Psychology
2012. Vol.3, Special Issue, 75 8-761
Published Online September 2012 in SciRes (http://www.SciRP.org/journal/psych) http://dx.doi.org/10.4236/psych.2012.329115
758
Working Memory in Patients with Major Depressive Disorder
Yasuhiro Ka neda
Department of Psychiatry, Iwaki Clinic, Tokushima, Japan
Received May 2nd, 2012; revised June 4th, 2012; accepted July 7th, 2012
Objective: Patients with major depressive disorder (MDD) have been reported to perform less well in
neurocognitive tests than normal control subjects. The author tested the hypotheses that verbal working
memory (WM) is predictive of the functional outcome in patients with MDD. Methods: In this naturalis-
tic longitudinal study, the subjects consisted of 22 adult outpatients receiving paroxetine as antidepressant
therapy. Functional outcome was rated on a scale of 0 (non-impaired) to 3 (severely impaired). Results: 1)
At 12 weeks, nine of the 22 patients currently experiencing MDD exhibited full remission; 2) signifi-
cantly decreased 7-item Hamilton Rating Scale for Depression (HAM-D7) scores were observed during
the 12-week study period, while Digit Sequencing Task (DST) scores increased significantly; 3) at base-
line, functional outcomes correlated significantly with HAM-D7 scores, but, at 12 weeks, correlated sig-
nificantly with bo th HAM-D7 and DST scores. Furthermore, when looking at only pa tients in full or partial
remission (mild depression), functional outcome correlated more strongly with DST than with HAM-D7
scores. Conclusions: A deficit of verbal WM correlated with the functional outcome after treatment in pa-
tients with MDD. Antidepressant therapy with paroxetine might contribute to improvement of verbal WM.
Keywords: Depression; Functional Outcome; Paroxetine; Symptomatlogy; Working Memory
Introduction
Major depressive disorder (MDD) is a significant health
problem, with major economic implications, and estimates of
the economic burden of depression range from $52 billion in
1990 to $83 billion in 2000 (Malone, 2007). Among others, the
effect on employment is considered to have a great impact on
the societal costs of depression, due to lost income, lost produc-
tivity, and disability income payments.
In previous studies (Kaneda et al., 2009, 2010), the author
demonstrated that neurocognitive performance was more im-
portant than the clinical symptoms in predicting the future em-
ployment status in patients with schizophrenia, and among the
neurocognitive functions, verbal working memory (WM) was
found to be the most important for determining the employment
outcome. Patients with MDD also have been reported to per-
form less well in neurocognitive tests than normal control sub-
jects, even after their depression is successfully treated with
newer-generation antidepressants (Gualtieri et al., 2006, Rep-
permund et al., 2009, Porter et al., 2003). Recently, Gualtieri &
Morgan (2008) reported that substantial numbers of patients
with depression exhibit cognitive impairment. However, until
date, little attention has been paid to the relation between neu-
rocognitive performance and the psychosocial or functional
outcomes in studies of depression compared to those of bipolar
disorder (Kennedy et al., 2007). Kitagawa & Koyama (2009)
have suggested a relationship between functional outcomes and
neurocognitive functions in depression, but their result has not
been duplicated.
Thus, the author first tested the hypothesis that the deficit of
neurocognitive performance, namely verbal WM, existed even
after remission (Kaneda, 2009b). In this naturalistic cross-sec-
tional study, the subjects consisted of 54 clinic adult out-patients
and 54 age- and sex-equated healthy comparison subjects. The
assessments were performed using the 7-item Hamilton Rating
Scale for Depression (HAM-D7) for the severity of depression,
and the Digit Sequencing Task (DST) for evaluation of verbal
WM (methodological details are described below). The DST
scores were significantly less in both patients with current epi-
sode of MDD and in full remitted or partial remitted (mildly
depressed) patients than in controls. Also there were no signify-
cant correlations between DST scores and the dose of antide-
pressants or benzodiazepines in full remitted or partial remitted
(mildly depressed) patients.
Since verbal WM was demonstrated to be poorly performed
in patients with MDD, another study was conducted to verify
the suggestion that verbal WM is predictive of the functional
outcome in patients with MDD after remission.
Methods
Subjects
Twenty-two consecutive adult outpatients (21 - 64 years of
age) were enrolled in this prospective, open-label study. All
subjects satisfied the DSM-IV (American Psychiatric Associa-
tion, 1994) criteria for a current episode of unipolar MDD (non-
psychotic) and were receiving paroxetine with or without sul-
piride or benzodiazepines. Patients had no co-morbid psychiat-
ric disorders, nor any medical, neurological, or developmental
conditions that could affect cognition (e.g., attention deficit
hyperactivity disorder, brain injury, mild cognitive impairment,
chronic pain). The investigation was carried out in accordance
with the Declaration of Helsinki, and informed consent was
obtained from all subjects.
Assessments
The assessments were performed using the following instru-
ments: 1) the 7-item Hamilton Rating Scale for Depression

Y. KANEDA
(HAM-D7; McIntyre et al., 2005) to determine the severity of
depression and remission status; full remission was defined as a
score of 3 or less, partial remission (mild depression) as a score
of 10 or less on the HAM-D7, and 2) the Brief Assessment of
Cognition in Schizophrenia (BACS; Keefe et al., 2004) Digit
Sequencing Task (DST) to assess verbal WM (the patients were
presented with clusters of numbers in random order of increas-
ing length, and asked to recount these numbers in the right or-
der, from lowest to highest, to the investigator). The BACS
DST has been validated in normal control subjects (Keefe et al.,
2004). The DST scores in each depression group were normal-
ized for the respective age- and sex-matched control groups
(data available on request). Functional outcome (productivity),
including the ability to go to work, do household chores, go to
school and so on, was assessed by the author based on inter-
views with patients and their partners/parents/children, and was
graded on a 0 - 3 scale: 0 = non-impaired, 1 = mildly impaired,
2 = moderately impaired, 3 = severely impaired. The clinical
assessments were performed on two occasions: 1) on the day of
entry and 2) after approximately 12 weeks (a mean period of
90.6 days, SD = 14.2).
Analyses
The JMI software (Version 8.0. 1) for Macintosh was used to
perform the analyses. For numerical variables, the t-test for
independent group comparisons was used to compare differ-
ences in variables between the two groups. The clinical assess-
ment scores were compared between the two assessments by re-
peated-measure analysis of variance (ANOVA). Pearson’s cor-
relation was used to examine the relationships between two
numerical variables. A logistic regression model with forward
selection criteria was used to predict functional outcomes using
functional outcome as a dependent variable and depressive and
verbal WM scores as independent variables. The level of sig-
nificance was set at p < .05.
Results
The demographic characteristics at baseline are presented in
Table 1. At baseline, eight of the 22 patients (36%) were taking
paroxetine, and two were taking sulpiride. In addition, 12 of the
22 patients (55%) were taking benzodiazepines. At the time of
the second assessment, all patients were receiving treatment
with paroxetine (mean dose, 28.8 mg), and one patient was tak-
ing sulpiride in addition to paroxetine. Fourteen of the 22 pa-
tients (64%) were being treated with benzodiazepines (mean
dose, 4.7 mg).
At the second assessment conducted after 12 weeks, accord-
ing to symptom severity scores, nine of the 22 patients (41%)
with a current episode of MDD were in full remission. In addi-
tion, normalization of work function was seen in eight patients
of these nine patients (89%).
A significant decrease in HAM-D7 scores from 11.0 ± 3.8 to
5.5 ± 4.9 (p < .0001) was observed during the 12-week study
period. In addition, DST scores (z-score) increased significantly
during the same period (from –1.1 ± 1.1 to –.7 ± 1.4, p < .05).
At baseline, functional outcomes were correlated signifi-
cantly with HAM-D7 scores (r = .56, p < .01), but not DST
scores, whereas at 12 weeks, functional outcomes correlated
significantly with both DST (r = –.46, p < .05) and HAM-D7 (r
= .72, p < .001) scores. When looking only at patients in full or
partial remission (mild depression), functional outcome corre-
lated more strongly with DST (r = –.74, p < .001) than with
HAM-D7 (r = .51, p < .05) scores.
There was no significant relation between DST and HAM-
D7 scoresat either baseline or at 12 weeks.
According to multiple regressi on analysis with a forward step-
wise procedure, the DST scores at baseline contributed more
strongly to the prediction of functional outcome at 12 weeks (F
= 1.9, df = 1, p = .19) than did HAM-D7 scores (F = .1, df = 1,
p = .74). On the other hand, HAM-D7 scores at baseline contri-
buted more strongly to the prediction of HAM-D7 scores at 12
weeks (F = 4.8, df = 1, p < .05) than did DST scores (F = 1.2, df
= 1, p = .29).
Discussion
The findings of this study suggest a correlation between a
verbal WM deficit associated with MDD and post-treatment
functional outcome. When looking at only patients in full or
partial remission (mild depression), the correlation was more
apparent. These findings are consistent with those of a previous
cross-sectional study carried out by the author (Kaneda, 2009a).
Kitagawa & Koyama (2009) have also suggested a relationship
between functional outcomes and neurocognitive functions in
depression. In their review, Kennedy et al. (2007) suggested
that residual symptoms after remission of depression may lead
to enduring psychosocial impairment, as many subtle neurocog-
nitive deficits. While the findings of this study do not under-
score the importance of clinical remission from depression, a
defined objective outcome indicated by a quantifiable score
with a depressive symptom measurement tool, symptomatic full
remission should always be the primary goal of treatment, since
it is the optimal outcome for patients with depression (McIntyre
et al., 2005). However, the present findings do suggest that en-
hancement of verbal WM function, e.g., via cognitive rehabili-
tation, may facilitate normalization of functioning which is an
important component of remission (Zimmerman et al., 2006).
In terms of predictive factors, previous studies have de-
scribed the value of neurocognitive functions for predicting
symptomatic improvement, mainly in geriatric patients with
depression (Marcos et al., 2005). However, in this study,
HAM-D7 scores were found to be more important for predict-
ing symptomatic improvement than DST scores used to assess
verbal WM, a finding well in line with that reported by Biringer
et al. (2007), who showed that neurocognitive function at base-
line did not predict amelioration of depressive symptoms over
time in young adults. The present results, however, suggest
verbal WM to be more important for predicting functional out-
Table 1.
Demographic Data.
N (F/M) Age (yr.) Education (yr.) Age at onset (y r. )Antidepressa n t dose (mg/d)a Benzodiazepines dose (mg/d)b
22 (9/13) 37.9 (13.3) 12.5 (2.9) 36.8 (13.0) 7.7 (13.3) 2.0 (3.2)
Note: Data are expressed as mea ns (SD); aParoxetine equivalent; bDiazepam equivalent.

Y. KANEDA
comes than depressive symptoms. Therefore, in order to obtain
a better functional outcome, it may be important to place more
emphasis on pretreatment verbal WM than depressive symp-
toms.
The findings of this study also suggest again that MDD-as-
sociated deficits in verbal WM exist both in acute depression
and after the treatment of depression. These findings are con-
sistent with those of previous studies conducted by the author
(Kaneda, 2009b) and by Nebes et al. (2003), who found that
verbal WM dysfunction persisted in older depressed patients
even after their mood disorder had responded to antidepressant
medications. The observations made herein may be explained
by an impairment of WM/central executive functions in MDD,
as suggested by Rose & Ebmeier (2006), since executive func-
tion impairment is considered to be, at least to some degree,
trait-related (Porter et al., 2003).
Regarding the effects of antidepressant medications on cog-
nitive function, a significant increase in DST scores was ob-
served during 12-week treatment with paroxetine in the present
study. Although a potential effect which would be maximized
by testing twice within 5 days (Keefe et al., 2004), cannot be
completely ruled out, paroxetine is suggested to possibly im-
prove cognitive function in depressed subjects. Antidepressant
medications may also have some negative effects on cognitive
function (Gorenstein et al., 2006, Hindmarch, 2009). Cassano et
al. (2002), in their double-blind, randomized, parallel-group,
multicenter study comparing paroxetine and fluoxetine treat-
ment for 1 year, found that improvement, rather than deteriora-
tion, was observed in most of the tested cognitive functions. It
has been speculated that paroxetine may have indirect effects
which improve cognitive function via amelioration of depres-
sion, and that this could outweigh the mildly toxic effects of
paroxetine (Hindmarch, 2009).
A limitation of this study is that the possible influence of
benzodiazepines (Stewart, 2005) on verbal WM function could
not be completely ruled out. Another limitation of this study is
that patients with full and partial remission (mild depression)
were analyzed together, mainly because there were few patients
with full remission. Also, a longer observation may be needed
to allow sufficient recovery of verbal WM function. Therefore,
a further long-term study of patients showing remission of de-
pression, who are no longer taking medication, might be neces-
sary to confirm the present results.
Acknowledgements
This work was supported in part by grants from Daido Life
Welfare Foundation (2009), the Japanese Association of Neuro-
Psychiatric Clinics (Ken TANAKA Memorial Grant, 2008),
and the Charitable Trust Kimi IMAI Memorial Stress Associ-
ated Diseases Research Aid Fund (2009, 2010). This study was
presented in part at the 20th Annual Meeting of the Japanese
Society of Clinical Neuropsychopharmacology, held jointly with
the 40th Annual Meeting of the Japanese Society of Neuropsy-
chopharmacology, in Sendai, Japan, and 164th Annual Meeting
of the American Psychiatric Association in Honolulu, USA.
REFERENCES
American Ps ych iat ric Asso ciat ion (19 94). Diagnostic and statistical man-
ual of mental disorders, Fourth Edition (DSM-IV). Washington DC:
American Psychiatric Association.
Biringer, E., Mykletun, A. , Sundet, K., Kr oken, R., Stordal, K. I., & L und,
A. (2007). A longitudinal analysis of neurocognitive function in uni-
polar depression. Journal of Clinical and Experimental Neuropsy-
chology, 29, 879-891. doi:10.1080/13803390601147686
Cassano, G. B., Puca, F., Scapicchio, P. L., & Trabucchi, M. (2002).
Paroxetine and fluoxetine effects on mood and cognitive functions in
depressed nondemented elderly patients. The Journal of Clinical P sy -
chiatry, 63, 396-40 2. doi:10.4088/JCP.v63n0504
Gorenstein, C., de Carvalho, S. C., Artes, R., Moreno, R. A., & Marcour-
akis, T. (2006). Cognitive performance in depressed patients after
chronic use of antidepressants. Psychopharmacology, 185 , 84-92.
doi:10.1007/s00213-005-0274-2
Gualtieri, C. T., Johnson, L. G., & Benedict, K. B. (2006). Neurocogni-
tion in depression: Patients on and off medication versus healthy
comparison subjects. The Journal of neuropsychiatry and clinical neu-
rosciences, 18, 217-225. doi:10.1176/appi.neuropsych.18.2.217
Gualtieri, C. T., & Morgan, D. W. (2008). The frequency of cognitive
impairment in patients with anxiety, depression, and bipolar disorder:
An unaccounted source of variance in clinical trials. The Journal of
Clinical Psychiatry, 69, 1122-1130. doi:10.4088/JCP.v69n0712
Hindmarch, I. (2009). Cognitive toxicity of pharmacotherapeutic agents
used in social anxiety disorder. International Journal of Clinical Prac-
tice, 63, 1085-1094. doi:10.1111/j.1742-1241.2009.02085.x
Kaneda, Y. (2009a). Verbal working memory and functional outcome
in patients with unipolar major depressive disorder. World Journal of
Biological Psychiatry, 10, 591-594.
doi:10.1080/15622970903183705
Kaneda, Y. (2009b). Verbal working memory impairment in patients
with current episode of unipolar major depressive disorder and in
remission. Clinical Neurop harmacology, 32, 346-347.
doi:10.1080/15622970903183705
Kaneda, Y., Jayathilak, K., & Meltzer, H. (2010). Determinants of work
outcome in neuroleptic-resistant schizophrenia and schizoaffective
disorder: Cognitive impairment and clozapine treatment. Psychiatry
Research, 178, 57-62. doi:10.1016/j.psychres.2009.04.001
Kaneda, Y., Jayathilak, K., & Meltzer, H. Y. (2009). Determinants of
work outcome in schizophrenia and schizoaffective disorder: Role of
cognitive function. Psychiatry Research, 169, 178- 179.
doi:10.1016/j.psychres.2008.08.003
Keefe, R. S., Goldberg , T. E., Harv ey, P. D., Gold, J. M., Poe, M. P ., &
Coughenour, L. (2004). The brief assessment of cognition in schizo-
phrenia: Reliability, sensitivity, and comparison with a standard neu-
rocognitive battery. Schizophrenia Research, 68, 283-297.
doi:10.1016/j.schres.2003.09.011
Kennedy, N., Foy, K., Sherazi, R., McDonough, M., & McKeon, P. (2007).
Long-term social functioning after depression treated by psychiatrists:
A review. Bipolar Disorders, 9, 25-37.
doi:10.1111/j.1399-5618.2007.00326.x
Kitagawa, N., & Koyama, T. (2009). Assessment of neurocognitiv e func-
tion in mood disorders and its utilization to clinical practice. Rinsho
Seishin Igaku (Japanese Journal of Clinical Psychiatry), 38, 437-445.
Malone, D. C. (2007). A budget-impact and cost-effectiveness model
for second-line treatment of major depression. Journal of Managed
Care Pharmacy, 13, S8-18.
Marcos, T., Portella, M. J., Navarro, V., Gasto, C., Rami, L., Lazaro, L.,
& Salamero, M. (2005). Neuropsychological prediction of recovery
in late-onset major depression. Intern ational Journal of Geriatric Psy-
chiatry, 20, 790-79 5. doi:10.1002/gps.1363
McIntyre, R. S., Konarski, J. Z., Mancini, D. A., Fulton, K. A., Parikh,
S. V., Grigoriadis, S., Grupp, L. A., Bakish, D., Filteau, M. J., Gor-
man, C., Nemeroff, C. B., & Kennedy, S. H. (2005). Measuring the
severity of depression and remission in primary care: Validation of
the HAMD-7 scale. Canadian Medical Association Journal, 173,
1327-1334. doi:10.1503/cmaj.050786
Nebes, R. D., Pollock, B. G., Ho uck, P. R., Butters, M. A., Mu lsant, B.
H., Zmuda, M. D., & Reynolds III, C. F., (2003). Persistence of cog-
nitive impairment in geriatric patients following antidepressant treat-
ment: A randomized, double-blind clin ica l trial with nortriptyline and
paroxetine. Journal of Psychiatric Research, 37, 99-108.
doi:10.1016/S0022-3956(02)00085-7
760

Y. KANEDA
Porter, R. J., Gallagher, P., Thompson, J. M., & Young, A. H. (2003).
Neurocognitive impairment in drug-free patients with major depress-
sive disorder. The British Journal of Psychiatr y, 182, 214-220.
doi:10.1192/bjp.182.3.214
Reppermund, S., Ising, M., Lucae, S., & Zihl, J. (2009). Cognitive im-
pairment in unipolar depression is persistent and non-specific: Fur-
ther evidence for the final common pathway disorder hypothesis.
Psychological Medicine, 39 , 603-614.
doi:10.1017/S003329170800411X
Rose, E. J., & Ebmeier, K. P. (2006). Pattern of impaired working
memory during major depression. Journal of Affective Disorders, 90,
149-161. doi:10.1016/j.jad.2005.11.003
Stewart, S. A. (2005). The effects of benzodiazepines on cognition. The
Journal of Clinical Psychia t ry , 66, 9-13.
Zimmerman, M., McGlinchey, J. B., Posternak, M. A., Friedman, M.,
Attiullah, N., & Boerescu, D. (2006). How should remission from
depression be defined? The depressed patient’s perspective. The Ameri-
can Journal of Psychiatry, 163, 148-150.
doi:10.1176/appi.ajp.163.1.148