Evolution of cognitive impairment in bipolar disorder: A systematic review of cross-sectional evidence
The notion that sufferers of bipolar disorder achieve complete syndromal and functional recovery between illness episodes has been brought into question by evidence that a large proportion of patients fail to regain premorbid levels of functioning after the resolution of major affective symptoms. A growing body of evidence suggests that bipolar patients exhibit neuropsychological impairment that persists even during the euthymic state, which may be a contributory factor to poor psychosocial outcome. However, the aetiology of such impairment and its relation to progression of illness are not well understood. This review aims to consider evidence from studies investigating both the relationship between cognitive impairment and clinical outcome and studies of neurocognitive function in unaffected first-degree relatives (FDRs) of bipolar sufferers to address issues of the temporal evolution of cognitive impairment in bipolar disorder. Systematic literature review. The weight of evidence suggests that greater neuropsychological dysfunction in bipolar disorder is associated with a worse prior course of illness, particularly the number of manic episodes, hospitalizations and length of illness. The most consistent finding was a negative relationship between the number of manic episodes and verbal declarative memory performance. Impairment in unaffected FDRs was reported in verbal declarative memory and some facets of executive function. Cognitive impairment may be a trait vulnerability factor for bipolar disorder that is present before illness onset and worsens as the illness progresses. Further investigation into the causal relationship between cognitive impairment and illness course is essential.
Evolution of cognitive impairment in bipolar
disorder: a systematic review of cross-sectional
Cognitive impairment has been documented in a
variety of neuropsychological domains during the
major mood disturbances associated with the acute
episodes of bipolar disorder (1–4). However, many
studies have shown that suﬀerers of bipolar disor-
der continue to exhibit cognitive impairment after
the resolution of major mood symptoms (5–20).
The signiﬁcance of these cognitive deﬁcits is
unknown, although it has been suggested that they
may impact negatively on social functioning and be
at least partially responsible for the poor rate of
inter-episode functional recovery seen in a high
proportion of patients (21–23).
The aetiology of cognitive impairment in bipolar
disorder is not understood. Exploring the poten-
tially complex interaction between neuropsy-
chological impairment and aﬀective symptoms
Robinson LJ, Ferrier IN. Evolution of cognitive impairment in bipolar
disorder: a systematic review of cross-sectional evidence.
Bipolar Disord 2006: 8: 103–116. ª Blackwell Munksgaard, 2006
Objectives: The n otion that suﬀerers of bipolar disorder achieve
complete syndromal and functional recovery between illness episodes has
been brought into question by evidence that a large proportion of
patients fail to regain premorbid levels of functioning after the resolution
of major aﬀective symptoms. A growing body of evidence suggests that
bipolar patients exhibit neuropsychological impairment that persists
even during the euthymic state, which may be a contributory factor to
poor psychosocial outcome. However, the aetiology of such impairment
and its relation to progression of illness are not well understood. This
review aims to consider evide nce from studies investigating both the
relationship between cognitive impairment and clinical outcome and
studies of neurocognitive function in unaﬀected ﬁrst-degree relatives
(FDRs) of bipolar suﬀerers to address issues of the temporal evolution
of cognitive impairment in bipolar disorder.
Methods: Systematic literature review.
Results: The weight of evidence suggests that greater neuropsychological
dysfunction in bipolar disorder is associated with a worse prior course
of illness, particularly the number of manic episodes, hospitalizations
and length of illness. The most consistent ﬁnding was a negative
relationship between the number of manic episodes and verbal
declarative memory performance. Impairment in unaﬀected FDRs was
reported in verbal declarative memory and some facets of executive
Conclusions: Cognitive impairment may be a trait vulnerability factor
for bipolar disorder that is present before illness onset and worsens as the
illness progresses. Further investigation into the causal relationship
between cognitive impairment and illness course is essential.
Lucy J Robinson and I Nicol Ferrier
School of Neurology, Neurobiology & Psychiatry
(Psychiatry), University of Newcastle upon Tyne,
Newcastle upon Tyne, UK
Key words: aetiology – bipolar disorder – ﬁrst-
degree relatives – neurocognitive impairment –
Received 28 February 2005, revised and accepted
for publication 10 October 2005
Corresponding author: Professor I Nicol Ferrier,
School of Neurology, Neurobiology & Psychiatry
(Psychiatry), University of Newcastle upon Tyne,
Leazes Wing, Royal Victoria Inﬁrmary, Queen
Victoria Road, Newcastle upon Tyne, NE1 4LP, UK.
The authors of this paper do not have any commercial associations
that might pose a conﬂict of interest in connection with this manu-
Bipolar Disorders 2006: 8: 103–116
Copyright ª Blackwell Munksgaard 2006
requires a better understanding of the temporal
evolution of cognitive deﬁcits. The ﬁrst part of this
review evaluates the ﬁndings of studies which
investigated whether course of illness variables
(e.g. number of aﬀective episodes, age at illness
onset, duration of illness) are associated with
cognitive impairment in bipolar patients. However,
this approach fails to ascertain whether cognitive
impairment precedes illness onset. Furthermore,
the direction of causality is unclear – mood
episodes may adversely aﬀect neuropsychological
performance, or impaired cognitive function may
contribute to the development and exacerbation of
mood symptoms. Similarly, cognitive dysfunction
may be an expression of the disease phenotype and
not necessarily bear any direct relation to mood
symptoms. Investigations into the neurocognitive
function of bipolar patients are also confounded
by several factors such as the use of medication, the
presence of sub-syndromal symptoms (6, 12), and
the possibility that patients have suﬀered lasting
neuroanatomical changes due to prior episodes (24,
25). Investigating cognitive function in healthy
individuals at high risk of developing bipolar
disorder avoids many of these confounds and
sheds light on whether this aspect of the disease
phenotype is shared by those at genetic risk for the
disorder. First-degree relatives (FDRs) of bipolar
suﬀerers are a population at especially high risk of
themselves developing a mood disorder – oﬀspring
of a bipolar parent are four times more likely to
develop an aﬀective disorder than oﬀspring of
healthy parents (26) and FDRs face a 10- to 20-
fold increase in risk for bipolar disorder (24).
Results from studies investigating the neuropsy-
chological performance of FDRs of bipolar
patients are therefore able to address whether
cognitive impairment is a trait-like feature of the
disorder independent of symptoms. The second
part of the review focuses on studies of cognitive
function in FDRs of bipolar probands.
To minimize any biases due to aﬀective symp-
toms, this review focuses on only those studies that
included a well-characterized sample of euthymic
bipolar patients or a FDR group determined to be
free of personal history of bipolar illness.
To identify studies relating neuropsychological
performance in euthymic bipolar patients to
illness characteristics, the title, abstract and key
word ﬁelds of three major bibliographic data-
bases (Medline, EMBASE and PsychINFO) were
searched using the terms Ôbipolar disorderÕ or the
wild card Ômanic depress*Õ along with each of the
search terms Ôneuropsych*Õ, Ôneurocognit*Õ and
Ôcognit*Õ. To identify studies investigating neuro-
psychological function in FDRs of bipolar pro-
bands, the same search strategy was used plus
the terms ÔfamilyÕ, ÔfamilialÕ, Ôﬁrst degree relativeÕ,
Ôhigh riskÕ and ÔprobandÕ. Each of the databases
was searched on the same day and the searches
were limited to English language articles avail-
able from 1st January 1980 to 31st January 2005.
Across all three databases, the strategy yielded
approximately 1350 references. Abstracts and
titles were used to determine whether the refer-
ence might be relevant to the review and full
texts of potentially relevant articles were retrieved
to assess the article for inclusion. The reference
lists of retrieved articles were checked for any
further relevant citations.
To be included as a study relating neuropsycho-
logical performance to illness characteristics the
following criteria had to be met: (i) the study
included patients diagnosed with bipolar disorder
selected using a well-established criterion-based
diagnostic system; (ii) the study included a distinct
group of bipolar patients who were all clinically
euthymic at the time of testing; (iii) the study
included a measure of neuropsychological function
(studies only employing very general measures of
cognitive function, such as the Mini-Mental State
Examination, were excluded); (iv) where studies
included more than one patient group or included
symptomatic bipolar patients, it had to be clear
that analyses between neurocognitive performance
and illness features were performed with the
euthymic bipolar patients as a distinct group.
Where this was unclear (n ¼ 4) authors were
contacted for clariﬁcation. If no reply was received
the study was excluded (n ¼ 2).
To be included as a study investigating neuro-
psychological performance in FDRs of bipolar
patients the following criteria had to be met: (i) the
study included a comparison group of individuals
with no family history of bipolar disorder; (ii)
relatives of a bipolar proband were included as a
distinct group; (iii) the study only included relatives
who did not show signs of bipolar or psychotic
In total, 150 full references were retrieved, from
which 26 were determined to meet the inclusion
criteria (14 cross-sectional studies of illness history
and neuropsychological function, 11 FDR studies,
and 1 which looked at both).
Robinson and Ferrier
Neuropsychological function and course of bipolar
Table 1 provides the sample characteristics of the
included studies and Table 2 reports the relation-
ships between illness features and neuropsycho-
logical measures detailed in each study. Table 3
reports the speciﬁc tests used by each study. The
ﬁndings related to each of the illness features most
frequently included in the studies – number of
aﬀective episodes, number of hospitalizations, age
at illness onset, duration of illness, and length of
time euthymic before neuropsychological testing –
are reported in turn.
Aﬀective episodes. Thirteen studies correlated the
number of aﬀective episodes with various measures
of neurocognitive function – ten of which tested the
impact of depressive and manic episodes sepa-
Manic episodes. Seven of the ten studies reported
signiﬁcant negative relationships between lifetime
number of manic episodes and cognitive function.
Examination of the pattern of relationships
reveals that manic episodes related consistently
to impairment on tests of verbal memory and
executive function. Five of six studies that includ-
ed a measure of verbal declarative memory
reported at least one signiﬁcant correlation
between number of manic episodes and poorer
performance (5, 12, 13, 15, 16). Speciﬁcally, the
most common ﬁnding (reported in four of these
ﬁve studies) was a relationship between manic
episodes and poorer delayed verbal memory
(recall or recognition of a word list after 30 min
or more). Although initial learning (encoding) is
generally more impaired in euthymic bipolar
patients than retention, the experience of manic
episodes appears to be associated with poorer
retention. With regard to executive function, both
studies including a test of concept formation
reported a relationship between poorer perfor-
mance and a greater number of manic episodes
(5, 11). Clark et al. (12) found that sustained
attention performance suﬀered with increasing
burden of manic episodes. Thompson et al. (18)
reported that performance on a mental manipu-
lation task related positively to the number of
manic episodes suﬀered and was the only study to
report signiﬁcant correlations in this unexpected
direction. Other measures of executive function
(such as executive shifting, mental planning and
mental manipulation) showed no relationship with
One study reported a relationship between manic
episodes and impairment on a visual memory task
(17), although this contrasts with two studies that
found no signiﬁcant relationships between manic
episodes and performance on tests of visual or
spatial memory (8, 10).
Depressive episodes. Six of ten studies reported
signiﬁcant negative relationships between lifetime
number of depressive episodes and various meas-
ures of neuropsychological function. Correlations
were generally smaller than those found for manic
episodes (Table 2).
The pattern of signiﬁcant ﬁndings showed
relationships between depressive episodes and
performance on tests of executive function
(11, 12, 18), verbal learning (12, 15), visual
memory (10, 17) and spatial working memory
(12). One small study failed to report signiﬁcant
correlations with delayed verbal recall, but the
correlation coeﬃcients were of similar magnitude
to studies reporting signiﬁcant relationships (5).
However, three studies failed to ﬁnd a relation-
ship with verbal learning performance (13, 16,
18), one reported relatively large non-signiﬁcant
positive relationships (13), and one study reported
no association with deﬁcits of executive function
Total episodes. El-Badri et al. (9) found that total
number of episodes was negatively related to
executive function and visual memory perfor-
mance. In contrast, Krabbendam et al. (7) and
et al. (20) found no signiﬁcant associa-
tions between total episodes and any neuropsycho-
logical measure. Ferrier et al. (6) found no
diﬀerences in the neuropsychological performance
of good and poor outcome patients (good outcome
patients had suﬀered no more than two major
mood episodes in the preceding 5 years, whereas
poor outcome patients had suﬀered at least three
major mood episodes in the last 2 years).
Number of hospitalizations. Five of six studies
reported at least one signiﬁcant negative associ-
ation between hospitalizations and cognitive func-
tion. However, there is no clear pattern to the
signiﬁcant results due to diﬀerences in both the
measure of hospitalizations used and the neuro-
psychological domains assessed. Rubinsztein et al.
(8) reported a signiﬁcant association between the
number of months hospitalized and visual memory
n et al. (16) found sig-
niﬁcant relationships between the total number
of admissions and poorer verbal memory
performance. Thompson et al. (18) documented
Evolution of cognitive impairment in BD
Table 1. Summary of bipolar patient study samples and euthymic criteria
(% male) Age
episodes Euthymic criteria
Van Gorp et al. (5) 13 NR 100 51.84 (13.4) 22.31 (10.1) 27.15 (11.77) 4.83 (4.75) 7.75 (8.62) HRSD < 7 & YMRS < 6 at 3 consecutive monthly
Ferrier et al. (6) 21 21 BDI 23.8 45.7 (10.0) NR NR NR NR HRSD £ 8 & MSS-6 < 20 3
Krabbendam et al. (7) 22 12 BDI, 10 BDII 22.7 47.7 (8.3) NR 32.2 (9.4) 3.9 (3.7) 6.2 (5.1) DSM-IV criteria for full remission were met 4
Rubinsztein et al. (8) 18 18 BDI NR 42 (2) NR 24.7 (NR) 4.5 (NR) 3.7 (NR) HRSD < 8 & YMRS < 8 for 4 months 1
El Badri et al. (9) 25
25 BDI 34.5 30.7 (6.1) 9.3 (5.7) 20.9 (4.2) 4.35
(1.84) ÔClinical examination used to exclude hypomania
…Mood stable and living in the community for
at least 8 weeksÕ
MacQueen et al. (10) 28 28 BDI 46.4 42.2 (12.3) NR NR 8.1 (3.7) 9.8 (5.9) HRSD £ 7 & YMRS £ 5 for 1 month 1
Zubieta et al. (11) 15 15 BDI 60 39 (13) 13 (8) 27 (9)
6.7 (6.5) 4.1 (5.0) HRSD-21 < 6 & YMRS < 4 for at least 6 months 1
Clark et al. (12) 30 30 BDI 56.7 35.9 (11.0) NR NR NR NR HRSD < 8 & YMRS < 8 2
Cavanagh et al. (13) 20 20 BDI 50 43.6 (14.2) 16 (12.5) NR 6 (7) 6 (6) HRSD-21 £ 7 & MMS £ 21
Fleck et al. (14) 14 14 BDI 42.9 31 (9) NR 24 (9) 1.3 (0.6) 1.0 (1.1) HRSD < 10 & YMRS < 10 & SAPS < 2 2
Deckersbach et al. (15) 30 30 BDI 40 37.3 (14.1) 21.0 (14.0) 16.2 (4.5) 7.3 (6.3) 12.8 (7.1) Patients described as euthymic 1
Deckersbach et al. (17) 25 25 BDI 40 39.9 (13.3) 22.6 (13.4) 16.9 (5.1) 4.8 (4.8) 10.8 (7.9) DSM-IV criteria for full remission were met 1
n et al. (16) 40 NR 42.5 38.5 (10.1) 15.0 (8.4) 23.6 (7.1) NR NR HRSD-17 £ 8 & YMRS £ 6 for 6 months 1
Thompson et al. (18) 63 54 BDI, 9 BDII 61.7 44.4 (8.6) 19.5 (10.0) 25.3 (7.2) 10.3 (17.4) 12.0 (16.4) HRSD < 7 & YMRS < 7 for 1 month
Clark et al. (19) 15 14 BDI, 1 BDII 20 37.8 (14.6) 13.3 (9.3) NR NR NR HRSD < 9 & YMRS < 9 1
et al. (20) 26 26 BDI 42.3 44.2 (1.6) 23.6 (7.8) NR NR NR DSM-IV criteria for full remission were met 3
NR ¼ value not reported.
Values represent mean (SD).
HRSD ¼ Hamilton Rating Scale for Depression; YMRS ¼ Young Mania Rating Scale; MSS ¼ Manic State Scale; MMS ¼ Modiﬁed Manic State rating scale; SAPS ¼ Schedule for the
Assessment of Positive Symptoms.
1 ¼ analyses involved only neuropsychological tests on which bipolar patients demonstrated impairment relative to healthy comparison subjects; 2 ¼ analyses included some neuro-
psychological indices not showing signiﬁcant differences between patients and controls; 3 ¼ analyses involved comparisons of neuropsychological performance between groups of patients
differing on illness features; 4 ¼ strategy unclear.
Only 25 patients completed neuropsychological testing but statistics are provided for 29 patients.
Value reported is the mean age at onset of mania. Mean age at onset of depression ¼ 27 (12) years.
Robinson and Ferrier
Table 2. The range of correlation coefﬁcients and the number of signiﬁcant relationships reported between clinical variables and any measure of neuropsychological function
Study Manic episodes Depressed episodes
Age at onset Illness duration Time euthymic Strategy
Total Manic Depressed
Van Gorp et al. (5) )0.01 to )0.8 (3/12)** +0.24 to )0.43 (0/12) – – – – – – 2
Krabbendam et al. (7) ns (0/11?) – – – – – – 3
Rubinsztein et al. (8) ns (0/3) ns (0/3) )0.5 (1/3)*
––– – – 1
El-Badri et al. (9) )0.49 to )0.66 (4/11?)***
– – – ns (0/11?) ns (0/11?) ns (0/11) 3
Zubieta et al. (11) +0.18 to )0.6 (2/6)* ) 0.03 to )0.55 (2/6)* – ) 0.03 to
+0.49 to )0.41 (0/12) ns (0/6) – 1
MacQueen et al. (10) ns (0/24) )0.43 to )0.62 (6/24) – – – – – ns (0/24) 1
Clark et al. (12) ) 0.09 to )0.48 (2/6)* )0.12 to )0.48 (4/6)* +0.01 to )0.22 (0/6) )0.06 to
– )0.17 to ) 0.53 (3/6)** )0.28 to
Cavanagh et al. (13) )0.5 to )0.9 (4/5)***
+0.5 to +0.2 (0/5)
+0.2 to )0.1 (0/5) – – – +0.01 to )0.5 (1/5)* – 1
Fleck et al. (14) – – – – – – ns (0/7) – 2
Deckersbach et al. (15) )0.56 to )0.57 (2/2)*** )0.51 to )0.63 (2/2)*** – – – )0.32 to )0.34 (0/2) )0.04 to )0.57 (1/2)*** – 1
Deckersbach et al. (17) )0.42 to )0.5 (2/2)* )0.44 to )0.52 (2/2)** – – – )0.09 to )0.26 (0/2) )0.16 to )0.28 (0/2) – 1
n et al. (16) +0.03 to )0.35 (2/9)* ns (0/9) +0.07 to )0.38 (3/9)* – – +0.12 to )0.47 (2/9)** +0.19 to )0.44 (5/9)** – 1
Thompson et al. (18) +0.35 to )0.17 (1/15)** +0.15 to )0.30 (1/15)* +0.1 to )0.48 (6/15)*** – – +0.15 to )0.17 (0/15) +0.05 to )0.38 (5/15)** +0.23 to
Clark et al. (19) – – – – – – )0.09 to )0.1 (0/2) – 1
et al. (20) ns (0/16?)
– – – – ns (0/16?) – 3
All signs have been transformed where necessary so that a negative correlation indicates a negative relationship between that illness variable and cognitive function.
Bold values are statistically signiﬁcant. When both values in the range are bold, this indicates the study reported r-values for signiﬁcant correlations only.
ns ¼ no signiﬁcant relationships (p > 0.05) when r-values were not reported.
Stars indicate the signiﬁcance level of the most signiﬁcant relationship. In the case where more than one signiﬁcant relationship was reported, they were not necessarily all signiﬁcant to the same level.
*p < 0.05; **p < 0.01; ***p < 0.001.
Values in parentheses: (the number of statistically signiﬁcant relationships/total number of relationships tested); ? ¼ a value was unclear or not reported.
1 ¼ analyses involved only neuropsychological tests on which bipolar patients demonstrated impairment relative to healthy comparison subjects; 2 ¼ analyses included some neuropsychological indices not showing signiﬁcant
differences between patients and controls; 3 ¼ strategy unclear.
Total months hospitalized.
Total number of episodes.
Correlations controlled for level of depressive symptoms.
Evolution of cognitive impairment in BD
signiﬁcant relationships between total hospitaliza-
tions and executive function, verbal ﬂuency, spatial
memory, psychomotor speed and visual recogni-
tion memory. In contrast, Clark et al. (12) reported
no signiﬁcant relationships with the total number
of admissions, but found that the number of
admissions for depressive episodes was positively
related to errors on a spatial working memory task.
Zubieta et al. (11) reported a signiﬁcant relation-
ship between the number of manic admissions and
Age at illness onset. Only one of six studies
examining this reported any signiﬁcant relation-
ships between age at illness onset and neuropsy-
chological function and correlation coeﬃcients
were generally low (Table 2). Martı
n et al.
(16) reported that those with a later onset of illness
performed more poorly on measures of psycho-
motor speed and executive shifting. However, this
ﬁnding was not consistent across other studies.
El-Badri et al. (9) found no signiﬁcant relationship
between performance on the same executive shift-
ing test and age at illness onset in a younger sample
of bipolar patients. None of the other domains
investigated [verbal learning (11, 15, 16, 18), verbal
ﬂuency (9, 18), executive function (9, 11, 18) and
visual memory (9, 17, 18)] showed any relationship
to age at illness onset.
Zubieta et al. (11) investigated the relationship
between neuropsychological function and the age
at onset of mania or the age at onset of depression
separately. There were no signiﬁcant relationships
with either on tests of verbal memory, executive
function or psychomotor speed in this relatively
Duration of illness. The impact of duration of
illness on neuropsychological function was inves-
tigated by 11 studies. The ﬁndings were somewhat
equivocal – ﬁve studies reported at least one
signiﬁcant negative relationship between the length
of time since the ﬁrst mood episode and a measure
of neuropsychological function, whilst six found no
Scores on tests of executive function (12, 18),
psychomotor speed (16, 18), visuospatial memory
(18) and verbal memory (12, 13, 15, 16) were
negatively related to length of illness. The most
extensively investigated domain was verbal mem-
ory – four of six studies including a verbal memory
measure reported signiﬁcant associations between
degree of impairment and duration of illness (12,
13, 15, 16). As with number of manic episodes, it
Table 3. Neuropsychological tests included in correlations with illness features
Study Tests employed
Van Gorp et al. (5) California Verbal Learning Test; Controlled Oral Word Association Test; Stroop colour-word test;
Trail Making Test B; Wisconsin Card Sorting Test
Krabbendam et al. (7) Concept Shifting Test; Letter Digit Substitution Test; Rey Auditory Verbal Learning Test; Stroop
colour-word test; word ﬂuency
Rubinsztein et al. (8) CANTAB Simultaneous/Delayed Match to Sample
El-Badri et al. (9) CANTAB Simultaneous/Delayed Match to Sample & Tower of London; Controlled Oral Word
Association Test; WAIS-R Digit Span & Digit Symbol Substitution Test
Zubieta et al. (11) Bead Tap Test; Manual Imitation Test; Stroop colour-word test; WMS Paired Associate Learning;
Wisconsin Card Sorting Test
MacQueen et al. (10) Visual Backward Masking
Clark et al. (12) CANTAB Intra-Dimensional-Extra-Dimensional Set Shift, Rapid Visual Information Processing,
Spatial Working Memory & Tower of London; California Verbal Learning Test; Iowa Gambling Task
Cavanagh et al. (13) California Verbal Learning Test
Fleck et al. (14) California Verbal Learning Test
Deckersbach et al. (15) California Verbal Learning Test
Deckersbach et al. (17) Rey-Osterrieth Complex Figures Test
n et al. (16) California Verbal Learning Test; Trail Making Test A & B; WAIS-R Reverse Digit Span
Thompson et al. (18) CANTAB Tower of London, Spatial Span, Spatial Working Memory, Spatial Recognition, Delayed
Match to Sample & Paired Associates Learning; Controlled Oral Word Association Test; Rey
Auditory Verbal Learning Test; Self-Ordered Pointing Test; Stroop colour-word test; Trail Making
Test A; Vigil; WAIS–R Reverse Digit Span & Digit Symbol Substitution Test
Clark et al. (19) Rapid Visual Information Processing
et al. (20) California Verbal Learning Test; Continuous Performance Test; Facial Recognition Task; Reverse
Visual Span; WAIS-R Vocabulary, Digit Symbol Substitution Test & Reverse Digit Span; WMS-R
Story Recall & Visual Reproduction
Tests listed are only those included in correlations with illness features. Any other tests used by the above studies are not included in this
table. CANTAB ¼ Cambridge Neuropsychological Test Automated Battery; WAIS-R ¼ Wechsler Adult Intelligence Scale-Revised;
WMS-R ¼ Wechsler Memory Scale-Revised.
Robinson and Ferrier
was impairment on long-delay verbal memory that
was most consistently associated with a longer
illness (13, 15, 16).
The six studies reporting no positive ﬁndings
investigated smaller patient samples, suggesting
statistical power may have been an issue. All of the
studies reporting no signiﬁcant relationships had
samples less than 27, whereas four of the ﬁve
remaining studies had samples of 30 or more.
Length of time euthymic. Four studies investigated
the relationship between the length of time in
clinical remission prior to neuropsychological test-
ing and test performance. None of the studies
found any signiﬁcant associations between dura-
tion of euthymia and performance on tests of
verbal ﬂuency (9, 18), attention (12, 18), executive
function (9, 12, 18), visual memory (9, 10, 18) or
verbal memory (18) and all correlation coeﬃcients
were low (Table 2).
Neuropsychological function in unaffected ﬁrst-degree
Table 4 includes the characteristics of the samples
included in the FDR studies.
Child studies. All three studies that included a
sample of children less than 16 years old with at
least one bipolar parent (27–29) reported some
evidence of impairment in the at-risk children. In a
large sample of children with one parent hospital-
ized for a manic episode, Winters et al. (27) found
that the high-risk children showed increased reac-
tion time on a visual search task. Performance on
tests of word communication, digit span, ability to
maintain attention, and object sorting did not
diﬀer from that of controls. Decina et al. (28)
investigated discrepancies between verbal IQ (VIQ)
and performance IQ (PIQ) in a sample of oﬀspring
of bipolar parents [a VIQ-PIQ discrepancy is
thought to indicate hemispheric processing diﬀer-
ences suggesting dominance of the left hemisphere,
although this may be an indicator of impaired
executive functioning (29)]. Full-scale IQ and VIQ
did not diﬀer between the control group and the
high-risk children, but PIQ was signiﬁcantly lower
in the latter. There was a signiﬁcantly higher
incidence of VIQ > PIQ discrepancy (deﬁned as
VIQ-PIQ ‡ 15 points) among the high-risk chil-
dren. However, in this sample, roughly half
(51.6%) of the children had psychopathology of
their own, either fulﬁlling criteria for an Research
Diagnostic Criteria diagnosis or suﬀering clinical
symptoms severe enough to impact on overall
functioning. Finally, McDonough-Ryan et al. (29)
also reported a signiﬁcantly higher incidence of
VIQ-PIQ discrepancy (deﬁned in this instance as
VIQ-PIQ ‡ 10 points) in a group of high-risk
children. Furthermore, the high-risk group per-
formed signiﬁcantly worse than control subjects on
measures of reading, spelling and arithmetic.
Adult studies. Five of seven studies investigating
the neuropsychological function of adult relatives
of bipolar probands reported some impairment in
the FDR group (30–36).
Pierson et al. (30) administered a sustained
auditory attention task to relatives of multiple
aﬀected bipolar families (at least two members with
a bipolar diagnosis) and reported signiﬁcantly
longer response latencies in the FDR group relative
to control subjects. This diﬀerence remained highly
signiﬁcant once relatives with their own history of
psychopathology were excluded from the analyses.
The same study reported evidence of electrophys-
iological abnormalities consistent with slower
processing speed in the high-risk subjects relative
to controls. Keri et al. (32) reported signiﬁcantly
impaired performance in long-delay verbal declar-
ative memory by relatives of bipolar patients
without any personal history of psychopathology.
Tests of executive function, verbal ﬂuency and
visuospatial working memory showed no impair-
ment. Prior to an acute tryptophan depletion
procedure, Sobczak et al. (33) reported baseline
deﬁcits in relatives of bipolar I disorder patients on
planning (executive function) and several measures
of delayed verbal declarative memory. Other
measures, including working memory, verbal ﬂu-
ency, divided attention and executive set shifting,
were in tact. Relatives of patients with bipolar II
disorder showed no baseline impairments in any
domain. Zalla et al. (34) found an isolated deﬁcit in
one executive function task in their sample of
bipolar relatives. Despite moderate eﬀect sizes,
diﬀerences in performance on other executive
measures did not reach the stringent Bonferroni-
corrected level of statistical signiﬁcance set in this
study (p < 0.006), but were statistically signiﬁcant
at conventional levels (p < 0.05). Ferrier et al.
(35) reported signiﬁcant deﬁcits in executive func-
tion and spatial recognition in relatives of bipolar
patients. Although some of the subjects had
psychopathology, excluding them from the analysis
did not change the results.
In contrast to the previous studies, Kremen et al.
(31) failed to ﬁnd signiﬁcant deﬁcits in the bipolar
relative group on any measure of a broad neuro-
psychological test battery. One notable diﬀerence
in this study is that all the participants were female.
In a small pilot study of seven subjects, MacQueen
Evolution of cognitive impairment in BD
Table 4. Summary of ﬁndings of studies investigating neuropsychological function in the ﬁrst-degree relatives of bipolar patients
Study n (relatives/controls) Age
Gender (% male) Sig/tested
Signiﬁcant impairments in high-risk group
Winters et al. (27) 76/134 6–16
42.1 3/13 Slower reaction time on the ﬁnal three quadrants of a visual search task
Decina et al. (28) 31/18 11.0 45.2 1/4 Lower PIQ; higher incidence of VIQ-PIQ discrepancy
Ryan et al. (29)
28/24 10.2 (2.7) 46.4 5/11 Higher incidence of VIQ-PIQ discrepancy; poorer scores on Wide Range
Achievement Test-3 reading, spelling and arithmetic subtests; higher
incidence of spelling impairment signiﬁcant enough to be classed as a
Pierson et al. (30) 18/17 26.2 (6.6) 36.8 1/1 Slower reaction time on a sustained auditory attention task
Kremen et al. (31) 15/44 36.3 (11.5) 0 0/22 None
Keri et al. (32) 20/20 35.1 (9.5) 40 1/12? Fewer words recalled on a long-delay verbal recall task
Sobczak et al. (33) 30/15 41.4 (2.6) 26.7 6/22? Relatives of BDI patients had longer planning time on an executive function
task and poorer recall on 5 measures of delayed verbal declarative memory;
relatives of BDII patients showed no impairments
Zalla et al. (34) 33/20 37.3 (11) 39.4 1/8 Greater interference on the Stroop colour-word task
Ferrier et al. (35) 17/17 34.8 (11.2) 41.2 3/20 Smaller reverse digit span; lower spatial span and poorer spatial recognition
MacQueen et al. (36) 7/7 22.29 57.1 0/2 No differences in error rate or reaction time on a visual backward masking task
Gourovitch et al. (37) 7/15 32.7 14.3 7/38 Poorer verbal recall on several subscales of a verbal learning test; lower recall
and greater interference on the Brown–Peterson Test; poorer performance on
mental control and memory quotient subscales of the Wechsler Memory Scale
et al. (20) 19/114 45.8 (1.7) 31.6 0/15? No differences on tests of attention, working memory, non-verbal memory,
verbal memory and verbal learning; signiﬁcant impairment in verbal learning
in the female non-bipolar co-twins only
PIQ ¼ Performance IQ; VIQ ¼ Verbal IQ; BDI ¼ bipolar disorder type I; BDII ¼ bipolar disorder type II.
Values represent mean (SD) in years.
Range (mean not reported).
Values represent the number of statistically signiﬁcant differences/number of differences tested. ? indicates a value that was unclear or not reported.
Robinson and Ferrier
et al. (36) did not ﬁnd any signiﬁcant impairments
on a visual memory procedure.
Twin studies. Gourovitch et al. (37) investigated
neuropsychological function in seven monozygotic
twin pairs discordant for bipolar disorder. This
design potentially controls for both genetic and
psychosocial factors and indicates which neuro-
psychological deﬁcits are related to the experience
of aﬀective symptoms versus those associated with
the broader bipolar genotype. The unaﬀected twins
with an aﬀected sibling performed signiﬁcantly
worse than control twins on measures of working
memory and attention/mental control. Impairment
was marked on several measures of long-delay
verbal memory, where both discordant twins
performed signiﬁcantly worse than control twins
but not signiﬁcantly diﬀerent to one another,
whereas performance on short-delay verbal mem-
ory was impaired only in the aﬀected twin.
In a larger twin study, Kieseppa
et al. (20) found
no signiﬁcant deﬁcits between unaﬀected co-twins
and healthy comparison twin pairs in tests of
attention, working memory, verbal or non-verbal
memory, or verbal learning. However, in account-
ing for the unequal gender distribution between the
high-risk twins and the control twins, the authors
did ﬁnd a signiﬁcant deﬁcit in verbal learning in the
unaﬀected female co-twins – once again impair-
ment was shown in delayed verbal recall. Unlike
the previous twin study, this study included non-
identical twin pairs. The greater genetic variation
between dizygotic twin pairs may explain why this
sample of co-twins showed less impairment relative
to the healthy comparison group than the mono-
zygotic twins in the earlier study.
Evidence from the ﬁrst section of this review
indicates that cognitive function in patients with
bipolar disorder is negatively related to features of
illness, particularly the number of episodes suf-
fered, the number of hospital admissions and
duration of illness. Episodes of both depression
and mania related negatively to neuropsychologi-
cal function, but manic episodes were related more
consistently to delayed verbal memory and some
measures of executive function, whereas depressed
episodes were related less consistently to a broader
range of impairments. Verbal memory perfor-
mance as measured using list-learning tasks was
most highly related to illness features, showing a
negative relationship with both number of manic
episodes, duration of illness and hospital admis-
sions. Evidence from the FDR studies documented
subtle signs of impairment in those at high genetic
risk for bipolar disorder in several components of
executive function (planning, response inhibition
and mental manipulation) and verbal memory.
One possible interpretation of these ﬁndings is that
patients with greater cognitive impairment are less
able to manage their illness and suﬀer a poorer
illness course as a result. However, the presence of
subtle deﬁcits in FDRs provides an indication that
cognitive impairment may represent a trait vulner-
ability factor in the development of bipolar dis-
order which is present before illness onset but
worsens as the illness progresses.
It is necessary to consider a number of limi-
tations. First, with regard to the studies of illness
outcome, one of the biggest limitations is the cross-
sectional nature of the ﬁndings. None of the studies
documented a decline in cognitive function from
premorbid (or perimorbid) levels. Thus it remains
unknown whether each successive episode of illness
resulted in further cognitive impairment. A few
test–retest studies have investigated cognitive func-
tion during and after an acute episode of mania or
depression. Most have documented improvement
as symptoms abate (38–42), but some have noted
sustained impairment in speciﬁc sub-groups of
patients [e.g. older patients (40)] or on speciﬁc
measures [e.g. executive and working memory (42,
43)]. However, the interval between tests has
generally been short, ranging from 1 to 8 months,
and practice eﬀects rarely accounted for by retest-
ing the control groups. Of the three studies
investigating periods longer than this, Englesman
et al. (44) reported stable memory performance in
a 6-year follow up of bipolar patients taking
lithium. Dhingra and Rabins (45) found that
almost one-third of a small sample of elderly
bipolar patients showed marked cognitive deteri-
oration 5–7 years after admission for mania, which
was more than the 1–2% expected by chance.
Finally, a more recent study investigated the
stability of cognitive impairment in euthymic
bipolar patients (46). Three years after ﬁrst assess-
ment, there were no signiﬁcant changes in neuro-
psychological function except a slight improvement
on 1 of 13 measures reported (Wisconsin Card
Sorting Test perseverative errors) and the bipolar
patients still exhibited deﬁcits relative to controls
on most indices. The persistence and worsening of
cognitive impairment reported in these latter two
studies lends some support to the suggestion that
cognitive impairment is a trait marker of illness
that worsens over time. Cross-sectional data from
a large case-register study of patients with bipolar
disorder reported a 6% increase in risk of dementia
with every episode leading to admission (47),
Evolution of cognitive impairment in BD
suggesting the experience of episodes may be
associated with cognitive decline. However, more
studies are needed, especially those which follow
patients from an early stage of illness and detail
illness episodes, hospitalization and medication
Second, all of the studies relied on retrospec-
tively reported life histories. It is diﬃcult to chart
the course of a long illness accurately, especially for
those who have suﬀered a large number of
episodes. Cavanagh et al. (13) suggested that there
may be a recall bias in favour of mania, which is
often more likely to be recorded in case notes and
to result in hospital admission. Relationships
between cognitive impairment and manic episodes
were more consistently found than those for
depression, which may simply reﬂect greater accu-
racy in documenting manic episodes.
Third, many of the clinical outcome measures
are highly intercorrelated. In a factor analysis of
illness features in depressed patients, Fossati et al.
(48) found that duration of illness, number of
episodes and number of hospitalizations all loaded
onto a single Ôlongitudinal severity of illnessÕ factor,
indicating they are highly related to one another.
Indeed it seems unlikely they are three independent
features of illness. This high degree of intercorre-
lation was poorly controlled for, as were other
potential covariates, such as age [although the only
study to control for current age statistically still
reported a signiﬁcant relationship between verbal
learning and length of illness (12)]. One study
found that both length of illness and number of
episodes related to verbal memory impairment
when considered alone, but when they were both
entered into a regression analysis as independent
variables, only the number of episodes remained as
a signiﬁcant predictor of verbal memory perfor-
mance (15). This begs the question of which
speciﬁc features of episodes or illness relate to
cognitive function. Two studies addressed this issue
further, examining whether the experience of dis-
crete episodes or length of time spent acutely
aﬀectively ill was more related to cognitive func-
tion (5, 18). In one instance the number of months
manic and number of months depressed were more
strongly related to a larger number of neurocog-
nitive domains than number of episodes per se (5).
There remain other possible inﬂuences on cognitive
function that have yet to be suﬃciently investigated
such as prior history of psychotic symptoms, rapid
cycling, comorbidity, and family history of psy-
chotic or aﬀective disorders.
Fourth, there was variation in statistical
approach such that some studies correlated illness
features with only those neuropsychological indices
that had shown signiﬁcant diﬀerences between
patients and controls, whereas others included
indices not showing such diﬀerences (Table 2,
bottom row). Given that wide variation is often
observed in the neuropsychological test scores of
bipolar patient samples, this ﬁrst method may have
omitted relevant domains from the analyses,
whereas the latter technique increases the likeli-
hood of type I error. Likewise there could be a
reporting bias in favour of statistically signiﬁcant
ﬁndings. Investigators may be more likely to report
the ﬁndings of correlations with illness features if
they ﬁnd signiﬁcant results. Fifth, many of the
studies had small sample sizes which may have
limited power or led to restricted range on some
indices. Finally, the criteria for euthymia were not
identical across studies and low-level mood symp-
toms were evident in most patient samples [in all
studies with a healthy control group the patient
group scored higher than controls on mood scales
and in several instances the diﬀerence reached
statistical signiﬁcance (6, 12, 13, 16, 18)]. These
low-level mood symptoms may impact on cognitive
function thus blurring the distinction between state
and trait dysfunction. There is no consensus on a
clinically meaningful deﬁnition of euthymia and a
variety of deﬁnitions have been used in the studies
included here (see Table 1). Only one study in this
review used partial correlation to control for the
presence of residual symptoms (13) and still
reported signiﬁcant relationships between number
of manic and depressive episodes and various
verbal learning indices. Several studies correlated
neuropsychological test scores with residual symp-
toms and reported no relationship (7, 10, 11, 15,
17), although some found signiﬁcant relationships
(6, 14, 16) and did not control for symptoms in
subsequent correlations with illness features. Com-
plete absence of symptoms is not a realistic target
for all patients with bipolar disorder, but estab-
lishing a consensus deﬁnition of euthymia that can
be readily applied will increase comparability of
Despite these limitations, the evidence points
towards a robust association between impaired
long-delay verbal memory and a greater burden of
illness. Studies of cognitive impairment in euthymic
bipolar subjects suggest that verbal memory and
executive functions are the areas showing greatest
impairment (5, 6, 11, 12, 15, 16, 18, 34, 49).
Cognitive ﬁndings have indicated that the acquisi-
tion of information is more impaired in bipolar
patients than retention. The present results suggest
that the experience of manic episodes is associated
with poorer recall performance, which may
reﬂect a negative inﬂuence of mania on retention.
Robinson and Ferrier
Executive dysfunction was less consistently related
to illness features than verbal memory. There are a
number of possible explanations for this. First, the
relationship between executive function and illness
factors has been tested less consistently due to the
broad range of executive tests available. In the case
of verbal memory tests, investigators have con-
verged on the use of a few highly similar list-
learning-based tests, which has provided a larger
body of comparable results to look for a relation-
ship. Second, executive dysfunction may be an
early feature of bipolar illness that does not
progressively worsen with increased burden of
illness. Third, the verbal memory impairment
may be due to executive dysfunction. The test
most consistently related to number of manic
episodes was the California Verbal Learning Test
(CVLT), which incorporates a strategic (executive)
component in that subjects can use semantic
organization of the list to aid performance. Inter-
estingly, the two studies that failed to ﬁnd a
relationship between manic episodes and verbal
learning (7, 18) employed the Rey Auditory Verbal
Learning Test, which does not incorporate this
strategic element. Signiﬁcant overlap between ver-
bal learning and executive function has been
documented in other clinical groups – Duﬀ et al.
found that these two cognitive domains shared 50–
60% of variance (50) – and this overlap was
especially marked for the CVLT (51). Declining
executive function over the course of illness may
therefore produce a relationship between number
of episodes and verbal memory impairment that is
more apparent than real. Alternatively, non-pro-
gressive executive dysfunction occurring early in
the illness may be at least partially compensated for
by the use of alternative ways of performing verbal
learning tests, and perhaps the functions underly-
ing these support strategies are vulnerable to illness
progression. This could explain why verbal mem-
ory deﬁcits are relatively moderate when compared
with executive deﬁcits in euthymic patients (49),
but also explains why they worsen with illness
progression more so than executive functions.
Executive impairment in patients with bipolar
disorder may reﬂect underlying dysfunction in the
structural or functional neuroanatomy of the
prefrontal cortex (PFC). In general there is no
evidence of diﬀerences in the overall size or volume
of the PFC in patients with bipolar disorder (52–
54). Of the few studies that measured the volume of
speciﬁc subregions of the PFC, the general pattern
of ﬁndings suggests there are abnormalities in the
dorsolateral PFC and anterior cingulate (52, 55–
58). However, most studies to date have been
conducted in symptomatic patients. Drevets et al.
(55) reported lower volume of the subgenual PFC
in a small group of euthymic patients. Newer
techniques such as diﬀusion tensor imaging (DTI)
provide a more sensitive measure of the integrity of
neural systems and microstructural diﬀerences that
may impair neural communication between diﬀer-
ent brain regions. Two recent studies in sympto-
matic bipolar patients employing DTI have shown
evidence of abnormalities in prefrontal white
matter tracts (59, 60). Only a small number of
studies have combined functional imaging with
executive tasks in euthymic bipolar patients. So far
evidence indicates reduced activity in frontal areas
necessary for successful task performance (61–63).
More data are required in euthymic patients before
stronger conclusions can be drawn.
Speculation as to the underlying pathophysiol-
ogy has centred on the role of the hypothalamic-
pituitary-adrenal (HPA) axis. Hyper-responsiveness
of the HPA axis to speciﬁc integrated challenges,
such as the dexamethasone/corticotrophin-releasing
hormone test, has been documented in both the
manic and depressed phases of illness (64, 65), and a
recent study in euthymic patients also found this
abnormality in remission (66). Rush et al. (67)
reported that 43% of bipolar depressed patients
were non-suppressors on the dexamethasone sup-
pression test. A number of studies have suggested
this incidence is higher (between 40% and 70%) in
patients with a manic episode, and especially high
(between 71% and 100%) in patients with mixed
mania (68). The potentially neurotoxic eﬀects of
hypercortisolaemia may instigate or worsen damage
to neural tissue. This may account for a worsening of
illness with each successive episode and possibly
explain why manic episodes are more strongly
associated with cognitive impairment. Treatment
with agents targeting the HPA axis, such as the
glucocorticoid receptor antagonist mifepristone,
has been shown to alleviate both mood and some
aspects of cognitive impairment in depressed bipolar
patients (69) and represents an important potential
Turning to the limitations of the studies of
FDRs of bipolar patients, some relatives had
personal psychopathology (including anxiety,
mood and personality disorders, although none
had a history of bipolar disorder). Generally this
was controlled for and only applied to a subset of
the sample, which makes it unlikely to be the sole
explanation for the reported diﬀerences. In studies
of high-risk adolescents it is not known whether
any reported deﬁcits are premorbid indicators of
illness. Evidence of cognitive impairment in unaf-
fected high-risk adults past the age of greatest risk
for developing bipolar disorder suggests on the one
Evolution of cognitive impairment in BD
hand that cognitive dysfunction could be an
endophenotypic marker for bipolar disorder, but
on the other hand it also suggests a degree of
independence between the experience of cognitive
dysfunction and the development of psychopathol-
ogy. However, evidence from three studies that
each included a patient group, an FDR group and
a healthy control group indicated that FDRs of
bipolar probands tend to show less severe and less
pervasive impairment than their aﬀected relatives
(20, 34, 37). In general, the FDR group showed
impairment in fewer domains than the patient
group and tended to exhibit intermediate perfor-
mance between the patients and controls.
Future studies should address the above issues.
The present results indicate a need to narrow
down the speciﬁc features of mood episodes that
are responsible for the relationship with cognitive
dysfunction. This could be aided by further
research tapping into the particular processes
showing impairment. Verbal declarative memory
and executive function are both broad domains
involving several diﬀerent processes which interact
with one another and with other cognitive sys-
tems. A more targeted approach focusing on these
areas is necessary to clarify the particular nature
and speciﬁcity of the reported deﬁcits. There is
also a great need for prospective, longitudinal
studies to document how cognitive function
changes over time in individuals with bipolar
The studies in FDRs have indicated some
cognitive domains show impairment before or
without the onset of symptoms. Further investiga-
tion into cognitive or emotional processing diﬀer-
ences between aﬀected individuals and their
unaﬀected relatives could provide insights into
the psychological risk factors for bipolar disorder
and could enhance the search for speciﬁc vulner-
Cognitive impairment in bipolar patients is not
benign – it is associated with signiﬁcant psychoso-
cial impairment (11, 16, 21–23). Improving func-
tional recovery for those who continue to
experience impairment in their everyday lives
should be a major goal of research and treatment
in this area. A much better understanding of the
role of neuropsychological dysfunction in clinical,
social and occupational impairment is an integral
part of reaching that goal.
1. Clark L, Iversen SD, Goodwin GM. A neuropsychological
investigation of prefrontal cortex involvement in acute
mania. Am J Psychiatry 2001; 158: 1605–1611.
n 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.
3. Murphy FC, Rubinsztein JS, Michael A et al. Decision-
making cognition in mania and depression. Psychol Med
2001; 31: 679–693.
4. Borkowska A, Rybakowski J. Neuropsychological frontal
lobe tests indicate that bipolar depressed patients are more
impaired than unipolar. Bipolar Disord 2001; 3: 88–94.
5. Van Gorp WG, Altshuler L, Theberge DC, Wilkins J,
Dixon W. Cognitive impairment in euthymic bipolar
patients with and without prior alcohol dependence: a
preliminary study. Arch Gen Psychiatry 1998; 55: 41–46.
6. Ferrier IN, Stanton BR, Kelly TP, Scott J. Neuropsycho-
logical function in euthymic patients with bipolar disorder.
Br J Psychiatry 1999; 175: 246–251.
7. Krabbendam L, Honig A, Wiersma J et al. Cognitive
dysfunctions and white matter lesions in patients with
bipolar disorder in remission. Acta Psychiatr Scand 2000;
8. Rubinsztein JS, Michael A, Paykel ES, Sahakian BJ.
Cognitive impairment in remission in bipolar aﬀective
disorder. Psychol Med 2000; 30: 1025–1036.
9. El-Badri S, Ashton C, Moore P, Marsh V, Ferrier I.
Electrophysiological and cognitive function in young
euthymic patients with bipolar aﬀective disorder. Bipolar
Disord 2001; 3: 79–87.
10. MacQueen GM, Young LT, Galway TM, Joﬀe RT.
Backward masking task performance in stable, euthymic
out-patients with bipolar disorder. Psychol Med 2001; 31:
11. Zubieta JK, Huguelet P, O’Neil RL, Giordani BJ. Cogni-
tive function in euthymic bipolar I disorder. Psychiatry Res
2001; 102: 9–20.
12. Clark L, Iversen SD, Goodwin GM. Sustained attention
deﬁcit in bipolar disorder. Br J Psychiatry 2002; 180: 313–
13. Cavanagh JTO, Van Beck M, Muir W, Blackwood DHR.
Case–control study of neurocognitive function in euthymic
patients with bipolar disorder: an association with mania.
Br J Psychiatry 2002; 180: 320–326.
14. Fleck DE, Shear PK, Zimmerman ME et al. Verbal
memory in mania: eﬀects of clinical state and task
requirements. Bipolar Disord 2003; 5: 375–380.
15. Deckersbach T, Savage CR, Reilly-Harrington N, Clark L,
Sachs G, Rauch SL. Episodic memory impairment in
bipolar disorder and obsessive-compulsive disorder: the
role of memory strategies. Bipolar Disord 2004; 6: 233–
n A, Vieta E, Colom F et al. Cognitive
impairment in euthymic bipolar patients: implications for
clinical and functional outcome. Bipolar Disord 2004; 6:
17. Deckersbach T, McMurrich S, Ogutha J, Savage CR,
Sachs G, Rauch SL. Characteristics of non-verbal memory
impairment in bipolar disorder: the role of encoding
strategies. Psychol Med 2004; 34: 823–832.
18. Thompson JM, Gallagher P, Hughes JH et al. Neurocog-
nitive impairment in euthymic patients with bipolar aﬀec-
tive disorder. Br J Psychiatry 2005; 186: 32–40.
19. Clark L, Kempton MJ, Scarna A, Grasby PM, Goodwin
GM. Sustained attention-deﬁcit conﬁrmed in euthymic
bipolar disorder but not in ﬁrst-degree relatives of bipolar
patients or euthymic unipolar depression. Biol Psychiatry
2005; 57: 183–187.
Robinson and Ferrier
T, Tuulio-Henriksson A, Haukka J et al. Mem-
ory and verbal learning functions in twins with bipolar I
disorder, and the role of information-processing speed.
Psychol Med 2005; 35: 205.
21. Atre-Vaidya N, Taylor MA, Seidenberg M, Reed R,
Perrine A, Glick-Oberwise F. Cognitive deﬁcits, psycho-
pathology, and psychosocial functioning in bipolar mood
disorder. Neuropsychiatry Neuropsychol Behav Neurol
1998; 11: 120–126.
n A, Penades R, Vieta E et al. Executive
function in patient s with remitted bipolar disorder and
schizophrenia and its relationship with functional out-
come. Psychother Psychosom 2002; 71: 39–46.
23. Zarate CA Jr, Tohen M, Land M, Cavanagh S. Functional
impairment and cognition in bipolar disorder. Psychiatr Q
2000; 71: 309–329.
24. Glahn DC, Bearden CE, Niendam TA, Escamilla MA. The
feasibility of neuropsychological endophenotypes in the
search for genes associated with bipolar aﬀective disorder.
Bipolar Disord 2004; 6: 171–182.
25. Ali SO, Denicoﬀ KD, Altshuler LL et al. Relationship
between prior course of illness and neuroanatomic struc-
tures in bipolar disorder: a preliminary study. Neuropsy-
chiatry Neuropsychol Behav Neurol 2001; 14: 227–232.
26. Chang K, Steiner H, Ketter T. Studies of oﬀspring of
parents with bipolar disorder. Am J Med Genet C Semin
Med Genet 2003; 123C: 26–35.
27. Winters KC, Stone AA, Weintraub S, Neale JM. Cognitive
and attentional deﬁcits in children vulnerable to psycho-
pathology. J Abnorm Child Psychol 1981; 9: 435–453.
28. Decina P, Kestenbaum CJ, Farber S et al. Clinical and
psychological assessment of children of bipolar probands.
Am J Psychiatry 1983; 140: 548–553.
29. McDonough-Ryan P, DelBello M, Shear PK, Ris MD,
Soutullo C, Strakowski SM. Academic and cognitive
abilities in children of parents with bipolar disorder: a
test of the nonverbal learning disability model. J Clin
Exp Neuropsychol 2002; 24: 280–285.
30. Pierson A, Jouvent R, Quintin P, Perez-Diaz F, Leboyer
M. Information processing deﬁcits in relatives of manic
depressive patients. Psychol Med 2000; 30: 545–555.
31. Kremen WS, Faraone SV, Seidman LJ, Pepple JR,
Tsuang MT. Neuropsychological risk indicators for
schizophrenia: a preliminary study of female relatives of
schizophrenic and bipolar probands. Psychiatry Res 1998;
32. Keri S, Kelemen O, Benedek G, Janka Z. Diﬀerent trait
markers for schizophrenia and bipolar disorder: a neuro-
cognitive approach. Psychol Med 2001; 31: 915–922.
33. Sobczak S, Riedel WJ, Booij I, Aan Het Rot M, Deutz
NEP, Honig A. Cognition following acute tryptophan
depletion: diﬀerence between ﬁrst-degree relatives of bipo-
lar disorder patients and matched healthy control volun-
teers. Psychol Med 2002; 32: 503–515.
34. Zalla T, Joyce C, Szoke A et al. Executive dysfunctions as
potential markers of familial vulnerability to bipolar
disorder and schizophrenia. Psychiatry Res 2004; 121:
35. Ferrier IN, Chowdhury R, Thompson JM, Watson S,
Young AH. Neurocognitive function in unaﬀected ﬁrst-
degree relatives of patients with bipolar disorder: a
preliminary report. Bipolar Disord 2004; 6: 319–323.
36. MacQueen GM, Grof P, Alda M, Marriott M, Young LT,
Duﬀy A. A pilot study of visual backward masking
performance among aﬀected versus unaﬀected oﬀspring
of parents with bipolar disorder. Bipolar Disord 2004; 6:
37. Gourovitch ML, Torrey EF, Gold JM, Randolph C,
Weinberger DR, Goldberg TE. Neuropsychological per-
formance of monozygotic twins discordant for bipolar
disorder. Biol Psychiatry 1999; 45: 639–646.
38. Henry GM, Weingartner H, Murphy DL. Inﬂuence of
aﬀective states and psychoactive drugs on verbal learning
and memory. Am J Psychiatry 1973; 130: 966–971.
39. Donnelly EF, Murphy DL, Goodwin FK, Waldman IN.
Intellectual function in primary aﬀective disorder. Br J
Psychiatry 1982; 140: 633–636.
40. Savard RJ, Rey AC, Post RM. Halstead–Reitan Category
Test in bipolar and unipolar aﬀective disorders. Relation-
ship to age and phase of illness. J Nerv Ment Dis 1980; 168:
41. Liu SK, Chiu CH, Chang CJ, Hwang TJ, Hwu HG,
Chen WJ. Deﬁcits in sustained attention in schizophrenia
and aﬀective disorders: stable versus state-dependent
markers. Am J Psychiatry 2002; 159: 975–982.
42. McGrath J, Scheldt S, Welham J, Clair A. Performance on
tests sensitive to impaired executive ability in schizophre-
nia, mania and well controls: acute and subacute phases.
Schizophr Res 1997; 26: 127–137.
43. McGrath J, Chapple B, Wright M. Working memory in
schizophrenia and mania: correlation with symptoms
during the acute and subacute phases. Acta Psychiatr
Scand 2001; 103: 181–188.
44. Engelsmann F, Katz J, Ghadirian AM, Schachter D.
Lithium and memory: a long-term follow-up study. J Clin
Psychopharmacol 1988; 8: 207–212.
45. Dhingra U, Rabins PV. Mania in the elderly: a 5–7 year
follow-up. J Am Geriatr Soc 1991; 39: 581–583.
46. Balanza-Martinez V, Tabares-Seisdedos R, Selva-Vera G
et al. Persistent cognitive dysfunctions in bipolar I disorder
and schizophrenic patients: a 3-year follow-up study.
Psychother Psychosom 2005; 74: 113–119.
47. Kessing LV, Andersen PK. Does the risk of developing
dementia increase with the number of episodes in patients
with depressive disorder and in patients with bipolar
disorder? J Neurol Neurosurg Psychiatry 2004; 75: 1662–
48. Fossati P, Harvey P-O, Le Bastard G, Ergis A-M, Jouvent
R, Allilaire J-F. Verbal memory performance of patients
with a ﬁrst depressive episode and patients with unipolar
and bipolar recurrent depression. J Psychiatr Res 2004; 38:
49. Robinson LJ, Goswami U, Gallagher P et al. Cognitive
deﬁcits in bipolar subjects: Delhi and UK data. Acta
Psychiatr Scand Suppl 2004; 110: 34.
50. Duﬀ K, Schoenberg MR, Scott JG, Adams RL. The
relationship between executive functioning and verbal and
visual learning and memory. Arch Clin Neuropsychol
2005; 20: 111–122.
51. Tremont G, Halpert S, Javorsky DJ, Stern RA. Diﬀeren-
tial impact of executive dysfunction on verbal list learning
and story recall. Clin Neuropsychol 2000; 14: 295–302.
52. Lopez-Larson MP, DelBello MP, Zimmerman ME,
Schwiers ML, Strakowski SM. Regional prefrontal gray
and white matter abnormalities in bipolar disorder. Biol
Psychiatry 2002; 52: 93–100.
53. Coﬀman JA, Bornstein RA, Olson SC, Schwarzkopf SB,
Nasrallah HA. Cognitive impairment and cerebral struc-
ture by MRI in bipolar disorder. Biol Psychiatry 1990; 27:
54. Strakowski SM, DelBello MP, Sax KW et al. Brain
magnetic resonance imaging of structural abnormalities
in bipolar disorder. Arch Gen Psychiatry 1999; 56: 254–
Evolution of cognitive impairment in BD
55. Drevets WC, Price JL, Simpson JR et al. Subgenual
prefrontal cortex abnormalities in mood disorders. Nature
1997; 386: 824–827.
56. Hirayasu Y, Shenton ME, Salisbury DF et al. Subgenual
cingulate cortex volume in ﬁrst-episode psychosis. Am J
Psychiatry 1999; 156: 1091–1093.
57. Brambilla P, Nicoletti MA, Harenski K et al. Anatomical
MRI study of subgenual prefrontal cortex in bipolar and
unipolar subjects. Neuropsychopharmacology 2002; 27:
58. Frangou S, Hadjulis M, Chitnis X, Baxter D, Donaldson
S, Raymont V. The Maudsley Bipolar Disorder Project:
brain structural changes in bipolar I disorder. Bipolar
Disord 2002; 4 (Suppl. 1): 123–124.
59. Beyer JL, Taylor WD, Macfall JR et al. Cortical white
matter microstructural abnormalities in bipolar disorder.
Neuropsychopharmacology 2005; 12: 2225–2229.
60. Adler CM, Holland SK, Schmithorst V et al. Abnormal
frontal white matter tracts in bipolar disorder: a diﬀusion
tensor imaging study. Bipolar Disord 2004; 6: 197–203.
61. Monks PJ, Thompson JM, Bullmore ET et al. A func-
tional MRI study of working memory task in euthymic
bipolar disorder: evidence for task-speciﬁc dysfunction.
Bipolar Disord 2004; 6: 550–564.
62. Strakowski SM, Adler CM, Holland SK, Mills N, DelBello
MP. A preliminary fMRI study of sustained attention in
euthymic, unmedicated bipolar disorder. Neuropsycho-
pharmacology 2004; 29: 1734–1740.
63. Strakowski SM, Adler CM, Holland SK, Mills NP,
DelBello MP, Eliassen JC. Abnormal fMRI brain activa-
tion in euthymic bipolar disorder patients during a
counting stroop interference task. Am J Psychiatry 2005;
64. Schmider J, Lammers CH, Gotthardt U, Dettling M,
Holsboer F, Heuser IJ. Combined dexamethasone/corti-
cotropin-releasing hormone test in acute and remitted
manic patients, in acute depression, and in normal
controls: I. Biol Psychiatry 1995; 38: 797–802.
65. Rybakowski JK, Twardowska K. The dexamethasone/
corticotropin-releasing hormone test in depression in
bipolar and unipolar aﬀective illness. J Psychiatr Res
1999; 33: 363–370.
66. Watson S, Gallagher P, Ritchie JC, Ferrier IN, Young
AH. Hypothalamic-pituitary-adrenal axis function in
patients with bipolar disorder. Br J Psychiatry 2004; 184:
67. Rush AJ, Giles DE, Schlesser MA et al. The dexametha-
sone suppression test in patients with mood disorders.
J Clin Psychiatry 1996; 57: 470–484.
68. Cassidy F, Carroll BJ. The clinical epidemiology of pure
and mixed manic episodes. Bipolar Disord 2001; 3: 35–40.
69. Young AH, Gallagher P, Watson S, Del-Estal D, Owen
BM, Ferrier IN. Improvements in neurocognitive function
and mood following adjunctive treatment with mifepri-
stone (ru-486) in bipolar disorder. Neuropsychopharma-
cology 2004; 29: 1538–1545.
Robinson and Ferrier