Directory UMM :Data Elmu:jurnal:B:Biological Psichatry:Vol48.Issue7.2000:

Neuropsychologic Impairments in Bipolar and
Unipolar Mood Disorders on the CANTAB
Neurocognitive Battery
John A. Sweeney, Julie A. Kmiec, and David J. Kupfer
Background: Cognitive deficits associated with mood
disorders, especially bipolar disorder, have been the focus
of limited systematic investigation.
Methods: We tested 35 bipolar (21 in depressed state and
14 in mixed or manic state) and 58 nonbipolar depressed
consecutively admitted young adult inpatients and 51
matched healthy individuals on the Cambridge Neuropsychological Test Automated Battery, a computerized neurocognitive battery.
Results: The mixed/manic bipolar patients demonstrated
robust deficits in episodic and working memory, spatial
attention, and problem solving. In contrast, depressed
bipolar and nonbipolar patients demonstrated impairments only in episodic memory.
Conclusions: Neuropsychologic findings with the Cambridge Neuropsychological Test Automated Battery indicate widely distributed deficits in cognitive domains subserved by temporal, parietal, and frontostriatal systems in
bipolar patients during mixed/manic states of illness.
Significant deficits in bipolar and nonbipolar depressed
patients were restricted to episodic memory, suggesting a
more selective dysfunction in mesial temporal lobe function during episodes of depression. These findings highlight the different cognitive profiles of mania and depression, demonstrate similar patterns of neuropsychologic
deficits in bipolar and nonbipolar depression, and point to

a need for further research investigating the characteristics, causes, course, and treatment of severe cognitive
deficits associated with mixed/manic phases of bipolar
disorder. Biol Psychiatry 2000;48:674 – 685 © 2000 Society of Biological Psychiatry
Key Words: Working memory, problem solving, cognition, depression, bipolar disorder, neuropsychology

From the Neurobehavioral Studies Program and the Mental Health Intervention
Research Center, Department of Psychiatry, University of Pittsburgh, Pittsburgh, Pennsylvania.
Address reprint requests to John A. Sweeney, Ph.D., WPIC, 3811 O’Hara Street,
Pittsburgh PA 15213.
Received February 4, 2000; revised April 19, 2000; accepted April 20, 2000.

© 2000 Society of Biological Psychiatry

Introduction

I

n contrast to the extensive investigation of neuropsychologic deficits in schizophrenia, relatively few investigators have characterized changes in neurocognitive
status associated with mood disorders. Although there has
been some growing interest in cognitive effects of depression in recent years (Murphy et al 1998; Sweeney et al

1998), surprisingly few studies have examined the neuropsychologic status of patients with bipolar disorder.
In unipolar depression, neuropsychologic deficits have
been reported in the domains of psychomotor speed (Sobin
and Sackeim 1997), memory (Burt et al 1995; Cohen et al
1982; Golinkoff and Sweeney 1989), sustained attention
(Cornblatt et al 1989), and executive functioning in domains
including working memory and complex problem solving
(Cassens et al 1990; Merriam et al 1999; Sweeney et al
1998). Several clinical and psychobiological correlates of
these changes have been demonstrated. Memory deficits
have been linked to dysregulation of the hypothalamic–
pituitary–adrenal axis, an effect believed to result from
adverse stress hormone effects on the hippocampus (Bemelmans et al 1996; Faustman et al 1990; Silberman et al 1985).
Clinical correlates of cognitive deficits include various indices of episode severity including psychotic (Albus et al 1996;
Gruzelier et al 1988; Nelson et al 1998), melancholic (Austin
et al 1999; Cornell et al 1984), and endogenous (Rush et al
1983) features. Deficits also appear to be more robust in
elderly patients (Deptula et al 1993; Harvey et al 1997;
Rubinow et al 1984). Because of widely reported but generally modest correlations between symptom severity and
neuropsychologic deficits in depressed patients, as well as

studies showing improvement in function after treatment
(Goldberg et al 1993), it has traditionally been accepted that
cognitive deficits in mood disorders are related to the acute
state of illness; however, there is growing consensus now that
some neuropsychologic deficits in depressed patients may
persist after symptom remission (Trichard et al 1995).
In bipolar patients, deficits in executive functions,
psychomotor skills, and memory have been reported
(Coffman et al 1990; Morice 1990; van Gorp et al 1998).
0006-3223/00/$20.00
PII S0006-3223(00)00910-0

Neuropsychology of Mood Disorders

In the relatively few studies that directly compared unipolar and bipolar patients, bipolar patients generally performed more poorly (Savard et al 1980; Wolfe et al 1987),
though this has not always been the case (Kessing 1998).
As in nonbipolar depression, some cognitive deficits
associated with bipolar disorder appear to persist into
remission (Clark et al 1999; Savard et al 1980).
Powell and Miklowitz (1994) and Veiel (1997) have

proposed that cognitive deficits in mood disorders primarily reflect frontal lobe dysfunction. A recently developed
computerized test battery, the Cambridge Neuropsychological Test Automated Battery (CANTAB), was designed
with a significant focus on neuropsychologic functions
subserved by frontostriatal circuitry (Fray et al 1997). This
test battery has procedural advantages such as the separation of mnemonic and strategic components of working
memory, as well as relatively strong data linking specific
performance deficits to focal brain abnormalities in animal
and human lesion studies. Whereas the CANTAB battery
emphasizes assessment of frontostriatal functions using
tests of spatial working memory, complex problem solving, and attentional set shifting, it also includes tests
sensitive to temporal lobe (Pattern Recognition and Delayed Match to Sample tests) and parietal lobe (Spatial
Span test) function.
When the CANTAB battery was used previously in
studies of depression, selected tests rather than the entire
test battery were administered. Deficits have been demonstrated in delayed recognition memory, problem solving,
spatial recognition memory, and attentional set shifting.
These effects have been most robust in elderly patients
(Beats et al 1996; Elliott et al 1997).
Impairments in Pattern Recognition Memory, Spatial
Recognition Memory, Delayed Matching to Sample, and

problem solving have been reported in acutely ill manic
patients (Murphy et al 1999). Murphy and colleagues
noted that these deficits were more severe than those that
they reported on previously in a study of unipolar depressed patients. Clark et al (1999) recently compared
euthymic bipolar patients with healthy subjects using
selected CANTAB tests, and reported disturbances in
sustained attention and attentional set shifting, but no
impairments in complex problem solving or spatial working memory.
The primary aim of our study was to directly compare
the cognitive function of bipolar patients in mixed/manic
or depressed phases of illness with nonbipolar depressed
patients and healthy subjects. We utilized the complete
main CANTAB battery, as well as its working memory
battery, to broadly characterize neuropsychologic deficits
associated with mood disorders in our sample of consecutively admitted inpatients.

BIOL PSYCHIATRY
2000;48:674 – 685

675

Methods and Materials
Subjects
Fifty-eight inpatients meeting DSM-IV (American Psychiatric
Association 1994) criteria for nonbipolar major depression, 35
inpatients meeting criteria for bipolar disorder (21 in depressed
phase and 14 in mixed or manic phase), and 51 healthy
comparison subjects were studied. The Structured Clinical Interview for DSM-IV Disorders (First et al 1997) sections relevant to
diagnosing mood and psychotic disorders were administered to
all subjects to determine diagnoses. Patients were recruited from
consecutive admissions to subspecialty inpatient units at the
Western Psychiatric Institute and Clinic, Pittsburgh, Pennsylvania. Healthy subjects were recruited from the community by
newspaper advertisements so as to maintain group matching on
age and IQ. These healthy individuals had never met criteria for
a mood or psychotic disorder, and they reported no knowledge of
affective disorders in first-degree relatives. No subject had a
history of electroconvulsive therapy, significant systemic or
neurologic disease, or recent substance abuse that might affect
cognitive performance. Clinical ratings were obtained close to
the time of testing using the Hamilton Rating Scale for Depression (17-item version; Hamilton 1960), the Bech–Rafaelsen

Mania Scale (Bech et al 1979), the Brief Psychiatric Rating Scale
(Overall and Gorham 1962), and the Scale for the Assessment of
Positive Symptoms (SAPS; Andreasen 1984), which assesses
psychotic symptoms. Demographic and clinical characteristics of
the subject groups are presented in Table 1.

Cognitive Testing
Tests from the CANTAB battery were presented on a highresolution touch-screen monitor under computer control. Patients
were tested as soon after admission as they were judged able to
participate in the cognitive assessments, typically within 3 days
(median time). The CANTAB tests have been described in detail
previously (Lawrence and Sahakian 1996; Robbins et al 1994)
and are described here briefly. The battery includes control tasks
designed to assess psychomotor speed, the ability to follow
instructions, and gross spatial perceptual skills, fundamental
abilities that can influence performance on the cognitive tasks.
The Big Circle/Little Circle task tests reaction time and the
ability to follow and then reverse rule-based responses. Subjects
are shown a pair of circles and instructed to touch the smaller
circle each time it appears for the first 20 trials, and then are told

to touch the bigger circle for the remaining trials. The Five Stage
Reaction Time task assesses psychomotor speed. Subjects are
told to hold down a press pad and release it and touch a yellow
dot on the screen as soon as it appears.

Tests Sensitive to Frontostriatal Dysfunction
STOCKINGS OF CAMBRIDGE. This task, a modified version of the Tower of London task, requires subjects to rearrange
colored balls in vertical columns to match a desired final
arrangement in a specified minimum number of moves. Subjects
are told to plan their sequence of moves before starting to move
the balls shown on the monitor. The time to plan the sequence of

676

J.A. Sweeney et al

BIOL PSYCHIATRY
2000;48:674 – 685

Table 1. Demographic and Clinical Characteristics of Mixed or Manic Bipolar Patients, Depressed Bipolar Patients, Nonbipolar

Depressed Patients, and Healthy Comparison Subjects

Age
Gender (M/F)
IQ
BR Mania
HRSD
BPRS
SAPS
Medications
(no. subjects)
Antipsychotics
Antidepressants
Lithium
Anticonvulsants
Benzodiazepines

Bipolar mixed/manic
patients [n 5 14,
mean (SD)]

Bipolar depressed
patients [n 5 21,
mean (SD)]

Nonbipolar depressed
patients [n 5 58,
mean (SD)]

Healthy individuals
[n 5 51,
mean (SD)]

F

p

36.14 (11.01)
6/8
102.08 (16.63)

16.00 (5.19)a,b
12.93 (9.09)a
32.93 (7.10)
1.93 (2.76)

31.90 (1.36)
12/9
100.60 (12.21)
1.35 (2.60)
17.30 (5.46)
29.47 (4.09)
0.32 (0.95)

32.29 (9.10)
19/39
102.93 (10.38)
0.07 (0.26)
21.64 (4.30)b
30.02 (5.95)
0.56 (2.18)

36.31 (9.69)
12/39
103.81 (8.93)
—
—
—
—

2.06
x2 5 8.03
0.43
257.69
15.64
1.68
2.84

.11
.05
.73
.001
.001
.19
.06

6
3
6
8
7

7
11
8
6
9

4
46
0
6
23

0
0
0
0
0

The number of patients receiving psychotropic medications is listed for each group. BR Mania, Bech–Rafaelsen Mania Scale; HRSD, Hamilton Rating Scale for
Depression; BPRS, Brief Psychiatric Rating Scale; SAPS, Scale for the Assessment of Positive Symptoms (sum of global ratings).
a
Significantly different from nonbipolar depressed patients.
b
Significantly different from depressed bipolar patients.

moves and the total number of moves made to solve the problem
are recorded.

Tests Sensitive to Temporal and Parietal Lobe
Dysfunction

In this self-ordered
searching task, subjects are asked to search through boxes on the
screen to discover which hides a token. Subjects are instructed
that once a token is discovered on a trial that box will not hide a
token on subsequent trials. Between-search errors (searching a
box in which a token was found on a previous trial) and
within-search errors (searching the same box twice in a trial) are
tabulated. Subjects’ ability to adopt a consistent search strategy
is also evaluated. Task difficulty is manipulated by increasing the
number of boxes presented in a block of trials (i.e., hiding targets
behind four, six, or eight boxes).

TASK.

SIMULTANEOUS AND DELAYED MATCHING TO SAMPLE
SPATIAL WORKING MEMORY TASK.

INTRADIMENSIONAL/EXTRADIMENSIONAL (ID/ED) AT-

This test was designed to
examine component processes evaluated in aggregate by the
Wisconsin Card Sorting Test (WCST). It assesses the ability to
maintain attention to different examples within a stimulus dimension (ID stages) and the ability to shift attention to a
previously irrelevant stimulus dimension (ED shifts as required
in the WCST). The number of stages/categories completed and
the number of errors before and during the ED set shift are
recorded.

TENTIONAL SET-SHIFTING TASK.

SPATIAL RECOGNITION MEMORY TASK. This task assesses the ability to remember the spatial location of visual
stimuli. Five squares are presented in sequence at different
locations on the screen, and then subjects are presented a pair of
squares and asked to identify which is at a location where a
square was previously presented.

This task assesses subjects’ ability to recognize complex visual designs after different delay intervals (0, 4, and 12
sec). Subjects are shown a target at screen center, and after the
delay interval four surrounding stimuli are presented. Subjects
are told to identify the design identical to the one presented
previously. In the Match to Sample simultaneous version of this
test, subjects are instructed to select the one of one, two, four, or
eight peripheral patterns that matches the pattern shown simultaneously at the center of the screen.

PATTERN RECOGNITION MEMORY TASK. This task assesses the ability to recognize abstract patterns. After a series of
patterns is displayed, subjects are presented pairs of patterns, one
shown previously and one being a novel pattern. They are asked
to indicate the pattern they were shown previously.
PAIRED ASSOCIATE LEARNING TASK. This task requires
subjects to learn the location of specific visual patterns. Designs
are presented in boxes on the screen at varying locations. The
designs are then presented sequentially in the center of the screen
and subjects are instructed to indicate the box in which each
design was initially presented.

SPATIAL SPAN TEST. This test of spatial short-term memory, similar to the Corsi block test, examines the ability to
remember the order in which visual stimuli are presented. Nine
white squares are presented on the monitor and the squares
change color one by one. Subjects are to indicate the order in
which the squares changed color. Subjects complete three trials

Neuropsychology of Mood Disorders

BIOL PSYCHIATRY
2000;48:674 – 685

677

Table 2. Performance of Mixed/Manic and Depressed Bipolar Patients, Nonbipolar Depressed Patients, and Healthy Comparison
Subjects on the CANTAB Neurocognitive Test Battery
Bipolar mixed/
manic patients
[mean (SD)]
Big Circle/Little Circle
Response latency (msec)
Five Stage Reaction Time
Reaction time (msec)
Movement time (msec)
Stockings of Cambridge
Problems solved in minimum moves
Spatial Working Memory
Strategy
ID/ED Shift
Stages completed
Pre-ED errors
EDS errors, adjusted
Spatial/Pattern Recognition Memory
Spatial percent correct
Pattern percent correct
Delayed Match to Sample
Latency, simultaneous (msec)
Latency, all delays (msec)
Match to Sample
Percent correct
Response latency (msec)
Paired Associate Learning
Total errors, adjusted
Spatial Span
Span length recalled

Bipolar depressed
patients
[mean (SD)]

Nonbipolar
depressed patients
[mean (SD)]

Healthy
individuals
[mean (SD)]

F

881 (162)

809 (152)

820 (176)

758 (129)

1.89

.13

404 (104)
595 (130)

385 (117)
570 (124)

380 (91)
620 (170)

354 (67)
602 (132)

1.07
0.36

.37
.78

6.79 (1.76)

7.19 (1.63)

7.67 (1.83)

8.16 (1.88)

3.44

.02

38.21 (3.40)a,b

36.10 (4.41)

34.50 (4.88)

33.20 (4.99)

5.58

.001

7.86 (1.66)
12.86 (10.65)
8.79 (11.68)

8.43 (1.72)
9.24 (5.84)
5.38 (7.19)

8.23 (1.85)
8.46 (5.80)
6.82 (8.32)

9.00 (.57)
7.44 (6.09)
4.40 (6.13)

1.45
2.10
1.35

.23
.10
.26

68.9 (16.5)a,b
81.0 (14.0)a,b

77.4 (15.4)
88.7 (8.9)

81.0 (12.1)
89.4 (10.4)

82.3 (11.0)
90.6 (7.9)

5.69
3.38

.001
.02

3,675 (1,277)a,b
3,542 (758)

3,340 (1,046)
3,460 (742)

2,969 (854)
3,562 (1,083)

2,834 (561)
3,549 (808)

3.82
0.18

.02
.91

95.2 (3.5)
2,442 (1,461)b,c

95.0 (5.5)
1,400 (803)

94.5 (5.5)a
1,338 (859)a

97.2 (2.6)
1,656 (546)

3.15
7.00

.03
.001

26.00 (32.33)

21.67 (40.39)

14.37 (21.54)

9.58 (14.33)

2.36

.08

5.36 (1.01)a

5.67 (1.43)

6.26 (1.60)

6.58 (1.26)

4.67

.004

p

CANTAB, Cambridge Neuropsychological Test Automated Battery; ID, intradimensional; ED, extradimensional.
Significantly different from healthy individuals.
Significantly different from nonbipolar depressed patients.
c
Significantly different from bipolar depressed patients.
a
b

each of sequences in which progressively more squares change
color (from two to nine squares). If subjects fail to correctly
remember the order of color change in three trials with a given
number of stimuli, then the test is terminated. The highest
number of items remembered in sequence is recorded to assess
performance.

Statistical Analysis
The primary statistical analyses were analyses of variance
(ANOVAs) contrasting subject groups. Repeated-measures
ANOVAs were conducted when tasks had difficulty manipulations (e.g., varying delay periods before memory was tested).
The Huynh–Feldt correction was used in these analyses. The
Tukey honestly significant difference post hoc test was used to
clarify group effects. Because there were group differences in
gender ratio, we treated gender as an independent factor in all
ANOVAs. Gender effects are reported in Results when significant. Skewed distributions (response latency data and Paired
Associate errors) were normalized (square root transformation)
before statistical analysis. The CANTAB battery calculates
average response latencies using the time to begin correct
responses, and that convention was followed in our analyses. For

the ID/ED Set Shift and the Paired Associate Learning tasks, we
followed the CANTAB method of adjusting the number of errors
such that if subjects failed to complete stages at a particular level
of task difficulty, more difficult test levels were not administered, and performance at those levels was assumed to be random
(50% correct). Exploratory correlational analyses were conducted to assess associations between test performance and both
clinical symptom ratings and demographic characteristics.

Results
Control Tasks
There were no group differences in response latencies on
the Big Circle/Little Circle test [F(3,134) 5 3.97, ns], and
there were no group differences in reaction time
[F(3,134) 5 2.62, ns] or movement time [F(3,134) 5
0.36, ns] on the Five Stage Reaction Time test (Table 2).
Thus, group differences on cognitive tests can be attributed primarily to differences in cognitive processing time
rather than to slow motor speed.

678

BIOL PSYCHIATRY
2000;48:674 – 685

J.A. Sweeney et al

Figure 1. Number of excess moves made when solving problems of varying task difficulty on the Stockings of Cambridge test by
mixed/manic and depressed bipolar patients, nonbipolar depressed patients, and healthy comparison subjects. Hatch marks represent
standard errors of measurement. aSignificantly different from healthy individuals. bSignificantly different from nonbipolar depressed
patients.

Tests Sensitive to Frontostriatal Dysfunction
STOCKINGS OF CAMBRIDGE. A diagnostic group by
gender by task difficulty repeated-measures ANOVA was
performed using the number of moves greater than the
minimum needed to solve problems to characterize test
performance. There was a significant group effect
[F(3,135) 5 3.74, p , .05] and group by task difficulty
interaction [F(5.7,258.3) 5 2.41, p , .05]. Post hoc
group comparisons indicated that the mixed/manic bipolar patients made significantly more excess moves to
solve problems than healthy individuals (p , .05). The
significant interaction term indicated that group differences became more pronounced as task difficulty increased (Figure 1). In analyses of initial thinking/
planning time, there were no group differences
[F(3,135) 5 1.10, ns] nor was there a significant task
difficulty by group interaction [F(7.0,314.7) 5 2.02,
ns]. There was a significant group by gender by task

difficulty interaction [F(7.0,314.7) 5 2.20, p , .05].
This interaction resulted from progressively longer
reaction times in male subjects, with increasing task
difficulty in healthy individuals and nonbipolar depressed patients, whereas in mixed/manic bipolar patients female subjects had progressively longer reaction
times as task difficulty increased.
SPATIAL WORKING MEMORY. A diagnostic group
by gender by task difficulty (number of boxes to be
remembered over sequential trials) repeated-measures
ANOVA was performed on the number of between search
errors. There was a significant group effect [F(3,135) 5
5.76, p , .01] and group by task difficulty interaction
[F(3.04,203.3) 5 3.26, p , .01]. Post hoc group comparisons indicated that mixed/manic bipolar patients performed significantly more poorly than healthy individuals
(p , .001) and nonbipolar depressed patients (p , .05).

Neuropsychology of Mood Disorders

BIOL PSYCHIATRY
2000;48:674 – 685

679

Figure 2. Between-search errors on the Spatial Working Memory test over varying levels of task difficulty by mixed/manic and
depressed bipolar patients, nonbipolar depressed patients, and healthy comparison subjects. Hatch marks represent standard errors of
measurement. aSignificantly different from healthy individuals. bSignificantly different from nonbipolar depressed patients.

The significant interaction term indicated that group differences became more pronounced as task difficulty increased (Figure 2). Significant group differences were also
observed on the strategy score, which reflects the consistency of planning responses over trials [F(3,135) 5 5.58,
p , .01]. Mixed/manic bipolar patients were less consistent in their search strategies than were nonbipolar depressed patients (p , .05) and healthy subjects (p , .01).
Neither the group effect [F(3,135) 5 0.97, ns] nor the
group by task difficulty interaction [F(4.26,191.7) 5 0.40,
ns] was significant for within search errors.
ID/ED ATTENTIONAL SET SHIFT. There were no
group differences in the overall number of stages completed [F(3,133) 5 1.45, ns] or in the number of errors

made before the ED shift condition [F(3,133) 5 2.10, ns].
There were also no significant differences between groups
in the number of errors made in the ED set shift condition
[F(3,133) 5 1.35, ns].
SPATIAL RECOGNITION MEMORY. There were significant group differences in spatial recognition accuracy
across groups [F(3,135) 5 5.69, p , .01]. Bipolar patients
in the mixed/manic phase of illness performed significantly more poorly than both nonbipolar depressed patients (p , .01) and healthy subjects (p , .01). A group by
gender interaction indicated that male subjects in both
bipolar groups made more errors than female subjects
[F(3,135) 5 4.18, p , .01], whereas minimal gender
differences were present in the other subject groups.

680

BIOL PSYCHIATRY
2000;48:674 – 685

J.A. Sweeney et al

Figure 3. Percent correct trials on the Delayed Match to Sample test over varying delay intervals achieved by mixed/manic and
depressed bipolar patients, nonbipolar depressed patients, and healthy comparison subjects. Hatch marks represent standard errors of
measurement. aSignificantly different from healthy individuals. bSignificantly different from nonbipolar depressed patients.

Tests Sensitive to Temporal and Parietal Lobe
Dysfunction
DELAYED MATCH TO SAMPLE. A diagnostic group
by gender by task difficulty (delay duration) repeated-measures ANOVA was performed using the percentage of trials
performed correctly to assess task performance. There was a
significant group effect [F(3,135) 5 10.88, p , .001] and
group by task difficulty interaction [F(9.0,405.0) 5 4.28, p ,
.001]. Post hoc group comparisons indicated that nonbipolar
depressed patients (p , .05), depressed bipolar patients (p ,
.01), and mixed/manic bipolar patients (p , .01) all performed more poorly than healthy subjects, and that the
mixed/manic phase patients performed more poorly than
nonbipolar depressed patients (p , .01). Group differences
became more pronounced as task difficulty increased (Figure
3). A significant group by gender interaction [F(3,136) 5
4.67, p , .01] indicated that, in healthy subjects and
nonbipolar depressed patients, male subjects performed better than female subjects, whereas male subjects performed
more poorly than female subjects in both bipolar groups.

On the Match to Sample task, groups differed in the
percentage of trials performed correctly [F(3,134) 5 3.15,
p , .05]. Post hoc tests revealed that the only significant
pairwise group difference was between the nonbipolar
depressed patients and healthy subjects (p , .01); however, these differences were very modest in size, with
depressed patients performing 94.5% of trials correctly, as
compared with 97.2% accuracy in healthy subjects. Also,
the reaction times of the depressed patients were relatively
fast, suggesting a speed/accuracy tradeoff rather than
significant visual perceptual disturbances.
PATTERN RECOGNITION MEMORY. There were significant group differences in the accuracy of encoding and
recognizing visual designs [F(3,135) 5 3.38, p , .05],
such that mixed/manic bipolar patients were less accurate
than nonbipolar depressed patients (p , .05) and healthy
subjects (p , .01).
PAIRED ASSOCIATE LEARNING. On the spatial
Paired Associate Learning test there was a significant

Neuropsychology of Mood Disorders

group difference in the number of errors made
[F(3,134) 5 2.77, p , .05], with mixed/manic patients
performing more poorly than healthy subjects (p , .05).
There also was a significant group by gender interaction
[F(3,134) 5 3.13, p , .05] due to male subjects in both
bipolar groups performing more poorly than female
subjects.
SPATIAL SPAN. Significant group differences in span
length were observed [F(3,135) 5 4.67, p , .01]. Mixed/
manic bipolar patients had significantly shorter spatial
spans than healthy subjects (p , .05).

Clinical and Demographic Status and Cognitive
Functioning
For mixed/manic bipolar patients, Bech–Rafaelsen mania
ratings were not significantly correlated with any
CANTAB measures. Task performance of depressed bipolar patients was not correlated with ratings of depression
(Hamilton depression ratings) or mania severity. In contrast, the severity of depression in nonbipolar depressed
patients was correlated with several aspects of cognitive
performance, including the total number of errors on the
Paired Associate Learning task (r 5 .40, p , .01), percent
correct on the Spatial Recognition Memory task (r 5
2.29, p , .05), problems solved in the minimum number
of moves required on the Stockings of Cambridge task
(r 5 2.41, p , .01), movement time on the Reaction Time
test (r 5 .41, p , .01), and response latency on the Match
to Sample test (r 5 .28, p , .05).
Severity of psychosis (sum of SAPS global ratings) was
correlated with response latency on several tasks, including the Match to Sample test in both the mixed/manic
bipolar group (r 5 .60, p , .05) and the depressed bipolar
group (r 5 2.47, p , .05), and the Big Circle/Little Circle
task (r 5 .35, p , .01) and the Delayed Match to Sample
task in the simultaneous condition (r 5 .29, p , .05) for
nonbipolar depressed patients. Also, in the nonbipolar
depressed patients psychosis ratings were related to the
number of errors on the ID/ED Shift task (r 5 .42, p ,
.01), number of errors on the Paired Associate Learning
task (r 5 .70, p , .001), percent correct on the Spatial
Recognition Memory task (r 5 2.37, p , .01), span
length on the Spatial Span task (r 5 2.40, p , .01),
between-trial errors on the Spatial Working Memory task
(r 5 .34, p , .01), and the number of problems solved in
the minimum moves required on the Stockings of Cambridge task (r 5 2.33, p , .05).
There was no pattern of significant correlations between
test performance and treatment with different classes of
psychotropic medication. Those treated with lithium did
not perform more poorly on any test in either bipolar

BIOL PSYCHIATRY
2000;48:674 – 685

681

group. The observation that bipolar depressed patients
receiving anticonvulsants made more errors on the Paired
Associate Learning task [t(19) 5 2.43, p , .05] was the
only association between anticonvulsant therapy and test
performance in any patient group.
Age was associated with test performance, especially in
the nonbipolar depressed patients. Excluding measures of
response latency, which were associated with age as
expected across groups, cognitive parameters associated
with age in the healthy subjects included only Spatial
Recognition Memory errors (r 5 2.33, p , .05) and
Spatial Working Memory strategy score (r 5 .28, p ,
.05). In contrast, in the nonbipolar depressed patients age
was associated with errors on Paired Associate Learning
(r 5 .44, p , .001), Spatial Span length (r 5 2.59, p ,
.001), Spatial Working Memory strategy score (r 5 .44,
p , .001), within-search errors (r 5 .29, p , .05) and
between-search errors (r 5 .35, p , .01) on the Spatial
Working Memory task, and Stockings of Cambridge
problems solved in the minimum number of moves (r 5
2.36, p , .01). In bipolar depressed patients age was
correlated with errors on the Paired Associate Learning
task (r 5 .46, p , .05), Spatial Span length (r 5 2.48,
p , .05), and the number of problems solved in the
minimum number of moves on the Stockings of Cambridge task (r 5 2.43, p , .05). There were no significant
correlations between age and test performance in the
mixed/manic bipolar patients.

Discussion
Results of our study indicate a pattern of widespread
neurocognitive disturbances in patients with bipolar disorder during mixed/manic phases of illness in domains
subserved by multiple cortical association areas. These
deficits were seen in executive functions, episodic memory, and spatial span performance, which suggest dysfunctions in frontostriatal systems, mesial temporal lobe systems, and the posterior parietal cortex, respectively. The
pattern of deficits in the manic/mixed bipolar patients
stands in contrast to the more limited and generally less
severe deficits demonstrated by both bipolar and nonbipolar depressed patients. Deficits in both groups of depressed
patients were confined to episodic memory functions,
suggesting a relatively restricted abnormality in mesial
temporal lobe memory systems during episodes of
depression.
Findings from this study provide important new evidence indicating that neuropsychologic deficits are significantly more pronounced during mixed/manic phases than
during the depressed phrase of bipolar illness, and influence a wide array of neocortical functions. In addition, our
data indicate that in bipolar patients, regardless of whether

682

BIOL PSYCHIATRY
2000;48:674 – 685

they were in a mixed/manic or a depressed phase of
illness, neuropsychologic deficits were greater in male
patients on tests including Spatial Recognition Memory,
Delayed Match to Sample, and Paired Associate Learning.
Thus, gender-related factors appear to modulate the severity of neuropsychologic deficits in bipolar disorder, a
pattern we did not detect in nonbipolar depression.
The wide array of neurobehavioral deficits observed in
the mixed/manic bipolar patients, in addition to the deficits
on the Delayed Match to Sample test that were present in
both groups of depressed patients, included performance
on tests of spatial and pattern recognition memory, spatial
span of attention, spatial working memory (both in retrieval and the consistent use of efficient response strategies), complex problem solving, and paired associate
learning. Their performance in the areas of spatial recognition memory, delayed matching to sample, pattern recognition memory, and response accuracy and use of
efficient strategies on the spatial working memory test was
impaired relative both to healthy subjects and to nonbipolar depressed patients. The one test of executive functions
on which bipolar patients did not demonstrate impairments
was on the ED attentional set-shifting task, indicating an
absence of significant inflexibility or perseveration in
problem-solving efforts.
Our findings of diverse cognitive impairments in bipolar patients are consistent with those reported by Murphy
and colleagues (1999) with the selected CANTAB tests
they administered. Murphy et al (1999) reported impairments in Delayed Match to Sample, on the Stockings of
Cambridge, and in both Spatial and Pattern Recognition
Memory during the manic phase of illness. The fact that
we observed these and other deficits in patients during the
mixed/manic phase but not in patients during the depressed phase of bipolar illness suggests that pronounced
neuropsychologic deficits of the mixed/manic group may
be related primarily to their state of illness at the time of
testing rather than to persistent stable neurocognitive
deficits. Consistent with the view that a significant component of the neuropsychologic deficits shown by mixed/
manic patients is related to their acute state of illness, a
recent study of euthymic bipolar patients performing
CANTAB tests found no impairments on the Stockings of
Cambridge and Spatial Working Memory tests (Clark et al
1999); however, Clark and colleagues did report an
impairment in attentional set-shifting in the ID/ED task in
their euthymic bipolar patients, a finding that is inconsistent with our own findings because we observed no
deficits on this test in our acutely ill sample. The ability of
acutely ill mixed/manic and depressed bipolar patients to
perform the attentional set-shifting task at a level not
significantly worse than healthy subjects in our study

J.A. Sweeney et al

suggests that deficits in this domain are not commonly
associated with bipolar disorder; however, longitudinal
studies of bipolar patients are needed to fully clarify the
discrepancy in findings on this particular test.
Previous studies of cognitive deficits in unipolar depression using the CANTAB battery have demonstrated impairments in psychomotor speed, memory, and executive
functions, but findings across studies have been somewhat
inconsistent. Robust patterns of deficits in multiple domains of function were reported in elderly depressed
patients (Beats et al 1996) and in depressed patients
recruited within the age range of 40 to 70 years (Elliott et
al 1996). In contrast, findings from a large sample of
young adult depressed patients (Purcell et al 1997) and
from our study indicate few cognitive deficits in younger
adult patients with nonbipolar depression.
It is noteworthy that our sample of young to midlife
nonbipolar depressed patients only demonstrated deficits on the Delayed Match to Sample task, indicating a
selective episodic memory disturbance. This finding is
consistent with those of previous neuropsychologic
studies suggesting a relatively selective disturbance of
episodic memory function during episodes of depression (Austin et al 1999; Calev and Erwin 1985;
Golinkoff and Sweeney 1989). An adverse stress steroid
effect on the hippocampus may mediate this selective
effect in depressed patients. Our results indicate that a
similar specific neurocognitive impairment affecting
episodic memory is associated with the depressed phase
of bipolar illness, suggesting that the effects of depression on memory function, and by inference on mesial
temporal lobe function, are similar in unipolar and
bipolar disorders.
Deficits in executive functioning were evident in bipolar patients in the mixed/manic state. The absence of such
deficits in both nonbipolar and bipolar depressed patients
was somewhat surprising because studies using other
neurophysiologic and neuropsychologic procedures have
demonstrated deficits in this domain (Freedman 1994;
Rush et al 1983; Sweeney et al 1998). The reasons for this
difference are not clear, but they most likely result from
the fact that executive function deficits associated with
depression are not common in younger adult patients,
except perhaps in those with particularly severe episodes
of illness.
Despite the minimal overall cognitive impairments in
the nonbipolar depressed patients, we did observe that age,
severity of overall depression, and ratings of psychotic
features were all correlated with performance of several
tests in these patients. This pattern of results suggest that
although generalized neuropsychologic deficits are not
typically prominent in young adults with nonbipolar depression, increasing age and more severe depression dur-

Neuropsychology of Mood Disorders

ing midlife may cause the pattern of more generalized
neuropsychologic deficits that have been observed in
elderly depressed patients. These data suggest an important interaction of aging and severity of illness effects on
the neuropsychologic function in midlife nonbipolar depression. This could be due to synergistic adverse effects
of depression severity and aging processes on brain
function, or to cumulative effects of more frequent prior
episodes of illness in older patients.
One important future research direction in this area is
the application of functional neuroimaging methods to
better clarify and localize the neural system disturbances
underlying cognitive deficits in mood disorders. Most
brain imaging studies of mood disorders have focused on
anatomy, resting-state brain physiology, or receptor physiology (Agren et al 1993; Drevets et al 1992; Mayberg et
al 1997), and few have used cognitive activation strategies. Investigating the interaction of biochemical, anatomic, and functional changes in cortical systems during
episodes of mood disorder may be a particularly useful
strategy for developing an integrated systems-level understanding of brain abnormalities in mood disorders. Another important line of work that needs to be pursued in
this relatively understudied area is longitudinal investigations of the course and persistence of neurocognitive
deficits in mood disorders, especially following mixed/
manic phases of bipolar disorder through acute episodes of
illness and remission (Altshuler 1993). Such neuropsychologic studies will not only help to differentiate state- and
trait-related deficits, but also to determine whether there is
any accelerated cognitive decline over the age span associated with the occurrence of more frequent or severe
episodes of illness, whether achieving prolonged states of
remission is associated with a gradual recovery of cognitive abilities, and whether different treatments have differential benefits for cognitive functioning.

This research was funded by Grants Nos. MH42969, MH45156,
MH01433, and MH30915 from the National Institute of Mental Health,
and by the John D. and Catherine T. MacArthur Foundation Research
Network on Psychopathology and Development.
The authors thank Lawrence Glanz, Ph.D., for recruiting and assessing
patients; Dawn Boarts, Jill Brown, Steve Hegedus, and Eli Merriam for
performing cognitive testing; and Tomika Cohen for administrative
assistance.

References
Agren H, Reibring L, Hartvig P, Tedroff J, Bjurling P, Lundqvist
H, et al (1993): Monoamine metabolism in human prefrontal
cortex and basal ganglia: PET studies using [b-11C]1-5hydroxytryptophan and [b-11C]L-dopa in healthy volunteers
and patients with unipolar major depression. Depression
1:71– 81.

BIOL PSYCHIATRY
2000;48:674 – 685

683

Albus M, Hubmann W, Wahlheim C, Sobizack N, Franz U,
Mohr F (1996): Contrasts in neuropsychological test profile
between patients with first-episode schizophrenia and firstepisode affective disorders. Acta Psychiatr Scand 94:87–93.
Altshuler LL (1993): Bipolar disorder: Are repeated episodes
associated with neuroanatomic and cognitive changes? Biol
Psychiatry 33:563–565.
American Psychiatric Association (1994): Diagnostic and Statistical Manual of Mental Disorders IV. Washington, DC:
American Psychiatric Association.
Andreasen NC (1984): Scale for the Assessment of Positive
Symptoms. Iowa City: University of Iowa Press.
Austin MP, Mitchell P, Wilhelm K, Parker G, Hickie I, Brodaty
H, et al (1999): Cognitive function in depression: A distinct
pattern of frontal impairment in melancholia. Psychol Med
29:73– 85.
Beats BC, Sahakian BJ, Levy R (1996): Cognitive performance
in tests sensitive to frontal lobe dysfunction in the elderly
depressed. Psychol Med 26:591– 603.
Bech P, Bolwig TG, Kramp P, Rafaelsen OJ (1979): The
Bech-Rafaelsen mania scale and the Hamilton depression
scale. Acta Psychiatr Scand 59:420 – 430.
Bemelmans KJ, Goekoop JG, van Kempen GM (1996): Recall
performance in acutely depressed patients and plasma cortisol. Biol Psychiatry 39:750 –752.
Burt DB, Zembar MJ, Niederehe G (1995): Depression and
memory impairment: A meta-analysis of the association, its
pattern, and specificity. Psychol Bull 117:285–305.
Calev A, Erwin PG (1985): Recall and recognition in depressives: Use of matched tasks. Br J Clin Psychol 24:127–128.
Cassens G, Wolfe L, Zola M (1990): The neuropsychology of
depression. J Neuropsychiatry Clin Neurosci 2:202–213.
Clark L, Goodwin GM, Iversen SD (1999): Frontal lobe function
in the euthymic phase of bipolar disorder. Soc Neurosci Abstr
390.
Coffman JA, Bornstein RA, Olson SC, Schwarzkopf SB, Nasrallah HA (1990): Cognitive impairment and cerebral structure by MRI in bipolar disorder. Biol Psychiatry 27:1188 –
1196.
Cohen RM, Weingartner H, Smallberg SA, Pickar D, Murphy
DL (1982): Effort and cognition in depression. Arch Gen
Psychiatry 39:593–597.
Cornblatt BA, Lenzenweger MF, Erlenmeyer-Kimling L (1989):
The continuous performance test, identical pairs version: II.
Contrasting attentional profiles in schizophrenic and depressed patients. Psychiatry Res 29:65– 85.
Cornell DG, Suarez R, Berent S (1984): Psychomotor retardation
in melancholic and nonmelancholic depression: Cognitive
and motor components. J Abnorm Psychol 93:150 –157.
Deptula D, Singh R, Pomara N (1993): Aging, emotional states,
and memory. Am J Psychiatry 150:429 – 434.
Drevets WC, Videen TO, Price JL, Preskorn SH, Carmichael T,
Raichle ME (1992): A functional anatomical study of unipolar depression. J Neurosci 12:3628 –3641.
Elliott R, Sahakian BJ, Herrod JJ, Robbins TW, Paykel ES
(1997): Abnormal response to negative feedback in unipolar
depression: Evidence for a diagnosis specific impairment.
J Neurol Neurosurg Psychiatry 63:74 – 82.

684

BIOL PSYCHIATRY
2000;48:674 – 685

Elliott R, Sahakian BJ, McKay AP, Herrod JJ, Robbins TW,
Paykel ES (1996): Neuropsychological impairments in unipolar depression: The influence of perceived failure on subsequent performance. Psychol Med 26:975–989.
Faustman WO, Faull KF, Whiteford HA, Borchet C, Csernansky
JG (1990): CSF 5-HIAA, serum cortisol, and age differentially predict vegetative and cognitive symptoms in depression. Biol Psychiatry 27:311–318.
First M, Spitzer RL, Gibbon M, Williams JBW (1997): Structured Clinical lnterview for DSM-IV Axis I Disorders—
Clinician Version (SCID-CV). Washington, DC: American
Psychiatric Press.
Fray PJ, Robbins TW, Sahakian BJ (1997): Neuropsychiatric
applications of CANTAB. Int J Geriatr Psychiatry 11:329 –336.
Freedman M (1994): Frontal and parietal lobe dysfunction in
depression: Delayed alternation and tactile learning deficits.
Neuropsychologia 32:1015–1025.
Goldberg TE, Gold JM, Greenberg R, Griffin S, Schulz SC,
Pickar D, et al (1993): Contrasts between patients with
affective disorders and patients with schizophrenia on a
neuropsychological test battery. Am J Psychiatry 150:1355–
1362.
Golinkoff M, Sweeney J (1989): Memory impairment in depression. J Affect Disord 17:105–112.
Gruzelier J, Seymour K, Wilson L, Jolley A, Hirsch S (1988):
Impairments on neuropsychologic tests of temporohippocampal and frontohippocampal functions and word fluency in
remitting schizophrenia and affective disorders. Arch Gen
Psychiatry 45:623– 629.
Hamilton M (1960): A rating scale for depression. J Neurol
Neurosurg Psychiatry 23:56 – 62.
Harvey PD, Powchik P, Parrella M, White L, Davidson M
(1997): Symptom severity and cognitive impairment in
chronically hospitalized geriatric patients with affective disorder. Br J Psychiatry 170:369 –374.
Kessing LV (1998): Cognitive impairment in the euthymic phase
of affective disorder. Psychol Med 28:1027–1038.
Lawrence AD, Sahakian BJ (1996): The neuropsychology of frontostriatal dementias. In: Woods RT, editor. Handbook of the
Clinical Psychology of Aging. New York: Wiley, 243–265.
Mayberg HS, Brannan SK, Mahurin RK, Jerabek PA, Brickman
JS, Tekell JL, et al (1997): Cingulate function in depression:
A potential predictor of treatment response. Neuroreport
8:1057–1061.
Merriam EP, Thase ME, Haas GL, Keshavan MS, Sweeney JA
(1999): Prefrontal cortical dysfunction in depression determined by Wisconsin Card Sorting Test performance. Am J
Psychiatry 156:780 –782.
Morice R (1990): Cognitive inflexibility and prefrontal dysfunction in schizophrenia and mania. Br J Psychiatry 157:50 –54.
Murphy FC, Sahakian BJ, O’Carroll RE (1998): Cognitive
impairment in depression: Psychological models and clinical
issues. Adv Biol Psychiatry 19:1–33.

J.A. Sweeney et al

Murphy FC, Sahakian BJ, Rubinsztein JS, Michael A, Rogers
RD, Robbins TW, et al (1999): Emotional bias and inhibitory
control processes in mania and depression. Psychol Med
29:1307–1321.
Nelson EB, Sax KW, Strakowski SM (1998): Attentional performance in patients with psychotic and nonpsychotic major
depression and schizophrenia. Am J Psychiatry 155:137–139.
Overall JE, Gorham DR (1962): The Brief Psychiatric Rating
Scale. Psychol Rep 10:799 – 812.
Powell KB, Miklowitz DJ (1994): Frontal lobe dysfunction in the
affective disorders. Clin Psychol Rev 14:525–546.
Purcell R, Maruff P, Kyrios M, Pantelis C (1997): Neuropsychological function in young patients with unipolar major depression. Psychol Med 27:1277–1285.
Robbins TW, James M, Owen AM, Sahakian BJ, McInnes L,
Rabbitt P (1994): Cambridge Neuropsychological Test Automated Battery (CANTAB): A factor analytic study of a large
sample of normal elderly volunteers. Dementia 5:266 –281.
Rubinow DR, Post RM, Savard R, Gold PW (1984): Cortisol
hypersecretion and cognitive impairment in depression. Arch
Gen Psychiatry 41:279 –283.
Rush AJ, Weissenburger J, Vinson DB, Giles DE (1983):
Neuropsychological dysfunctions in unipolar nonpsychotic
major depression. J Affect Disord 5:281–287.
Savard RJ, Rey AC, Post RM (1980): Halstead-Reitan Category
Test in bipolar and unipolar affective disorders. J Nerv Ment
Dis 168:297–304.
Silberman EK, Weingartner H, Targum SD, Byrnes S (1985):
Cognitive functioning in biological subtypes of depression.
Biol Psychiatry 20:654 – 661.
Sobin C, Sackeim HA (1997): Psychomotor symptoms of depression. Am J Psychiatry 154:4 –17.
Sweeney JA, Strojwas MH, Mann JJ, Thase ME (1998): Prefrontal
and cerebellar abnormalities in major depression: evidence from
oculomotor studies. Biol Psychiatry 43:584 –594.
Trichard C, Martinot JL, Alagille M, Masure MC, Hardy P,
Ginestet D, et al (1995): Time course of prefrontal lobe
dysfunction in severely depressed in-patients: A longitudinal
neuropsychological study. Psychol Med 25:79 – 85.
van Gorp WG, Altshuler L, Theberge DC, Wilkins J, Dixon W
(1998): Cognitive impairment in euthymic bipolar patients
with and without prior alcohol dependence: a preliminary
study. Arch Gen Psychiatry 55:41– 46.
Veiel HO (1997): A preliminary profile of neuropsychological
deficits associated with major depression. J Clin Exp Neuropsychol 19:587– 603.
Wolfe J, Granholm E, Butters N, Sanders E, Janowsky D (1987):
Verbal memory deficits associated with major affective disorders: A comparison of unipolar and bipolar patients. J
Affect Disord 13:83–92.