Journal of Affective Disorders 203 (2016) 281291

Contents lists available at ScienceDirect

Journal of Affective Disorders

journal homepage: www.elsevier.com/locate/jad

Research paper

Diffusion tensor imaging in euthymic bipolar disorder A tract-based

spatial statistics study
Bartholomeus C.M. (Benno) Haarman a,b,n, Rixt F. Riemersma Van der Lek a,
Huibert Burger a,c, Jan Cees de Groot d, Hemmo A. Drexhage e, Willem A. Nolen a,
Leonardo Cerliani f,g
a

University of Groningen, University Medical Center Groningen, Department of Psychiatry, Groningen, The Netherlands
Radiology Morphological Solutions, Berkel en Rodenrijs, The Netherlands
c
University of Groningen, University Medical Center Groningen, Department of General Practice, Groningen, The Netherlands
d
University of Groningen, University Medical Center Groningen, Department of Radiology, Groningen, The Netherlands
e
Erasmus MC, Department of Immunology, Rotterdam, The Netherlands
f
University of Groningen, Neuroimaging Center, Groningen, The Netherlands
g
Institute du Cerveau et de la Moelle pinire, Paris, France
b

art ic l e i nf o

a b s t r a c t

Article history:
Received 4 February 2016
Accepted 22 May 2016
Available online 1 June 2016

Background: In the current DTI study we compared euthymic bipolar I disorder (BD-I) patients and
healthy controls (HC). We subsequently divided the total patient group into lithium-users and non-lithium-users and estimated differences across the three groups.
Methods: Twenty-one euthymic BD-I patients and twenty-two HC participants were included in psychiatric interviews and MRI image acquisition (diffusion-weighted (DW) and T1-weighted scans). Fractional anisotropy (FA), radial, mean and axial diffusivity (RD, MD, AD) were estimated from the DW data,
using DTI. These measures were then compared between groups using FSL Tract Based Spatial Statistics
(TBSS). Correlations with age at onset, number of episodes and depression score were analyzed.
Results: A difference in FA, MD, RD and AD between the whole sample of euthymic BD-I patients and
healthy controls could not be detected. Amongst others, lithium-using patients demonstrated a higher FA
and lower RD when compared to non-lithium-using BD-I patients in the corpus callosum and left
anterior corona radiata. Widespread clusters demonstrated negative FA associations and positive RD and
MD associations with minor depressive symptoms.
Limitations: Patients were naturalistically treated. Although the sample size is comparable to several
other DTI studies, a larger sample size would have been benicial. TBSS and DTI have their own limitations.
Conclusion: Our ndings support the theory that previously described DTI-based microstructural differences between HC and BD patients could be less pronounced in euthymic BD patients. Differences in
FA between patients using and not using lithium suggest a counteracting effect of lithium on white
matter microstructural disturbances.
& 2016 Elsevier B.V. All rights reserved.

Keywords:
Bipolar disorder
Diffusion tensor imaging
Tbss
Lithium
Gsk-3

1. Introduction
Abbreviations: BD, bipolar disorder; BD-I, bipolar I disorder; BDNF, brain derived
neurotrophic factor; cAMP, adenosine monophosphate; DTI, diffusion tensor imaging; FA, fractional anisotropy; HC, healthy control; GSK-3), glycogen synthase
kinase 3 beta; IDS-C30, inventory of depressive symptoms - clinician version;
MARCKS, myristoylated alanine-rich c kinase substrate; MD, mean diffusivity; mI,
myoinositol; MINI, Mini-international neuropsychiatric interview; MNI, Montreal
Neurological Institute; MRI, magnetic resonance imaging; PKC, protein kinase C;
RD, radial diffusivity; TBSS, tract based spatial statistics; YMRS, Young mania rating
scale
n
Correspondence to: Department of Psychiatry, CC44, University of Groningen,
University Medical Center Groningen, P.O. Box 30.001, 9700 RB Groningen, The
Netherlands.
E-mail address: b.c.m.haarman@rug.nl (B.C.M. (Benno) Haarman).
http://dx.doi.org/10.1016/j.jad.2016.05.040
0165-0327/& 2016 Elsevier B.V. All rights reserved.

Over the years neuroimaging techniques have demonstrated

important neuroanatomical differences between patients diagnosed with bipolar disorder (BD) and healthy controls (HC). Since
group differences were reported in several brain regions (Gonul
et al., 2009; Haarman et al., 2014a, 2014b; Kupferschmidt and
Zakzanis, 2011), recent neuroimaging research began to focus on
the network level, to assess to what extent the symptomatology of
BD might be related to an atypical functional and anatomical
connectivity between sensible brain regions. In the context of
anatomical
connectivity,
technical
and
methodological

282

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

developments in magnetic resonance imaging (MRI) allowed to

estimate various parameters related to the microstructural properties of the white matter bundles connecting different brain regions, providing additional substance and complexity to BD
pathophysiology.
Diffusion-weighted (DW) imaging is based on the observation
that in white matter tracts, the coherent geometrical orientation of
the axons and of the myelin sheet causes the water to diffuse more
along the direction of the bers than perpendicular to it. Water
diffusion yields an attenuation of the net spin-echo signal in a
white matter voxel. Therefore, by measuring this attenuation
along different directions, it is possible to reconstruct a three-dimensional diffusivity prole reecting the organization of white
matter bers in a voxel. Modelling the diffusivity prole using
diffusion tensor imaging (DTI) allows to separately estimate the
amount of water diffusion along the ber tract (axial diffusivity,
AD) and perpendicular to it (radial diffusivity, RD), as well as the
mean diffusivity (MD) of water in that voxel along all the recorded
directions (Basser et al., 1994). Another widely used measure of
water diffusion in a white matter tract is fractional anisotropy (FA),
which reects the relationship between axial and radial diffusivity,
quantifying how strongly directional is the diffusion of water in a
voxel. While these measures are related to each other, their
combined examination is particularly informative since they can
be differentially modulated by several microstructural features,
including axonal density or caliber, myelination, axonal membrane
permeability, ber complexity and orientation (Beaulieu, 2002;
Benedetti et al., 2013; Jones et al., 2013; Lebel et al., 2012a). Differences in these microstructural features, and in the corresponding DTI measurement, across subjects can be associated with
development, or relate to other stratications of neurotypical
participants. However, these measures can also reect progressed
or ongoing pathophysiological processes affecting white matter
microstructure in different categories of neurological or neuropsychiatric patients.
In BD, results from whole-brain DTI studies suggest widespread
white matter abnormalities. Signicant widespread decreases in
FA have been predominantly reported so far in BD patients compared with healthy controls, with differences encompassing all
major white matter tracts (Nortje et al., 2013). These widespread
white matter diffusion abnormalities, including fronto-temporal
and ventral striatal regions, were marked by differences in diffusivity measures other than FA as well, leading to the conclusion
that myelination problems rather than axonal loss is implicated in
BD (Harsan et al., 2006; Song et al., 2002; Travers et al., 2012).
Despite these advancements, a general consensus is still lacking, since other studies reported either lower, higher or no difference in FA between BD patients and healthy controls (Nortje
et al., 2013). Inconsistency across ndings might reect differences
in data acquisition methods as well as in patient characteristics.
Importantly, in several studies (Liu et al., 2010; Nortje et al., 2013;
Sussmann et al., 2009; Versace et al., 2008; Zanetti et al., 2009) the
BD sample includes patients in different mood states, which contributes to increase inter-individual variability, and is likely associated with differential activity within and across brain networks,
as well as with different patterns of white matter connectivity
(Bellani et al., 2016). In addition to mood state heterogeneity, BD
patients samples may differ for medication status, which could be
considered important for interpreting DTI ndings. Indeed, lithium
seems to act as a promyelinating factor (Bartzokis, 2012; Makoukji
et al., 2012) and the administration of lithium was found to be
associated with increased axial connectivity (Benedetti et al.,
2012).
In the current study we rst compared estimates of white
matter microstructure (FA, MD, RD, AD) between euthymic bipolar
I disorder (BD-I) patients and HC. Based on previous studies

pointing toward myelination problems in BD, we expected to nd

a widespread decrease in FA in several white matter tracts of BD-I
patients compared to HC, associated with reciprocal alterations of
other white matter microstructural parameters. Subsequently, we
divided the patient group into lithium-users and non-lithiumusers and analyzed the estimates of white matter microstructure
across these three groups (non-lithium-users, lithium-users,
healthy controls). In this analysis, supposing a restoring effect of
lithium on myelination, we expected FA to be increased -and
consistently a decrease of other white matter microstructural
parameters- in lithium-using patients compared to non-lithiumusing patients, possibly even attaining healthy control values. Finally, we investigated whether the estimates of white matter microstructure were associated with patient characteristics.
As in some of the mentioned studies, we employed the analytical framework of tract-based spatial statistics (TBSS), which focuses on the centers of the major ber bundles. With respect to
whole-brain investigations, this approach increases the statistical
power of the analysis and limits the chance of misalignment across
subjects, as well as the issue of partial volume contamination
(Smith et al., 2006).

2. Material and methods

2.1. Participants
For the present cross-sectional case-control study we included
22 BD-I patients and 24HC frequency matched on age (in 5-year
age groups) and gender that participated in the MOODINFLAME
study (http://www.moodiname.eu). All subjects underwent a
MRI-scan and a subgroup of the subjects also received a PET-scan
(15 patients, 12 controls) (Haarman et al., 2015, 2014b).
In the MOODINFLAME study we included adult male and female subjects who were free of inammation-related symptoms
including fever and current or recent infectious or inammatory
disease, uncontrolled systemic disease, uncontrolled metabolic
disease or other signicant uncontrolled somatic disorders known
to affect mood. They did not use somatic medication known to
affect mood or the immune system, such as corticosteroids, nonsteroid anti-inammatory drugs and statins. Female candidates
who were pregnant or recently gave birth were excluded. Patients
and controls did not have a contraindication for MRI scanning.
Patients were allowed to continue their regular psychopharmacological treatment. They were euthymic at the time of
scanning as indicated by an Inventory of Depressive Symptoms Clinician Version (IDS-C30) score o22 and a Young Mania Rating
Scale (YMRS) score o 12, respectively. Patients with any other
current primary major psychiatric diagnosis were excluded including: schizophrenia, schizoaffective disorder, anxiety disorder
and substance use disorders. HC did not have any current or lifetime psychiatric diagnosis.
2.2. Ethical considerations
The Medical Ethical Review Committee of the University
Medical Center Groningen approved the protocol, which was
performed in accordance with the Helsinki Declaration of 2013
(World Medical Association, 2013). Written informed consent was
obtained from all participants.
2.3. Assessments
All subjects underwent a Mini-International Neuropsychiatric
Interview (MINI) to conrm the diagnosis in the patient group and
the absence of psychiatric disorders in the HC (Sheehan et al.,

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

1998). Clinical features were extracted from the interviews held

according to the general MOODINFLAME protocol. This protocol
included the Patient Questionnaire of the former Stanley Foundation Bipolar Network, the YMRS, the IDS-C30 and a somatic illness questionnaire (Leverich et al., 2001; Rush et al., 1986; Young
et al., 1978). The somatic illness questionnaire is a MOODINFLAME
specic checklist exploring all the organ systems for current and
lifetime medical symptoms. In the event of a mismatch of results
from the MINI in relation to the Patient Questionnaire, diagnoses
were checked with the treating physician.
2.4. Magnetic resonance imaging acquisition
Diffusion-weighted (DW) data were acquired using a singleshot pulsed gradient spin echo EPI sequence (TR 8041 ms,
TE 59.98 ms) on a 3-T MR scanner (Intera, Philips Medical Systems, Best, the Netherlands) equipped with an eight-channel head
coil. For each subject, 32 DW images were collected along noncollinear directions, with a maximum gradient strength of 40 mT/
m and a b-value of 1000 s/mm2. In addition, one non-DW volume
(b E0, referred in the following as a b0 image) was acquired. Each
volume consisted of 55 transverse slices (FOV 240  240 mm,
voxel size 2.5 mm isotropic, no gap). The total acquisition time
for this sequence was less than 6 min.
In addition, we acquired also one high-resolution (1 mm isotropic) T1-weighted image for each participant (SPGR,
TR 9758 ms, TE 4.59 ms, ip angle 8, FOV 220  220 mm,
voxel size 0.859  0.859, slice thickness 1.2 mm) covering the entire cerebrum. The total acquisition time for this sequence was less
than 5 min. PAR-REC les collected from scanning were transferred into NIFTI format data using MRIcroX software (version 1.2,
http://www.mccauslandcenter.sc.edu/CRNL/tools/mricro) which
also provided the text les containing the diffusion vectors (bvecs)
and the b-values (bvals) needed to t the diffusion tensor to each
image voxel.
The acquisition of all MRI data was performed according to our
protocol for limiting head movement: the participant's head was
restrained with foam pads on either sides, and further contained
with a Velcro strap.
2.5. Image processing and analysis
Data preprocessing and analysis were performed with the FSL
FMRIB Software Library v5.0 (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/,
FMRIB, Oxford Center for Functional MRI of the Brain, University of
Oxford, Department of Clinical Neurology, John Radcliffe Hospital,
Oxford, United Kingdom), a comprehensive library of analysis tools
for MRI data (Jenkinson et al., 2012; Smith et al., 2004; Woolrich
et al., 2009).
2.5.1. Preprocessing and diffusion tensor estimation
For every participant, and in the native space, DW images were
rst corrected for motion and eddy-current distortions using the
eddy correct utility in FSL. The b0 image was used as the reference
image for realignment of all the DW images. The log of this procedure provided information to estimate the mean displacement
for each subject across the acquisition. A two-samples t-test failed
to reveal a signicant difference between the mean displacement
between groups (t  1.51, p 0.13). Subsequently, we extracted a
brain mask on the b0 image by means of automated skull stripping
(Smith, 2002) and FSL Fdt was used to perform diffusion tensor
estimation on the voxels encompassed by the brain mask (Behrens
et al., 2003). This yielded an estimation of fractional anisotropy
(FA), mean diffusivity (MD), radial diffusivity (RD) and axial diffusivity (AD) for each brain voxel. Quality control was performed
by visual inspection of the raw DTI and FA images (Soares et al.,

283

2013; Tournier et al., 2011), performed independently by two researchers (BH, LC). Due to technical problems DW acquisition
failed to provide workable images in one patient and two healthy
controls, effectively yielding 22 HC DW scans and 21 BD-I DW
scans.
2.5.2. Tract based spatial statistics
Group differences in white matter measurements from DTI can
be assessed for evey voxel in the white matter, however they are
particularly reliable for the inner section of large white matter
tracts, where the effect of partial voluming (i.e. contamination of
the diffusion signal by intravoxel grey matter or CSF) can be ruled
out, and inaccuracies in image registration across individuals are
reduced with respect to the outer (radial) white matter, the latter
being more problematic due to the highly variable gyral
morphology.
In this perspective, we estimated group differences in white
matter microstructural parameters exclusively on the voxel of a
white matter 'skeleton' approximating the center of the ber tracts
common to all the participants. To estimate the spatial location of
this skeleton, we used Tract-Based Spatial Statistics to prepare our
FA, MD, RD, AD images. The details of this method are provided in
the article of Smith et al. (2006).
In summary, each subject's FA image is nonlinearly registered
to the FMRIB58 FA template and checked for distortions generated
by the alignment procedure. A low degrees-of-freedom nonlinear
registration is implemented, to prevent altering the subject-specic white matter topology (e.g. by merging two distinct ber
tracts into one). The mean FA image across all participants is calculated and thresholded to a minimum value of 0.2. The thresholded mean FA image denes the sample-specic white matter
skeleton. The white matter skeleton is overlaid onto each subjects
registered FA image. The maximum FA value is searched along the
normals to the skeleton - and weighted for increasing distance
from it - and dragged onto the skeleton for group-level comparison. This step is required to take care of potential residual misalignments from step 1.
TBSS has several advantages over whole-brain approaches when
it comes to estimate between-group differences in white matter
microstructure indices derived from diffusion tensor imaging: no
smoothing is required either in the two-stages registration or in the
creation of the white matter skeleton, thereby relaxing the requirement of prior hypotheses on the spatial extent of the effect,
which is hard to dene a priori. In addition, performing the group
comparisons only on the white matter skeleton increases the statistical power of the analysis and minimizes the chance that the
results are driven by partial volume effects - another potential
drawback of spatial smoothing - or confounding morphological
differences such as ventricles enlargement or atrophy.
Two separate group comparisons were performed. In the rst,
the mean FA, MD, RD and AD values for each voxel on the skeleton
were compared between healthy control subjects (HC) and the
whole group of bipolar patients (BD-I). To test a model in which
lithium corrects white matter microstructure disturbances, we
subsequently split up the BD-I subjects into two subgroups according to the presence or absence of lithium use. To maximize
statistical power, we compared voxelwise values of white matter
microstructure across the three groups (non-lithium-users, lithium-users, healthy controls). In both cases, parameter estimates
for group-level analysis were obtained from a general linear model
(GLM). Potential artifacts due to age (Lebel et al., 2012b; Madden
et al., 2007) or residual motion effects on DWI images were controlled in the group-level analysis, using a procedure similar to
that described in Yendiki et al. (2013), that is by adding subjectspecic estimates of motion - derived from the eddy_correct
procedure - as a covariate of no interest in assessing differences in

284

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

FA, MD, RD and AD. Within the clusters found by the second GLM
model we compared the white matter estimates between the
subgroups using a two-independent samples Students t-test.
Finally, correlation analyses were performed to estimate the
correlation between FA, MD, RD, AD and age at onset, number of
depressive and manic episodes, and IDS-C30 score within the patient group. These analyses were done using a GLM including age
and head motion as a covariate.
Inference was carried out by means of nonparametric permutation testing (Nichols and Holmes, 2002; Winkler et al., 2014)
implemented in FSL's randomize software. Five-thousands permutations were calculated for each group comparison, in order to
estimate the null distribution of either FA, MD, RD and AD differences. Threshold-Free Cluster Enhancement (TFCE) (Smith and
Nichols, 2009) was used to correct for multiple comparisons using
the default values provided by the T2 options of FSL randomize,
which are optimized for TBSS analysis. Cluster size and MNI coordinates at signal peak were derived with FSL Cluster and the
corresponding white matter tract retrieved from the ICBM-DTI-81
white-matter labels, JHU White-Matter Tractography and HarvardOxford cortical structural atlases (Desikan et al., 2006; Hua et al.,
2008; Wakana et al., 2007).
2.5.3. Demographic data analyses
Statistical analyses on the demographic and FA data were performed using Stata Statistical Software, release 14 (StataCorp.
2015, College Station, Texas). The differences in demographic and
FA data between the groups were investigated with Students ttest (age, FA), Pearson's chi-squared test (gender, medication use)
and Wilcoxon-Mann-Whitney rank-sum test (IDS-C30 score, YMRS
score). Pearsons correlation coefcient was used in the analysis of
the association between FA and IDS-C30 score.

Table 1
Characteristics of the subjects.

Group size
Gender
Male
Female
Age (mean (range),
yr)
IDS-C30 score (mean
(range))
YMRS score (mean
(range))
Duration of illness
(mean (range), yr)
Age at onset (mean
(range), yr)
Medication
Citalopram
Trazodon
Lithium
Valproate
Carbamazepine
Lamotrigine
Levetiracetam
Quetiapine
L-Thyroxine
Benzodiazepines

Healthy
controls

Bipolar disorder whole

group

Bipolar disorder nonlithium

Bipolar disorder
lithium

22

21

12

11 (50%)
11 (50%)
38.2 (19
67)
1.1 (04)

9 (43%)
12 (57%)
44.7 (2461)

5 (56%)
4 (44%)
41.2 (2459)

4 (33%)
8 (67%)
47.4 (3661)

5.0 (014)n

6.2 (014)

3.6 (012)

0 (00)

0.2 (02)

0.2 (02)

0.2 (02)

25.2(239)

23.2 (239)

26.7 (1237)

19.5 (1243)

18.0 (1230)

20.7 (1243)

3 (14%)
3 (14%)
12 (57%)
5 (24%)
1 (5%)
6 (29%)
1 (5%)
6 (29%)
4 (19%)
1 (5%)

1 (11%)
1 (11%)
- (0%)
3 (33%)
1 (11%)
2 (22%)
1 (11%)
4 (44%)
1 (11%)
0 (0%)

2 (16%)
2 (16%)
12 (100%)
2 (17%)
0 (0%)
4 (33%)
0 (0%)
2 (17%)
3 (25%)
1 (8%)

Characteristics of the subjects. With exception of the IDS-C30 score, there were no
signicant differences between BD and HC groups nor within the lithium use
subgroups. * statistical signicant p o 0.05.

3. Results

lithium-using BD-I patients and healthy controls were not signicant (Fig. 1). Signicant differences were not found between
any of the groups in MD and AD.

3.1. Demographics

3.3. Correlation analyses within the patient group

Subject characteristics are displayed in Table 1. With exception

of the IDS-C30 score, there were no signicant differences between
BD and HC groups nor within the lithium use subgroups. Although
within the normal (i.e. euthymic) range (between 0 and 14), mean
IDS-C30 scores of BD-I patients and of both subgroups using and
not using lithium were higher than in HCs (z  3.68; p o0.001;
z 2.90, p 0.004; and z  3.33, p 0.001, respectively).

Analyses of the white matter microstructural estimates within

the BD-I patient group while controlling for age and head motion,
demonstrated a negative association between IDS-C30 and FA in 23
clusters, complemented with 2 clusters demonstrating a positive
association with MD and 1 large cluster demonstrating a positive
association with RD (Table 3). The localization and an association
plot of the ve largest FA clusters are depicted in Fig. 2. Signicant
differences were not found between AD and IDS-C30 and any of the
estimates and age at onset, number of depressive or number of
manic episodes.

3.2. Group differences in white matter microstructure

In the rst model comparing the whole BD-I patient group to
the HC group, while controlling for age and head motion, no signicant clusters were found between the groups in the FA, MD, RD
and AD analyses.
In the second model analyzing the estimates of white matter
microstructure across non-lithium-using bipolar disorder I (BD-I)
patients, lithium-using BD-I patients and healthy controls, while
controlling for age and head motion, signicant clusters with a
positive FA association were found in the genu, body and splenium
of the corpus collosum, the left anterior corona radiata, as well as
two small peripheral tracts in the frontal orbital cortex (Table 2,
Fig. 1). Signicant clusters with a negative RD association in this
model were found in the corpus callosum, left and right anterior
corona radiata and the anterior thalamic radiation (Table 2).
Within the genu, body, splenium of the corpus collosum and the
left anterior corona radiata, FA was found to be signicantly lower
in non-lithium-using BD-I patients compared to lithium-using BDI patients, as well as healthy controls. Differences between

4. Discussion
In the present study we could not demonstrate a difference in
white matter microstructure estimates between BD-I patients and
healthy controls. Across non-lithium-using BD-I patients, lithiumusing BD-I patients and healthy controls a positive FA association
with a concomitant negative RD association was found in the
corpus collosum and the (left) anterior corona radiata. Within
these regions, FA was found to be signicantly lower in non-lithium-using BD-I patients compared to lithium-using BD-I patients, as well as healthy controls, but differences between lithium-using BD-I patients and healthy controls were not signicant. Within the BD-I patient group, negative associations between minor depressive symptoms and FA, partly overlapping
with regions with a positive RD or MD association, were found to
be widespread.

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

285

Table 2
Clusters with signicant white matter microstructure estimate associations across non-lithium-using BD-I patients, lithium-using BD-I patients and healthy controls.
No

FA

1
2
3
4
5

RD

1
2
3

White matter tract

Genu and body of corpus callosum

Anterior corona radiate R & L
Body and splenium of corpus callosum
Anterior corona radiata L
White matter of the frontal orbital cortex Lnn
White matter of the frontal orbital cortex Lnn
Genu and body of corpus callosum
Anterior corona radiata R & L
Body and splenium of corpus callosum
Anterior thalamic radiationn

MNI coordinates at signal peak

Number of voxels

Minimal corrected p

Direction of effect

78

156

81

6427

0.020

non-li o li o hc

82
116
113
112

90
155
145
141

89
85
56
55

3993
687
4
3

0.019
0.044
0.050
0.050

non-li
non-li
non-li
non-li

78

156

81

7624

0.026

hc 4 li 4 non-li

82
119

91
164

89
92

2419
222

0.038
0.049

hc 4 li 4 non-li
hc 4 li 4 non-li

o
o
o
o

li
li
li
li

o
o
o
o

hc
hc
hc
hc

Clusters with signicant white matter microstructure estimate associations across non-lithium-using bipolar disorder I (BD-I) patients, lithium-using BD-I patients and
healthy controls. The clusters are ordered on size and presented with the MNI coordinates at signal peak and the corresponding white matter tract retrieved from the ICBMDTI-81 white-matter labels, JHU White-Matter Tractography (*) and Harvard-Oxford cortical structural (**) atlases (Desikan et al., 2006; Hua et al., 2008; Wakana et al., 2007).
Voxelwise values of white matter microstructure were compared across non-lithium-users, lithium-users and healthy controls in a general linear model, including age and
head motion as covariates. Voxelwise levels of signicance, corrected for multiple comparisons, were calculated with a standard permutation testing (5000 permutations) by
building up the null distribution (across permutation of the input data) of the maximum (across voxels) threshold-free cluster-enhancement (TFCE) scores and then using the
95th percentile of the null distribution to threshold signals at corrected p 0.05. FA fractional anisotropy, MD mean diffusivity, RD radial diffusivity, hc healthy control,
li lithium-using BD-I patients, non-li non-lithium-using BD-I patients, R right, L left.

4.1. Euthymic patients

As far as we are aware this is the rst DTI study in BD that did
not nd a difference in DTI-measures between patients and HC.
There are several explanations for this. First, as suggested by recent meta-analytic reviews on DTI in BD an effect of publication
bias reporting positive but not negative ndings cannot be ruled
out (Nortje et al., 2013; Vederine et al., 2011). However, our negative nding may also be related to the fact that in the present
study all patients were euthymic, thereby presenting with less
symptoms than patients do during an episode and could be less
affected in a pathophysiological sense. Yet, some previous studies
did report a difference between euthymic patients and HC, albeit
some with a higher FA and others with a lower FA compared to HC
(Barysheva et al., 2013; Canales-Rodrguez et al., 2014; Chaddock
et al., 2009; Haller et al., 2011; Magioncalda et al., 2016; Mahon
et al., 2012; Oertel-Knchel et al., 2014; Scholz et al., 2016; Sussmann et al., 2009; Wessa et al., 2009). In this regard, it is conspicuous that the studies investigating BD patients while in an
episode reported more regions with altered FA than studies investigating euthymic BD patients (Vederine et al., 2011),
strengthening the case that white matter microstructure disturbances might indeed be more increased or more widespread in
BD patients experiencing more mood symptoms than in euthymic
patients.
Indeed, this theory is supported by our nding of widespread
regions with a negative association between FA and minor depressive symptoms, as well as partly overlapping regions with a
positive association between RD, MD, and minor depressive
symptoms. As far as we know, this correlation has never been
described before in BD. Interestingly however, in a study investigating euthymic older patients a negative relation between FA
and depression score has been reported (Lamar et al., 2010) and in
another study with women with subclinical depression a positive
relation with RD was described (Hayakawa et al., 2013), suggesting
that the association between white matter microstructure aberrations and depression score may not be specic for BD.
Additionally supporting this theory, Magioncalda et al. recently
demonstrated white matter microstructure abnormalities to be
more pronounced in groups of BD patients in a depressive or
manic mood state, compared to euthymic patients (Magioncalda
et al., 2016). Although this, together with the present study merely
suggests a stronger association between mood state and white

matter microstructure disturbances, it does emphasize the importance of longitudinal DTI studies in BD patients, while cycling
through different mood states.
4.2. Lithium
The major regions that we report to have a decreased FA and
increased RD in non-lithium-using patients, compared to lithiumusing patients and to healthy controls, have all been reported to be
implicated in BD DTI studies before (Nortje et al., 2013). It is important to mention that while alterations of DTI parameters (FA,
MD, RD) account for microstructural differences, it is difcult to
further elaborate on precisely what aspect of white matter microstructural integrity is affected in BD. In several of the anatomical locations where group differences were detected, such as
the corpus callosum and the anterior corona radiata (see
Tables 2 and 3), the FA decrease is associated with an increase in
perpendicular diffusivity (RD), while parallel diffusivity (AD) was
not affected. This prole is usually associated with demyelination
or dysmyelination (Benedetti et al., 2013; Song et al., 2002).
From our results it seems that the commissural and projection
bers are more affected by lithium use. When affected, these tracts
are known to be associated with impaired neurocognitive functioning. White matter leftward asymmetry of the anterior corona
radiate is known to relate to the executive control function of attention (Yin et al., 2013), a function that is frequently disturbed in
BD. Implication of the corpus callosum is of specic interest, since
this structure is thought to be important for adequate cognitive
functioning (Hinkley et al., 2012) and several studies describe
cognitive functioning to be signicantly associated with white
matter integrity of the corpus callosum (Ajilore et al., 2015; Bauer
et al., 2015; Poletti et al., 2015), whereas in another study lithiumuse was found to be associated with lower prevalence of dementia
in an elder BD population (Nunes et al., 2007).
The regions demonstrating a higher FA in lithium-using BD-I
patients compared to non-lithium-using BD-I patients is corresponding with a previous study by Benedetti et al. demonstrating
lithium-treatment to be associated with increased DTI measures
(Benedetti et al., 2012). In a recent study Gildengers et al. demonstrated a longer duration of lithium treatment to be associated
with a higher integrity of the white matter microstructure (Gildengers et al., 2015).
Building on this evidence and our results it may seem obvious

286

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

Fig. 1. Localization and fractional anisotropy plot of clusters demonstrating a signicant association across non-lithium-using bipolar disorder I (BD-I) patients, lithium-using
BD-I patients and healthy controls. Sagittal, coronal and axial MRI locations of the signal peak (cross hair) of the identied clusters with a signicant positive fractional
anisotropy (FA) association across non-lithium-using bipolar disorder I (BD-I) patients, lithium-using BD-I patients and healthy controls. The mean FA skeleton has a green
color. Voxels with a signicant association (threshold corrected p 0.05) have a blue light blue color. Comparisons between the subgroups were analyzed using Students
independent samples t-test. A anterior, P posterior, S superior, I inferior, L left, R right, BD bipolar disorder. (For interpretation of the references to color in this
gure legend, the reader is referred to the web version of this article.)

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

287

Table 3
Clusters with a signicant association between white matter estimates and IDS-C30 score in BD-I patients.
No. White matter tract

FA

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

MD 1

2
RD

Corpus callosum and Anterior corona radiata R & L

Posterior thalamic radiation R and Sagittal stratum R
Sagittal stratum L
Cerebral peduncle R & L
Superior longitudinal fasciculus L
Inferior longitudinal fasciculus L and Anterior thalamic
radiation Ln
Corticospinal tract Ln
Corticospinal tract Ln
Inferior longitudinal fasciculus Ln
Body of corpus callosum
Posterior thalamic radiation R and Retrolenticular part of
internal capsule R
Cingulumn
Cingulumn
Cingulumn
Superior and Inferior longitudinal fasciculus L
Inferior fronto-occipital fasciculus Rn
Anterior thalamic radiation Ln
White matter of Precentral Gyrusnn
Inferior longitudinal fasciculus Ln
Anterior thalamic radiation Ln
Anterior thalamic radiation Ln
Anterior thalamic radiation Ln
Cerebral peduncle R & L
Superior cerebellar peduncle R & L
Retrolenticular part of internal capsule R
Fornix
Posterior limb of internal capsule R
Genu, body and splenium of corpus callosum
Posterior thalamic radiation R & L
Internal capsule R & L
Corona radiata R & L
Cingulum
Superior longitudinal fasciculus R & L

MNI coordinates at signal peak

Number of
voxels

Minimal corrected
p

Direction of effect

144
70
129
102
136
106

97
43
116
101
131
67

106
107
34
66
84
120

18,810
6130
1706
1046
570
195

0.017
0.029
0.029
0.041
0.037
0.049

Negative
Negative
Negative
Negative
Negative
Negative

107
103
138
81
108

82
96
107
145
91

131
128
70
99
80

175
147
113
110
96

0.048
0.049
0.048
0.049
0.049

Negative
Negative
Negative
Negative
Negative

104
79
102
138
57
117
131
136
122
93
122

73
69
75
98
84
166
129
120
168
113
169

128
129
121
52
86
90
114
57
89
86
86

46
29
27
26
19
10
8
4
3
3
2

0.049
0.044
0.050
0.049
0.050
0.050
0.049
0.050
0.050
0.050
0.050

Negative
Negative
Negative
Negative
Negative
Negative
Negative
Negative
Negative
Negative
Negative

94

113

70

2674

0.018

Positive

79

123

68

23

0.046

Positive

79

95

97

25,445

0.006

Positive

Clusters with a signicant association between fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD) and inventory of depressive symptoms (IDS-C30) score
in bipolar disorder I (BD-I) patients. The clusters are ordered on size and presented with the MNI coordinates at signal peak and the corresponding white matter tract
retrieved from the ICBM-DTI-81 white-matter labels, JHU White-Matter Tractography (n) and Harvard-Oxford cortical structural (**) atlases(Desikan et al., 2006; Hua et al.,
2008; Wakana et al., 2007). Voxelwise associations between values of white matter microstructure and IDS-C30 score were analyzed in a general linear model, including age
and head motion as covariates. Voxelwise levels of signicance, corrected for multiple comparisons, were calculated with a standard permutation testing (5000 permutations) by building up the null distribution (across permutation of the input data) of the maximum (across voxels) threshold-free cluster-enhancement (TFCE) scores and
then using the 95th percentile of the null distribution to threshold signals at corrected p 0.05. FA fractional anisotropy, MD mean diffusivity, RD radial diffusivity,
hc healthy control, li lithium-using BD-I patients, non-li non-lithium-using BD-I patients, R right, L left.

to assume that lithium has a counteracting effect on white matter

microstructural disturbances, perhaps being one of its therapeutic
mechanisms (Benedetti and Bollettini, 2014; Lenox and Hahn,
2000). From this point of view, in a prognostic way it may be
promising that in the present study the FA of lithium-using patients in the determined clusters was indistinguishable from
healthy controls.
Besides being acknowledged for its efcacy in treating BD, lithium is also known for its multiple points of action (Grunze et al.,
2013). Summarizing, lithium acts on a several second messenger
systems that underpin its regulatory effects on neurotransmission
and its neuroprotective properties. It modulates neurotransmission by moderating adenyl cyclase and cyclic adenosine monophosphate (cAMP) uctuations, and by limiting myoinositol (mI),
protein kinase C (PKC) and myristoylated alanine-rich c kinase
substrate (MARCKS). Over time, these constraints modify gene
transcription within the cells and yield long-lasting mood stabilization. Additionally, lithium reduces the oxidative burden caused
by mood episodes and protects against apoptosis by promoting
neuroprotective pathways such as Akt and facilitating the actions
of neuroprotective proteins such as brain derived neurotrophic
factor (BDNF) and bcl-2. Furthermore, it inhibits glycogen synthase

kinase 3 beta (GSK-3), which besides regulating glycogen

synthesis, is also involved in gene transcription, synaptic plasticity,
cell structure and resilience. Finally, lithium also inhibits proapoptotic proteins such as p53 and processes such as autophagy
(Malhi et al., 2013).
In white matter microstructure, lithium may primarily assert its
effects by moderation of GSK-3 in glial cells, either directly or via
the Akt pathway (Marlinge et al., 2014). In mice, injection with
GSK-3-inhibitors was found to regulate oligodendrocyte differentiation and enhance myelination (Azim and Butt, 2011). Compared to other species, the human brain is exceptionally myelinated (Smaers et al., 2011). This extensive myelination has imposed exceptionally high metabolic demands and is associated
with vulnerabilities that are thought to make humans highly
susceptible to brain disorders throughout their lifespan (Bartzokis,
2012; Haroutunian et al., 2014). The present study, in addition to
the above mentioned DTI studies by Benedetti et al. and Gildengers et al., provides in vivo support for the theory that lithium in
BD has a counteracting effect on white matter microstructural
disturbances by enhancing myelination of neuronal tracts via interaction with GSK-3 in glial cells. However, so far no longitudinal
studies exist looking at DTI measures in patients both before and

288

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

Fig. 2. Localization and fractional anisotropy plot of clusters demonstrating a signicant negative association between fractional anisotropy and IDS-C30 score within BD-I
patients. Sagittal, coronal and axial MRI locations of the signal peak (cross hair) of the identied clusters with a signicant positive fractional anisotropy (FA) association
across non-lithium-using bipolar disorder I (BD-I) patients, lithium-using BD-I patients and healthy controls. The mean FA skeleton has a green color. Voxels with a
signicant association (threshold corrected p 0.05) have a blue light blue color. Pearsons correlation coefcient was used in the analysis of the association between FA
and IDS-C30 score. A anterior, P posterior, S superior, I inferior, L left, Rright, IDS-C30 inventory of depressive symptoms - clinician version. (For interpretation of
the references to color in this gure legend, the reader is referred to the web version of this article.)

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

during or both during and after lithium treatment.

4.3. Limitations
Several limitations have to be mentioned. First, all patients
were naturalistically treated and none of them was medication
nave possibly leading to other medication effects in the observations revealing false phenomena or obfuscating true ones
(positive or false negative ndings). In a review of structural and
function imaging studies in BD, Hafeman et al. observed a limited
effect of medication on functional fMRI and DTI study. Although
rigorous analyses were typically not possible, they noted that in
general the effects were normalizing in nature (Hafeman et al.,
2012).
Second, the study would have taken advantage of a larger
sample size thereby increasing its statistical power. However, the
sample size of the whole BD-I patient group and the HC group is
comparable to several other DTI studies in BD (ranging from 18 to
42 subjects per group) that did report differences between the
groups(Vederine et al., 2011). Since the lithium-using BD and nonlithium-using BD subgroups are even smaller, we analyzed the
effect of lithium across the three subgroups, instead of comparing
it solely between lithium-using patients and non-lithium-using
patients. Furthermore, a post-hoc power calculation of the study of
Benedetti reveals that with the sample size used in that study, a
2550% probability of type II error is still present (see Supplemental information), which could also explain our null nding
(see Supplemental information).
Third, although TBSS represents the state-of-the-art technique
for voxel-wise DTI analyses, it has its own limitations: (1) the
mentioned advantages of considering only the center of the ber
tract limit the possibility to detect group differences in tract
thickness; (2) the white matter skeleton does not include many
white matter structures in the most radial (i.e. outer) white matter,
closest to the cortex, where short-range connections subserved by
U-shaped bers are located. (3) precise hypotheses about the expected spatial location of the group difference would allow to
perform a group comparison based on FA values averaged over a
region of interest, thereby benetting from the corresponding
decrease of noise in the subject- and group-level parameter
estimate.
Finally, although in our study the more larger clusters display a
FA decrease associated with an increase in perpendicular diffusivity, typical for demyelination or dysmyelination (Benedetti
et al., 2013; Song et al., 2002), the anatomical interpretation becomes more difcult for regions where only a difference in FA was
detected. In this regions changes in FA can be due to a number of
factors, including axonal diameter, axonal packing, or even to nonpathological situations (e.g. complex ber organization such as
crossing, kissing or fanning bers) where the diffusion tensor is
not able to adequately model the diffusion-weighted signal (Jbabdi
et al., 2010). The latter in particular represents a limitation of all
in vivo studies on microstructural integrity in BD published so far.
A better anatomical characterization of the group differences that
can be estimated using diffusion-weighted imaging therefore
awaits future methodological developments, which can better
characterize the properties of the diffusion signal with respect to
their anatomical origin.

5. Conclusions
In conclusion, in the present study we did not nd a difference
in DTI-based white matter microstructural estimates beyond an
alpha level of 0.05, for the comparison between euthymic BD-I
patients and HC. Our null nding might suggest that white matter

289

microstructure disturbances could be less outspoken in BD patients not experiencing mood symptoms, supported by our nding
of widespread clusters demonstrating negative FA associations and
positive RD and MD associations with minor depressive symptoms. Alternatively, our negative nding, as well as previous positive ndings, could be explained by issues related to statistical
power.
Differences in FA and RD across non-lithium-using BD-I patients, lithium-using BD-I patients and healthy controls could be
demonstrated in the genu, body, splenium of the corpus collosum
and the anterior corona radiata, suggesting a counteracting effect
of lithium on myelination problems that are associated with BD.
In our opinion, there is need for longitudinal DTI studies of BD
patients that would allow for the investigation of white matter
microstructural disturbances over the course of the illness and
across mood episodes, as well as the evaluation of the effect of
different treatments including lithium, deepening our understanding of this obvious important intermediate mechanism in BD
pathophysiology.

Acknowledgements
We thank Julitte Kalkman for accompanying the patients;
Anita Sibeijn-Kuiper and Judith Streurman for their assistance with
the acquisition of MRI images and Annemiek Oldenziel for het
assistance with the TBSS analysis.

Appendix A. Supplementary material

Supplementary data associated with this article can be found in
the online version at http://dx.doi.org/10.1016/j.jad.2016.05.040.

References
Ajilore, O., Vizueta, N., Walshaw, P., Zhan, L., Leow, A., Altshuler, L.L., 2015. Connectome signatures of neurocognitive abnormalities in euthymic bipolar I
disorder. J. Psychiatr. Res. 68, 3744. http://dx.doi.org/10.1016/j.
jpsychires.2015.05.017.
Azim, K., Butt, A.M., 2011. GSK3 negatively regulates oligodendrocyte differentiation and myelination in vivo. Glia 59, 540553. http://dx.doi.org/10.1002/
glia.21122.
Bartzokis, G., 2012. Neuroglialpharmacology: myelination as a shared mechanism
of action of psychotropic treatments. Neuropharmacology 62, 21372153. http:
//dx.doi.org/10.1016/j.neuropharm.2012.01.015.
Barysheva, M., Jahanshad, N., Foland-Ross, L., Altshuler, L.L., Thompson, P.M., 2013.
White matter microstructural abnormalities in bipolar disorder: a whole brain
diffusion tensor imaging study. NeuroImage Clin. 2, 558568. http://dx.doi.org/
10.1016/j.nicl.2013.03.016.
Basser, P.J., Mattiello, J., LeBihan, D., 1994. MR diffusion tensor spectroscopy and
imaging. Biophys. J. 66, 259267. http://dx.doi.org/10.1016/S0006-3495(94)
80775-1.
Bauer, I.E., Ouyang, A., Mwangi, B., Sanches, M., Zunta-Soares, G.B., Keefe, R.S.E.,
Huang, H., Soares, J.C., 2015. Reduced white matter integrity and verbal uency
impairment in young adults with bipolar disorder: a diffusion tensor imaging
study. J. Psychiatr. Res. 62, 115122. http://dx.doi.org/10.1016/j.
jpsychires.2015.01.008.
Beaulieu, C., 2002. The basis of anisotropic water diffusion in the nervous system - a
technical review. NMR Biomed. 15, 435455. http://dx.doi.org/10.1002/
nbm.782.
Behrens, T.E.J., Woolrich, M.W., Jenkinson, M., Johansen-Berg, H., Nunes, R.G., Clare,
S., Matthews, P.M., Brady, J.M., Smith, S.M., 2003. Characterization and propagation of uncertainty in diffusion-weighted MR imaging. Magn. Reson. Med. 50,
10771088. http://dx.doi.org/10.1002/mrm.10609.
Bellani, M., Boschello, F., Delvecchio, G., Dusi, N., Altamura, C.A., Ruggeri, M.,
Brambilla, P., 2016. DTI and myelin plasticity in bipolar disorder: integrating
neuroimaging and neuropathological nding. Front. Psychiatry 7, 110. http:
//dx.doi.org/10.3389/fpsyt.2016.00021.
Benedetti, F., Bollettini, I., 2014. Recent ndings on the role of white matter pathology in bipolar disorder. Harv. Rev. Psychiatry 22, 338341. http://dx.doi.org/
10.1097/HRP.0000000000000007.
Benedetti, F., Bollettini, I., Barberi, I., Radaelli, D., Poletti, S., Locatelli, C., Pirovano, A.,

290

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

Lorenzi, C., Falini, A., Colombo, C., Smeraldi, E., 2012. Lithium and GSK3-
promoter gene variants inuence white matter microstructure in bipolar disorder. Neuropsychopharmacology 38, 313327. http://dx.doi.org/10.1038/
npp.2012.172.
Benedetti, F., Giacosa, C., Radaelli, D., Poletti, S., Pozzi, E., Dallaspezia, S., Falini, A.,
Smeraldi, E., 2013. Widespread changes of white matter microstructure in
obsessive-compulsive disorder: effect of drug status. Eur. Neuropsychopharmacol. 23, 581593. http://dx.doi.org/10.1016/j.euroneuro.2012.07.002.
Canales-Rodrguez, E.J., Pomarol-Clotet, E., Radua, J., Sarr, S., Alonso-Lana, S., Del
Mar Bonnn, C., Goikolea, J.M., Maristany, T., Garca-lvarez, R., Vieta, E.,
McKenna, P., Salvador, R., 2014. Structural abnormalities in bipolar euthymia: a
multicontrast molecular diffusion imaging study. Biol. Psychiatry 76, 239248.
http://dx.doi.org/10.1016/j.biopsych.2013.09.027.
Chaddock, C.A., Barker, G.J., Marshall, N., Schulze, K., Hall, M.H., Fern, A., Walshe, M.,
Bramon, E., Chitnis, X.A., Murray, R., McDonald, C., 2009. White matter microstructural impairments and genetic liability to familial bipolar I disorder. Br.
J. Psychiatry 194, 527534. http://dx.doi.org/10.1192/bjp.bp.107.047498.
Desikan, R.S., Sgonne, F., Fischl, B., Quinn, B.T., Dickerson, B.C., Blacker, D., Buckner,
R.L., Dale, A.M., Maguire, R.P., Hyman, B.T., Albert, M.S., Killiany, R.J., 2006. An
automated labeling system for subdividing the human cerebral cortex on MRI
scans into gyral based regions of interest. Neuroimage 31, 968980. http://dx.
doi.org/10.1016/j.neuroimage.2006.01.021.
Gildengers, A.G., Butters, M.A., Aizenstein, H.J., Marron, M.M., Emanuel, J., Anderson, S.J., Weissfeld, L.A., Becker, J.T., Lopez, O.L., Mulsant, B.H., Reynolds, C.F.,
2015. Longer lithium exposure is associated with better white matter integrity
in older adults with bipolar disorder. Bipolar Disord. 17, 248256. http://dx.doi.
org/10.1111/bdi.12260.
Gonul, A.S., Coburn, K., Kula, M., 2009. Cerebral blood ow, metabolic, receptor, and
transporter changes in bipolar disorder: the role of PET and SPECT studies. Int.
Rev. Psychiatry 21, 323335. http://dx.doi.org/10.1080/09540260902962131.
Grunze, H., Vieta, E., Goodwin, G.M., Bowden, C., Licht, R.W., Mller, H.-J., Kasper, S.,
2013. The World Federation of Societies of Biological Psychiatry (WFSBP)
guidelines for the biological treatment of bipolar disorders: update 2012 on the
long-term treatment of bipolar disorder. World J. Biol. Psychiatry 14, 154219.
http://dx.doi.org/10.3109/15622975.2013.770551.
Haarman, B.C.M., Burger, H., Doorduin, J., Renken, R.J., Sibeijn-Kuiper, A.J., Marsman,
J.-B.C., de Vries, E.F.J., de Groot, J.C., Drexhage, H.A., Mendes, R., Nolen, W.A.,
Riemersma-Van der Lek, R.F., 2015. Volume, metabolites and neuroinammation of the hippocampus in bipolar disorder - a combined magnetic resonance
imaging and positron emission tomography study. Brain. Behav. Immun. 60,
15. http://dx.doi.org/10.1016/j.bbi.2015.09.004.
Haarman, B.C.M., Riemersma - Van der Lek, R.F., Ruh, H.G., Groot, J.C., Nolen, W.A.,
Doorduin, J., 2014a. Bipolar disorders. In: Dierckx, R.A.J.O., Otte, A., de Vries, E.F.
J., Waarde, A., den Boer, J.A. (Eds.), PET and SPECT in Psychiatry. Springer Berlin
Heidelberg, Heidelberg,
pp. 223251. http://dx.doi.org/10.1007/978-3-642-40384-2_9.
Haarman, B.C.M., Riemersma-Van der Lek, R.F., de Groot, J.C., Ruh, H.G., (Eric),
Klein, H.C., Zandstra, T.E., Burger, H., Schoevers, R. a, de Vries, E.F.J., Drexhage, H.
a, Nolen, W. a, Doorduin, J., 2014b. Neuroinammation in bipolar disorder a
[11C]-(R)-PK11195 positron emission tomography study. Brain. Behav. Immun.
40, 219225. http://dx.doi.org/10.1016/j.bbi.2014.03.016.
Hafeman, D.M., Chang, K.D., Garrett, A.S., Sanders, E.M., Phillips, M.L., 2012. Effects
of medication on neuroimaging ndings in bipolar disorder: an updated review. Bipolar Disord. 14, 375410. http://dx.doi.org/10.1111/
j.1399-5618.2012.01023.x.
Haller, S., Xekardaki, A., Delaloye, C., Canuto, A., Lvblad, K.O., Gold, G., Giannakopoulos, P., 2011. Combined analysis of grey matter voxel-based morphometry
and white matter tract-based spatial statistics in late-life bipolar disorder. J.
Psychiatry Neurosci. 36, 391401. http://dx.doi.org/10.1503/jpn.100140.
Haroutunian, V., Katsel, P., Roussos, P., Davis, K.L., Altshuler, L.L., Bartzokis, G., 2014.
Myelination, oligodendrocytes, and serious mental illness. Glia 62, 18561877.
http://dx.doi.org/10.1002/glia.22716.
Harsan, L.A., Poulet, P., Guignard, B., Steibel, J., Parizel, N., de Sousa, P.L., Boehm, N.,
Grucker, D., Ghandour, M.S., 2006. Brain dysmyelination and recovery assessment by noninvasive in vivo diffusion tensor magnetic resonance imaging. J.
Neurosci. Res. 83, 392402. http://dx.doi.org/10.1002/jnr.20742.
Hayakawa, Y.K., Sasaki, H., Takao, H., Mori, H., Hayashi, N., Kunimatsu, A., Aoki, S.,
Ohtomo, K., 2013. Structural brain abnormalities in women with subclinical
depression, as revealed by voxel-based morphometry and diffusion tensor
imaging. J. Affect. Disord. 144, 263268. http://dx.doi.org/10.1016/j.
jad.2012.10.023.
Hinkley, L.B.N., Marco, E.J., Findlay, A.M., Honma, S., Jeremy, R.J., Strominger, Z.,
Bukshpun, P., Wakahiro, M., Brown, W.S., Paul, L.K., Barkovich, A.J., Mukherjee,
P., Nagarajan, S.S., Sherr, E.H., 2012. The role of corpus callosum development in
functional connectivity and cognitive processing. PLoS One, 7. http://dx.doi.org/
10.1371/journal.pone.0039804.
Hua, K., Zhang, J., Wakana, S., Jiang, H., Li, X., Reich, D.S., Calabresi, P.A., Pekar, J.J.,
van Zijl, P.C.M., Mori, S., 2008. Tract probability maps in stereotaxic spaces:
analyses of white matter anatomy and tract-specic quantication. Neuroimage
39, 336347. http://dx.doi.org/10.1016/j.neuroimage.2007.07.053.
Jbabdi, S., Behrens, T.E.J., Smith, S.M., 2010. Crossing bres in tract-based spatial
statistics. Neuroimage 49, 249256. http://dx.doi.org/10.1016/j.
neuroimage.2009.08.039.
Jenkinson, M., Beckmann, C.F., Behrens, T.E.J., Woolrich, M.W., Smith, S.M., 2012.
FSL. Neuroimage 62, 782790. http://dx.doi.org/10.1016/j.
neuroimage.2011.09.015.

Jones, D.K., Knsche, T.R., Turner, R., 2013. White matter integrity, ber count, and
other fallacies: the dos and don'ts of diffusion MRI. Neuroimage 73, 239254.
http://dx.doi.org/10.1016/j.neuroimage.2012.06.081.
Kupferschmidt, D. a, Zakzanis, K.K., 2011. Toward a functional neuroanatomical
signature of bipolar disorder: quantitative evidence from the neuroimaging
literature. Psychiatry Res. 193, 7179. http://dx.doi.org/10.1016/j.
pscychresns.2011.02.011.
Lamar, M., Charlton, R.A., Morris, R.G., Markus, H.S., 2010. The impact of subcortical
white matter disease on mood in euthymic older adults: a diffusion tensor
imaging study. Am. J. Geriatr. Psychiatry 18, 634642. http://dx.doi.org/10.1097/
JGP.0b013e3181cabad1.
Lebel, C., Benner, T., Beaulieu, C., 2012a. Six is enough? Comparison of diffusion
parameters measured using six or more diffusion-encoding gradient directions
with deterministic tractography. Magn. Reson. Med. 68, 474483. http://dx.doi.
org/10.1002/mrm.23254.
Lebel, C., Gee, M., Camicioli, R., Wieler, M., Martin, W., Beaulieu, C., 2012b. Diffusion
tensor imaging of white matter tract evolution over the lifespan. Neuroimage
60, 340352. http://dx.doi.org/10.1016/j.neuroimage.2011.11.094.
Lenox, R.H., Hahn, C.G., 2000. Overview of the mechanism of action of lithium in
the brain: fty-year update. J. Clin. Psychiatry 61 (Suppl 9), 515.
Leverich, G.S., Nolen, W.A., Rush, A.J., McElroy, S.L., Keck, P.E., Denicoff, K.D., Suppes,
T., Altshuler, L.L., Kupka, R., Kramlinger, K.G., Post, R.M., 2001. The Stanley
foundation bipolar treatment outcome network. I. Longitudinal methodology. J.
Affect. Disord. 67, 3344.
Liu, J.-X., Chen, Y.-S., Hsieh, J.-C., Su, T.-P., Yeh, T.-C., Chen, L.-F., 2010. Differences in
white matter abnormalities between bipolar I and II disorders. J. Affect. Disord.
127, 309315. http://dx.doi.org/10.1016/j.jad.2010.05.026.
Madden, D.J., Spaniol, J., Whiting, W.L., Bucur, B., Provenzale, J.M., Cabeza, R., White,
L.E., Huettel, S.A., 2007. Adult age differences in the functional neuroanatomy of
visual attention: a combined fMRI and DTI study. Neurobiol. Aging 28, 459476.
http://dx.doi.org/10.1016/j.neurobiolaging.2006.01.005.
Magioncalda, P., Martino, M., Conio, B., Piaggio, N., Teodorescu, R., Escelsior, A.,
Marozzi, V., Rocchi, G., Roccatagliata, L., Northoff, G., Inglese, M., Amore, M.,
2016. Patterns of microstructural white matter abnormalities and their impact
on cognitive dysfunction in the various phases of type I bipolar disorder. J.
Affect. Disord. 193, 3950. http://dx.doi.org/10.1016/j.jad.2015.12.050.
Mahon, K., Burdick, K.E., Wu, J., Ardekani, B.A., Szeszko, P.R., 2012. Relationship
between suicidality and impulsivity in bipolar I disorder: a diffusion tensor
imaging study. Bipolar Disord. 14, 8089. http://dx.doi.org/10.1111/
j.1399-5618.2012.00984.x.
Makoukji, J., Belle, M., Meffre, D., Stassart, R., Grenier, J., Ghjacumu, G., 2012. Lithium enhances remyelination of peripheral nerves. PNAS, 27. http://dx.doi.
org/10.1073/pnas.1121367109.
Malhi, G.S., Tanious, M., Das, P., Coulston, C.M., Berk, M., 2013. Potential mechanisms of action of lithium in bipolar disorder: current understanding. CNS Drugs
. http://dx.doi.org/10.1007/s40263-013-0039-0.
Marlinge, E., Bellivier, F., Houenou, J., 2014. White matter alterations in bipolar
disorder: potential for drug discovery and development. Bipolar Disord. 16,
97112. http://dx.doi.org/10.1111/bdi.12135.
Nichols, T.E., Holmes, A.P., 2002. Nonparametric permutation tests for functional
neuroimaging: a primer with examples. Hum. Brain Mapp. 15, 125. http://dx.
doi.org/10.1002/hbm.1058.
Nortje, G., Stein, D.J., Radua, J., Mataix-Cols, D., Horn, N., 2013. Systematic review
and voxel-based meta-analysis of diffusion tensor imaging studies in bipolar
disorder. J. Affect. Disord. 150, 192200. http://dx.doi.org/10.1016/j.
jad.2013.05.034.
Nunes, P.V., Forlenza, O.V., Gattaz, W.F., 2007. Lithium and risk for Alzheimers
disease in elderly patients with bipolar disorder. Br. J. Psychiatry 190, 359360.
http://dx.doi.org/10.1192/bjp.bp.106.029868.
Oertel-Knchel, V., Reinke, B., Alves, G., Jurcoane, A., Wenzler, S., Prvulovic, D.,
Linden, D., Knchel, C., 2014. Frontal white matter alterations are associated
with executive cognitive function in euthymic bipolar patients. J. Affect. Disord.
155, 223233. http://dx.doi.org/10.1016/j.jad.2013.11.004.
Poletti, S., Bollettini, I., Mazza, E., Locatelli, C., Radaelli, D., Vai, B., Smeraldi, E.,
Colombo, C., Benedetti, F., 2015. Cognitive performances associate with measures of white matter integrity in bipolar disorder. J. Affect. Disord. 174,
342352. http://dx.doi.org/10.1016/j.jad.2014.12.030.
Rush, A.J., Weissenburger, J., Giles, E., Schlesser, M.A., Fulton, C.L., 1986. Inventory
Prelim. Depress. Find. Symptomatology ( IDS ): Depress., 6587.
Scholz, V., Houenou, J., Kollmann, B., Duclap, D., Poupon, C., Wessa, M., 2016.
Dysfunctional decision-making related to white matter alterations in bipolar I
disorder. J. Affect. Disord. 194, 7279. http://dx.doi.org/10.1016/j.
jad.2015.12.019.
Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorim, P., Janavs, J., Weiller, E., Hergueta, T., Baker, R., Dunbar, G.C., 1998. The Mini-International Neuropsychiatric
Interview (M.I.N.I.): the development and validation of a structured diagnostic
psychiatric interview for DSM-IV and ICD-10. J. Clin. Psychiatry 59 (Suppl. 2),
2233, quiz 3457.
Smaers, J.B., Steele, J., Case, C.R., Cowper, A., Amunts, K., Zilles, K., 2011. Primate
prefrontal cortex evolution: human brains are the extreme of a lateralized ape
trend. Brain. Behav. Evol. 77, 6778. http://dx.doi.org/10.1159/000323671.
Smith, S.M., 2002. Fast robust automated brain extraction. Hum. Brain Mapp. 17,
143155. http://dx.doi.org/10.1002/hbm.10062.
Smith, S.M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T.E., Mackay, C.
E., Watkins, K.E., Ciccarelli, O., Cader, M.Z., Matthews, P.M., Behrens, T.E.J., 2006.
Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data.

B.C.M. (Benno) Haarman et al. / Journal of Affective Disorders 203 (2016) 281291

Neuroimage 31, 14871505. http://dx.doi.org/10.1016/j.

neuroimage.2006.02.024.
Smith, S.M., Jenkinson, M., Woolrich, M.W., Beckmann, C.F., Behrens, T.E.J., Johansen-Berg, H., Bannister, P.R., De Luca, M., Drobnjak, I., Flitney, D.E., Niazy, R.
K., Saunders, J., Vickers, J., Zhang, Y., De Stefano, N., Brady, J.M., Matthews, P.M.,
2004. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 23 (Suppl. 1), S208S219. http://dx.doi.org/
10.1016/j.neuroimage.2004.07.051.
Smith, S.M., Nichols, T.E., 2009. Threshold-free cluster enhancement: addressing
problems of smoothing, threshold dependence and localisation in cluster inference. Neuroimage 44, 8398. http://dx.doi.org/10.1016/j.
neuroimage.2008.03.061.
Soares, J.M., Marques, P., Alves, V., Sousa, N., 2013. A Hitchhiker's guide to diffusion
tensor imaging. Front. Neurosci. 7, 114. http://dx.doi.org/10.3389/
fnins.2013.00031.
Song, S.-K., Sun, S.-W., Ramsbottom, M.J., Chang, C., Russell, J., Cross, A.H., 2002.
Dysmyelination revealed through MRI as increased radial (but unchanged axial)
diffusion of water. Neuroimage 17, 14291436.
Sussmann, J.E., Lymer, G.K.S., McKirdy, J., Moorhead, T.W.J., Muoz Maniega, S., Job,
D., Hall, J., Bastin, M.E., Johnstone, E.C., Lawrie, S.M., McIntosh, A.M., 2009.
White matter abnormalities in bipolar disorder and schizophrenia detected
using diffusion tensor magnetic resonance imaging. Bipolar Disord. 11, 1118.
http://dx.doi.org/10.1111/j.1399-5618.2008.00646.x.
Tournier, J.-D., Mori, S., Leemans, A., 2011. Diffusion tensor imaging and beyond.
Magn. Reson. Med. 65, 15321556. http://dx.doi.org/10.1002/mrm.22924.
Travers, B.G., Adluru, N., Ennis, C., Tromp, D.P.M., Destiche, D., Doran, S., Bigler, E.D.,
Lange, N., Lainhart, J.E., Alexander, A.L., 2012. Diffusion tensor imaging in autism spectrum disorder: a review. Autism Res. 5, 289313. http://dx.doi.org/
10.1002/aur.1243.
Vederine, F.-E., Wessa, M., Leboyer, M., Houenou, J., 2011. A meta-analysis of whole-brain
diffusion tensor imaging studies in bipolar disorder. Prog. Neuropsychopharmacol.
Biol. Psychiatry . http://dx.doi.org/10.1016/j.pnpbp.2011.05.009.
Versace, A., Almeida, J.R.C., Hassel, S., Walsh, N.D., Novelli, M., Klein, C.R., Kupfer, D.
J., Phillips, M.L., 2008. Elevated left and reduced right orbitomedial prefrontal
fractional anisotropy in adults with bipolar disorder revealed by tract-based

291

spatial statistics. Arch. Gen. Psychiatry 65, 10411052. http://dx.doi.org/

10.1001/archpsyc.65.9.1041.
Wakana, S., Caprihan, A., Panzenboeck, M.M., Fallon, J.H., Perry, M., Gollub, R.L., Hua,
K., Zhang, J., Jiang, H., Dubey, P., Blitz, A., van Zijl, P., Mori, S., 2007. Reproducibility of quantitative tractography methods applied to cerebral white matter.
Neuroimage 36, 630644. http://dx.doi.org/10.1016/j.neuroimage.2007.02.049.
Wessa, M., Houenou, J., Leboyer, M., Chanraud, S., Poupon, C., Martinot, J.-L., Paillre-Martinot, M.-L., 2009. Microstructural white matter changes in euthymic
bipolar patients: a whole-brain diffusion tensor imaging study. Bipolar Disord.
11, 504514. http://dx.doi.org/10.1111/j.1399-5618.2009.00718.x.
Winkler, A.M., Ridgway, G.R., Webster, M.A., Smith, S.M., Nichols, T.E., 2014. Permutation inference for the general linear model. Neuroimage 92, 381397.
http://dx.doi.org/10.1016/j.neuroimage.2014.01.060.
Woolrich, M.W., Jbabdi, S., Patenaude, B., Chappell, M., Makni, S., Behrens, T.,
Beckmann, C., Jenkinson, M., Smith, S.M., 2009. Bayesian analysis of neuroimaging data in FSL. Neuroimage 45. http://dx.doi.org/10.1016/j.
neuroimage.2008.10.055.
World Medical Association, 2013. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA
310, 21912194. http://dx.doi.org/10.1001/jama.2013.281053.
Yendiki, A., Koldewyn, K., Kakunoori, S., Kanwisher, N., Fischl, B., 2013. Spurious
group differences due to head motion in a diffusion MRI study. Neuroimage
88C, 7990. http://dx.doi.org/10.1016/j.neuroimage.2013.11.027.
Yin, X., Han, Y., Ge, H., Xu, W., Huang, R., Zhang, D., Xu, J., Fan, L., Pang, Z., Liu, S.,
2013. Inferior frontal white matter asymmetry correlates with executive control
of attention. Hum. Brain Mapp. 34, 796813. http://dx.doi.org/10.1002/
hbm.21477.
Young, R.C., Biggs, J.T., Ziegler, V.E., Meyer, D. a, 1978. A rating scale for mania:
reliability, validity and sensitivity. Br. J. Psychiatry 133, 429435. http://dx.doi.
org/10.1192/bjp.133.5.429.
Zanetti, M.V., Jackowski, M.P., Versace, A., Almeida, J.R.C., Hassel, S., Duran, F.L.S.,
Busatto, G.F., Kupfer, D.J., Phillips, M.L., 2009. State-dependent microstructural
white matter changes in bipolar I depression. Eur. Arch. Psychiatry Clin. Neurosci. 259, 316328. http://dx.doi.org/10.1007/s00406-009-0002-8.