression (UP). However, identifying these individuals is of scientific as well as clinical importance because they, compared to individuals with UP, are more likely to have severe role impairment (1), commit suicide (13), suffer more recurrent depressive episodes (10), and convert to bipolar disorders (1, 9). Kretschmer (14) previously described a continuum of affective states ranging from extreme temperaments to full-blown mania and further proposed that cyclothymic disposition constituted the core feature of the classic manic-depressive illness. Recently, Akiskal and colleagues (15, 16) suggested that temperamental dysregulation represented the fundamental pathology of mood disorders and proposed a construct of soft bipolar spectrum (SBP) beyond bipolar I and bipolar II disorders, which, along with bipolar II disorder, form a bipolar II spectrum, as opposed to bipolar I disorder (17). SBP incorporates temperaments or traits with stated features, which they suggest helps to improve the validity of current diagnosis. SBP consists of bipolar II1/2 [depression with the cyclothymic temperament (CT)], bipolar III (depression with hypo/mania associated with antidepressant), and bipolar IV (depression with the hyperthymic temperament). These soft bipolar subtypes were subsequently validated in the French National epidemiology of depression (EPIDEP) study (8, 18).

Cyclothymia, which falls short of hypo/manic and depressive episodes and lasts for at least two years, pertains to bipolar disorder on Axis I in the current official nosology of DSM-IV (and DSM-V) as an attenuated variant. On the other hand, patients with UP superimposed with the CT (bipolar II1/2) overlap largely with bipolar II disorder, and it is not easy to differentiate them (19). Furthermore, the ‘up-and-down’ trait, which is a characteristic of the CT, has been suggested as a potential marker for bipolar II disorder (20). Finally, the CT may help to predict switching to bipolar II disorder from depression (21). The CT thus may represent the nature of bipolarity. As for bipolar III disorder, in a follow-up study for an average of three years, Akiskal et al. (22) found that the specificity of antidepressant-associated mania was 100% as a marker for bipolar disorder, as now reflected in the DSM-V, wherein such a response to an antidepressant is sufficient to establish a bipolar diagnosis. Cassano and colleagues (23) found that individuals with UP who carried hyperthymic temperament had a higher rate of familial bipolarity than those who did not carry the temperament. The evidence above supports the validity of SBP. However, SBP is still understudied and debated; more studies are needed to further validate it and explore its clinical utilities. Another approach, by Ghaemi et al. (24), to characterize the softer expressions of bipolar syndrome is to weight a certain type (and number) of indicators of bipolarity in the context of a MDE, requiring two ‘positive’ responses from Criterion C and nine from Criterion D. Accordingly, individuals with UP who meet the definition of Ghaemi et al. are re-defined as having bipolar spectrum disorder (BSD). Cognitive deficits are a core feature of UP. Consistent results have been reported in such domains as set shifting and visual working memory, whereas both impairment and non-impairment results have been reported in such domains as attention, processing speed, and verbal working memory (25–29). Among the confounding factors accountable for the discrepancy are heterogeneous clinical samples and the deleterious effect of medication on neuropsychological performance. With regard to the cognitive deficits in bipolar disorder, these have been documented not only during depressive and manic states, but also in the euthymic period. These findings nonetheless may be limited to bipolar I disorder as most of the studies recruited individuals with bipolar I disorder only or included a mixed sample without separating bipolar I and bipolar II disorders (30–34). A few neuropsychological studies examining cognitive

195

Lin et al. deficits in the unaffected relatives of probands with bipolar disorder suggested that deficits in some domains, such as verbal learning, memory, verbal fluency, and planning, may serve as cognitive endophenotypes (35, 36). However, little is known about cognitive deficits in individuals with SBP and the extent to which they perform differently from those with UP, bipolar I disorder, or bipolar II disorder. We hypothesized that individuals with SBP may perform differently from those with UP but similarly to those with bipolar II disorder, and that the profiles of cognitive functioning in ‘bipolar II spectrum’ that subsumes bipolar II and SBP were distinct from those in bipolar I disorder. The primary aims, then, were to (i) validate Akiskal’s SBP using external validators (e.g., family history of bipolar disorder, earlier age at onset, recurrent depressive episodes) as well as past hypomanic signs/symptoms; (ii) compare neuropsychological performance in SBP and UP; (iii) compare neuropsychological performance between Akiskal’s SBP and Ghaemi’s BSD; and, finally, (iv) compare cognitive profile deficits between bipolar I disorder and the ‘bipolar II spectrum’. Besides the necessity of delineating cognitive deficits for SBP, such a strategy is important because if SBP is distinct from UP in cognitive deficits, then the heterogeneous samples may be responsible for the incongruent results of cognitive deficits in UP. More importantly, a dimensional approach (e.g., cognitive function) to characterizing mood disorders may assist in classifying them and eventually identifying differentiating biomarkers.

Methods Study setting and procedures

The study was conducted in Guangzhou Psychiatric Hospital – the oldest psychiatric hospital in China, which was established by an American doctor, J. G. Kerr, in 1898 (37) – and the First Affiliated Hospital of Jinan University. Both hospitals are tertiary medical centers (e.g., national centers and university teaching hospitals) from which all of the patients were recruited. The present study presents the neuropsychological data derived from the project – Clinical and Biological Characteristics and Optimizing treatment in Bipolar depressive disorders (CBCOB) – the aims of which were to improve the detection of bipolar disorder in the context of a MDE and to optimize treatment and functional outcomes for patients with bipolar disorder. More details can be found elsewhere (29, 38).

196

Inpatients or outpatients who received psychiatric services in the two hospitals were referred to the study by their first-contact psychiatrist when suffering from a MDE. After obtaining written consent, one of the study authors, a full-time research psychiatrist of this study (who had been in practice for more than ten years) collected demographic and clinical data from the patients and very often from their significant others (e.g., family members), conducted subsequent clinical interviews, and applied the Chinese version of the Structured Clinical Interview for DSM-IV-TR Axis I Disorders Patient Edition (SCID-I/P) to confirm the diagnoses. At this stage, another of the study authors (also a senior psychiatrist) conducted independent clinical interviews. The inter-rater reliability for the diagnoses between the two interviewers was high (kappa value > 0.9). To improve the detection of bipolar II disorder, as well as to probe sub-hypomania in depressed patients whenever possible, we collected clinical data from the significant others (often family members) who lived with the patients and accompanied them to the hospitals, which is not uncommon in China. This procedure helps to identify clinical variables as well as clarify some diagnoses that require functional impairment outcome – that is, Criterion D for bipolar II disorder (symptoms observable by others) (19). In addition, we applied two self-rated instruments emphasizing the signs and symptoms of bipolar disorder – that is, the Chinese version of the 15-item hypomania checklist (HCL-15), which we have found to show fairly high sensitivity and high specificity for bipolar depressive disorders as opposed to unipolar depression (the sensitivity of HCL-15 was 0.78 in the detection of bipolar II disorder and 0.46 for bipolar I disorder; the specificity was 0.9 and 0.69, respectively) (manuscript in press), and the Chinese translation of the Bipolar Spectrum Diagnostic Scale originally developed by Ghaemi et al. (39). As a result, we diagnosed the patients on the basis of the combination of the consensus of clinical impression, the SCID-I/P interview, and a review of self-rated scales and medical records. After the one-week screening and evaluating period, we followed the patients for six weeks. During this period, all of the patients were evaluated each week using the 17-item Hamilton Depression Rating Scale (HAM-D) (40), Young Mania Rating Scale (YMRS) (41), Hamilton Anxiety Rating Scale (HAM-A) (42), and the Brief Psychiatric Rating Scale (BPRS) (43). As an additional quality control measure, the Guangzhou Psychiatric Hospital assigned a group of three senior psychiatrists to conduct random

Cognitive deficits in bipolar spectrum disorders inspections. In brief, all of the patients in the study underwent systematic assessments and repeated, in-depth evaluations, and their diagnoses were prospectively validated. The project was approved by the Ethics Committee of the Guangzhou Psychiatric Hospital and administered by the Chinese Clinical Trial Registry (ChiCTR-TNRC-10001112; http://www.chictr.org/). Participants

The present data analysis involved 300 patients with DSM-IV-defined UP, 98 patients with bipolar I disorder, and 138 patients with bipolar II disorder, aged between 18 and 60 years. According to the exclusion criteria, patients were free from the following conditions: pregnancy, serious general medical illness, history of seizure disorder, DSMIV-TR-defined organic mental disorders, dementia, schizophrenia, delusional disorder, schizoaffective disorder, active substance use disorder, and history of mental retardation. Based on the definition of SBP by Akiskal and Pinto (17), we categorically re-assigned those with UP into 81 SBP (36 bipolar II1/2, nine bipolar III, and 36 bipolar IV) and 219 strict UP patients. The CT was evaluated by means of a semi-structured interview (44). The hyperthymic temperament was diagnosed on the basis of clinical interviews and using the criteria of Akiskal et al. (45), which require four out of six habitual traits: (i) cheerful, over-optimistic, or exuberant; (ii) extroverted and people-seeking; (iii) over-talkative, eloquent, and jocular; (iv) uninhibited, stimulus-seeking, and sexually driven; (v) vigorous, full of plans, and improvident; and (vi) overconfident, self-assured, and boastful. These temperament instruments were derived from the Temperament Evaluation of Memphis, Pisa, Paris, and San Diego-Autoquestionnaire (TEMPS-A) (46), which has been translated into many main languages and validated, including its Chinese version (47). In addition, 41 patients with BSD were determined using the definition of Ghaemi et al. (24), such that they met the DSM-IV criteria for a MDE and presented a certain number of indicators of bipolarity. We note that in our UP samples, 99 patients were experiencing their first episode, which may have resulted in a relatively small number of patients with BSD by Ghaemi’s criteria, which are weighted largely on illness course variables and the history of treatment responses, such as recurrent major depressive episodes (> 3), brief major depressive episodes (on average, three months), and lack of response to antidepressant treatment trials (≥ 3).

The healthy control group consisted of 200 healthy subjects (100 males and 100 females), aged 18–60 years, with no personal or family history (first-degree relatives) of mental disorders, as documented in our previous report (29). To ascertain that they were free of mental illness, we conducted clinical interviews with them and applied the following self- and clinician-administered rating scales: (i) the self-rated scale (Chinese translation) of the Symptom Checklist 90, with a cut-off score of ≤ 160 in the total scores and ≤ 2 in the subscale scores (48); (ii) the Chinese version of HCL-15, with a cut-off of < 8 for symptoms/signs; and (iii) HAM-D with a cut-off score of ≤ 6. Neuropsychological assessment

We administered a battery of neuropsychological tests to all the patients and healthy controls: (i) processing speed: the Trail Making Test–Part A (TMT-A) (49) and the Digit Symbol–Coding subtest of the Wechsler Adult Intelligence Scale– Revised by China (WAIS-RC) (50); (ii) attention: the Digit Span Forward subtest of the WAIS-RC (50); (iii) set shifting: the Modified Wisconsin Card Sorting Test (WCST-M) (51) and Trail Making Test–Part B (TMT-B) (52); (iv) planning: the Tower of Hanoi (TOH) (53); (v) verbal fluency: the animal naming test (54); (vi) visual-spatial memory: the Immediate Visual Reproduction subtest of the Wechsler Memory Scale–Revised by China (WMS-RC) (55); and (vii) verbal working memory: the Digit Span Backward subtest of the WAIS-RC (50). For some tests that measure similar domains (e.g., set shifting by WCST-M and TMT-B), we combined these measures of similar domains by creating Z scores (described below). The rationale for such grouping is that it may help to generate relatively conservative results, thus avoiding the issue of whether the deficits in a single test can represent those of the domain that can be assessed by some similar tests. On the other hand, it is worth noting that there are no identical measures for specific domains, particularly executive functioning. A trained rater conducted all of the neuropsychological tasks within the one-week screening and evaluating period. As all of the patients were either newly diagnosed cases or had discontinued psychiatric medications for at least two weeks, they were free of these medications when the neuropsychological assessments were administered. Participants who had missing data on two or more cognitive domains were excluded from the analysis. Seven patients with UP did not complete the TMT-B, and six patients with bipolar I disorder did not

197

Lin et al. complete the TOH. Four patients with bipolar II disorder did not complete the TOH, and another three patients with bipolar II disorder did not complete the TMT-B. Statistical analysis

All of the data were analyzed by using the Statistical Package for Social Sciences (IBM SPSS statistics for Windows, Version 20.0; IBM Corp., Armonk, NY, USA) software. One-way analysis of variance (ANOVA) or the chi-square test was used to compare demographic and clinical variables among the groups. Neurocognitive measures were standardized using the mean values and standard deviations from the control group as norms. Neurocognitive measures with negative scale properties were inverted, with high scores indicating better performance. For the cognitive domains – processing speed, set shifting, and planning – that had more than two measuring scores, a composite score was computed for each. Multivariate analysis of variance (MANOVA) was used to compare neuropsychological performance among the groups, with age, years of education, and psychotic features as covariates. One-way ANOVA was then applied to clarify the significant findings identified by MANOVA, using Bonferroni correction for multiple comparisons. To assess the differences in cognitive profiles across the diagnostic groups, repeated-measures MANOVA, with age, years of education, and psychotic features as covariates, was performed to examine the group 9 cognitive domain interactions.

Results Sample characteristics

Table 1 summarizes the comparisons of SBP with strict UP, bipolar I disorder, and bipolar II disorder. Patients with SBP were younger than those with strict UP, but of a similar age to those with bipolar I disorder or bipolar II disorder. The percentage of females in the SBP group was similar to that in the strict UP group and in the bipolar II disorder group, but was higher than that in the bipolar I disorder group. There were no significant differences between the SBP group and the strict UP, bipolar I disorder, and bipolar II disorder groups in terms of years of education, HAM-D scores at baseline, YMRS scores at baseline, and the rate of psychotic features at baseline. Clinically validating SBP

We considered the following as clinical validators for SBP: family history of bipolar disorder, age at onset, greater depressive recurrence (≥ 3 episodes), and past hypomanic symptoms measured by HCL15. Table 1 compares patients who were defined as SBP to those who were diagnosed with strict UP, bipolar I disorder, or bipolar II disorder. Compared to strict UP, patients with SBP had a higher incidence of a family history of bipolar disorder (17.3% versus 8.6%, p = 0.032), earlier age at onset (24.7 years versus 31.1 years, p < 0.001), and were more likely to have had more (≥ 3) depressive recurrences (28.4% versus 9%, p = 0.042). In all of these characteristic features of

Table 1. Sample characteristics and clinical validity of the soft bipolar spectrum p-valueb Characteristics

Gender, female, n (%) Age, years, mean (SD) Education, years, mean (SD) HAM-D at baseline, mean (SD) YMRS at baseline, mean (SD) Psychotic features, n (%) HCL-15, mean (SD) Family history of bipolar disorder, n (%) Age at onset, years, mean (SD) ≥ 3 major depressive episodes, n (%)

SBPa (n = 81)

49 (60.5) 30.0 (11.0) 11.6 (4.1) 26.9 (5.8) 1.6 (2.3) 29 (35.8) 3.1 (3.9) 14 (17.3) 24.7 (11.2) 23 (28.4)

UP (n = 219)

118 (53.9) 36.3 (13.0) 11.0 (3.9) 26.5 (6.4) 1.0 (1.8) 61 (27.6) 1.5 (2.8) 19 (8.6) 31.1 (12.7) 20 (9.0)

BP-II (n = 138)

72 (52.2) 31.1 (11.6) 12.5 (4.0) 26.9 (6.5) 1.6 (2.4) 35 (25.4) 9.0 (3.4) 29 (21.0) 24.1 (10.7) 40 (28.8)

BP-I (n = 98)

39 (39.8) 31.4 (11.0) 11.9 (3.8) 25.8 (6.7) 1.9 (3.3) 39 (39.8) 6.0 (5.2) 14 (14.3) 25.3 (10.0) 39 (39.8)

SBP versus UP

BP-II

BP-I

0.306 0.000 1.000 1.000 1.000 0.167 0.005 0.032 0.000 0.043

0.333 1.000 0.501 1.000 1.000 0.101 0.000 0.502 1.000 0.926

0.006 1.000 1.000 1.000 1.000 0.584 0.000 0.583 1.000 0.111

BP-I = bipolar I disorder; BP-II = bipolar II disorder; HAM-D = Hamilton Depression Rating Scale; HCL-15 = 15-item hypomania symptom checklist; SBP = soft bipolar spectrum; SD = standard deviation; UP = unipolar depression; YMRS = Young Mania Rating Scale. a No differences were found between the subgroups of SBP. b Chi-square tests were applied for qualitative data, and one-way analysis of variance for quantitative data, using Bonferroni correction.

198

Cognitive deficits in bipolar spectrum disorders bipolar disorder, there were no statistically significant differences between the SBP group and the bipolar I disorder and bipolar II disorder groups. With regard to past hypomanic signs/symptoms, patients with SBP scored significantly higher than patients with strict UP (3.1 versus 1.5, p = 0.005), but scored lower than patients with bipolar II disorder (9.0, p < 0.001), as well as patients with bipolar I disorder (6.0, p < 0.001). In addition, there were no significant differences in any of these characteristics between the subgroups of SBP (p > 0.05). Comparing SBP with strict UP and Ghaemi’s BSD

Table 2 describes how the two patient groups and the control group performed in the seven cognitive domains. The MANOVA revealed that six out of seven cognitive domains had significant main effects [Pillai’s F = 8.747, df = (2,467), p < 0.001]. Post-hoc comparisons using the Bonferroni correction showed that patients with SBP and those with strict UP were cognitively impaired when compared to the healthy controls (p < 0.001), but this

was not the case with verbal working memory in patients with SBP (p = 0.201). Compared to patients with strict UP, patients with SBP were better at processing speed (p = 0.002), visual-spatial memory (p = 0.017), and verbal working memory (p = 0.017). MANOVA did not reveal any main effect between SBP and Ghaemi’s BSD groups [Pillai’s F = 0.381, df = (1,112), p = 0.912]. Comparing SBP and bipolar I and bipolar II disorders

Table 3 shows how the patient groups and the control group performed in the seven cognitive domains. Among the three patient groups and the control group, MANOVA revealed that there were main effects for all of the cognitive domains, except for attention [Pillai’s F = 8.234, df = (3,495), p < 0.001]. Post-hoc comparisons showed that bipolar I and bipolar II depression patients had impairments in these domains when compared to the healthy controls (p < 0.05). Patients with SBP performed significantly better than those with bipolar I disorder in set shifting (p < 0.001) and

Table 2. Different performance on neuropsychological tests between patients with soft bipolar spectrum, unipolar depression, and healthy controls

Domains and measures Processing speed Trail Marking Test–Part A WAIS-RC Digit Symbol–Coding subtest Attention Digit Span Forward subtest of the WAIS-RC Set shifting Modified Wisconsin Card Sorting Test Categories completed Total errors Perseverative errors Trail Marking Test–Part B Planning Tower of Hanoi Total scores Average planning time Average executive time Verbal fluency Animal naming Visual-spatial memory Immediate Visual Reproduction subtest of the WMS-RC Verbal working memory Digit Span Backward subtest of the WAIS-RC

SBP (n = 81) Mean (SD)

UP (n = 221) Mean (SD)

HC (n = 200) Mean (SD)

50.0 (20.3) 42.8 (15.4)

58.1 (32.9) 35.5 (15.1)

39.6 (12.0) 55.9 (13.5)

8.2 (1.6)

7.7 (1.4)

8.0 (1.5)

3.6 (2.0) 21.5 (11.9) 7.8 (7.5) 83.4 (41.0)

3.3 (2.0) 22.5 (12.6) 9.9 (11.1) 100.7 (57.9)

MANOVAa F-value

p-value

Post hocb

34.594

0.000

2.038

0.131

19.156

0.000

SBP, UP < HC

30.358

0.000

SBP, UP < HC

11.032

0.000

SBP, UP < HC

30.504

0.000

UP < SBP < HC

3.486

0.031

UP < SBP, HC

UP < SBP < HC

4.6 (1.6) 14.8 (9.4) 4.4 (4.5) 62.4 (21.2)

43.6 (20.0) 7.2 (7.0) 31.4 (17.0)

42.0 (20.1) 7.2 (6.2) 31.3 (14.0)

56.4 (9.2) 5.6 (6.1) 22.1 (7.5)

16.8 (6.4)

15.2 (5.8)

20.2 (5.7)

8.0 (3.0)

6.8 (3.4)

10.3 (2.8)

4.9 (1.7)

4.3 (1.5)

5.3 (1.7)

HC = healthy controls; MANOVA = multivariate analysis of variance; SBP = soft bipolar spectrum; SD = standard deviation; UP = unipolar depression; WAIS-RC = Wechsler Adult Intelligence Scale–Revised by China; WMS-RC = Wechsler Memory Scale–Revised by China. a MANOVA (and post-hoc comparisons) using Z-scores, controlling for years of education, age, and psychotic features. b The threshold for significance was p = 0.05, with Bonferroni correction.

199

Lin et al. Table 3. Different performance on neuropsychological tests between patients with bipolar I disorder, bipolar II disorder, and healthy controls

Domains and measures

BP-I (n = 98) Mean (SD)

BP-II (n = 138) Mean (SD)

SBP (n = 81) Mean (SD)

HC (n = 200) Mean (SD)

Processing speed Trail Marking Test–Part A WAIS-RC Digit Symbol–Coding subtest Attention Digit Span Forward subtest of the WAIS-RC Set shifting Modified Wisconsin Card Sorting Test Categories completed Total errors Perseverative errors Trail Marking Test–Part-B Planning Tower of Hanoi Total scores Average planning time Average executive time Verbal fluency Animal naming Visual-spatial memory Immediate Visual Reproduction subtest of the WMS-RC Verbal working memory Digit Span Backward subtest of the WAIS-RC

55.5 (30.0) 36.8 (15.3)

49.4 (23.4) 42.4 (16.6)

50.0 (20.3) 42.8 (15.4)

39.6 (12.0) 55.9 (13.5)

7.9 (1.4)

7.8 (1.4)

8.2 (1.6)

8.0 (1.5)

2.6 (2.2) 26.2 (12.9) 11.7 (11.9) 115.5 (89.9)

3.4 (2.2) 21.8 (12.9) 8.3 (8.8) 82.9 (35.9)

3.6 (2.0) 21.5 (11.9) 7.8 (7.5) 83.4 (41.0)

MANCOVAa F-value

p-value

Post hocb

28.108

0.000

1.543

0.202

29.098

0.000

BP-I < BP-II, SBP < HC

30.227

0.000

BP-I, BP-II, SBP < HC BP-I < BP-II

11.327

0.000

BP-I, BP-II, SBP < HC

27.176

0.000

BP-I < BP-II, SBP < HC

7.215

0.000

BP-I, BP-II < HC

BP-I, BP-II, SBP < HC BP-I < BP-II

4.6 (1.6) 14.8 (9.4) 4.4 (4.5) 62.4 (21.2)

36.9 (18.3) 7.2 (4.7) 32.8 (13.9)

45.4 (18.0) 6.5 (5.6) 29.2 (12.5)

43.6 (20.0) 7.2 (7.0) 31.4 (17.0)

56.4 (9.2) 5.6 (6.1) 22.1 (7.5)

15.4 (6.0)

17.0 (5.9)

16.8 (6.4)

20.2 (5.7)

6.8 (3.1)

8.6 (3.1)

8.0 (3.0)

10.3 (2.8)

4.3 (1.4)

4.7 (1.7)

4.9 (1.7)

5.3 (1.7)

BP-I = bipolar I disorder; BP-II = bipolar II disorder; HC = healthy controls; MANOVA = multivariate analysis of variance; SBP = soft bipolar spectrum; SD = standard deviation; WAIS-RC = Wechsler Adult Intelligence Scale–Revised by China; WMS-RC = Wechsler Memory Scale–Revised by China. a MANOVA (and post-hoc comparisons) using Z-scores, controlling for years of education, age, and psychotic features. b The threshold for significance was p = 0.05, with Bonferroni correction.

visual-spatial memory (p = 0.042). Compared to those with bipolar II, patients with SBP performed similarly in all of the cognitive domains tested (p > 0.05). Finally, patients with bipolar I disorder were significantly more impaired than those with bipolar II disorder in the domains of processing speed (p = 0.036), set shifting (p < 0.001), planning (p = 0.003), and visual-spatial memory (p < 0.001). Cognitive profile comparisons

To examine the profiles of cognitive functioning between the diagnostic groups, we applied repeatedmeasures MANOVA, with the diagnostic group 9 cognitive domain interaction being the key statistical test. This revealed a significant group 9 domain interaction for the four patient groups [Pillai’s F = 2.231, df = (18,1437), p = 0.002], indicating that the cognitive profiles differed across

200

the groups (Fig. 1). We then excluded the most impaired group (bipolar I disorder) and found that the group 9 domain interaction effect for the remaining three groups (bipolar II disorder, SBP, and strict UP groups) was not significant [Pillai’s F = 1.545, df = (12,784), p = 0.103]. Discussion

To the best of our knowledge, this was the first study to compare cognitive deficits in SBP, strict UP, bipolar I disorder, and bipolar II disorder in the context of major depression. We began by clinically validating SBP, as proposed by Akiskal and Pinto (17), using external validators (e.g., family history of bipolar disorder, age at onset, and recurrent depressive episodes ≥ 3), as well as past hypomanic symptoms. The results suggested that SBP is bipolar in nature, in accordance with previous research (4, 8), thereby supporting the

Cognitive deficits in bipolar spectrum disorders

0.2

Bipolar I Bipolar II

Soft bipolar Unipolar depression

Estimated marginal means

–0.3

–0.8

–1.3

–1.8 Processing speed

Set shifting

Verbal fluency

Verbal working memory

Fig. 1. Neuropsychological profiles for patients with bipolar I disorder, bipolar II disorder, soft bipolar spectrum, and unipolar depression.

conceptualization of SBP that incorporates depression with cyclothymic and hyperthymic traits and antidepressant-associated hypo/mania. This finding reinforces the concern that individuals with UP who manifest bipolarity but do not meet the criteria for bipolar disorder by the DSM system per se may yet be ‘bipolar enough’ to be at risk of adverse reactions to antidepressant treatments, as are individuals with SBP. Unlike bipolar I disorder, which is defined by discrete and full-blown manic episodes, the conceptualization of SBP (and ‘bipolar II spectrum’) is based on an unstable temperamental substrate in the context of which protracted or brief hypomanic episodes occur with alterations in depression (16, 21). Indeed, a growing body of evidence suggests that affective temperaments that have genetic underpinnings (56) play a big role in the relationship of morningness-eveningness to mood fluctuation (57) and in the treatment response (58), and are associated with comorbid conditions such as alcohol dependence (59) and the metabolic syndrome (60). Furthermore, such temperamental attributes as mood labile, energy activity, and day dreaming – characteristics of the CT described by Kretschmer (14) – are commonly seen in bipolar II disorder and may be the predictors of switching to bipolar II from depression (21). The main finding is that, under DSM-IV, a special group of patients with UP, who are redefined as SBP owing to their bipolar nature, perform significantly differently from those with strict UP in several cognitive domains in the context of depression – processing speed, verbal working memory, and visual-spatial memory – whereas their neuropsychological performance is similar to that

of bipolar II disorder. We found one study by Smith et al. (61), comparing patients who were diagnosed with Ghaemi’s BSD to those who were diagnosed with UP, to which our findings are comparable. Smith et al. found that patients with BSD (n = 21) performed significantly worse than those with UP (n = 42) in a TMT-B test and a California Verbal Learning Test that measures verbal learning and memory. However, we found that patients with SBP showed a trend to perform better in comparison to those with strict UP in all of the tested domains, with some domains reaching significant difference. Given the differences in definition between SBP and Ghaemi’s BSD, we directly compared patients with these conditions and found that they each performed similarly in all the neuropsychological tasks. Further studies examining how patients with SBP perform when they clinically recover from depression are needed to fully understand the deficits in SBP and their relationships with mood states. In the latest version of DSM–DSM-5, the criteria for bipolar disorder are similar to those in DSM-IV, which implies that a substantial proportion of individuals with SBP may be still under the umbrella of UP. For future neuropsychological studies, however, our data suggest a need to differentiate SBP from strict UP in order to improve the validity of studied samples. In addition, our study showed, for the first time, that the profiles of cognitive functioning during a MDE may be different for bipolar I disorder and ‘bipolar II spectrum’ – subsuming bipolar II disorder and SBP (Fig. 1). In fact, family data (62–64) have shown that the relatives of bipolar II probands are more likely to have a bipolar II disorder than a bipolar I disorder diagnosis, and that the prevalence of mood disorders (bipolar and unipolar) in the relatives of bipolar II probands is higher than that in the relatives of bipolar I probands. Furthermore, some neuroimaging studies separating bipolar subtypes demonstrated neurobiological differences between bipolar I and bipolar II disorders (65, 66). Li and colleagues (65) observed a differential correlation of regional glucose uptakes and executive performance between bipolar I and bipolar II disorder, and concluded that the anterior cingulate cortex and caudate but not the prefrontal cortex (PFC) played a role in executive functioning in bipolar I disorder, as opposed to the critical role of PFC in bipolar II disorder. Finally, a recent study by Lu et al. (67) found that traits of novelty seeking and harm avoidance – both traits are positively associated with the CT and the former characterizes the hyperthymic temperament (68) – modified differently the effects of the serotonin transporter gene-linked

201

Lin et al. polymorphic region gene (5-HTTLPR) on bipolar I and bipolar II disorders, suggesting that the mechanism of gene–temperament interaction might assist in distinguishing bipolar II and bipolar I disorders. Based on the finding of distinct cognitive profiles across the diagnostic groups, our data support the notion that the soft forms of the disorder (e.g., bipolar II spectrum) may be distinct from the classical manic–depressive illness. It is nonetheless possible that for such domains as processing speed and visual-spatial memory, which were less impaired in SBP but more impaired in bipolar I disorder, bipolarity (or bipolar tendency) may serve as a protective factor at the softer end of the bipolar spectrum but as the disease severity progresses, the illness overwhelms the protective effects on these domains. As mentioned in the introduction, deficits in some domains, such as verbal fluency and verbal learning memory, were found in the unaffected relatives of bipolar disorder, indicating that they have the potential to be endophenotypes for bipolar disorder. Further studies are needed to investigate whether, or to what extent, the unaffected relatives of SBP patients perform differently to those of bipolar disorder (bipolar I and bipolar II) as well as UP patients in terms of cognitive functioning. Some limitations of the present study should be noted. The gender ratio and ages of the patients were not comparable between the groups. However, these factors are unlikely to explain our findings. First, we controlled for age in the analysis. Second, we observed that patients with bipolar I disorder were younger but performed worse than patients with strict UP, and age is believed to be negatively associated with cognitive performance. Third, there is little evidence of a difference in neuropsychological performance between genders. There were several strengths to the study. As all of the patients were either newly diagnosed cases or had discontinued psychiatric medications for more than two weeks, the deleterious effects of medication were minimized. Moreover, the diagnosis was validated longitudinally and prospectively, and the sample sizes were relatively large. Finally, cognitive performance was examined in the context of a MDE this could have minimized the influences of different mood states on neuropsychological performance, which might have been a confounding factor for many of the previous studies. In summary, SBP differs from strict UP not only in the external clinical validators and hypomanic signs/symptoms, but also in cognitive deficits during major depression. Our data provide evidence for the growing sense that a spectrum model of mood disorders more closely reflects the observed

202

phenomena than a dichotomous model such as the DSM. In accordance with the distinctions between bipolar I disorder and Akiskal’s ‘bipolar II spectrum’, which subsumes bipolar II and SBP, the profiles of cognitive functioning between them are differential, supporting the notion that they may have a distinct neurobiological basis. We suggest that a multiple approach (including temperament and cognitive aspects) to characterizing major mood disorders may assist in improving the validity of diagnosis. Acknowledgements Funding was provided by (i) Guangzhou Scientific and Technological Bureau, Guangzhou, China; (ii) Guangdong Provincial Department of Science and Technology; (iii) The National Key Clinical Medical Specialties of Ministry of Health, China; (iv) The May Endowed Professorship of The University of Hong Kong; and (v) The Research Grant Council Humanities and Social Sciences Prestigious Fellowship (Ref: HKU703HSS-13).

Disclosures The authors of this paper do not have any commercial associations that might pose a conflict of interest in connection with this manuscript.

References 1. Merikangas KR, Akiskal HS, Angst J et al. Lifetime and 12-month prevalence of bipolar spectrum disorder in the National Comorbidity Survey replication. Arch Gen Psychiatry 2007; 64: 543–552. 2. Weissman MM, Bland RC, Canino GJ et al. Crossnational epidemiology of major depression and bipolar disorder. JAMA 1996; 276: 293–299. 3. Pini S, de Queiroz V, Pagnin D et al. Prevalence and burden of bipolar disorders in European countries. Eur Neuropsychopharmacol 2005; 15: 425–434. 4. Akiskal HS, Bourgeois ML, Angst J et al. Re-evaluating the prevalence of and diagnostic composition within the broad clinical spectrum of bipolar disorders. J Affect Disord 2000; 59 (Suppl. 1): S5–S30. 5. Angst J, Gamma A, Benazzi F et al. Toward a re-definition of subthreshold bipolarity: epidemiology and proposed criteria for bipolar-II, minor bipolar disorders and hypomania. J Affect Disord 2003; 73: 133–146. 6. Benazzi F, Akiskal HS. Refining the evaluation of bipolar II: beyond the strict SCID-CV guidelines for hypomania. J Affect Disord 2003; 73: 33–38. 7. Angst J, Cassano G. The mood spectrum: improving the diagnosis of bipolar disorder. Bipolar Disord 2005; 7 (Suppl. 4): 4–12. 8. Akiskal HS, Akiskal KK, Lancrenon S et al. Validating the bipolar spectrum in the French National EPIDEP Study: overview of the phenomenology and relative prevalence of its clinical prototypes. J Affect Disord 2006; 96: 197–205. 9. Zimmermann P, Bruckl T, Nocon A et al. Heterogeneity of DSM-IV major depressive disorder as a consequence of

Cognitive deficits in bipolar spectrum disorders

10.

11. 12.

13.

14. 15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

subthreshold bipolarity. Arch Gen Psychiatry 2009; 66: 1341–1352. Angst J, Cui L, Swendsen J et al. Major depressive disorder with subthreshold bipolarity in the National Comorbidity Survey Replication. Am J Psychiatry 2010; 167: 1194–1201. Nusslock R, Frank E. Subthreshold bipolarity: diagnostic issues and challenges. Bipolar Disord 2011; 13: 587–603. Phelps J, Angst J, Katzow J, Sadler J. Validity and utility of bipolar spectrum models. Bipolar Disord 2008; 10: 179– 193. Rihmer Z, Pestality P. Bipolar II disorder and suicidal behavior. Psychiatr Clin North Am 1999; 22: 667–673, ix–x. Kretschmer E. Physique and Character. London: Kegan Paul, Trench Trubner and Co., Ltd., 1936. Akiskal HS. Toward a temperament-based approach to depression: implications for neurobiologic research. Adv Biochem Psychopharmacol 1995; 49: 99–112. Perugi G, Akiskal HS. The soft bipolar spectrum redefined: focus on the cyclothymic, anxious-sensitive, impulsedyscontrol, and binge-eating connection in bipolar II and related conditions. Psychiatr Clin North Am 2002; 25: 713–737. Akiskal HS, Pinto O. The evolving bipolar spectrum. Prototypes I, II, III, and IV. Psychiatr Clin North Am 1999; 22: 517–534, vii. Akiskal HS, Akiskal KK, Lancrenon S et al. Validating the soft bipolar spectrum in the French National EPIDEP Study: the prominence of BP-II 1/2. J Affect Disord 2006; 96: 207–213. Hantouche EG, Akiskal HS, Lancrenon S et al. Systematic clinical methodology for validating bipolar-II disorder: data in mid-stream from a French national multi-site study (EPIDEP). J Affect Disord 1998; 50: 163– 173. Benazzi F. Validating Angst’s “ups & downs” personality trait as a new marker of bipolar II disorder. Eur Arch Psychiatry Clin Neurosci 2004; 254: 48–54. Akiskal HS, Maser JD, Zeller PJ et al. Switching from ‘unipolar’ to bipolar II. An 11-year prospective study of clinical and temperamental predictors in 559 patients. Arch Gen Psychiatry 1995; 52: 114–123. Akiskal HS, Walker P, Puzantian VR et al. Bipolar outcome in the course of depressive illness. Phenomenologic, familial, and pharmacologic predictors. J Affect Disord 1983; 5: 115–128. Cassano GB, Akiskal HS, Savino M et al. Proposed subtypes of bipolar II and related disorders: with hypomanic episodes (or cyclothymia) and with hyperthymic temperament. J Affect Disord 1992; 26: 127–140. Ghaemi SN, Ko JY, Goodwin FK. “Cade’s disease” and beyond: misdiagnosis, antidepressant use, and a proposed definition for bipolar spectrum disorder. Can J Psychiatry 2002; 47: 125–134. Lee RS, Hermens DF, Porter MA, Redoblado-Hodge MA. A meta-analysis of cognitive deficits in first-episode Major Depressive Disorder. J Affect Disord 2012; 140: 113–124. Reppermund S, Ising M, Lucae S et al. Cognitive impairment in unipolar depression is persistent and non-specific: further evidence for the final common pathway disorder hypothesis. Psychol Med 2009; 39: 603–614. Castaneda AE, Suvisaari J, Marttunen M et al. Cognitive functioning in a population-based sample of young adults with a history of non-psychotic unipolar depressive disorders without psychiatric comorbidity. J Affect Disord 2008; 110: 36–45.

28. Snyder HR. Major depressive disorder is associated with broad impairments on neuropsychological measures of executive function: a meta-analysis and review. Psychol Bull 2013; 139: 81–132. 29. Xu G, Lin K, Rao D et al. Neuropsychological performance in bipolar I, bipolar II and unipolar depression patients: a longitudinal, naturalistic study. J Affect Disord 2012; 136: 328–339. 30. Martinez-Aran A, Vieta E, Reinares M et al. Cognitive function across manic or hypomanic, depressed, and euthymic states in bipolar disorder. Am J Psychiatry 2004; 161: 262–270. 31. Bearden CE, Shih VH, Green MF et al. The impact of neurocognitive impairment on occupational recovery of clinically stable patients with bipolar disorder: a prospective study. Bipolar Disord 2011; 13: 323–333. 32. McKenna BS, Sutherland AN, Legenkaya AP et al. Abnormalities of brain response during encoding into verbal working memory among euthymic patients with bipolar disorder. Bipolar Disord 2014; 16: 289–299. 33. Levy B, Weiss RD. Neurocognitive impairment and psychosis in bipolar I disorder during early remission from an acute episode of mood disturbance. J Clin Psychiatry 2010; 71: 201–206. 34. Malhi GS, Ivanovski B, Hadzi-Pavlovic D, Mitchell PB, Vieta E, Sachdev P. Neuropsychological deficits and functional impairment in bipolar depression, hypomania and euthymia. Bipolar Disord 2007; 9: 114–125. 35. Drysdale E, Knight HM, McIntosh AM, Blackwood DH. Cognitive endophenotypes in a family with bipolar disorder with a risk locus on chromosome 4. Bipolar Disord 2013; 15: 215–222. 36. Kulkarni S, Jain S, Reddy YCJ, Kumar KJ, Kandavel T. Impairment of verbal learning and memory and executive function in unaffected siblings of probands with bipolar disorder. Bipolar Disord 2010; 12: 647–656. 37. Zhang L, Ning Y. Guangzhou psychiatric hospital: the oldest psychiatric hospital in china. Psychiatry (Edgmont) 2010; 7: 53–54. 38. Wang Y, Jia Y, Xu G et al. Frontal white matter biochemical abnormalities in first-episode, treatment-naive patients with major depressive disorder: a proton magnetic resonance spectroscopy study. J Affect Disord 2012; 136: 620– 626. 39. Nassir Ghaemi S, Miller CJ, Berv DA et al. Sensitivity and specificity of a new bipolar spectrum diagnostic scale. J Affect Disord 2005; 84: 273–277. 40. Hamilton M. Development of a rating scale for primary depressive illness. Br J Soc Clin Psychol 1967; 6: 278–296. 41. Young RC, Biggs JT, Ziegler VE et al. A rating scale for mania: reliability, validity and sensitivity. Br J Psychiatry 1978; 133: 429–435. 42. Hamilton M. The assessment of anxiety states by rating. Br J Med Psychol 1959; 32: 50–55. 43. Overall JE, Gorham DR. The Brief Psychiatric Rating Scale. Psychol Rep 1962; 10: 799–812. 44. Akiskal HS, Hantouche EG, Allilaire JF. Bipolar II with and without cyclothymic temperament: “dark” and “sunny” expressions of soft bipolarity. J Affect Disord 2003; 73: 49–57. 45. Akiskal HS, Placidi GF, Maremmani I et al. TEMPS-I: delineating the most discriminant traits of the cyclothymic, depressive, hyperthymic and irritable temperaments in a nonpatient population. J Affect Disord 1998; 51: 7–19. 46. Akiskal HS, Akiskal K, Allilaire JF et al. Validating affective temperaments in their subaffective and socially

203

Lin et al.

47.

48. 49.

50. 51. 52.

53.

54.

55. 56.

57.

58.

positive attributes: psychometric, clinical and familial data from a French national study. J Affect Disord 2005; 85: 29–36. Lin K, Xu G, Miao G et al. Psychometric properties of the Chinese (Mandarin) TEMPS-A: a population study of 985 non-clinical subjects in China. J Affect Disord 2013; 147: 29–33. Wang X, Wang X, Ma H. Symptom Checklist 90. Chin Ment Health J 1993; 7: 31–36. Lezak MD, Howeison DB, Loring DW. Neuropsychological Assessment, 4th edn. New York, NY: Oxford University Press, 2004. Gong Y. Wechsler Adult Intelligence Scale-Revised in China. Hunan: Hunan Medical School Press, 1982. Nelson HE. A modified card sorting test sensitive to frontal lobe defects. Cortex 1976; 12: 313–324. Lu L, Bigler ED. Performance on original and a Chinese version of Trail Making Test Part B: a normative bilingual sample. Appl Neuropsychol 2000; 7: 243–246. Welsh MC, Satterlee-Cartmell T, Stine M. Towers of Hanoi and London: contribution of working memory and inhibition to performance. Brain Cogn 1999; 41: 231–242. Chen YL, Chen YH, Lieh-Mak F. Semantic verbal fluency deficit as a familial trait marker in schizophrenia. Psychiatry Res 2000; 2: 133–148. Gong Y. Wechsler Memory Scale-Revised in China. Hunan: Hunan Medical School Press, 1987. Greenwood TA, Akiskal HS, Akiskal KK et al. Genomewide association study of temperament in bipolar disorder reveals significant associations with three novel loci. Biol Psychiatry 2012; 72: 303–310. Jankowski KS. The role of temperament in the relationship between morningness-eveningness and mood. Chronobiol Int 2014; 31: 114–122. Rybakowski JK, Dembinska D, Kliwicki S et al. TEMPSA and long-term lithium response: positive correlation

204

59.

60.

61.

62. 63. 64.

65.

66.

67.

68.

with hyperthymic temperament. J Affect Disord 2013; 145: 187–189. Pombo S, Figueira ML, da Costa NF et al. The burden of cyclothymia on alcohol dependence. J Affect Disord 2013; 151: 1090–1096. Altinbas K, Guloksuz S, Oral ET. Metabolic syndrome prevalence in different affective temperament profiles in bipolar-I disorder. Rev Bras Psiquiatr 2013; 35: 131–135. Smith DJ, Muir WJ, Blackwood DH. Neurocognitive impairment in euthymic young adults with bipolar spectrum disorder and recurrent major depressive disorder. Bipolar Disord 2006; 8: 40–46. Coryell W, Endicott J, Reich T et al. A family study of bipolar II disorder. Br J Psychiatry 1984; 145: 49–54. Heun R, Maier W. Bipolar II disorders in six first-degree relatives. Biol Psychiatry 1993; 34: 274–276. Gershon ES, Hamovit J, Guroff JJ et al. A family study of schizoaffective, bipolar I, bipolar II, unipolar, and normal control probands. Arch Gen Psychiatry 1982; 39: 1157– 1167. Li CT, Hsieh JC, Wang SJ et al. Differential relations between fronto-limbic metabolism and executive function in patients with remitted bipolar I and bipolar II disorder. Bipolar Disord 2012; 14: 831–842. Ketter TA, Kimbrell TA, George MS et al. Effects of mood and subtype on cerebral glucose metabolism in treatment-resistant bipolar disorder. Biol Psychiatry 2001; 49: 97–109. Lu YA, Lee SY, Chen SL et al. Gene-temperament interactions might distinguish between bipolar I and bipolar II disorders: a cross-sectional survey of Han Chinese in Taiwan. J Clin Psychiatry 2012; 73: 339–345. Akiskal HS, Mendlowicz MV, Jean-Louis G et al. TEMPS-A: validation of a short version of a self-rated instrument designed to measure variations in temperament. J Affect Disord 2005; 85: 45–52.