Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49

Contents lists available at ScienceDirect

Progress in Neuropsychopharmacology
& Biological Psychiatry
journal homepage: www.elsevier.com/locate/pnp

Neural correlates of working memory in first episode and recurrent T

depression: An fMRI study
⁎
Dilara Yüksela, , Bruno Dietschea, Carsten Konrada,b, Udo Dannlowskia,c, Tilo Kirchera, Axel Kruga
a
Department of Psychiatry and Psychotherapy, Philipps-University Marburg, Rudolf-Bultmann-Str. 8, 35039 Marburg, Germany
b
Agaplesion Diakonieklinikum Rotenburg, Centre for Psychosocial Medicine, Elise-Averdieck-Straße 17, 27356 Rotenburg (Wümme), Germany
c
Department of Psychiatry, University of Münster, Albert-Schweitzer-Campus 1, 48149 Münster, Germany

A R T I C L E I N F O A B S T R A C T

Keywords: Background: Patients suffering from major depressive disorder (MDD) show deficits in working memory (WM)
Depression performance accompanied by bilateral fronto-parietal BOLD signal changes. It is unclear whether patients with a
First episode first depressive episode (FDE) exhibit the same signal changes as patients with recurrent depressive episodes
Recurrent depression (RDE).
fMRI
Methods: We investigated seventy-four MDD inpatients (48 RDE, 26 FDE) and 74 healthy control (HC) subjects
Working memory
performing an n-back WM task (0-back, 2-back, 3-back condition) in a 3T-fMRI.
Results: FMRI analyses revealed deviating BOLD signal in MDD in the thalamus (0-back vs. 2-back), the angular
gyrus (0-back vs. 3-back), and the superior frontal gyrus (2-back vs. 3-back). Further effects were observed
between RDE vs. FDE. Thus, RDE displayed differing neural activation in the middle frontal gyrus (2-back vs. 3-
back), the inferior frontal gyrus, and the precentral gyrus (0-back vs. 2-back). In addition, both HC and FDE
indicated a linear activation trend depending on task complexity.
Conclusions: Although we failed to find behavioral differences between the groups, results suggest differing
BOLD signal in fronto-parietal brain regions in MDD vs. HC, and in RDE vs. FDE. Moreover, both HC and FDE
show similar trends in activation shapes. This indicates a link between levels of complexity-dependent activation
in fronto-parietal brain regions and the stage of MDD. We therefore assume that load-dependent BOLD signal
during WM is impaired in MDD, and that it is particularly affected in RDE. We also suspect neurobiological
compensatory mechanisms of the reported brain regions in (working) memory functioning.

1. Introduction Particularly, impaired working memory functions are common in MDD

Major depressive disorder (MDD) is characterized by affective, be-
havioral, and physical symptoms. Common accessory symptoms are e.g.
an enduring feeling of sadness, lack of motivation and drive or inter-
ferences of sleep and appetite. According to ICD-10, MDD is categorized
into recurrent depressive disorder (RDE) and a single, first episode
(FDE).
1.1. Working memory deficits in MDD
Patients suffering from acute MDD display cognitive deficits
(Hasselbalch et al., 2011; Marazziti et al., 2010; Porter et al., 2003;
Roca et al., 2015) in domains such as attention, episodic memory, ex-
ecutive functioning, or processing speed (Evans et al., 2014; Zakzanis
et al., 1998; Zaninotto et al., 2014). Studies investigating cognitive
impairments in FDE indicated similar effects (Lee et al., 2012).
(Airaksinen et al., 2004; Bourke et al., 2012; Christopher and
MacDonald, 2005; Dumas and Newhouse, 2014; Ilsley et al., 1995;
Marazziti et al., 2010; McIntyre et al., 2013; Rose and Ebmeier, 2006).
1.2. Neural correlates of working memory
Working memory (WM), which is defined as “the capability to
memorize, retrieve and utilize information for a limited period of time”
(Rottschy et al., 2012) is required for diverse activities in daily- or job
routines. The n-back version of the Continuous Performance Task (CPT)
is a valid and reliable instrument to determine WM, especially during
fMRI (Cohen et al., 1994; Jacola et al., 2014). Working-memory-load is
mediated by bilateral fronto-parietal networks, as these regions have
consistently been associated with WM-tasks (e.g. Kondo et al., 2004;
Krug et al., 2008; Owen et al., 2005; Rottschy et al., 2012; Wager and
Smith, 2003; Yüksel et al., 2017). Commonly reported areas required

⁎
Corresponding author at: Department of Psychiatry and Psychotherapy, Philipps-University Marburg, Rudolf-Bultmann-Str. 8, 35039 Marburg, Germany
E-mail address: dilara.yueksel@med.uni-marburg.de (D. Yüksel).

https://doi.org/10.1016/j.pnpbp.2018.02.003
Received 20 September 2017; Received in revised form 1 February 2018; Accepted 3 February 2018
Available online 05 February 2018
0278-5846/ © 2018 Elsevier Inc. All rights reserved.
D. Yüksel et al.
for WM-functioning are for example the middle frontal gyrus (e.g. Lee
et al., 2013; Vasic et al., 2009), the inferior frontal gyrus (e.g. Kondo
et al., 2004), or the precentral gyrus (e.g. Wang et al., 2015). Studies
investigating neuroimaging data from MDD patients and healthy con-
trol (HC) subjects confirmed the concept of a fronto-parietal network-
pattern associated with WM demand and also revealed BOLD signal
differences between these groups.
1.3. Activation load – shape and direction
Though enhanced BOLD signal is commonly observed during WM
tasks in patients as well as in healthy subjects, studies show different
and partly conflicting results especially in relation to the direction of
the WM associated brain activation. Thus, hypoactivation during WM
tasks was frequently reported in various patient samples – such as
acute, remitted, unipolar or bipolar depressive patients (Bartova et al.,
2015; Brooks et al., 2015; Goethals et al., 2005; Kerestes et al., 2012; Pu
et al., 2012) – e.g. in areas comprising the bilateral thalamus, right
precentral gyrus, right parietal cortex (Barch et al., 2003), precuneus,
cerebellum or the left dorsolateral prefrontal cortex (including the
precentral gyrus, frontal operculum, bilateral thalamus and posterior
part of the left caudate) (Rodríguez-Cano et al., 2014).
Nevertheless, some studies reported opposite results as they re-
vealed higher levels of activation in MDD in respective regions (e.g.
Fitzgerald et al., 2008; Harvey et al., 2005; Matsuo et al., 2007; Walsh
et al., 2007). Referring to these results, Harvey et al. (2005) suggested a
hyperfrontality during WM tasks in MDD, stating that greater activation
of the underlying networks is required for comparable performance in
MDD patients during WM, which is inconsistent with the above studies.
Furthermore, studies began to examine the nature (shape) of acti-
vation during WM load by stating that increasing BOLD-signal is asso-
ciated with WM-complexity in a linear relationship in healthy subjects
(Braver et al., 1997; Gould et al., 2003; Jansma et al., 2000; Kirschen
et al., 2005; Veltman et al., 2003). Nevertheless, some studies came to
the conclusion that BOLD signal is represented as an “inverted U-shape”
(Callicott et al., 1999; Manoach, 2003) so that task complexity mod-
ulates activation in a reverse U-shape (decreased BOLD-signal at the
lowest and highest WM demand). However, while the most of the
studies that reported a linear relationship between WM demand and
BOLD signal increase investigated healthy subjects, Manoach (2003)
also observed the inverted U-shape in schizophrenic individuals.
Nevertheless, results are hardly comparable as previous studies in-
vestigated other (patient) samples than the present study does. The
current state of research is varying and in some cases contradictory and
therefore requires further clarification and discussion.
1.4. RDE vs. FDE
So far, existing studies rarely differentiated between RDE and FDE.
To our knowledge there is no present study investigating (neural)
correlates of WM in chronic vs. first-episode depression. Nevertheless,
there is evidence that the stage of disorder, respectively severity of
symptoms or the course of illness [in terms of the number of (previous)
depressive episodes or illness duration], impairs executive and cogni-
tive functioning in MDD (e.g. Fossati et al., 2004; Hasselbalch et al.,
2013; Karabekiroğlu et al., 2010; Lawrence et al., 2013; Maes, 2015;
Snyder, 2014; Talarowska et al., 2015), and also BOLD signal during
fMRI (e.g. Dichter et al., 2012; Holt et al., 2016; Kerestes et al., 2015;
Liu et al., 2017; Meng et al., 2014; Zhou et al., 2017). Therefore we
suggest an association between altered BOLD signal during WM load in
RDE vs. FDE.
Considering the lack of research investigating the distinction be-
tween FDE and RDE, further aspects of previous studies complicate the
comparison of results, such as the diversity of paradigms (e.g. Rottschy
et al., 2012; Wang et al., 2015), demographic and clinical factors (Wang
et al., 2015), small sample sizes, and variance in inclusion criteria [e.g.
40
Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49
unmedicated patients only, particular age of subjects, and state of de-
pression (e.g. Dumas and Newhouse, 2014; Fernández-Corcuera et al.,
2013; Garrett et al., 2011; Korsnes et al., 2013; Matsuo et al., 2007;
Opmeer et al., 2013; Vasic et al., 2009; Vilgis et al., 2014; Walsh et al.,
2007; Walter et al., 2007)].
1.5. Objective
We investigated functional and behavioral differences in depressed
patients as compared to healthy subjects during a WM task. We expect
WM impairments in MDD mediated by differing BOLD signal in fronto-
parietal regions. In addition, we focus on the differences within the
MDD sample taking into account their stage of disorder (RDE vs. FDE).
Since there are only few studies investigating WM performance be-
tween RDE and FDE (e.g. Fossati et al., 2004) - and differences in neural
correlates of WM in FDE and RDE are still unknown - we conclude from
previous studies indicating a link between the duration of illness or the
number of depressive episodes, respectively the illness history, and
functional activation during WM load. Thus, we hypothesize greater
cognitive impairments in RDE as compared FDE and HC. Furthermore,
we expect to find greater deviation of BOLD signal in respective –
mainly fronto-parietal – regions such as the middle frontal gyrus or
precentral gyrus in MDD patients in general compared to HC. With
regard to the shape of activation, we assume that HC has an increasing
activation pattern depending on WM demand, as current studies in-
vestigating healthy subjects report this finding (Braver et al., 1997;
Gould et al., 2003; Jansma et al., 2000; Kirschen et al., 2005; Veltman
et al., 2003). Since there is little research on this subject in MDD, we
assume that MDD exhibits a divergent activation - possibly in the re-
verse U-form and left-shifted as observed in e.g. schizophrenic subjects
(Manoach, 2003). In addition, we assume that the activation patterns of
FDE are more similar to those of HC than of RDE.
2. Experimental procedures/methods
2.1. Subjects
Seventy-four healthy subjects and 74 patients with MDD (for char-
acteristics see Table 1) were included. All participants were native
German speakers and recruited in Marburg, Germany. Healthy subjects
were selected from a pool of 147 HC by using the MatchIt package (Ho
et al., 2011) for R (version 2.13.1, http://www.r-project.org). HC were
matched to MDD by considering sex and age.
All MDD patients were diagnosed for current and lifetime psychia-
tric disorders and comorbid personality disorders using the German
version of the Structured Clinical Interview for DSM-IV (SKID-I/ SKID-
II) (Wittchen et al., 1997). MDD subjects were inpatients diagnosed
with a depressive episode according to DSM-IV criteria. Within these
subjects we distinguished between patients with their first depressive
episode (FDE, n = 26, 0 previous depressive episodes) and patients
with recurrent depressive episodes with at least one depressive episode
before the present one [RDE, n = 48; M = 9.65 previous depressive
episodes (SD = 10.5)].
2.2. Exclusion criteria
In order to exclude psychiatric disorders in healthy subjects, HC also
underwent the German version of the Structured Clinical Interview for
DSM-IV (SKID-I) (Wittchen et al., 1997). Furthermore, HC did not re-
port any psychiatric disorder in first-degree relatives. Exclusion criteria
for MDD patients were bipolar disorder or current substance abuse. Yet,
patients with comorbid anxiety, obsessive-compulsive, somatoform or
personality disorder (avoidant, narcissistic or dependent) were in-
cluded, if these disorders were not the main reason for present hospi-
talization. In our case, 14 of the 74 patients were additionally diag-
nosed with one or two of these disorders. Furthermore, 22 of the
D. Yüksel et al. Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49

Table 1
Sociodemographic, neuropsychological, clinical and behavioral data of the sample (N = 148), standard deviations in parentheses.

MDD (n = 74) HC (n = 74) t df p

Sex ratio (men/women) 33/41 33/41 χ2 = 0.00 1 n.s.

Age 36.8 (10.7) 35.8 (10.8) 0.26 147 n.s.
BDI 22.9 (10.9) 3.8 (3.8) 200.1 144 < 0.01
Neuropsychological Testing
Spatial-span 8.8 (1.7) 8.9 (2.0) 0.32 145 n.s.
LNS 15.3 (3.2) 15.8 (2.9) 0.89 145 n.s.
DSC 53.7 (11.23) 62.5 (11.3) 21.91 142 < 0.01
TMT 31.0 (17.3) 26.7 (15.9) 2.42 144 n.s
Verbal-IQ 109.7 (12.0) 113.8 (13.8) 3.76 147 < 0.05

RDE (n = 48) FDE (n = 26) t df p

Sex ratio (men/women) 21/27 12/14 χ2 = 0.04 1 n.s.
Age 37.5 (10.7) 35.5 (10.7) 0.54 72 n.s.
BDI 23.0 (11.1) 22.6 (10.8) 0.02 69 n.s.
Number of episodes 9.65 (10.5) 0 19.69 52 < 0.01
Comorbid disorders
Comorbid dysthymia (double depression) 18 4 χ2 = 3.94 1 < 0.05
Comorbid anxiety, obsessive-compulsive, somatoform or personality disorder 6 8 χ2 = 3.66 1 n.s.
Medication
One SA 28 17 χ2 = 0.35 1 n.s.
Two ore more SA 14 4 χ2 = 1.74 1 n.s.
AP 17 8 χ2 = 0.16 1 n.s.
MS 2 2 χ2 = 0.41 1 n.s.
Drug naive 3 4 χ2 = 0.20 1 n.s.
Neuropsychological Testing
Spatial-span 8.6 (1.7) 9.1 (1.6) 0.66 73 n.s.
LNS 15.6 (3.2) 14.8 (3.1) 0.93 73 n.s.
DSC 54.2 (10.9) 52.7 (11.9) 0.29 74 n.s.
TMT 30.3 (17.0) 32.3 (17.9) 0.24 73 n.s.
Verbal-IQ 110.2 (11.2) 108.6 (13.5) 0.28 73 n.s.

Results of sociodemographic, clinical and and behavioral data of the sample. SA = standard antidepressant, AP = Antipsychotics, MS = Mood stabilizers. BDI = Becks Depression
Inventory, DSC = digit-symbol-coding, LNS = Letter-number-span, TMT = Trail-making test A – Trail-making test B, Verbal IQ = verbal IQ score derived from the MWT-B. n.s. = non
significant (all p > .1).

patients also suffered from dysthymia (double depression), mainly in
the RDE group (18 RDE patients). Sixty-five patients received either one
(n = 45) or a combination of two standard antidepressants (n = 18),
whereas 7 patients were drug naïve at time of inclusion. 25 patients
were taking antipsychotics, and 4 patients were additionally treated
with mood stabilizers. For details and accurate apportionment see
Table 1.
Subjects with neurological disorders, serious head injury, mental
retardation, severe somatic diseases, or any condition that might have
an effect on cerebral metabolism or MR safety were excluded.
Anatomical (T1) images of all participants – which were assessed si-
multaneously for other research questions – were visually examined
and checked for brain morphometric abnormalities. Subjects with brain
morphometric lesions were excluded. The local ethics committee ap-
proved the protocol according to the declaration of Helsinki for the
present study. Before participation all subjects signed a letter of consent
after a comprehensive disclosure.
2.3. Neuropsychological data acquisition and analysis
In addition to fMRI data, several neuropsychological and behavioral
variables were assessed. All neuropsychological tests described below
are paper-pencil-tests. Besides the behavioral WM-performance in the
fMRI, neuropsychological tests can provide insights into the general
cognitive functioning in MDD.
The Trail Making Test (TMT) (Reitan, 1955) was used as a screening
instrument for detecting neuropsychological impairments in cognitive
domains such as executive functioning (Bowie and Harvey, 2006), but
also attention and WM (e.g. Deshpande et al., 2007). For the test pro-
cedure participants had to connect numbers in ascending order (TMT-
A), and then alternately connect numbers and characters in ascending
order (TMT-B). Part A was subtracted from Part B to calculate the TMT
score. For more information, e.g. on the psychometric norms, see:
Drobnik, 2014. Moreover, we used the letter-number-sequencing (LNS)
to examine verbal working memory (Wechsler, 1997). For this test, the
instructor reads a combination of random numbers and letters. Subse-
quently, participants must reproduce the content they have heard in
ascending order. Detailed information about the implementation and
further links are provided in: Laschkowski et al., 2010. The digit symbol
coding test (DSC) from the Wechsler Adult Intelligence Scale - Revised
(WAIS-R) (Wechsler, 1981) examines cognitive processing speed. A
series of numbers is presented to subjects; moreover, all numbers are
assigned to an individual symbol. Participants have to assign the re-
spective symbol to the respective number under time pressure. The DSC
is a valid instrument to measure speed, but also memory (Joy et al.,
2004). To examine visual working memory, the spatial-span (Block-
tapping-test) (Schelling, 1997) was used. During the execution, the
investigator presents a certain sequence of “finger-tips” on a board.
Afterwards, participants have to imitate this sequence. Standardization
and standard values are provided in: Kessels et al. (2000). The Verbal-
IQ was estimated using the “Multiple choice vocabulary test” (MWT-B)
(Lehrl et al., 1995). The test contains actual existing – but very rarely
used – words, and invented words. Participants have to identify the
existing words. The MWT-B is a short and valid test for the estimation of
premorbid intelligence or verbal-IQ (Lehrl et al., 1995). Finally, the
Becks-Depression-Inventory (BDI) (Hautzinger et al., 1995) was as-
sessed in all subjects. The BDI is a reliable instrument to assess de-
pressive symptoms (Beck et al., 1988; Kühner et al., 2007). Subjects in
the healthy control group were excluded if they had a BDI-score higher
than 9, as a value of 8 is considered the standard threshold for minimal
depressive symptoms.

41
D. Yüksel et al. Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49

2.4. fMRI data acquisition MDD and controls between the different conditions, the interaction
contrasts were assessed as follows: [(2b HC > 0b HC) & (0b
The fMRI data was acquired with a 3 Tesla Tim Trio MR scanner MDD > 2b MDD)], [(0b HC > 2b HC) & (2b MDD > 0b MDD)], [(3b
(Siemens Medical Systems) at the Department of Psychiatry and HC > 0b HC) & (0b MDD > 3b MDD)], [(0b HC > 3b HC) & (3b
Psychotherapy, Philipps University Marburg. MDD > 0b MDD)], [(2b HC > 3b HC) & (3b MDD > 2b MDD)], [(3b
We used a T2*-weighted echo-planar imaging (EPI) sequence HC > 2b HC) & (2b MDD > 3b MDD)].
(64 ∗ 64 matrix, 224 mm ∗ 224 mm FoV, 40 slices, 3.5 mm slice thick- Secondly, differences in brain activation between MDD groups were
ness, TR = 2.5 s, TE = 30 ms, flip angle = 90°) to collect functional analyzed. Nevertheless, we included HC as a comparative value in the
imaging data. The slices masking the whole brain were placed trans- calculations to improve the comparability of the results. Again, we were
axially parallel to the anterior–posterior line (AC–PC). In order to interested in all possible group by condition interactions between RDE
eliminate interferences of the T1 stabilization effects, we have excluded and FDE and used the following contrasts:
the initial three of 160 gathered functional images from further ana- [(2b RDE > 0b RDE) & (0b FDE > 2b FDE)], [(0b RDE > 2b
lysis. Although subjects' heads were fixed in the scanner, several sub- RDE) & (2b FDE > 0b FDE)], [(2b RDE > 3b RDE) & (3b FDE > 2b
jects with excessive head movement had to be excluded from the study. FDE)], [(3b RDE > 2b RDE) & (2b FDE > 3b FDE)], [(3b RDE > 0b
Thus, 12 HC, 8 RDE, and 4 FDE participants were excluded from the RDE) & (0b FDE > 3b FDE)], [(0b RDE > 3b RDE) & (3b FDE > 0b
original sample. The threshold value for head movement was 3 mm or FDE)].
3° in any direction. Since we were interested in the direction of the respective contrasts,
we calculated all analyses by means of F-tests and subsequent post-hoc
2.5. N-back task and fMRI-procedure t-tests which were not additionally adjusted for multiple comparisons,
thus, the p-value was set to .001 uncorrected, cluster extend = 43
The Presentation software package (version 14.1.09.21.09, voxels. This procedure prevented false positive findings at the equiva-
Neurobehavioral Systems Inc.) was used to present the stimuli in the lent level of p < .05 (FWE corrected). For a better presentation of the
MRI-scanner. We assessed three different conditions of the n-back task: results, we illustrated all conditions in the respective figures. However,
0-back (0b), 2-back (2b) and 3-back (3b). The 0-back condition implies only the relevant interaction contrast led to significant results.
the lowest WM load and is therefore the baseline-condition. 2-back and In both analyses we report only the respective interaction contrasts,
accordingly 3-back represent the higher – more demanding - conditions since the main effect of time and the main effect of condition go beyond
(for detailed description of the task and methods see: Yüksel et al., the scope of the current manuscript. Thus, the information given by the
2017). main effect of group or condition would not prevent further information
for the present research question. Nevertheless, in previous studies,
2.6. fMRI data analyses regular effects of activation during WM have been observed (e.g. Krug
et al., 2008; Owen et al., 2005).
Analogous to Yüksel et al. (2017), the SPM8 (v4977) standard The behavioral data collected outside the scanner (Table 1) were
routines and templates (Wellcome Trust Centre for Neuroimaging, analyzed using t-tests. The data collected inside the scanner (Table 2)
http://www.fil.ion.ucl.ac.uk/spm) running on MATLAB 7.9.0 (R2009b) were analyzed using a one-way ANOVA. All behavioral analyses were
(The MathWorks, Inc.) were used to analyze the data. Images were performed with the SPSS 21.0 statistics software.
realigned and normalized to standard MNI (Montreal Neurological In-
stitute) space (voxel size: 2 mm × 2 mm × 2 mm). All scans were spa- 3. Results
tially smoothed with an 8 mm full-width-at-half-maximum (FWHM)
Gaussian kernel to improve the signal-to-noise ratio and to correct for 3.1. Clinical, neuropsychological and social demographical results
anatomical variation between subjects. In order to preclude false po-
sitive findings as a result of multiple testing, a cluster extension Characteristics of the sample and behavioral results of the neu-
threshold of 43 resampled voxels was utilized. ropsychological testing are presented in Table 1 and in Section 2.2. Sex
and age did not significantly differ between all groups, but MDD ex-
2.7. First level analyses hibited significantly higher BDI scores than HC. Furthermore, HC
achieved a significantly better outcome in the DSC and verbal-IQ than
A block design was used to model the BOLD signal during the 6 MDD. Apart from that, however, there were no behavioral differences
blocks (2 blocks per condition) with 12 trials each. Every block con- between all groups.
sisted of 4 targets and 8 non-target stimuli besides the last block (3-
back), which only comprised 3 target stimuli. The blocks, but not the
individual trials, were modeled for the subject level analysis. Periods of 3.2. Behavioral fMRI-task results
time for the instructions were grouped together as a coherent individual
condition and excluded from further calculations. In addition, move- The behavioral performance of all participants is presented in
ment parameters were used as a covariate of no interest to control for Table 2. None of the behavioral measures displayed significant group
head movement.
Table 2
Behavioral performance of the n-back task (N = 148), standard deviations in parentheses.
2.8. Second level analyses
FDE (n = 26) RDE (n = 46) HC (n = 74) F p
For the group analyses, we used a full factorial design with the three
groups (HC vs. FDE vs. RDE) as the between-subject factor and the three Correct answers 0b 95.8% (6.9) 92.3% (10.1) 88.7% (21.8) 1.91 n.s.
Correct answers 2b 76.1% (21.1) 79.6% (16.5) 80.1% (16.1) 0.54 n.s.
conditions (0-back, 2-back, 3-back) as the within-subject factor. To Correct answers 3b 55.7% (24.9) 62.0% (25.5) 63.2% (21.3) 1.00 n.s.
control interference effects, the possible confounder variables (gender, Reaction time 0b 554.9 (20.2) 539.6 (15.3) 536.2 (11.9) 0.32 n.s.
age and medication in terms of whether patients were medicated or (ms)
not) were included as covariates for all calculations. Reaction time 2b 721.6 (33.6) 709.6 (25.5) 714.1 (19.9) 0.04 n.s.
(ms)
First, we analyzed differential brain activation between all MDD
Reaction time 3b 752.8 (37.8) 776.4 (28.8) 757.7 (22.4) 0.17 n.s.
patients and healthy controls for the varying task conditions. Since we (ms)
were interested in both an increased and a decreased BOLD signal in

42
D. Yüksel et al. Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49

Fig. 1. Brain areas with significant group (MDD, HC) by condition (0-back, 2-back, 3-back) interaction effect (p < .01, corrected by Monte Carlo cluster simulations; cluster extent
threshold = 43 voxels). Bar graphs display the mean signal intensity during the three conditions for particular clusters. Error bars represent 95% confidence interval. For better
comparability, all conditions are illustrated in the figure, although only the interaction contrast (*) was significant. HC showed greater activation with increasing WM load in the
respective brain regions, while MDD indicated no such general pattern. SFG = Superior frontal gyrus. AG = Angular gyrus. CB = Cerebellar vermis. TH = Thalamus.

differences.
3.3. fMRI results
3.3.1. MDD vs. HC
The results of the group by condition analysis for MDD vs. HC in-
dicated significant differences of BOLD signal in the right thalamus, the
left superior frontal gyrus (SFG), the cerebellar vermis (CB), and the left
angular gyrus (AG) as illustrated in Fig. 1 and Table 3. The first two
interaction analyses covered both advanced conditions, 2-back and 3-
back, in relation to the baseline condition. In particular, the first con-
trast [(2b HC > 0b HC)] & (0b MDD > 2b MDD)] revealed greater
activation in HC during 2-back > 0-back, and less activation in MDD
during 0-back < 2-back in the thalamus. The opposite contrast yielded
no significant results. Comparing the 3-back vs. 0-back condition [(3b
HC > 0b HC)] & (0b MDD > 3b MDD)] HC revealed greater activa-
tion in the left AG and the cerebellar vermis during 3-back > 0-back,
and less activation in MDD during 0-back > 3-back. For the reverse
contrast no significant activations were observed. The last contrast
compared the advanced conditions with each other [(2b HC > 3b HC)]
& (3b MDD > 2b MDD)] and revealed less activation of the superior
frontal gyrus in MDD during 2-back < 3-back, and greater activation in
HC during 3-back < 2-back. The opposite contrast revealed no sig-
nificant outcome. Overall, we observed a trend in HC towards in-
creasing activation depending on the complexity of the task, while MDD
did not show such a trend.
3.3.2. RDE vs. FDE
The results of the group by condition analysis for RDE vs. FDE in-
dicated differential activations in the right inferior frontal gyrus (IFG),
left precentral gyrus (PCG), and bilateral middle frontal gyri (MFG) (see
Table 4) as illustrated in Figs. 2 and 3. More precisely, the first contrast
[(2b RDE > 0b RDE) & (0b FDE > 2b FDE)] revealed greater activa-
tion in RDE during 2-back > 0-back, while FDE displayed decreased

43
D. Yüksel et al. Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49

Table 3
Results of the group (HC, MDD) by condition (0-back, 2-back, 3-back) interaction analyses [(2b HC > 0b HC) & (0b MDD > 2b MDD)], [(3b HC > 0b HC) & (0b MDD > 3b MDD)],
[2b HC > 3b HC) & (3b MDD > 2b MDD)]. Coordinates of the peak voxels of each cluster are listed in MNI atlas space. (p < .01, corrected by Monte Carlo cluster simulations; cluster
extent threshold = 43 voxels). KE = clustersize; Max. SPM (F) = calculated T-value.

Region Hemisphere x y z kE Max. SPM (F)

Interaction group × condition (2b HC > 0b HC) & (0b MDD > 2b MDD)
Cluster I 46 3.11
Thalamus R 20 -32 4 3.74
Interaction group × condition (0b HC > 2b HC) & (2b MDD > 0b MDD)
No significant results
Interaction group × condition (3b HC > 0b HC) & (0b MDD > 3b MDD)
Cluster I 139 > 3.11
Angular Gyrus L −38 −74 48 3.92
Cluster II 70 > 3.11
Cerebellar Vermis 2 −68 −38 3.49
Interaction group × condition (0b HC > 3b HC) & (3b MDD > 0b MDD)
No significant results
Interaction group × condition (2b HC > 3b HC) & (3b MDD > 2b MDD)
Cluster I 50 > 3.11
Superior Frontal Gyrus L −16 60 16 3.59
Interaction group × condition (3b HC > 2b HC) & (2b MDD > 3b MDD)
No significant results

Table 4 4. Discussion
Results of the group (FDE, RDE) by condition (0-back, 2-back, 3-back) interaction ana-
lyses [(2b RDE > 0b RDE) & (0b FDE > 2b FDE)], [(2b RDE > 3b RDE) & (3b
The goal of our study was to investigate working memory differ-
FDE > 2b FDE)]. Coordinates of the peak voxels of each cluster are listed in MNI atlas
space. (p < .01, corrected by Monte Carlo cluster simulations; cluster extent
ences in patients with major depressive disorder compared to healthy
threshold = 43 voxels). control subjects. In addition, we compared patients with a first de-
pressive episode to patients with recurrent depressive episodes. All
Region Hemisphere x y z kE Max. patients were in-patients with at least mild to moderate depressive
SPM (F)
symptoms, usually in partial remission. We used an n-back paradigm
Interaction group × condition (0-, 2-, and 3-back) to assess neural correlates of WM in a total sample
(2b RDE > 0b RDE) & (0b of 74 MDD patients and 74 HC. To our knowledge, this is the first study
FDE > 2b FDE) that investigates differences between these two patient subgroups on a
Cluster I 297 > 3.11
neuroimaging level investigating WM.
Middle Frontal Gyrus R 38 36 26 3.83
Cluster II 105 > 3.11 Consistent with previous studies (e.g. Krug et al., 2008; Wang et al.,
Precentral Gyrus L −32 −16 36 3.75 2015), all subjects engaged a fronto-parietal network during WM load.
Cluster III 85 > 3.11 Our results indicated different signal changes in MDD as compared to
Middle Frontal Gyrus L −30 8 32 3.98 HC in several brain regions comprising the left SFG, the left AG, the
Cluster IV 78 > 3.11
Inferior Frontal Gyrus R 36 10 34 3.78
cerebellum and the right thalamus. Moreover, our second analysis – the
Cluster VI > 3.11 comparison between two patient subgroups FDE vs. RDE – indicated a
Middle Frontal Gyrus L −36 52 4 58 3.62 different activation in brain areas comprising the bilateral middle
Cluster VII > 3.11 frontal gyri, right precentral gyrus, and right inferior frontal gyrus in
Middle Frontal Gyrus L −34 34 31 49 3.66
RDE as compared to FDE.
Interaction group × condition
(0b RDE > 2b RDE) & (2b
FDE > 0b FDE)
4.1. Behavioral results
No significant results
Interaction group × condition
(2b RDE > 3b RDE) & (3b Results of the n-back task of WM revealed no significant behavioral
FDE > 2b FDE) differences between MDD and HC. In addition, both patient groups -
Cluster I 50 > 3.11
FDE and RDE - indicated almost similar performance levels in accuracy
Middle Frontal Gyrus R 36 32 42 3.97
Interaction group × condition and reaction time compared to each other and to the healthy control
(3b RDE > 2b RDE) & (2b group. This outcome could be explainable by the degree of illness in the
FDE > 3b FDE) patient sample at the time of the survey, since we only included patients
No significant results - mostly in partial remission - who were able to cope with the stress of
our examination. Most patients were on the point of their hospital
dismissal and therefore only moderately depressed. Nevertheless, our
activation during 0-back > 2-back. The second contrast [(3b RDE >
findings are consistent with previous investigations (e.g. Harvey et al.,
0b RDE) & (0b FDE > 3b FDE)] revealed no significant outcome. The
2005; Matsuo et al., 2007; Norbury et al., 2014; Walsh et al., 2007),
third contrast compared the advanced conditions [(2b RDE > 3b RDE)
which also reported no significant behavioral differences between MDD
& (3b FDE > 2b FDE)] and indicated increased activation in the right
vs. HC. In addition, a recent meta-analysis did not confirm the as-
MFG in FDE during 3-back > 2-back, and decreased activation in RDE
sumption that the number of depressive episodes increases the risk of
during 2-back > 3-back. Hence, none of the possible opposite contrasts
cognitive impairments (Roca et al., 2015), which is also consistent with
in the group by condition analysis revealed any significant difference.
our results as we found no differences between FDE and RDE. In ac-
Overall, FDE indicated an increasing activation trend similar to HC. On
cordance with recent investigations (e.g. McIntyre et al., 2013) we
the opposite, RDE displayed no such trend. Thus, activation shapes in
detected differences in the digit symbol coding test (DSC) for processing
RDE partially tend to resemble a reversed U-shape in some cases.
speed between the patient- and healthy control group: HC showed a
significantly better performance than MDD. Although we failed to

44
D. Yüksel et al. Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49

Fig. 2. Brain areas with significant group (FDE, RDE) by condition (0-back, 2-back) interaction effect [(2b RDE > 0b RDE) & (0b FDE > 2b FDE)] (p < .01, corrected by Monte Carlo
cluster simulations; cluster extent threshold = 43 voxels). Bar graphs display the mean signal intensity during the three conditions for particular clusters. Error bars represent 95%
confidence interval. For better comparability, all conditions are illustrated in the figure, although only the interaction contrast (*) was significant. Furthermore, HC was added as a
comparative value for the two groups. FDE (and HC) showed greater activation with increasing WM load in the respective brain regions, while RDE indicated activations patterns
comparable to a reversed U-shape. IFG = Inferior frontal gyrus, PCG = Precentral gyrus, MFG = Middle frontal gyrus.

report any further behavioral differences in our sample, fMRI data
shows various differences on a neuroimaging level, which is described
below.
4.2. fMRI results
4.2.1. MDD vs. HC
Our first analysis aimed to identify differences between healthy
participants and MDD patients. We found lower levels of activation in
MDD in regions comprising the left angular gyrus, right thalamus, and
cerebellar vermis during the advanced conditions as compared to the
baseline condition (0-back > 2-back/3-back). In contrast, HC displayed
decreased activation during 0-back vs. 2-back, respectively rising
activation depending on task complexity in these regions. In the context
of WM, (de)activation of these regions has been observed in patients
and HC (thalamus: Barch et al., 2003; Walsh et al., 2007. AG:
Fernández-Corcuera et al., 2013; Fitzgerald et al., 2008; Rodríguez-
Cano et al., 2014; Schöning et al., 2009. cerebellum: Lee et al., 2013;
Rodríguez-Cano et al., 2014; Vasic et al., 2009; Walter et al., 2007).
Furthermore, there is a trend towards increasing BOLD signal in HC as a
function of WM-complexity, which partly corresponds to the theory of
an ‘inverted U-shape’ (Callicott et al., 1999), but is even more in line
with the approach of a linear relationship between BOLD signal in-
crease and WM load (Gould et al., 2003; Kirschen et al., 2005). Al-
though it could be assumed that the activation in HC may drop during a
more demanding condition not examined here, MDD showed no such

45
D. Yüksel et al.
Fig. 3. Brain areas with significant group (FDE, RDE) by condition (2-back, 3-back) interaction effect [(2b RDE > 3b RDE) & (3b FDE > 2b FDE)] (p < .01, corrected by Monte Carlo
cluster simulations; cluster extent threshold = 43 voxels). Bar graphs display the mean signal intensity during the three conditions for particular clusters. Error bars represent 95%
confidence interval. For better comparability, all conditions are illustrated in the figure, although only the interaction contrast (*) was significant. Furthermore, HC was added as a
comparative value for the two groups. FDE (and HC) showed greater activation with increasing WM load in the right middle frontal gyrus (MFG), while RDE indicated activations patterns
comparable to a reversed U-shape.
trend. The comparison of the advanced WM conditions indicated in-
creased brain activation in MDD during 3-back > 2-back, and respec-
tively decreased activation during 2-back as compared to HC in the left
SFG; a region which is commonly associated with WM (e.g. Barch et al.,
2003; Garrett et al., 2011). Beside the confirmation of WM associated
brain regions, our results suggest neurofunctional differences in MDD as
compared to HC in important brain areas, which are discussed below.
4.2.2. Context of the affected regions
The reported brain regions – SFG, TH, CB, and AG – are commonly
associated with WM load. The thalamus is an important component of
both sensory and motor mechanisms (Herrero et al., 2002), and also –
what is of interest for our findings – for memory functioning (Johnson
and Ojemann, 2000). The SFG is involved in a variety of cognitive and
motor control tasks (Li et al., 2013), and furthermore, also crucial for
WM performance (Du Boisgueheneuc et al., 2006). However, the AG
has been attributed central features in semantic processing, attention,
or memory retrieval (Seghier, 2012).
Dysfunctions or abnormalities of the mentioned regions can lead to
cognitive impairment, which is often observed in MDD (Lee et al., 2012;
Marazziti et al., 2010; Porter et al., 2003). As our results indicated
functional deviations of both the thalamus and the SFG in MDD during
WM, we speculate that these systems are particularly affected in
memory processing in depression. Furthermore, as behavioral results
suggested equal levels of performance in HC and MDD, these brain
regions may be especially involved in compensating for cognitive (WM)
impairments. In addition, we assume the same compensatory qualities
for the other reported regions AG and CB. Accordingly, we assume that
the often-reported alterations of cognitive functioning in depressed
patients (e.g. Hasselbalch et al., 2011; Marazziti et al., 2010) - which
again often lead to severe impairments of daily life - are balanced by
compensatory mechanisms of the reported brain networks.
4.2.3. FDE vs. RDE
The second objective of this study was to investigate and compare
the MDD groups FDE vs. RDE. Our results confirmed significant
46
Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49
differences between these groups, as we found increased activation for
RDE during 2-back > 0-back in the right inferior frontal gyrus, left
precentral gyrus and bilateral middle frontal gyri, and also decreased
activation in the right middle frontal gyrus during 2-back > 3-back.
The resulting areas are usually associated with WM (IFG: Barch et al.,
2003; Jonides et al., 1998; Vasic et al., 2009. PCG: Barch et al., 2003;
Fitzgerald et al., 2008; Lee et al., 2013; MFG: Harvey et al., 2005; Vasic
et al., 2009.)
Although current research did not find a consensus regarding the
direction of brain activation in MDD [(increased (e.g. Fitzgerald et al.,
2008) vs. decreased (e.g. Vasic et al., 2009; Wang et al., 2015) brain
activation)] within the mentioned areas, our results are not comparable
to any previous study due to the unique study design. Nevertheless, our
results indicated differing BOLD signal in RDE in respective regions,
and furthermore an increasing tendency of mean signal intensity de-
pending on WM load in FDE. Accordingly, FDE showed increasing ac-
tivation during 2-back as compared to 0-back, and during 3-back as
compared to 2-back. This leads to the assumption that both HC and FDE
exhibit (linearly) increasing activation depending on task complexity,
and that this mechanism is affected in MDD respectively further af-
fected in RDE. Moreover, activation patterns in RDE tended to resemble
an inverted U-shape as described in Manoach (2003). In terms of
Manoach (2003), the activation curve of schizophrenic individuals is
shifted to the left depending on task difficulty, which means that the
decrease in activation starts earlier and during more basic conditions.
The same mechanism can be assumed for RDE in the present study.
With regard to the neurobiology of MDD-subgroups, this finding
suggests that FDE (and HC) require greater activation for demanding
WM conditions, while RDE tends to indicate a trend towards declining
activation at the highest WM demand. Thus, RDE and FDE seem to
engage different activation-shapes for similar results. Since FDE in-
dicates similar activation patterns as HC, we assume that the neural
networks of RDE function differently or are more affected than those of
FDE.
D. Yüksel et al.
4.2.4. Context of the affected regions
Related to our findings the right IFG is important for detecting re-
levant cues (Hampshire et al., 2010) which is essential for the im-
plementation of the WM task in this study. Furthermore, consistent with
our findings, the MFG is known to be associated with high-level ex-
ecutive functions (e.g. Talati and Hirsch, 2005), whereas the PCG (as
part of the primary motor cortex) is associated with finger movements
(e.g. Rao et al., 1996). All of these areas are fundamental for performing
the specific task used in our investigation, and in addition, often im-
paired in MDD. With regard to the neurobiology of MDD-subgroups, our
results indicate different activation in RDE in brain regions which are
fundamental for executive functions and therefore of great importance
for the implementation of the WM task.
4.3. Summarization and conclusions
We failed to detect behavioral differences between all groups, MDD
vs. HC and FDE vs. RDE, as the WM performance did not differ sig-
nificantly. Nevertheless, we found distinct differences on a neuroima-
ging level. Thus, both, MDD and RDE, exhibited divergent signal
changes in fronto-parietal brain regions. Hence, we were able to detect
functional differences in brain regions that are known to be involved in
higher level executive functioning. We assume that our results indicate
a relation between the observed BOLD signal in the mentioned brain
areas and the stage/course of depression. Furthermore, both groups HC
and FDE, showed similar activation patterns as they exhibited a linear
trend to increase the BOLD signal in respective regions, depending on
the complexity of the task. Nevertheless, it is possible that the linear
activation curve in HC and FDE may drop during more sophisticated
WM conditions due to a too complex WM load – as it did in RDE during
the 3-back condition. Thus, this activation pattern could be an indicator
of the severity of MDD.
Moreover, we speculate that the observed activation differences in
respective areas could reflect a vulnerability factor for MDD and, in
addition, for recurrent forms of MDD with multiple episodes. Since we
have not found any behavioral differences in our sample, we suspect
that the reported brain regions moderate the cognitive – in particular
working memory – impairments in MDD (which are frequently reported
in other studies) or moreover counterbalance for the severity of these
impairments. As neither MDD nor FDE or RDE indicated behavioral WM
differences, we speculate that the referred brain regions might be in-
volved in underlying neuropsychological mechanisms, which compen-
sate or balance the often-observed impairments of (WM) memory
functioning in MDD. In addition, we conclude that the n-back task used
in our study is a suitable tool to investigate WM in MDD, as our results
replicate current research referring to a fronto-parietal activation
cluster during WM. This further validates our findings.
4.4. Limitations
Several limitations of our study must be noted. First of all, we did
not compare equal sized patient subgroups (26 FDE vs. 48 RDE) in our
analysis. Several patients exhibited comorbid disorders (such as per-
sonality disorders). Furthermore, almost 1/3 of the MDD sample was
taking antipsychotics at the time of measurement. However, we did not
have any influence on the respective medication, as the treating phy-
sician medicated all subjects according to their symptoms.
Furthermore, most subjects were unable to participate in the present
study without receiving medical treatment. However, the listed co-
morbidities and drug use are common and widespread in MDD. Our
sample therefore represents a common clinical population and is dis-
tributed evenly among the groups. The medication was comparable
between the two MDD groups and it is unlikely to have caused the
differences.
Our study is a further step towards a better understanding of major
depressive disorder and the associated cognitive impairments,
47
Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49
especially working memory. Since MDD subgroups differed in brain
activation during the task, future research should always report whe-
ther patients had first or recurrent depressive episodes. Furthermore,
prospective studies should include higher and more demanding WM
conditions to better illustrate possibly dropping activation curves in
participants. In addition, it would be of great interest to investigate the
trajectories of neural correlates of WM in different patients groups (e.g.
bipolar disorder or schizophrenia) with a longitudinal design to assess
how underlying activation patterns change as a function of future epi-
sodes or lack thereof. Furthermore, methods such as support vector
machines should investigate whether subsequent depressive episodes or
the course of disease/disease progression can be predicted.
Author disclosures
Carsten Konrad received fees for an educational program from
Aristo Pharma, Janssen-Cilag, Lilly, MagVenture, Servier, and
Trommsdorff as well as travel support and speakers honoraria from
Aristo Pharma, Janssen-Cilag, Lundbeck, and Servier.
Conflict of interest
All authors report no conflict of interest.
Role of funding source
This work was supported by grants from the German Research
Foundation (DFG, grant no. KR 3822/2-1 to AK, grant no. DA 1151/5-1
to UD, grant no. KI 588/14-1 and grant no. KO 4291/3-1 to CK and AK),
a grant by the University Medical Center Giessen and Marburg (UKGM,
grant no. 11/2010 MR) and the Von-Behring-Röntgen-Foundation
(grant no. 61-0030).
References
Airaksinen, E., Larsson, M., Lundberg, I., Forsell, Y., 2004. Cognitive functions in de-
pressive disorders: evidence from a population-based study. Psychol. Med. 34, 83–91.
Barch, D.M., Sheline, Y.I., Csernansky, J.G., Snyder, A.Z., 2003. Working memory and
prefrontal cortex dysfunction: Specificity to schizophrenia compared with major
depression. Biol. Psychiatry 53, 376–384. http://dx.doi.org/10.1016/S0006-
3223(03)01674-8.
Bartova, L., Meyer, B.M., Diers, K., Rabl, U., Scharinger, C., Popovic, A., Pail, G., Kalcher,
K., Boubela, R.N., Huemer, J., Mandorfer, D., Windischberger, C., Sitte, H.H., Kasper,
S., Praschak-Rieder, N., Moser, E., Brocke, B., Pezawas, L., 2015. Reduced default
mode network suppression during a working memory task in remitted major de-
pression. J. Psychiatr. Res. 64, 9–18. http://dx.doi.org/10.1016/j.jpsychires.2015.
02.025.
Beck, A.T., Steer, R.A., Carbin, M.G., 1988. Psychometric properties of the Beck
Depression Inventory: Twenty-five years of evaluation. Clin. Psychol. Rev. 8, 77–100.
http://dx.doi.org/10.1016/0272-7358(88)90050-5.
Bourke, C., Porter, R.J., Carter, J.D., Mcintosh, V.V., Jordan, J., Bell, C., Carter, F.,
Colhoun, H., Joyce, P.R., 2012. Comparison of neuropsychological functioning and
emotional processing in major depression and social anxiety disorder subjects, and
matched healthy controls. 46. http://dx.doi.org/10.1177/0004867412451502.
Bowie, C.R., Harvey, P.D., 2006. Administration and interpretation of the Trail Making
Test. Nat. Protoc. 1, 2277–2281. http://dx.doi.org/10.1038/nprot.2006.390.
Braver, T.S., Cohen, J.D., Nystrom, L.E., Jonides, J., Smith, E.E., Noll, D.C., 1997. A
parametric study of prefrontal cortex involvement in human working memory.
NeuroImage 5, 49–62. http://dx.doi.org/10.1006/nimg.1996.0247.
Brooks, J.O., Vizueta, N., Penfold, C., Townsend, J.D., Bookheimer, S.Y., Altshuler, L.L.,
2015. Prefrontal hypoactivation during working memory in bipolar II depression.
Psychol. Med. 45, 1731–1740. http://dx.doi.org/10.1017/S0033291714002852.
Callicott, J.H., Mattay, V.S., Bertolino, A., Finn, K., Coppola, R., Frank, J.A., Goldberg,
T.E., Weinberger, D.R., 1999. Physiological characteristics of capacity constraints in
working memory as revealed by functional MRI. Cereb. Cortex 9, 20–26. http://dx.
doi.org/10.1093/cercor/9.1.20.
Christopher, G., MacDonald, J., 2005. The impact of clinical depression on working
memory. Cogn. Neuropsychiatry 10, 379–399. http://dx.doi.org/10.1080/
13546800444000128.
Cohen, J.D., Forman, S.D., Braver, T.S., Casey, B.J., Servan-Schreiber, D., Noll, D.C.,
1994. Activation of the prefrontal cortex in a nonspatial working memory task with
functional MRI. Hum. Brain Mapp. 1, 293–304. http://dx.doi.org/10.1002/hbm.
460010407.
Deshpande, S., Bhatia, T., Shriharsh, V., Adlakha, S., Bisht, V., Garg, K., 2007. The trail
making test in India. Indian J. Psychiatry 49, 113. http://dx.doi.org/10.4103/0019-
D. Yüksel et al.
5545.33258.
Dichter, G.S., Kozink, R.V., McClernon, F.J., Smoski, M.J., 2012. Remitted major de-
pression is characterized by reward network hyperactivation during reward antici-
pation and hypoactivation during reward outcomes. J. Affect. Disord. 136,
1126–1134. http://dx.doi.org/10.1016/j.jad.2011.09.048.
Drobnik, L., 2014. Evaluation of the Trail Making Test and interval timing as measures of
cognition in healthy adults: Comparisons by age, education, and gender. Med. Sci.
Monit. 20, 173–181. http://dx.doi.org/10.12659/MSM.889776.
Du Boisgueheneuc, F., Levy, R., Volle, E., Seassau, M., Duffau, H., Kinkingnehun, S.,
Samson, Y., Zhang, S., Dubois, B., 2006. Functions of the left superior frontal gyrus in
humans: A lesion study. Brain 129, 3315–3328. http://dx.doi.org/10.1093/brain/
awl244.
Dumas, J.A., Newhouse, P.A., 2014. Impaired working memory in geriatric depression:
An fMRI study. Am. J. Geriatr. Psychiatry. http://dx.doi.org/10.1016/j.jagp.2014.09.
011.
Evans, V.C., Iverson, G.L., Yatham, L.N., Lam, R.W., 2014. The relationship between
neurocognitive and psychosocial functioning in major depressive disorder: a sys-
tematic review. J. Clin. Psychiatry 75, 1359–1370. http://dx.doi.org/10.4088/JCP.
13r08939.
Fernández-Corcuera, P., Salvador, R., Monté, G.C., Salvador Sarró, S., Goikolea, J.M.,
Amann, B., Moro, N., Sans-Sansa, B., Ortiz-Gil, J., Vieta, E., Maristany, T., McKenna,
P.J., Pomarol-Clotet, E., 2013. Bipolar depressed patients show both failure to acti-
vate and failure to de-activate during performance of a working memory task. J.
Affect. Disord. 148, 170–178. http://dx.doi.org/10.1016/j.jad.2012.04.009.
Fitzgerald, P.B., Srithiran, A., Benitez, J., Daskalakis, Z.Z., Oxley, T.J., Kulkarni, J., Egan,
G.F., 2008. An fMRI study of prefrontal brain activation during multiple tasks in
patients with major depressive disorder. Hum. Brain Mapp. 29, 490–501. http://dx.
doi.org/10.1002/hbm.20414.
Fossati, P., Harvey, P.-O., Le Bastard, G., Ergis, A.-M., Jouvent, R., Allilaire, J.-F., 2004.
Verbal memory performance of patients with a first depressive episode and patients
with unipolar and bipolar recurrent depression. J. Psychiatr. Res. 38, 137–144.
http://dx.doi.org/10.1016/j.jpsychires.2003.08.002.
Garrett, A., Kelly, R., Gomez, R., Keller, J., Schatzberg, A.F., Reiss, A.L., 2011. Aberrant
brain activation during a working memory task in psychotic major depression. Am. J.
Psychiatry 168, 173–182. http://dx.doi.org/10.1176/appi.ajp.2010.09121718.
Goethals, I., Audenaert, K., Jacobs, F., Van De Wiele, C., Ham, H., Pyck, H.,
Vandierendonck, A., Van Heeringen, C., Dierckx, R., 2005. Blunted prefrontal per-
fusion in depressed patients performing the Tower of London task. Psychiatry Res.
Neuroimaging 139, 31–40. http://dx.doi.org/10.1016/j.pscychresns.2004.09.007.
Gould, R.L., Brown, R.G., Owen, A.M., Ffytche, D.H., Howard, R.J., 2003. fMRI BOLD
response to increasing task difficulty during successful paired associates learning.
NeuroImage 20, 1006–1019. http://dx.doi.org/10.1016/S1053-8119(03)00365-3.
Hampshire, A., Chamberlain, S.R., Monti, M.M., Duncan, J., Owen, A.M., 2010. The role
of the right inferior frontal gyrus: inhibition and attentional control. NeuroImage 50,
1313–1319. http://dx.doi.org/10.1016/j.neuroimage.2009.12.109.
Harvey, P.-O., Fossati, P., Pochon, J.-B., Levy, R., Lebastard, G., Lehéricy, S., Allilaire, J.-
F., Dubois, B., 2005. Cognitive control and brain resources in major depression: an
fMRI study using the n-back task. NeuroImage 26, 860–869. http://dx.doi.org/10.
1016/j.neuroimage.2005.02.048.
Hasselbalch, B.J., Knorr, U., Kessing, L.V., 2011. Cognitive impairment in the remitted
state of unipolar depressive disorder: A systematic review. J. Affect. Disord. 134,
20–31. http://dx.doi.org/10.1016/j.jad.2010.11.011.
Hasselbalch, B.J., Knorr, U., Hasselbalch, S.G., Gade, A., Kessing, L.V., 2013. The cu-
mulative load of depressive illness is associated with cognitive function in the re-
mitted state of unipolar depressive disorder. Eur. Psychiatry 28, 349–355. http://dx.
doi.org/10.1016/j.eurpsy.2012.03.004.
Hautzinger, M., Bailer, M., Worall, H., Keller, F., 1995. Beck Depressions-Inventar (BDI).
Testhandbuch.
Herrero, M.T., Barcia, C., Navarro, J.M., 2002. Functional anatomy of thalamus and basal
ganglia. Childs Nerv. Syst. 18, 386–404. http://dx.doi.org/10.1007/s00381-002-
0604-1.
Ho, D.E., Imai, K., King, G., Stuart, E.A., 2011. MatchIt: nonparametric preprocessing for
parametric causal inference. J. Stat. Softw. 42, 1–28.
Holt, R.J., Graham, J.M., Whitaker, K.J., Hagan, C.C., Ooi, C., Wilkinson, P.O., Van
Nieuwenhuizen, A.O., Lennox, B.R., Sahakian, B.J., Goodyer, I.M., Bullmore, E.T.,
Suckling, J., 2016. Functional MRI of emotional memory in adolescent depression.
Dev. Cogn. Neurosci. 19, 31–41. http://dx.doi.org/10.1016/j.dcn.2015.12.013.
Ilsley, J.E., Moffoot, A.P., O'Carroll, R.E., 1995. An analysis of memory dysfunction in
major depression. J. Affect. Disord. 35, 1–9.
Jacola, L.M., Willard, V.W., Ashford, J.M., Ogg, R.J., Scoggins, M.A., Jones, M.M., Wu, S.,
Conklin, H.M., 2014. Clinical utility of the N-back task in functional neuroimaging
studies of working memory. J. Clin. Exp. Neuropsychol. 36, 875–886. http://dx.doi.
org/10.1080/13803395.2014.953039.
Jansma, J.M., Ramsey, N.F., Coppola, R., Kahn, R.S., 2000. Specific versus Nonspecific
Brain Activity in a Parametric N-Back Task. NeuroImage 12, 688–697. http://dx.doi.
org/10.1006/nimg.2000.0645.
Johnson, M.D., Ojemann, G.A., 2000. The role of the human thalamus in language and
memory: evidence from electrophysiological studies. Brain Cogn. 42, 218–230.
http://dx.doi.org/10.1006/brcg.1999.1101.
Jonides, J., Schumacher, E.H., Smith, E.E., Koeppe, R. a, Awh, E., Reuter-Lorenz, P.A,
Marshuetz, C., Willis, C.R., 1998. The role of parietal cortex in verbal working
memory. J. Neurosci. 18, 5026–5034.
Joy, S., Kaplan, E., Fein, D., 2004. Speed and memory in the WAIS-III digit symbol -
coding subtest across the adult lifespan. Arch. Clin. Neuropsychol. 19, 759–767.
http://dx.doi.org/10.1016/j.acn.2003.09.009.
Karabekiroğlu, A., Topçuoğlu, V., Gimzal Gönentür, A., Karabekiroğlu, K., 2010.
48
Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49
Executive function differences between first episode and recurrent major depression
patients. Turk Psikiyatri Derg. 21, 280–288.
Kerestes, R., Ladouceur, C.D., Meda, S., Nathan, P.J., Blumberg, H.P., Maloney, K., Ruf,
B., Saricicek, A., Pearlson, G.D., Bhagwagar, Z., Phillips, M.L., 2012. Abnormal pre-
frontal activity subserving attentional control of emotion in remitted depressed pa-
tients during a working memory task with emotional distracters. Psychol. Med.
http://dx.doi.org/10.1017/S0033291711001097.
Kerestes, R., Harrison, B.J., Dandash, O., Stephanou, K., Whittle, S., Pujol, J., Davey, C.G.,
2015. Specific functional connectivity alterations of the dorsal striatum in young
people with depression. NeuroImage Clin. 7, 266–272. http://dx.doi.org/10.1016/j.
nicl.2014.12.017.
Kessels, R.P.C., van Zandvoort, M.J.E., Postma, A., Kappelle, L.J., de Haan, E.H.F., 2000.
The Corsi Block-Tapping Task: Standardization and Normative Data. Appl.
Neuropsychol. 7, 252–258. http://dx.doi.org/10.1207/S15324826AN0704_8.
Kirschen, M.P., Chen, S.H.A., Schraedley-Desmond, P., Desmond, J.E., 2005. Load- and
practice-dependent increases in cerebro-cerebellar activation in verbal working
memory: An fMRI study. NeuroImage 24, 462–472. http://dx.doi.org/10.1016/j.
neuroimage.2004.08.036.
Kondo, H., Morishita, M., Osaka, N., Osaka, M., Fukuyama, H., Shibasaki, H., 2004.
Functional roles of the cingulo-frontal network in performance on working memory.
NeuroImage 21, 2–14. http://dx.doi.org/10.1016/j.neuroimage.2003.09.046.
Korsnes, M.S., Lövdahl, H., Andersson, S., Björnerud, A., Due-Tönnesen, P., Endestad, T.,
Malt, U.F., 2013. Working memory in recurrent brief depression: An fMRI pilot study.
J. Affect. Disord. 149, 383–392. http://dx.doi.org/10.1016/j.jad.2013.02.017.
Krug, A., Markov, V., Eggermann, T., Krach, S., Zerres, K., Stöcker, T., Shah, N.J.,
Schneider, F., Nöthen, M.M., Treutlein, J., Rietschel, M., Kircher, T., 2008. Genetic
variation in the schizophrenia-risk gene neuregulin1 correlates with differences in
frontal brain activation in a working memory task in healthy individuals.
NeuroImage 42, 1569–1576. http://dx.doi.org/10.1016/j.neuroimage.2008.05.058.
Kühner, C., Bürger, C., Keller, F., Hautzinger, M., 2007. Reliability and validity of the
Revised Beck Depression Inventory (BDI-II). Results from German samples.
Nervenarzt 78, 651–656. http://dx.doi.org/10.1007/s00115-006-2098-7.
Laschkowski, W., Dichtler, O., Schulpsychologe, S., Flessa, R., Käsdorf, M., Tharandt, K.,
Schulpsychologin, S., Walden, K., 2010. Arbeitsmaterialien zum HAWIK-IV.
Lawrence, C., Roy, A., Harikrishnan, V., Yu, S., Dabbous, O., 2013. Association Between
Severity of Depression and Self-Perceived Cognitive Difficulties Among Full-time
Employees. vol. 15. pp. 30–34. http://dx.doi.org/10.4088/PCC.12m01469.
Lee, R.S.C., Hermens, D.F., Porter, M.A., Redoblado-Hodge, M.A., 2012. A meta-analysis
of cognitive deficits in first-episode Major Depressive Disorder. J. Affect. Disord. 140,
113–124. http://dx.doi.org/10.1016/j.jad.2011.10.023.
Lee, T.W., Liu, H.L., Wai, Y.Y., Ko, H.J., Lee, S.H., 2013. Abnormal neural activity in
partially remitted late-onset depression: An fMRI study of one-back working memory
task. Psychiatry Res. Neuroimaging 213, 133–141. http://dx.doi.org/10.1016/j.
pscychresns.2012.04.010.
Lehrl, S., Triebig, G., Fischer, B., 1995. Multiple choice vocabulary test MWT as a valid
and short test to estimate premorbid intelligence. Acta Neurol. Scand. 91, 335–345.
http://dx.doi.org/10.1111/j.1600-0404.1995.tb07018.x.
Li, W., Qin, W., Liu, H., Fan, L., Wang, J., Jiang, T., Yu, C., 2013. Subregions of the human
superior frontal gyrus and their connections. NeuroImage 78, 46–58. http://dx.doi.
org/10.1016/j.neuroimage.2013.04.011.
Liu, C.-H., Ma, X., Yuan, Z., Song, L.-P., Jing, B., Lu, H.-Y., Tang, L.-R., Fan, J., Walter, M.,
Liu, C.-Z., Wang, L., Wang, C.-Y., 2017. Decreased Resting-State Activity in the
Precuneus Is Associated With Depressive Episodes in Recurrent Depression. J. Clin.
Psychiatry 78, e372–e382. http://dx.doi.org/10.4088/JCP.15 m10022.
Maes, M., 2015. Mechanisms Underlying Neurocognitive Dysfunctions in Recurrent Major
Depression. Med. Sci. Monit. 21, 1535–1547. http://dx.doi.org/10.12659/MSM.
893176.
Manoach, D.S., 2003. Prefrontal cortex dysfunction during working memory performance
in schizophrenia: reconciling discrepant findings. Schizophr. Res. 60, 285–298.
http://dx.doi.org/10.1016/S0920-9964(02)00294-3.
Marazziti, D., Consoli, G., Picchetti, M., Carlini, M., Faravelli, L., 2010. Cognitive im-
pairment in major depression. Eur. J. Pharmacol. 626, 83–86. http://dx.doi.org/10.
1016/j.ejphar.2009.08.046.
Matsuo, K., Glahn, D.C., Peluso, M.A., Hatch, J.P., Monkul, E.S., Najt, P., Sanches, M.,
Zamarripa, F., Li, J., Lancaster, J.L., Fox, P.T., Gao, J.-H., Soares, J.C., 2007.
Prefrontal hyperactivation during working memory task in untreated individuals
with major depressive disorder. Mol. Psychiatry 12, 158–166. http://dx.doi.org/10.
1038/sj.mp.4001894.
McIntyre, R.S., Cha, D.S., Soczynska, J.K., Woldeyohannes, H.O., Gallaugher, L.A.,
Kudlow, P., Alsuwaidan, M., Baskaran, A., 2013. Cognitive deficits and functional
outcomes in major depressive disorder: determinants, substrates, and treatment in-
terventions. Depress. Anxiety 30, 515–527. http://dx.doi.org/10.1002/da.22063.
Meng, C., Brandl, F., Tahmasian, M., Shao, J., Manoliu, A., Scherr, M., Schwerthöffer, D.,
Bäuml, J., Förstl, H., Zimmer, C., Wohlschläger, A.M., Riedl, V., Sorg, C., 2014.
Aberrant topology of striatum?s connectivity is associated with the number of epi-
sodes in depression. Brain 137, 598–609. http://dx.doi.org/10.1093/brain/awt290.
Norbury, R., Godlewska, B., Cowen, P.J., 2014. When less is more: a functional magnetic
resonance imaging study of verbal working memory in remitted depressed patients.
Psychol. Med. 44, 1197–1203. http://dx.doi.org/10.1017/S0033291713001682.
Opmeer, E.M., Kortekaas, R., van Tol, M.J., van der Wee, N.J. A, Woudstra, S., van
Buchem, M. a., Penninx, B.W., Veltman, D.J., Aleman, A., 2013. Influence of COMT
val158met Genotype on the Depressed Brain during Emotional Processing and
Working Memory. PLoS One 8, e73290. doi:https://doi.org/10.1371/journal.pone.
0073290.
Owen, A.M., McMillan, K.M., Laird, A.R., Bullmore, E., 2005. N-back working memory
paradigm: a meta-analysis of normative functional neuroimaging studies. Hum. Brain
D. Yüksel et al.
Mapp. 25, 46–59. http://dx.doi.org/10.1002/hbm.20131.
Porter, R.J., Gallagher, P., Thompson, J.M., Young, A.H., 2003. Neurocognitive impair-
ment in drug-free patients with major depressive disorder. Br. J. Psychiatry 182,
214–220. http://dx.doi.org/10.1192/bjp.182.3.214.
Pu, S., Yamada, T., Yokoyama, K., Matsumura, H., Mitani, H., Adachi, A., Kaneko, K.,
Nakagome, K., 2012. Reduced prefrontal cortex activation during the working
memory task associated with poor social functioning in late-onset depression: Multi-
channel near-infrared spectroscopy study. Psychiatry Res. Neuroimaging 203,
222–228. http://dx.doi.org/10.1016/j.pscychresns.2012.01.007.
Rao, S.M., Bandettini, P.A., Binder, J.R., Bobholz, J.A., Hammeke, T.A., Stein, E.A., Hyde,
J.S., 1996. Relationship between finger movement rate and functional magnetic re-
sonance signal change in human primary motor cortex. J. Cereb. Blood Flow Metab.
16, 1250–1254. http://dx.doi.org/10.1097/00004647-199611000-00020.
Reitan, R.M., 1955. The relation of the trail making test to organic brain damage. J.
Consult. Psychol. 19, 393–394. http://dx.doi.org/10.1037/h0044509.
Roca, M., Vives, M., López-navarro, E., García-campayo, J., 2015. Cognitive impairments
and depression: a critical review. 43, 187–193.
Rodríguez-Cano, E., Sarró, S., Monté, G.C., Maristany, T., Salvador, R., McKenna, P.J.,
Pomarol-Clotet, E., 2014. Evidence for structural and functional abnormality in the
subgenual anterior cingulate cortex in major depressive disorder. Psychol. Med. 44,
3263–3273. http://dx.doi.org/10.1017/S0033291714000841.
Rose, E.J., Ebmeier, K.P., 2006. Pattern of impaired working memory during major de-
pression. J. Affect. Disord. 90, 149–161. http://dx.doi.org/10.1016/j.jad.2005.11.
003.
Rottschy, C., Langner, R., Dogan, I., Reetz, K., Laird, A.R., Schulz, J.B., Fox, P.T., Eickhoff,
S.B., 2012. Modelling neural correlates of working memory: a coordinate-based meta-
analysis. NeuroImage 60, 830–846. http://dx.doi.org/10.1016/j.neuroimage.2011.
11.050.
Schelling, D., 1997. BTT Block-Tapping-Test. Swets Test Services, Frankfurt, Germany.
Schöning, S., Zwitserlood, P., Engelien, A., Behnken, A., Kugel, H., Schiffbauer, H., Lipina,
K., Pachur, C., Kersting, A., Dannlowski, U., Baune, B.T., Zwanzger, P., Reker, T.,
Heindel, W., Arolt, V., Konrad, C., 2009. Working-memory fMRI reveals cingulate
hyperactivation in euthymic major depression. Hum. Brain Mapp. 30, 2746–2756.
http://dx.doi.org/10.1002/hbm.20702.
Seghier, M., 2012. The Angular. Multiple Functions and Multiple Subdivisions. Neurosci,
Gyrus. http://dx.doi.org/10.1177/1073858412440596.
Snyder, H.R., 2014. NIH Public. Access 139, 81–132. http://dx.doi.org/10.1037/
a0028727.Major.
Talarowska, M., Zajaczkouska, M., Galecki, P., 2015. Cognitive function in first episode
depression and recurrent depressive disorder. Psychiatr. Danub. 27, 31–43.
Talati, A., Hirsch, J., 2005. Functional specialization within the medial frontal gyrus for
perceptual go/no-go decisions based on “what,” “when,” and “where” related in-
formation: an fMRI study. J. Cogn. Neurosci. 17, 981–993. http://dx.doi.org/10.
1162/0898929054475226.
Vasic, N., Walter, H., Sambataro, F., Wolf, R.C., 2009. Aberrant functional connectivity of
dorsolateral prefrontal and cingulate networks in patients with major depression
49
Progress in Neuropsychopharmacology & Biological Psychiatry 84 (2018) 39–49
during working memory processing. Psychol. Med. 39, 977–987. http://dx.doi.org/
10.1017/S0033291708004443.
Veltman, D.J., Rombouts, S.A.R.B., Dolan, R.J., 2003. Maintenance versus manipulation
in verbal working memory revisited: an fMRI study. NeuroImage 18, 247–256.
Vilgis, V., Chen, J., Silk, T.J., Cunnington, R., Vance, A., 2014. Frontoparietal function in
young people with dysthymic disorder (DSM-5: Persistent depressive disorder) during
spatial working memory. J. Affect. Disord. 160, 34–42. http://dx.doi.org/10.1016/j.
jad.2014.01.024.
Wager, T.D., Smith, E.E., 2003. Neuroimaging studies of working memory: a meta-ana-
lysis. Cogn. Affect. Behav. Neurosci. 3, 255–274. http://dx.doi.org/10.3758/CABN.3.
4.255.
Walsh, N.D., Williams, S.C.R., Brammer, M.J., Bullmore, E.T., Kim, J., Suckling, J.,
Mitterschiffthaler, M.T., Cleare, A.J., Pich, E.M., Mehta, M.A., Fu, C.H.Y., 2007. A
longitudinal functional magnetic resonance imaging study of verbal working memory
in depression after antidepressant therapy. Biol. Psychiatry 62, 1236–1243. http://
dx.doi.org/10.1016/j.biopsych.2006.12.022.
Walter, H., Wolf, R.C., Spitzer, M., Vasic, N., 2007. Increased left prefrontal activation in
patients with unipolar depression: An event-related, parametric, performance-con-
trolled fMRI study. J. Affect. Disord. 101, 175–185. http://dx.doi.org/10.1016/j.jad.
2006.11.017.
Wang, X.-L., Du, M.-Y., Chen, T.-L., Chen, Z.-Q., Huang, X.-Q., Luo, Y., Zhao, Y.-J., Kumar,
P., Gong, Q.-Y., 2015. Neural correlates during working memory processing in major
depressive disorder. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 56, 101–108.
http://dx.doi.org/10.1016/j.pnpbp.2014.08.011.
Wechsler, D., 1981. Wechsler Adult Intelligence Scale-Revised Manual (WAIS-R).
Wechsler, D., 1997. Wechsler Memory Scale - Third Edition. Manual. The Psychological
Corporation, San Antonio, TX.
Wittchen, H.-U., Zaudig, M., Frydrich, T., 1997. SKID-I. Strukturiertes klinisches
Interview für DSM-IV. Achse I: Psychische Störungen. Göttingen, Hogrefe.
Yüksel, D., Bruno, D., Forstner, A.J., Witt, S.H., Maier, R., Rietschel, Marcella, Konrad, C.,
Nöthen, Markus M., Dannlowski, U., Baune, B.T., Kircher, T., Krug, A., 2017.
Polygenic risk for depression and the neural correlates of working memory in healthy
subjects. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 67–76. Accepted. https://
doi.org/10.1016/j.pnpbp.2017.06.010.
Zakzanis, K.K., Leach, L., Kaplan, E., 1998. On the nature and pattern of neurocognitive
function in major depressive disorder. Neuropsychiatry Neuropsychol. Behav. Neurol.
11, 111–119.
Zaninotto, L., Guglielmo, R., Calati, R., Ioime, L., Camardese, G., Janiri, L., Bria, P.,
Serretti, A., 2014. Cognitive markers of psychotic unipolar depression: A meta-ana-
lytic study. J. Affect. Disord. 174C, 580–588. http://dx.doi.org/10.1016/j.jad.2014.
11.027.
Zhou, M., Hu, X., Lu, L., Zhang, L., Chen, L., Gong, Q., Huang, X., 2017. Intrinsic cerebral
activity at resting state in adults with major depressive disorder: A meta-analysis.
Prog. Neuro-Psychopharmacol. Biol. Psychiatry 75, 157–164. http://dx.doi.org/10.
1016/j.pnpbp.2017.02.001.