A pilot study of positive mood induction in euthymic
bipolar subjects compared with healthy controls
ANNE FARMER*, DOMINIC LAM, BARBARA SAHAKIAN, JONATHON ROISER,
AILBHE BURKE, NATHAN O’NEILL, SAM KEATING, GEORGIA POWELL SMITH
AND PETER MCGUFFIN
MRC Social, Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry, London, UK
Background. Demonstrating differences between euthymic bipolar subjects and healthy controls in
response to positive (happy) mood induction may help elucidate how mania evolves. This pilot
study evaluates the Go task in a reward paradigm as a method for inducing a happy mood state and
compares the response of euthymic bipolar subjects and healthy controls.
Method. The Sense of Hyperpositive Self Scale, the Tellegen positive and negative adjectives, the
Global-Local task and a visual analogue scale for measuring positive affect were administered to
15 euthymic bipolar subjects and 19 age-and-sex-matched healthy control subjects before and after
they had performed the Go task in a reward paradigm.
Results. Significant differences were found between subjects and controls on several measures at
each time-point but there were no differences across the groups across time except for the visual
analogue scales, where subjects had a more sustained duration in self-reported happiness compared
Conclusions. This pilot study has shown that a positive affect can be induced in bipolar subjects and
controls which can be demonstrated by changes in scores on several tasks. However, only the visual
analogue scales showed a significant difference between cases and controls over time. Such tests may
prove valuable in furthering understanding about the evolution of manic mood states.
Much research has focused on cognitive aspects
of depression but manic mood states, although
they have been far less intensively investigated,
also appear to be associated with altered
cognitions (Chamberlain & Sahakian, 2004).
Recent studies have shown that changes in
the fluency of thought, speech, learning and
memory impairment as well as disturbances in
association patterns and attentional processes
occur in the acute phase of mania, and per-
sist in remission in between one-quarter and
one-third of subjects (Martinez-Aran et al.
2004; Thompson et al. 2005). However, stan-
dard emotionally neutral tests of motor func-
tion, memory and executive function do not
differentiate between the cognitive changes
associated with mania from those associated
with depression. Consequently Sahakian and
colleagues (Roiser et al. 2003; Chamberlain &
Sahakian, 2004) have suggested that ‘hot’ or
emotion-dependent tasks may provide the key
to the cognitive changes specific to each mood
While there have been several studies ex-
amining the cognitive changes associated with
negative mood induction in both healthy
students and subjects with depression, exam-
ination of the cognitions associated with mania
* Address for correspondence: Professor Anne Farmer, Box 080,
MRC Social, Genetic and Developmental Psychiatry Research
Centre, Institute of Psychiatry, De Crespigny Park, Camberwell,
London SE5 8AF, UK.
Psychological Medicine, Page 1 of 6.
f 2006 Cambridge University Press
Printed in the United Kingdom
in response to positive mood induction has
been a relatively neglected area of research. One
exception is a study by Wright and colleagues
(Wright et al. 2005). The authors found that,
after positive mood induction using video clips,
bipolar patients scored significantly higher on
a 24-item dysfunctional attitudes scale than
remitted depressives and a normal control
group. However, one of the problems with a
video clips paradigm is that it can lead to cog-
nitive priming. Hence a more cognitively neutral
paradigm may be advantageous. The following
pilot study was undertaken to examine whether
it was possible to induce a positive mood using
the Go task in a reward paradigm and whether
once induced, significant differences in the cog-
nitions of euthymic bipolar subjects compared
with healthy controls could be demonstrated.
Fifteen subjects who had participated in a gen-
etic case-control study of bipolar disorder and
who fulfilled operational criteria for DSM-IV
bipolar 1 disorder volunteered to participate
in a pilot mood induction study. All bipolar
subjects had been previously interviewed using
the Schedules for the Clinical Assessment of
Neuropsychiatry (SCAN; Wing et al. 1990)
for psychopathology occurring in their worst
episode of mania and their worst episode of
depression and the results entered into the
CATEGO5 scoring program (Celik, 2003) to
assign a lifetime-ever diagnosis.
The Past History Schedule (McGuffin et al.
1986) was administered by phone to establish
whether 19 volunteer age- and sex-matched
control subjects had ever suffered or were cur-
rently suffering from any mental disorder.
All participants gave written informed con-
sent and were screened with the Beck De-
pression Inventory (BDI; Beck et al. 1961) and
the Altman Mania Scale (AMS; Altman et al.
1997) by telephone one week prior to testing,
as well as on the day of the mood induction
procedure. (The short-form BDI was admin-
istered by phone and the long-form completed
on the day of testing.)
On the day of testing, all participants also
completed the National Attainment Reading
Test (NART; Nelson, 1982) to provide a proxy
measure of IQ. In addition, they also undertook
the following before and after completing the
Go task (Johnson et al. 2005) in a reward para-
(1) The Sense of Hyper-Positive Self Scale
(SHPSS; Lam et al. 2005b). Subjects are
asked to describe, first how they are ‘most
of the time’ and second how they ‘would
like to be’ by responding to seven positive
adjectives: confident, dynamic, etc., on a six-
point scale from ‘not at all’ to ‘extremely’.
(Watson et al. 1988). Subjects rate 10 ad-
jectives of high positive affect and 10 of high
negative affect a five-point scale rating from
‘not true at all’ to ‘very true about me’.
(3) The Global-Local processing task (Witkin
et al. 1971; Shah & Frith, 1983). The ability
to identify whether a small figure is embed-
ded in a larger one.
(4) Visual analogue scales on four occasions
throughout the testing period (subjects
placed a cross on a 10 cm line to indicate
how happy or sad they felt.
The Go task was programmed using a serial
response box with five lamps and five corre-
sponding buttons. Subjects started each trial by
pressing a centre button, after which a randomly
selected lamp lit. The subject is then asked to
press the button below that lamp as fast as
possible. Subjects’ reaction times were recorded
in a computerized program. A practice trial,
consisting of five button pressings was followed
by two 2-minute button pressings without any
monetary rewards (neutral condition). Subjects
were then given feedback that their speed was
‘very fast’. After these trials in neutral con-
dition, subjects were encouraged to press as
many buttons as they could in 2 minutes, and
were paid 10 pence for each button pressed
correctly. The program was set so that subjects
were given feedback that they performed accu-
rately 70% of the time.
Age, gender, BDI, Altman Mania Scale (AMS),
NART and overall Go task scores for cases
The mean age for 15 cases was 44.4 years [stan-
dard deviation (S.D.)=13.42 years] and for 19
controls was 35.42 years (S.D.=13.60 years).
2A. Farmer et al.
Sixty-seven per cent of cases and 84% of
controls were female. The mean score on the
short-form BDI undertaken 1 week prior to
testing was 2.60 (S.D.=2.92) for cases and 1.00
(S.D.=1.77) for controls. Altman Mania Scale
(AMS) scores were 1.87 (S.D.=1.77) for cases
and 1.93 (S.D.=1.75) for controls. NART scores
were 39.80 (S.D.=5.88) for cases and 39.95
(S.D.=5.78) for controls. There were no signi-
ficant differences between cases and controls for
age, sex, BDI (short form), AMS or NART.
However, there were significant differences
between cases and controls for the long version
of the BDI undertaken on the day of testing.
Mean BDI for cases was 5.00 (S.D.=5.98)
and for controls was 1.37 (S.D.=2.67) [t test
(t)=2.23, degrees of freedom (df) 17.19, p=
0.04]. Repeated measure analysis of variance
(ANOVA) showed significant differences be-
tween cases andcontrols forBDI score over time
(F=8.87, p<0.01) but there was no significant
group by time interaction (F=3.19, p=N.S.).
For the Go task, cases made significantly
fewer responses compared with controls. Mean
number of responses for cases was 399.29
(S.D.=61.48) and for controls was 460.00
(S.D.=42.98) (t=x3.12, df=22.59, p<0.01).
However when age was added to the analysis
as a covariate, the significant group difference
disappeared (group: F=2.11, p=N.S.; age:
Cases were also significantly slower than
controls for both the non-reward as well as the
rewarded 2 minutes. Average response time, in
milliseconds, at the non-rewarded 2 minutes for
cases was 630.60 (S.D.=113.00) and for controls
was 542.13 (S.D.=61.74) (t=2.62, df=19.23,
p=0.02). For the rewarded 2 minutes average
response time, in milliseconds, for cases was
627.30 (S.D.=111.36) and for controls was
552.07 (S.D.=59.97) (t=2.27, df=19.07, p=
0.04). However, these significant group differ-
ences again disappeared when age was added
as a co-variate (non-reward 2 minutes: group:
F=1.15, p=N.S.; age: F=13.42, p<0.05; re-
ward 2 minutes: group: F=0.36, p=N.S.; age:
F=15.86, p<0.01). There were no significant
differences between cases and controls for per-
centage correct, percentage wrong or percentage
SHPSS, Tellegen adjectives and Global-Local
task in cases and controls pre- and post-mood
The mean scores on these measures pre- and
post-mood induction (MI) are shown in Table 1.
For the SHPSS, Table 1 shows that there was
no significant difference in overall scores be-
tween cases and controls for either pre- or post-
MI. Also repeated measures ANOVA with age
as a co-variate indicates that there was no effect
of MI on scores for both cases or controls
(F=1.13, p=N.S.) and no MI by group
(F=0.18; p=N.S.) or MI by age interaction
Table 1 shows that cases endorsed signi-
ficantly more Tellegen negative adjectives as
being true about themselves than controls both
pre- (t=2.54, df=20.48, p=0.02) and post-
(t=2.13, df=21.35, p=0.05) mood induction,
although there were no significant differences
adjectives. For negative adjectives there is an MI
effect of lowering scores (F=9.81, p<0.01) in
both cases and controls but there is no signifi-
cant MI by group (F=1.26, p=N.S.) or MI by
age interaction (F=0.16; p=N.S.).
On the Global-Local task, Table 1 shows that
there were significant differences between cases
and controls on both testing occasions for both
Table 1. Mean scores, standard deviations and significance levels for cases and controls for
measures taken pre- and post-mood induction
SHPSS overall score (S.D.)
Tellegen positive adjectives (S.D.)
Tellegen negative adjectives (S.D.)
Global-Local time taken (S.D.)
Global-Local number correct (S.D.)
Positive mood induction in euthymic bipolar subjects3
average time taken (pre-MI: t=2.67, df=20.90,
p=0.01; post-MI: t=2.78, df=19.71, p=0.01)
and average number correct (pre-MI: t=x2.11,
df=14.71, p=0.02). Both cases and controls
improved their speed at the task following MI
(F=4.36, p=0.05) and there was also a signifi-
cant effect of age on speed over time (F=15.00,
p<0.01) but there were no significant group by
MI (F=0.00, p=N.S.) or age by MI (F=3.48,
p=N.S.) interactions. Both cases and controls
improved on the number correct but this was
not significant (F=3.72, p=0.06), although the
effect due to age was (F=6.52, p=0.02). There
were no interactional effects for group by MI
(F=0.35, p=N.S.) or for MI by age (F=1.73,
Visual analogue scales pre- and post-mood
The mean scores on the visual analogue (VA)
scales 1 through 4 for cases and controls are
shown in Table 2. The Go task in the reward
paradigm came between VA ratings 1 and 2.
The table shows that cases rated themselves as
somewhat less happy than controls before MI
although this was not statistically significant.
Mean scores for both cases and controls
increased following MI, indicating that all sub-
jects were happier. However, again there were
no statistically significantly score differences
between cases and controls. After completion of
the second Global-Local task (VA4), the mean
VA scores for controls had returned to just be-
low that of the start of testing while mean scores
for cases remained elevated. This was reflected
in the repeated measures ANOVA that showed
that both cases and controls became happier
following MI (F=4.01, p=0.05) and that there
was a significant MI by group interaction
The SHPSS showed no differences in the scores
of cases and controls at either time of test-
ing although repeated measures ANOVA has
shown that scores for all subjects became
more positive following MI. However cases and
controls did not differ in the magnitude of
positive self-regard following MI.
In contrast, the Tellegen negative adjectives,
and the Global-Local task show significant dif-
ferences between cases and controls, at both
times of testing. Such case-control differences
have been shown in other studies of euthymic
bipolar subjects (Thompson et al. 2005). In
addition, in the Global-Local task, repeated
measures ANOVA showed that all subjects had
an enhanced ability to perform this task both
correctly and speedily following MI, which
could have been due to a practice effect rather
than the impact of mood elevation. However,
although both the Tellegen and Global-Local
tasks showed greater improvement in scores in
cases compared with controls following MI
(the effect predicted by our hypothesis), these
failed to achieve statistical significance, which
may have been due to the small sample size of
the pilot study.
The VA scales provide a simple self-rating
of ‘happiness’ which has shown a difference
in response to positive MI between cases and
controls. Table 2 shows that although cases
were unhappier at the start of the testing
procedure compared with controls, their scores
rose following MI and remained elevated
while controls show a rise in happiness ratings
which declined again back to (below) pre-
induction levels by the end of the testing
procedure. This finding of a prolonged happy
mood response after positive MI is consistent
with Goplerud & Depue’s (1985) finding that
cyclothymic participants had a more prolonged
duration of recovery of mood, behaviour and
cognitions than normal controls after a natu-
ralistic stressful event. Although we have not
shown any statistically significant differences
between cases and controls for individual visual
analogue ratings, repeated measures ANOVA
showed that the change over time in ratings was
The MI paradigm employed in this study has
demonstrated not only that both cases and
analogue (VA) scales for cases and controls
Mean scores on four visual
aGo task in reward condition occurred following this rating.
4A. Farmer et al.
controls respond to positive MI but also that the
response shown by euthymic bipolar subjects
is different to that of controls in being more
sustained. Thus our findings support Depue’s
Behavioural Activation theory (BAS) theory of
bipolar disorder (Depue et al. 1981). An
alternative explanation is that the euthymic
bipolar subjects started from a lower baseline
and consequently had more scope for mood
elevation. Indeed, VA1 shows that bipolar sub-
jects do have lower scores than the controls
although these do not differ significantly.
However, what is particularly notable about the
VA scores over time is that mood elevation
persists in bipolar subjects so that by VA4
bipolar subjects scores are remaining high while
scores for controls have come down to pre-MI
levels. Thus we suggest that it is the sustained
elevated mood following positive MI that
distinguishes bipolar subjects from controls
rather than the level of their self-reported mood
prior to MI.
However, our results need to be considered in
light of some limitations to the study design.
First, we have not included a psychiatric control
group, for example subjects with remitted uni-
polar depression, which would have allowed the
identification of vulnerability factors that are
unique to bipolar disorder. Secondly, in order to
distinguish between the effects of mood change
and a practice effect, a negative MI paradigm
could have been added to the positive MI,
counterbalancing the order of presentation.
Thirdly, we have not included a control con-
dition for unrewarded activity. Hence we cannot
say for certain that the mood improvement was
entirely due to the rewarding properties of the
Despite these caveats and the small size of a
pilot study, we suggest that our results show
that following positive MI, significant differ-
ences can be found between the response of
euthymic bipolar subjects and healthy controls.
A detailed analysis of the changes in mood,
behaviour and cognition that occur in an
elevated mood state will improve our under-
standing of how mania unfolds. Not only will
this have important implications for the devel-
opment of psychological treatments for bipolar
disorder (Lam et al. 2005a) but also the MI
paradigm used in this pilot study is relatively
simple and could be readily employed for
examining the brain function in mania using
The results of this study have shown that a
positive or ‘happy’ mood can be induced in
both control and euthymic bipolar subjects
using the Go Task in a reward paradigm which
can be demonstrated by increases in ‘happiness’
ratings on scales such as the SHPSS, Tellegen
adjectives, the Global-Local task and serial
visual analogue ratings. However, only the
visual analogue ratings showed significant dif-
ferences between cases and controls following
MI, with cases having a sustained elevation of
mood while in controls mood returns to normal.
DECLARATION OF INTEREST
The subjects who agreed to participate in this
pilot study had already participated in a case-
control genetic study funded by Glaxo Smith
Farmer and Peter McGuffin have received con-
sultancy fees and honoraria for presentations,
chairing, and participating in expert panels, plus
travel and subsistence costs from Glaxo Smith
Kline, Eli Lilly, Bristol Myers Squibb and
Altman, E. G., Hedeker, D., Peterson, J. L. & Davis, J. M. (1997).
The Altman Self-Rating Mania Scale. Biological Psychiatry 42,
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. & Erbaugh, J.
(1961). An inventory for measuring depression. Archives of
General Psychiatry 4, 561–571.
Celik, C. (2003). Computer assisted personal interviewing application
for the Schedules for Clinical Assessment in Neuropsychiatry,
Version 2.1, and diagnostic algorithms for WHO ICD-10 Chapter V
DCR and for statistical manual IV. Release 1. Ed. 126.96.36.199. Win 9x
NT. World Health Organisation: Geneva.
Chamberlain, S. R. & Sahakian, B. J. (2004). Cognition in mania and
depression: psychological models and implications. Current
Psychiatry Reports 6, 451–458.
Depue, R. A., Slater, J. F., Wolfstetter-Kausch, H., Klein, D.,
Goplerud, E. & Farr, D. (1981). A behavioral paradigm for iden-
tifying persons at risk for bipolar depressive disorder: a conceptual
framework and five validation studies. Journal of Abnormal
Psychology 90, 381–437.
Goplerud, E. & Depue, R. A. (1985). Behavioural response to natu-
rally occurring stress in cyclothymia and dysthymia. Journal of
Abnormal Psychology 94, 128–139.
Johnson, S. L., Ruggero, C. J. & Carver, C. S. (2005). Cognitive,
behavioral, and affective responses to reward: links with
Positive mood induction in euthymic bipolar subjects5
hypomanic symptoms. Journal of Social and Clinical Psychology
Lam, D. H., Hayward, P., Watkins, E. R., Wright, K. & Sham, P.
(2005a). Relapse prevention in patients with bipolar disorder:
cognitive therapy outcome after 2 years. American Journal of Psy-
chiatry 162, 324–329.
Lam, D. H., Wright, K. & Sham, P. (2005b). Sense of hyper-positive
self and response to cognitive therapy in bipolar disorder.
Psychological Medicine 35, 69–77.
Martinez-Aran, A., Vieta, A., Colom, F., Torrent, C., Sanchez-
Moreno, J., Reinares, M., Benabarre, A., Goikolea, J. M., Brugue,
E., Daban, C. & Salamero, M. (2004). Cognitive function across
manic or hypomanic depressed and euthymic states in bipolar
disorder. American Journal of Psychiatry 161, 262–270.
McGuffin, P., Katz, R. & Aldrich, J. (1986). Past and present state
examination: the assessment of ‘lifetime ever’ psychopathology.
Psychological Medicine 16, 461–465.
Nelson, H. E. (1982). National Adult Reading Test, NART. Nelson:
Roiser, J. P., Rubensztein, J. S. & Sahakian, B. J. (2003). Cognition
in depression. Psychiatry 2, 43–47.
Shah, A. & Frith, U. (1983). An islet of ability in autistic children: a
research note. Journal of Child Psychology and Psychiatry 24,
Thompson, J. M., Gallagher, P., Hughes, J. H., Watson, S., Gray,
J. M., Ferrier, I. N. & Young, A. H. (2005). Neurocognitive im-
pairment in euthymic patients with bipolar affective disorder.
British Journal of Psychiatry 186, 32–40.
Watson, D., Clark, L. A. & Tellegen, A. (1988). Development and
validation of brief measures of positive and negative affect: the
PANAS Scale. Journal of Personality and Social Psychology 54,
Wing, J. K., Babor, T., Brugha, T., Burke, J., Cooper, J. E., Giel, R.,
Jablenski, A., Regier, D. & Sartorius, N. (1990). SCAN. Schedules
for Clinical Assessment in Neuropsychiatry. Archives in General
Psychiatry 47, 589–593.
Witkin, H. A., Ottman, P. K., Raskin, E. & Karp, S. (1971). A Manual
for the Embedded Figure Test. Consulting Psychologists Press:
Mountain View, CA.
Wright, K., Lam, D. H. & Newsom-Davis, I. (2005). Induced mood
and dysfunctional attitudes in remitted bipolar I affective disorder.
Journal of Abnormal Psychology 114, 689–696.
6A. Farmer et al.