Available via license: CC BY-NC 3.0
Content may be subject to copyright.
Increased involvement of the parahippocampal gyri in a
sad mood predicts future depressive symptoms
Vera Zamoscik,
1
Silke Huffziger,
2
Ulrich Ebner-Priemer,
3
Christine Kuehner,
2,
* and Peter Kirsch
1,
*
1
Department of Clinical Psychology,
2
Research Group Longitudinal and Intervention Research, Department of Psychiatry and Psychotherapy,
Central Institute of Mental Health, Mannheim and Medical Faculty Mannheim, Heidelberg University, and
3
Chair of Applied Psychology,
Department of Sport and Sport Science and House of Competence, Karlsruhe Institute of Technology, Germany
Behavioral studies suggest a relationship between autobiographical memory, rumination and depression. The objective of this study was to determine
whether remitted depressed patients show alterations in connectivity of the posterior cingulate cortex (PCC, a node in the default mode network) with
the parahippocampal gyri (PHG, a region associated with autobiographical memory) while intensively recalling negative memories and whether this is
related to daily life symptoms and to the further course of depression. Sad mood was induced with keywords of personal negative life events in
participants with remitted depression (n¼29) and matched healthy controls (n¼29) during functional magnetic resonance imaging. Additionally,
daily life assessments of mood and rumination and a 6-month follow-up were conducted. Remitted depressed participants showed greater connectivity
than healthy controls of the PCC with the PHG, which was even stronger in patients with more previous episodes. Furthermore, patients with increased
PCC–PHG connectivity showed a sadder mood and more rumination in daily life and a worsening of rumination and depression scores during follow-up. A
relationship of negative autobiographical memory processing, rumination, sad mood and depression on a neural level seems likely. The identified
increased connectivity probably indicates a scar of recurrent depression and may represent a prognostic factor for future depression.
Keywords: depression; parahippocampal gyri; autobiographical memory; default mode network; rumination
INTRODUCTION
Major depressive disorder is a mental disorder characterized by lasting
pervasive feelings of low mood, guilt and worthlessness. One factor
that has been linked to cognitive vulnerability in depression is rumin-
ationrepeatedly thinking about one’s negative emotional state and its
possible causes and consequences (Nolen-Hoeksema et al., 2008;
Huffziger et al., 2009). According to the perseverative cognition hy-
pothesis by Brosschot and colleagues (Brosschot et al., 2005), rumin-
ation is a mediator between stressors and prolonged stress-related
processes by entailing an ongoing mental representation of the stressor
(cf. Huffziger et al., 2013). Such stressors could be negative life events,
that is, autobiographical experiences and related attention processes,
which play an important role in depression (Sumner et al., 2010;
Whalley et al., 2012) and rumination (Watson et al., 2012).
Emotions play a major formative role in episodic memory (see Dere
et al., 2010 for an overview) and particularly in autobiographical
memory (Williams and Broadbent, 1986), the part of episodic
memory which is highly self-referential. Therefore, it is not surprising
that emotional memories are generally remembered more often than
neutral memories (Howes et al., 1993;Moritz et al., 2005). Positive
autobiographical memories contain more sensory and contextual de-
tails than neutral or negative memories (D’Argembeau et al., 2003),
whereas negative autobiographical memories, which seem to be ‘more
central to identity’ in depressive patients (Watson, et al., 2012), appear
to be more salient and remembered in a more general way, possibly
sustained and boosted by rumination.
Rumination has also been linked to an increased activity in the
brain’s default mode network (DMN; Berman et al., 2011;Hamilton
et al., 2011). The DMN is an activity pattern of the brain, presumably
representing self-referential processes (Whitfield-Gabrieli et al., 2011).
Major brain regions involved are the posterior cingulate cortex (PCC),
the medial prefrontal cortex and the angular gyri. Increased connect-
ivity and activity in the DMN have been reported in depressed indi-
viduals (Greicius et al., 2007;Sheline et al., 2009,2010;Berman et al.,
2011;Li et al., 2013). Recently, Zhu and colleagues (2012) described
two dissociated DMNs in depressive patients in comparison with
healthy controls, one of increased functional connectivity between
the medial prefrontal cortex and anterior cingulate cortex and the
other one posterior, including PCC and precuneus, showing decreased
connectivity. Although these findings are partially inconsistent, report-
ing both increased and decreased connectivity, the overall conclusion is
that altered DMN connectivity and activity seem to occur in depres-
sion and may, for example, predict future relapses (Farb et al., 2011)or
help to classify participants with resting state functional connectivity
data as depressed or non-depressed (Zeng et al., 2012).
Interestingly, one of the four best regions for discriminating between
depressed and healthy participants in the aforementioned analyses by
Zeng and colleagues (2012) was the parahippocampal gyrus (PHG)
which showed, among others, altered connectivity to the PCC and
the anterior cingulate cortex in patients when compared with controls.
Furthermore, Cooney and colleagues found greater activity in the
PHG during rumination versus abstract distraction in depressive
patients in contrast to healthy controls (Cooney et al., 2010).
Moreover, the posterior default mode network seems to recruit
among others the parahippocampal gyri during episodic memory re-
trieval (Sestieri et al., 2011). These findings could implicate a connec-
tion between DMN aberrations in depressed patients, rumination and
autobiographical memory processes. In general, autobiographical
memory pathways in the brain seem to include the PHG. For example,
patients with transient epileptic amnesia, displaying specific autobio-
graphical memory problems, show decreased activity in the posterior
PHG (Milton et al., 2012). Furthermore, a recent meta-analysis showed
that the left posterior parahippocampal gyrus was more likely to be
recruited during recall or imagination of personal events compared
Received 8 May 2013; Revised 17 December 2013; Accepted 10 January 2014
Advance Access publication 3 February 2014
*These authors contributed equally to this work.
The project was supported by grants of the German Research Foundation (Deutsche Forschungsgemeinschaft,
DFG; KI 576/12-1 to P.K., KU 1464/4-1 to C.K.).
Correspondence should be addressed to Vera Zamoscik, Department of Clinical Psychology, Central Institute of
Mental Health, Mannheim and Medical Faculty Mannheim, Heidelberg University, ZI, J5, 68159 Mannheim,
Germany. E-mail: vera.zamoscik@zi-mannheim.de
doi:10.1093/scan/nsu006 SCAN (2014) 9, 2034^2040
ßThe Aut hor (2014). Published by Oxford University Press. For Permissio ns, please email: journals.permis sions@ oup.com
with recognition tasks (Viard et al., 2012). In a further quantitative
meta-analysis, Spreng and colleagues demonstrated a strong overlap of
the autobiographical brain network and the DMN, particularly in the
PHG (Spreng et al., 2009). Interestingly, increased gray matter volume
in the parahippocampal gyri was found in meditators (Leung et al.,
2013), whereas the cortex was thinner in this area in late life depressive
patients who did not respond to psychotherapy (Mackin et al., 2012).
Summing up, autobiographical memory seems to be linked to the
DMN, the PHG and to rumination. However, the role of autobio-
graphical memory processes, their neurobiological underpinning and
interaction with the DMN and their relation to daily life experiences in
depression are not yet fully understood. Recently, Hamilton and col-
leagues proposed that dysfunctional interactions between different
neural networks could be an important aspect in understanding
depressive pathology (Hamilton et al., 2013). They argue that an
altered inter-connection between DMN and task-positive networks
like the executive or the salience network might contribute to the
pathophysiology of depression. Following this argument, it is an inter-
esting issue whether in depression the DMN shows altered connections
to structures associated with autobiographical memory. Therefore, in
this study, we address the question whether the PHG are more strongly
connected with the PCC as a core region of the DMN during a negative
affective state induced by recalling and immersing negative autobio-
graphical life events in patients with remitted depression compared
with healthy controls. This induction procedure should increase self-
referential processes and therefore increase DMN connectivity.
Remitted patients were chosen as we expect the altered connectivity
to be vulnerability-related and therefore it should be present even
when the criteria of a major depressive episode are currently not ful-
filled. If a neural signature of autobiographical memory processing is
indicative of vulnerability, it should be related to both present and
future depressive symptoms and rumination. We, therefore, examined
whether connectivity between PCC and PHG is related to daily life
measures of mood and rumination and predicts changes in depression
and rumination scores in a 6-month follow-up.
METHODS
Participants
Participants were 29 remitted depressed patients (rMDD) with at least
two previous major depressive episodes (n¼27) or a previous chronic
major depressive episode of at least 2 years duration (n¼2) according
to DSM-IV and 29 healthy controls (HC), who were individually
matched to the patients by age, sex and education level. All patients
had to be in a state of partial or full remission, that is they, did not
fulfill the criteria of a major depressive episode according to DSM-IV
for at least the previous 2 months. Participants were recruited using
online (e.g. on the homepage of the Central Institute of Mental Health)
and newspaper call outs. The study initially included 64 individuals
aged between 26 and 55 years; however, six cases (three out of each
group) were excluded from analyses due to altered anatomical or other
physiological parameters (head motion exclusion criteria: rotation
>1.58, translation >2 mm). One patient fulfilled criteria for a
co-morbid current agoraphobia and another patient for a co-morbid
current social phobia. A detailed sample description is available in
Table 1.
Exclusion criteria for both groups were bipolar and psychotic dis-
orders, substance dependence, current substance abuse, generalized
anxiety disorder, current obsessive-compulsive, post-traumatic stress
and eating disorders according to DSM-IV as well as contraindications
for the functional magnetic resonance imaging (fMRI).
Psychopathology-related in- and exclusion criteria were assessed by a
trained clinical psychologist with the Structured Clinical Interview for
DSM-IV axis I (SCID; Wittchen et al., 1997) applied in an individual
session.
The study was approved by the local ethics committee of the
University of Heidelberg. All participants gave written informed con-
sent after complete description of the study.
Interview-, questionnaire-based and daily life measures
At baseline (t1), depressive symptoms during the previous 2 weeks
were assessed with the self-rated Beck Depression Inventory II-
Revised (BDI II; Beck et al., 1996; German version: Hautzinger et al.,
2006) and the Montgomery and Asberg Depression Rating Scale
(MADRS; Montgomery and Asberg, 1979; German version:
Schmidtke et al., 1988) rated by a trained clinical psychologist. As
both measures were highly correlated (r¼0.73) and to increase reli-
ability, a mean score of z-standardized sum scores of the BDI II and
MADRS was created. Trait rumination in response to depressed mood
was assessed with the two five-item subscales brooding and reflection
of the Ruminative Response Scale (RRS; Treynor et al., 2003; German
version: Huffziger and Ku
¨hner, 2012).
Daily life variables were measured using ambulatory assessment
(Trull and Ebner-Priemer, 2013) which was conducted over two con-
secutive weekdays with 10 pseudo-randomized assessments per day
using personal digital assistants (PDAs; Palm Tungsten E2, Palm
Inc.). At each subjective assessment, which was indicated by a beep,
participants rated momentary mood and ruminative self-focus
(Huffziger et al., 2013). Sadness was assessed with the item “At the
moment I feel sad”. Item scores ranged from 0 (not at all) to 6 (very
much). Momentary ruminative self-focus was assessed with the two
items “At the moment, I am thinking about my problems” and “At the
moment, I am thinking about my feelings” which were averaged
Table 1 Descriptive and psychometric variables of the depression group (rMDD) and the
control group (HC) for the time points t1 and follow-up 6 months later (t2)
Variable Mean SD P-value
rMDD HC
n29 29
sex (female/male) 20/9 21/8 0.842
a
age (years) 45.55 7.45 44.24 8.09 0.524
b
education (CSE/high school diploma/A levels) 4/7/18 2/8/19 0.684
a
age of illness onset (years) 23.14 11.36
number of depressive episodes 3.96 2.22
previous inpatient treatment 72%
current psychotropic medication 24%
current psychotherapy 28%
BDI II t1 10.14 8.35 3.32 4.04 0.001
b
BDI II t2 10.83 10.18 4.14 6.37 0.002
b
MADRS t1 5.83 5.29 1.38 2.44 0.001
b
MADRS t2 7.62 7.48 2.38 5.32 0.002
b
dep-score t1 (z) 0.51 1.11 0.46 0.49 0.001
b
dep-score t2 (z) 0.35 1.06 0.40 0.70 0.002
b
RRS-R t1 10.66 3.59 8.69 3.40 0.037
b
RRS-R t2 10.52 3.09 9.00 3.17 0.070
b
RRS-B t1 10.97 3.96 8.07 2.52 0.002
b
RRS-B t2 10.72 4.22 8.00 3.01 0.006
b
daily rumination (mean) 1.36 1.25 0.86 0.90 0.083
b
daily sadness (mean) 0.95 1.23 0.41 0.71 0.075
b
a
test
b
two sample t-test
CSE: Certificate of Secondary Education, 8 years
BDI II: Beck Depression Inventory revised, self-rated
MADRS: Montgomery and Asberg Depression Rating Scale, rated by a trained clinical psychologist
dep-score: z-standardized sum score of the BDI II and MADRS scores
RRS, Ruminative Response Scale; R: reflection subscale, B: brooding subscale
Brain networks and depression SCAN (2014) 2035
(Moberly and Watkins, 2008). These two items were rated on a scale
from 0 (not at all) to 7 (very much). For this analyses, both sadness
and ruminative self-focus scores were aggregated per person over 2
days. Compliance with the ambulatory assessment was high in both
groups (percentages of responded assessments rMDD: 93%, HC: 94%).
Patients responded on average to 18.5 assessments (SD ¼1.4, range:
15–20), control participants to 18.8 assessments (SD ¼1.6, range
14–20).
After 6 months (t2), a follow-up on depressive symptoms (MADRS,
BDI II) and rumination (RRS) was conducted via telephone interviews
and questionnaires. None of the participants dropped out between t1
and t2.
Functional magnetic resonance imaging
The fMRI session used built-in goggles and the Presentation software
package from Neurobehavioral Systems Inc. The procedure was carried
out within 2 weeks after the SCID interview and the ambulatory as-
sessment. Every participant underwent six phases, each of which was
4.5 min: two resting states, two sad mood inductions, one rumination
phase and one distraction phase (the order of the rumination and
distraction phases were counterbalanced across participants). During
the resting state phases, participants were told to keep their eyes open
and background pink noise was presented. During the sad mood in-
duction phase, key words to remind participants of personal negative
life events and background sad music were presented. The rumination
and distraction induction phases had corresponding induction state-
ments selected from Nolen-Hoeksema and Morrow (1993) and did not
include pink noise or music.
The individually matched patients and controls were scanned in a
timely manner. Positive and negative affect was assessed by the Positive
and Negative Affect Scale (PANAS; Watson et al., 1988; German ver-
sion: Krohne et al., 1996) before and after each phase with a built-in
key-pad. Participants also rated how well they could concentrate on the
negative life events (presented in the mood induction phase, see below)
at the end of the session. An exemplary schematic overview of the fMRI
paradigm is given in Figure 1.
In the following, we focus mainly on the sad mood induction
phases. Sad mood was induced with three negative life events that
were individually assessed for every participant immediately before
the fMRI session started and were later presented consecutively in
the scanner via a keyword (each for 1.5 min; resulting in a 4.5 min
phase). In parallel, participants listened to instrumental background
music (parts of Adagio in g-minor by Albinoni) during the mood
induction phase.
6180 T2* weighted EPI images (TR ¼1.5 s, ¼808,TE¼28 ms)
with 24 slices (slice thickness 4 mm, voxel size 3 34mm
3
, FOV
192 mm) were recorded with a 3T Trio TIM Scanner with a 12 channel
head coil (Siemens Medical Systems, Erlangen, Germany). In addition,
heart rate and respiration rate were sampled at 50 Hz with the scanner
built-in equipment. Initially, the first 20 images of each phase were
discarded to let participants get into the phase. Data were corrected for
physiological artefacts using the Aztec software tool (van Buuren et al.,
2009) including a high-pass filter of 1/512 Hz. Pre-processing included
motion correction, slice time correction (13th slice as reference), nor-
malization to an EPI template and smoothing with a 9-mm Gaussian
kernel.
For the analyses, the mean of both sad mood induction phases was
used for every participant. Seed region for the connectivity analysis was
the PCC (10 mm sphere around 7, 45, 24) as main posterior part of
the DMN (Berman et al., 2011). Pre-processing and statistics were
conducted with SPM8 (Wellcome Trust Centre for Neuroimaging,
University College London, UK), IBM SPSS20 (SPSS Inc., Chicago,
Il, USA) and G*Power 3.1.2 (Faul et al., 2007). First level analyses
were conducted for every single subject and second level analysis to
compare groups. In the whole brain analyses minimal cluster size was
set to 10 adjacent voxels at p
unc.
< 0.001. Using an uncorrected P-value
together with a cluster size threshold has been discussed as an appro-
priate way to control for type I errors but also to avoid an inflation of
type II errors (Lieberman and Cunningham, 2009). For further regres-
sion analyses, a mask including all connected voxels at p
unc.
¼0.01
during the mood induction phase was overlaid with an anatomical
mask for the PHG from the WFU pickatlas (Lancaster et al., 1997;
Maldjian et al., 2003). The resulting intersection mask was used to
extract the mean connectivity from the PCC to the PHG in each par-
ticipant during the mood induction phase. The latter were used for
regression analyses on the relationship of PCC–PHG connectivity with
interview- and questionnaire-based and daily life measures. To account
for the baseline scores in longitudinal analyses, stepwise regressions
were conducted.
RESULTS
Mood ratings
Both groups reported sadder mood after mood induction (increased
negative affect: rMDD: 12.84 2.61 vs 19.17 7.93, t(28) ¼4.57,
P< 0.001, dz ¼0.85, HC: 12.55 4.23 vs 14.40 5.18, t(28) ¼3.52,
P¼0.001, dz ¼0.65; reduced positive affect: rMDD: 25.55 6.13 vs
22.90 5.88, t(28) ¼2.61, P¼0.008, dz ¼0.48, small but not statistic-
ally significant effect in the HC: 27.76 7.99 vs 26.74 8.56,
t(28) ¼1.55, P¼0.067, dz ¼0.29; rMDD reported a significantly
more increased negative affect and a similarly decreased positive
affect compared with controls (ANOVA group time interactions):
negative: F(1,57)¼9.16, P¼0.004, f¼0.41; positive: F(1,57) ¼1.83,
P¼0.182, f¼0.18) and comparable concentration ratings on the nega-
tive life events (rMDD 73 14%, HC 75 19%; t(56) ¼0.65,
P¼0.519, d¼0.17).
Brain connectivity during mood induction
To test whether the DMN can be identified during sad mood induction
we calculated the main effect for PCC connectivity over all participants
after regressing out the mean signal. When doing so, we identified a
network including parietal, prefrontal and occipital regions, which is
strongly overlapping with the DMN as described by Smith et al.
Fig. 1 Exemplary scheme of the fMRI paradigm; 2 3 phases (overall 6 4.5 min): resting state phase (eyes open; with pink noise), sad mood induction phase (with key words of personal negative life events
and music, every key word presented for 1.5 min) and either rumination or distraction induction phase (counterbalanced), figure shows rumination induction phase first with corresponding induction statements
selected from the original item pool provided by Nolen-Hoeksema and Morrow (1993). PANAS, Positive and Negative Affect Scale.
2036 SCAN (2014) V. Z a m o s c i k et al.
(2009). See supplementary online material for details (Supplementary
Figures S1 and S2).
As the pattern of the functional connectivity results remained the
same when only the first sad mood induction phase was considered, we
averaged across both phases to increase power.
Group comparisons of brain connectivity
In a whole brain analysis of the sad mood induction phase, remitted
depressed patients showed greater connectivity between PCC and PHG
compared with healthy controls (Fig. 2 and Table 2). No regions
showed greater connectivity to the PCC in the controls compared
with patients. When looking only at the resting state, there were no
differences between groups at the same threshold (p
unc.
< 0.001).
The connectivity between PCC and PHG appeared stronger in the
left than right PHG [F(1,56) ¼14.30, P< 0.001, f¼0.50] but this did
not differ between groups [F(1,56) ¼0.13, P¼0.717, f¼0.04].
Relations of PCC–PHG connectivity to number of depressive
episodes and sadness and rumination in daily life
In the rMDD, more lifetime episodes of major depression were linked
to a stronger connectivity of the PCC and PHG during sad mood
induction (R
2
¼0.29, B¼0.01, SE < 0.01, t¼3.26, P¼0.003).
Furthermore, a stronger connectivity of PCC and PHG during sad
mood induction was significantly associated with higher levels of ru-
mination and sadness in daily life in the rMDD, but not in the HC
(rumination: rMDD: R
2
¼0.28, B¼10.76, SE ¼3.37, t¼3.19,
P¼0.004, HC: R
2
¼0.06, B¼5.44, SE ¼4.26, t¼1.28, P¼0.212;
sadness: rMDD: R
2
¼0.24, B¼10.74, SE ¼4.22, t¼2.54, P¼0.019,
HC: R
2
¼0.02, B¼2.80, SE ¼3.42, t¼0.82, P¼0.419). All regres-
sion coefficients remained significant when including t1 depression
scores into the models.
To investigate whether the relations between neural connectivity and
behavioral ratings were largely due to their associations with number
of previous episodes, a series of stepwise regressions were conducted
with daily life rumination and daily sadness as dependent variables. In
these models, depression score at t1 was entered in a first step, followed
by number of episodes and PCC–PHG connectivity score in a stepwise
manner, depending on their effect size. In all models PCC–PHG con-
nectivity yielded a significant effect, while number of episodes did not
(daily life rumination: PHG–PCC connectivity: R2
inc ¼0.16, B¼8.76,
SE ¼2.66, t¼3.28, P¼0.003, nof episodes: R2
inc < 0.01, B¼0.04,
SE ¼0.08, t¼0.50, P¼0.621; daily life sadness: PCC–PHG connect-
ivity: R2
inc ¼0.09, B¼6.81, SE ¼2.43, t¼2.80, P¼0.011, nof episodes:
R2
inc ¼0.02, B¼0.091, SE ¼0.065, t¼1.40, p¼.181).
Predictive effects of PCC–PHG connectivity on the course of
depressive symptoms and rumination
The stepwise regression analysis revealed a significant prediction of t2
depression scores by t1 depression scores in both groups (rMDD:
R
2
¼0.37, B¼0.58, SE ¼0.15, t¼3.95, P¼0.001; HC: R
2
¼0.32,
B¼0.80, SE ¼0.23, t¼3.53, P¼0.002). However, adding the
PCC–PHG connectivity to the model revealed that the latter signifi-
cantly predicted a significant worsening of depressive symptoms in the
rMDD (R2
inc ¼0.13, B¼6.66, SE ¼2.54, t¼2.62, P¼0.015; Fig. 3) but
not in the HC (t¼1.18, P¼0.249).
Furthermore, baseline habitual rumination predicted rumination
scores 6 months later in both groups (brooding: rMDD: R
2
¼0.52,
B¼0.77, SE ¼0.14, t¼5.37, P< 0.001; HC: R
2
¼0.35, B¼0.70,
SE ¼0.19, t¼3.78, P¼0.001; reflection: rMDD: R
2
¼0.52, B¼0.62,
SE ¼0.12, t¼5.43, P< 0.001; HC: R
2
¼0.32, B¼0.53, SE ¼0.14,
t¼3.59, P¼0.001). Adding the PCC–PHG connectivity to the step-
wise regression model revealed that the latter predicted significant
higher brooding scores in the rMDD but not in the HC (brooding:
rMDD: R2
inc ¼0.07, B¼18.71, SE ¼8.99, t¼2.08, P¼0.047; HC:
t¼0.05, P¼0.957; reflection: rMDD: t¼1.69, P¼0.104; HC:
t¼1.25, P¼0.224).
The overall pattern of the reported results remained the same, for
both connectivity and predictions, when excluding all patients with
psychotropic medication and their matched controls (n¼14).
However, due to reduced power the results did not reach significance
for the prediction of daily life sadness and brooding after 6 months.
DISCUSSION
The study highlights the potential role of connectivity between the
PCC (a node in DMN) and PHG (a node in the autobiographical
network) during negative autobiographical recollection in remitted
depression, its association with daily life mood and rumination and
its prognostic value for the further course of illness. To this end, we
induced sad mood in remitted recurrently or chronic depressed pa-
tients and healthy controls by presenting key words of personal nega-
tive life events in the scanner. Within this context, we investigated
whether the remitted depressed group would show a stronger connec-
tion between the PCC and structures of the autobiographical memory
network during recollection of negative life events and whether this
connectivity would be related to concurrent daily life assessments of
mood and rumination and to the prospective 6 months course of
depressive symptoms and habitual rumination.
Persons with remitted recurrent or chronic depression in compari-
son with healthy controls showed a stronger connectivity of the PCC
with the bilateral PHG which was further increased in patients with
more previous depressive episodes. Furthermore, in the patient group
a stronger connectivity of the PCC and PHG was related to a sadder
mood and more ruminative thoughts in daily life and a worsening of
rumination and depression scores over the 6-months period, even
when controlling for current depressive symptoms. Although former
Fig. 2 Increased connectivity of the PCC and PHG during the sad mood induction phase in the
depression group (n¼29) contrasted to the control group (n¼29); whole-brain main effect, min.
k¼10 voxels, p
unc.
< 0.001, colors indicate t-values.
Table 2 Functional brain imaging results of the sad mood induction for the whole brain
main effect of group (rMDD>HC), p
unc.
< .001; rMDD: depression group, HC: control group
Anatomical region MNI Cluster size Peak t-value
xyZ
left parahippocampal gyrus/PCC, BA29 15 43 4 62 4.54
right parahippocampal gyrus 24 43 4 33 3.98
left mid frontal cortex 33 26 25 20 4.19
Brain networks and depression SCAN (2014) 2037
studies of depression have casually reported data suggesting a poten-
tially important role of the PHG, this region has not previously been a
major focus of discussion. Our own results strengthen the assumption
that the PHG may have a key role in depression, especially in recurrent
and chronic depression.
Our findings are interesting in several respects. First, the stronger
correlation between PCC and PHG during mood induction in remitted
depressed patients suggests that the recollection of negative events in
patients was associated with a more intensive recruitment of autobio-
graphical memories. This could be a vulnerability mechanism for the
re-occurrence of depression. We assume that whenever these vulner-
able individuals are reminded of personally relevant negative events, a
broad associative network is being activated, thereby reinstating nega-
tive thoughts and rumination. This interpretation is supported by the
significant association of PCC–PHG connectivity with sad mood and
rumination in daily life. This might indicate that increased involve-
ment of the autobiographical memory network is linked to negative
mood and rumination. Furthermore, the observed correlation between
PHG involvement during mood induction and the number of anam-
nestic depressive episodes supports the idea that this mechanism not
only reflects vulnerability but also a neural signature of the disorder.
This signature could represent a pre-existing, possibly genetic, risk
factor but could also be acquired over the course of illness. Although
our design does not allow clear conclusions, we speculate that the
greater connectivity between PCC and PHG in patients compared
with controls might represent a kind of neural ‘scar’ (cf. Marchetti
et al., 2012) of former episodes in recurrent and chronic depression
which leads even in remission to enhanced reactions in demanding
situations such as recalling personal negative events. Furthermore, in
our study, this responsiveness of the PHG was not only related to
illness history and daily life emotional and cognitive experiences but
also predictive for the worsening of rumination and depressive symp-
toms during a 6-month period. Importantly, this relation cannot be
explained by the association between PHG involvement and current
depressive symptoms as we controlled for t1 depression scores. This
finding further emphasizes the importance of this system for the course
of illness.
In addition, our result of a stronger PCC–PHG connectivity in the
remitted depressed group during sad mood induction but not in the
resting state matches with findings from non-fMRI studies showing
that remitted depressed people perform similarly on measures of mal-
adaptive cognitions and information processing as never-depressed
individuals. However, if primed by the activation of maladaptive sche-
mata or sad mood induction, these patients show differences from
healthy controls in various types of information processing. This phe-
nomenon, labeled as cognitive reactivity, was shown by self-reports,
memory, attention and interpretive biases and found to be connected
with a poorer longitudinal course of the disorder (e.g. Scher et al.,
2005;Segal et al., 2006). In this context, our work is the first to
study DMN activity in relation to cognitive reactivity in remitted de-
pressed individuals, as suggested also by Marchetti et al. (2012).We
identified increased PCC–PHG connectivity specifically during sad
mood induction, which was furthermore linked to daily life rumin-
ation and to a poorer symptom course over the next months. So far,
only Farb et al. (2011) used sad mood induction to perform a neuroi-
maging study on risk of relapse. Using an experimental design different
from ours, these authors found greater medial prefrontal cortex re-
activity to emotional challenges which predicted rumination and seems
to be a marker for relapse risk during the following 18 months.
Together, these two studies are the first to demonstrate neuroimaging
phenotypes of cognitive reactivity in remitted depressed individuals
and lend further support for the diathesis-stress hypothesis of depres-
sion (Beck, 2008).
Depressed individuals have problems with remembering specific
personal past events and instead recall more general events (over gen-
eral memory, OGM; Williams and Broadbent, 1986). Studies indicate
that OGM predicts higher follow-up depressive symptoms (Sumner,
et al., 2010) as well as rumination and avoidance (Watson et al., 2013).
Furthermore, memory specificity was reduced after an induction of a
self-discrepancy focus in persons with high trait rumination (Raes
et al., 2012). Therefore, OGM is possibly modulated via self-discrep-
ancy (Raes et al., 2012;Schoofs et al., 2013) and related to rumination.
We can only speculate that the association between higher PHG in-
volvement during negative mood induction and a poorer course of
depressive symptoms and rumination is mediated by OGM for nega-
tive events, which might be associated with stronger autobiographical
involvement during mood induction in patients. However, this stron-
ger involvement might not result from a more specific memory for
Fig. 3 Prediction of t2 depression scores with t1 connectivity of PCC and PHG during sad mood induction; rMDD: depression group (n¼29); HC: control group (n¼29).
2038 SCAN (2014) V. Z a m o s c i k et al.
negative experiences but from repeatedly thinking of them with a ru-
minative focus that generally takes place at a rather abstract level of
thinking (Watkins et al., 2009). This cognitive bias is apparently not
state-dependent but can be observed in remission. Unfortunately, we
did not measure OGM directly in this study. Future studies on the
relation between neural correlates of negative mood, rumination and
outcome should include this construct to further elucidate possible
connections.
In conclusion, our results suggest a vicious circle in which remem-
bering and ruminating about negative events sensitizes a neural net-
work involved in self-reference and autobiographical memory, which
again facilitates the recollection of negative life events and rumination,
thereby leading to a poor outcome. Interestingly, this interpretation is
to a large extent compatible with a neurobiological model of cognitive
vulnerability recently presented by Marchetti and colleagues (2012).
Based on a review of the literature on DMN alterations in depression,
they argue that an altered DMN can be seen as a neural mechanism
involved in the cognitive risk for the recurrence of depressive episodes
leading to increased rumination, cognitive reactivity and impaired at-
tentional control.
A drawback of our study might be seen in the fact that the fMRI
results were not corrected for multiple testing. However, as mentioned
before, this procedure has been discussed as an appropriate way to
control for type I errors but also avoiding an inflation of type II
errors (Lieberman and Cunningham, 2009). Particularly for patient
studies, applying a whole head corrected significance level and there-
fore reaching an incredibly high probability to overlook true effects
is also, from an ethical point of view as then patients would have
been exposed in vain to a potentially stressful situation, a disputable
approach. Furthermore, when analyzing group differences, we
observed only three significant clusters for the whole brain, two of
them located in the PHG, which further validates our findings.
However, replication of these findings is clearly warranted to confirm
our conclusions.
In short, remitted patients with a recurrent or chronic depression
precourse showed a stronger connectivity between PCC and PHG
when they were cued to think about own negative life events. This
finding suggests that depression is associated with a greater salience
of negative autobiographical memories. This connectivity was asso-
ciated with the number of previous episodes, with daily life measures
of depression-related features such as more rumination and sadder
mood and with a worsening of depression and rumination over 6
months. These results highlight the importance of the autobiographical
memory network and dysfunctional cognitions related to it and em-
phasize current interventions that target rumination in the context of
negative autobiographical memories on a neurobiological level.
Further research on this specific pathway could be a possible oppor-
tunity to ameliorate interventions in depression. A promising candi-
date for such interventions might be mindfulness-based training which
has not only been found to reduce rumination (Jain et al., 2007;
van Vugt et al., 2012), but also to particularly reduce those neural
connections with the PHG under resting conditions (Taylor et al.,
2013) that we identified to be involved in mood, rumination and the
risk of recurrence of depression.
SUPPLEMENTARY DATA
Supplementary data are available at SCAN online.
REFERENCES
Beck, A.T. (2008). The evolution of the cognitive model of depression and its neurobio-
logical correlates. American Journal of Psychiatry,165(8), 969–77.
Beck, A.T., Steer, R.A., Brown, G.K. (1996). Manual for the Beck Depression Inventory-II.
San Antonio: The Psychological Corporation.
Berman, M.G., Peltier, S., Nee, D.E., Kross, E., Deldin, P.J., Jonides, J. (2011). Depression,
rumination and the default network. Social Cognitive and Affective Neuroscience,6(5),
548–55.
Brosschot, J.F., Pieper, S., Thayer, J.F. (2005). Expanding stress theory: prolonged activa-
tion and perseverative cognition. Psychoneuroendocrinology,30(10), 1043–49.
Cooney, R.E., Joormann, J., Eugene, F., Dennis, E.L., Gotlib, I.H. (2010). Neural correlates
of rumination in depression. Cognitive Affective & Behavioral Neuroscience,10(4),
470–478.
D’Argembeau, A., Comblain, C., Van der Linden, M. (2003). Phenomenal characteristics of
autobiographical memories for positive, negative, and neutral events. Applied Cognitive
Psychology,17(3), 281–294.
Dere, E., Pause, B.M., Pietrowsky, R. (2010). Emotion and episodic memory in neuro-
psychiatric disorders. Behavioural Brain Research,215(2), 162–71.
Farb, N.A., Anderson, A.K., Bloch, R.T., Segal, Z.V. (2011). Mood-linked responses in
medial prefrontal cortex predict relapse in patients with recurrent unipolar depression.
Biological Psychiatry,70(4), 366–72.
Faul, F., Erdfelder, E., Lang, A.G., Buchner, A. (2007). G*Power 3: a flexible statistical
power analysis program for the social, behavioral, and biomedical sciences. Behavior
Research Methods,39(2), 175–191.
Greicius, M.D., Flores, B.H., Menon, V., et al. (2007). Resting-state functional connectivity
in major depression: abnormally increased contributions from subgenual cingulate
cortex and thalamus. Biological Psychiatry,62(5), 429–437.
Hamilton, J.P., Chen, M.C., Gotlib, I.H. (2013). Neural systems approaches to understand-
ing major depressive disorder: an intrinsic functional organization perspective.
Neurobiology of Disease,52,4–11.
Hamilton, J.P., Furman, D.J., Chang, C., Thomason, M.E., Dennis, E., Gotlib, I.H. (2011).
Default-mode and task-positive network activity in major depressive disorder: implica-
tions for adaptive and maladaptive rumination. Biological Psychiatry,70(4), 327–333.
Hautzinger, M., Keller, F., Kuehner, C., Beck, A.T. (2006). Das Beck Depressionsinventar II.
Deutsche Bearbeitung und Handbuch zum BDI II. Frankfurt a.M.: Harcout Test Services.
Howes, M., Siegel, M., Brown, F. (1993). Early childhood memories: accuracy and affect.
Cognition,47(2), 95–119.
Huffziger, S., Ebner-Priemer, U., Zamoscik, V., Reinhard, I., Kirsch, P., Kuehner, C. (2013).
Effects of mood and rumination on cortisol levels in daily life: an ambulatory assessment
study in remitted depressed patients and healthy controls. Psychoneuroendocrinology,
38(10), 2258–67.
Huffziger, S., Ku
¨hner, C. (2012). Die Ruminationsfacetten Brooding und Reflection.
Zeitschrift Fur Klinische Psychologie Psychiatrie Und Psychotherapie.,41(1), 38–46.
Huffziger, S., Reinhard, I., Kuehner, C. (2009). A longitudinal study of rumination and
distraction in formerly depressed inpatients and community controls. Journal of
Abnormal Psychology,118(4), 746–56.
Jain, S., Shapiro, S.L., Swanick, S., et al. (2007). A randomized controlled trial of mind-
fulness meditation versus relaxation training: effects on distress, positive states of mind,
rumination, and distraction. Annals of Behavioral Medicine,33(1), 11–21.
Krohne, H.W., Egloff, B., Kohlmann, C.W., Tausch, A. (1996). Investigations with a
German version of the positive and negative affect schedule (PANAS). Diagnostica,
42(2), 139–156.
Lancaster, J.L., Rainey, L.H., Summerlin, J.L., et al. (1997). Automated labeling of the
human brain: a preliminary report on the development and evaluation of a forward-
transform method. Human Brain Mapping,5(4), 238–242.
Leung, M.K., Chan, C.C., Yin, J., Lee, C.F., So, K.F., Lee, T.M. (2013). Increased gray matter
volume in the right angular and posterior parahippocampal gyri in loving-kindness
meditators. Social Cognitive and Affective Neuroscience,8(1), 34–9.
Li, B., Liu, L., Friston, K.J., et al. (2013). A treatment-resistant default mode subnetwork in
major depression. Biological Psychiatry,74(1), 48–54.
Lieberman, M.D., Cunningham, W.A. (2009). Type I and Type II error concerns in fMRI
research: re-balancing the scale. Social Cognitive and Affective Neuroscience,4(4), 423–28.
Mackin, R.S., Tosun, D., Mueller, S.G., et al. (2012). Patterns of reduced cortical thickness
in late-life depression and relationship to psychotherapeutic response. American Journal
of Geriatric Psychiatry,21(8), 794–802.
Maldjian, J.A., Laurienti, P.J., Kraft, R.A., Burdette, J.H. (2003). An automated method
for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets.
Neuroimage,19(3), 1233–
9.
Marchetti, I., Koster, E.H., Sonuga-Barke, E.J., De Raedt, R. (2012). The default mode
network and recurrent depression: a neurobiological model of cognitive risk factors.
Neuropsychology Review,22(3), 229–51.
Milton, F., Butler, C.R., Benattayallah, A., Zeman, A.Z. (2012). The neural basis of auto-
biographical memory deficits in transient epileptic amnesia. Neuropsychologia,50(14),
3528–41.
Moberly, N.J., Watkins, E.R. (2008). Ruminative self-focus and negative affect: an experi-
ence sampling study. Journal of Abnormal Psychology,117(2), 314–23.
Montgomery, S.A., Asberg, M. (1979). A new depression scale designed to be sensitive to
change. British Journal of Psychiatry,134,382–9.
Moritz, S., Glascher, J., Brassen, S. (2005). Investigation of mood-congruent false and true
memory recognition in depression. Depression and Anxiety,21(1), 9–17.
Brain networks and depression SCAN (2014) 2039
Nolen-Hoeksema, S., Morrow, J. (1993). Effects of rumination and distraction on naturally
occurring depressed mood. Cognition & Emotion,7(6), 561–70.
Nolen-Hoeksema, S., Wisco, B.E., Lyubomirsky, S. (2008). Rethinking rumination.
Perspectives on Psychological Science,3(5), 400–24.
Raes, F., Schoofs, H., Griffith, J.W., Hermans, D. (2012). Rumination relates to reduced
autobiographical memory specificity in formerly depressed patients following a self-
discrepancy challenge: the case of autobiographical memory specificity reactivity.
Journal of Behavior Therapy and Experimental Psychiatry,43(4), 1002–7.
Scher, C.D., Ingram, R.E., Segal, Z.V. (2005). Cognitive reactivity and vulnerability: em-
pirical evaluation of construct activation and cognitive diatheses in unipolar depression.
Clinical Psychology Review,25(4), 487–510.
Schmidtke, A., Fleckenstein, P., Moises, W., Beckmann, H. (1988). Studies of the reliability
and validity of the German version of the Montgomery-Asberg Depression Rating Scale
(MADRS). Schweizer Archiv fur Neurologie und Psychiatrie,139(2), 51–65.
Schoofs, H., Hermans, D., Griffith, J.W., Raes, F. (2013). Self-discrepancy and reduced
autobiographical memory specificity in ruminating students and depressed patients.
Cognition & Emotion,27(2), 245–62.
Segal, Z.V., Kennedy, S., Gemar, M., Hood, K., Pedersen, R., Buis, T. (2006). Cognitive
reactivity to sad mood provocation and the prediction of depressive relapse. Archives of
General Psychiatry,63(7), 749–55.
Sestieri, C., Corbetta, M., Romani, G.L., Shulman, G.L. (2011). Episodic memory retrieval,
parietal cortex, and the default mode network: functional and topographic analyses.
Journal of Neuroscience,31(12), 4407–20.
Sheline, Y.I., Barch, D.M., Price, J.L., et al. (2009). The default mode network and self-
referential processes in depression. Proceedings of the National Academy of Sciences USA,
106(6), 1942–1947.
Sheline, Y.I., Price, J.L., Yan, Z., Mintun, M.A. (2010). Resting-state functional MRI in
depression unmasks increased connectivity between networks via the dorsal nexus.
Proceedings of the National Academy of Sciences USA,107(24), 11020–5.
Smith, S.M., Fox, P.T., Miller, K.L., et al. (2009). Correspondence of the brain’s functional
architecture during activation and rest. Proceedings of the National Academy of Sciences
USA,106(31), 13040–5.
Spreng, R.N., Mar, R.A., Kim, A.S. (2009). The common neural basis of autobiographical
memory, prospection, navigation, theory of mind, and the default mode: a quantitative
meta-analysis. Journal of Cognitive Neuroscience,21(3), 489–510.
Sumner, J.A., Griffith, J.W., Mineka, S. (2010). Overgeneral autobiographical memory as a
predictor of the course of depression: a meta-analysis. Behaviour Research and Therapy,
48(7), 614–625.
Taylor, V.A., Daneault, V., Grant, J., et al. (2013). Impact of meditation training on the
default mode network during a restful state. Social Cognitive and Affective Neuroscience,
8(1), 4–14.
Treynor, W., Gonzalez, R., Nolen-Hoeksema, S. (2003). Rumination reconsidered: a psy-
chometric analysis. Cognitive Therapy and Research,27(3), 247–259.
Trull, T.J., Ebner-Priemer, U.W. (2013). Ambulatory assessment. Annual Review of Clinical
Psychology,9, 4.1–4.27.
van Buuren, M., Gladwin, T.E., Zandbelt, B.B., et al. (2009). Cardiorespiratory effects on
default-mode network activity as measured with fMRI. Human Brain Mapping,30(9),
3031–42.
van Vugt, M.K., Hitchcock, P., Shahar, B., Britton, W. (2012). The effects of mindfulness-
based cognitive therapy on affective memory recall dynamics in depression: a mechan-
istic model of rumination. Frontiers in Human Neuroscience,6, 257.
Viard, A., Desgranges, B., Eustache, F., Piolino, P. (2012). Factors affecting medial temporal
lobe engagement for past and future episodic events: an ALE meta-analysis of neuroi-
maging studies. Brain Cognition,80(1), 111–125.
Watkins, E.R., Baeyens, C.B., Read, R. (2009). Concreteness training reduces dysphoria:
proof-of-principle for repeated cognitive bias modification in depression. Journal of
Abnormal Psychology,118(1), 55–64.
Watson, D., Clark, L.A., Tellegen, A. (1988). Development and validation of brief measures
of positive and negative affect: the PANAS scales. Journal of Personality and Social
Psychology,54(6), 1063–70.
Watson, L.A., Berntsen, D., Kuyken, W., Watkins, E.R. (2012). The characteristics of in-
voluntary and voluntary autobiographical memories in depressed and never depressed
individuals. Consciousness and Cognition,21(3), 1382–92.
Watson, L.A., Berntsen, D., Kuyken, W., Watkins, E.R. (2013). Involuntary and voluntary
autobiographical memory specificity as a function of depression. Journal of Behavior
Therapy and Experimental Psychiatry,44(1), 7–13.
Whalley, M.G., Rugg, M.D., Brewin, C.R. (2012). Autobiographical memory in depression:
an fMRI study. Psychiatry Research,201(2), 98–106.
Whitfield-Gabrieli, S., Moran, J.M., Nieto-Castanon, A., Triantafyllou, C., Saxe, R.,
Gabrieli, J.D. (2011). Associations and dissociations between default and self-reference
networks in the human brain. Neuroimage,55(1), 225–232.
Williams, J.M., Broadbent, K. (1986). Autobiographical memory in suicide attempters.
Journal of Abnormal Psychology,95(2), 144–9.
Wittchen, H.U., Wunderlich, U., Gruschwitz, S., Zaudig, M. (1997). SCID: Structured
Clinical Interview for DSM-IV axis I disorders. Goettingen: Hogrefe.
Zeng, L.L., Shen, H., Liu, L., et al. (2012). Identifying major depression using whole-brain
functional connectivity: a multivariate pattern analysis. Brain,135(Pt 5), 1498–507.
Zhu, X., Wang, X., Xiao, J., et al. (2012). Evidence of a dissociation pattern in resting-state
default mode network connectivity in first-episode, treatment-naive major depression
patients. Biol Psychiatry,71(7), 611–617.
204 0 SCAN (2014) V. Z a m o s c i k et al.
