Gray matter and intrinsic network changes in the posterior cingulate cortex after selective serotonin reuptake inhibitor intake

NeuroImage 84 (2014) 236–244

Contents lists available at ScienceDirect

NeuroImage journal homepage: www.elsevier.com/locate/ynimg

Gray matter and intrinsic network changes in the posterior cingulate cortex after selective serotonin reuptake inhibitor intake Christoph Kraus a,b, Sebastian Ganger a,b, Jan Losak a,b, Andreas Hahn a,b, Markus Savli a,b, Georg S. Kranz a,b, Pia Baldinger a,b, Christian Windischberger c,d, Siegfried Kasper a, Rupert Lanzenberger a,b,⁎ a

Department of Psychiatry and Psychotherapy, Medical University of Vienna, Austria Functional, Molecular and Translational Neuroimaging Lab — PET & MRI, Medical University of Vienna, Austria Center for Medical Physics and Biomedical Engineering, Medical University of Vienna, Austria d MR Center of Excellence, Medical University of Vienna, Austria b c

a r t i c l e

i n f o

Article history: Accepted 16 August 2013 Available online 26 August 2013

a b s t r a c t Preclinical studies have demonstrated that serotonin (5-HT) challenge changes neuronal circuitries and microarchitecture. However, evidence in human subjects is missing. Pharmacologic magnetic resonance imaging (phMRI) applying selective 5-HT reuptake inhibitors (SSRIs) and high-resolution structural and functional brain assessment is able to demonstrate the impact of 5-HT challenge on neuronal network morphology and functional activity. To determine how SSRIs induce changes in gray matter and neuronal activity, we conducted a longitudinal study using citalopram and escitalopram. Seventeen healthy subjects completed a structural and functional phMRI study with randomized, cross-over, placebo-controlled, double-blind design. Significant gray matter increases were observed (among other regions) in the posterior cingulate cortex (PCC) and the ventral precuneus after SSRI intake of 10 days, while decreases were observed within the pre- and postcentral gyri (all P b 0.05, family-wise error [FWE] corrected). Furthermore, enhanced resting functional connectivity (rFC) within the ventral precuneus and PCC was associated with gray matter increases in the PCC (all FWE Pcorr b 0.05). Corroborating these results, whole-brain connectivity density, measuring the brain's functional network hubs, was significantly increased after SSRI-intake in the ventral precuneus and PCC (all FWE Pcorr b 0.05). Short-term administration of SSRIs changes gray matter structures, consistent with previous work reporting enhancement of neuroplasticity by serotonergic neurotransmission. Furthermore, increased gray matter in the PCC is associated with increased functional connectivity in one of the brain's metabolically most active regions. Our novel findings provide convergent evidence for dynamic alterations of brain structure and function associated with SSRI pharmacotherapy. © 2013 Elsevier Inc. All rights reserved.

Introduction Magnetic resonance imaging (MRI) and voxel-based morphometry (VBM) studies in patients with depression and obsessive–compulsive disorder (OCD) showed gray matter enhancements after treatment with selective serotonin reuptake inhibitors (SSRIs) (Hoexter et al., 2012; Smith et al., 2012). Moreover, depressive patients homozygous for the LA-allele in the SERT gene (rs25531) seem to be more susceptible to gray matter atrophy (Frodl et al., 2008), this polymorphism also seems to affect gray matter in healthy subjects (Frodl et al., 2008). Remarkably, at least 3 of the 16 known 5-HT receptors (5-HT1A, 5-HT1B and 5-HT2A) (Gaspar et al., 2003) and the 5-HT transporter (SERT) (Benninghoff et al., 2012) are involved in neuroplasticity processes (Gould, 1999; ⁎ Corresponding author at: Functional, Molecular and Translational Neuroimaging Lab, Department of Psychiatry and Psychotherapy, Medical University of Vienna, Waehringer Guertel 18-20, 1090 Vienna, Austria. Fax: +43 1 40400 3099. E-mail address: [email protected] (R. Lanzenberger). 1053-8119/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.neuroimage.2013.08.036

Mogha et al., 2012; Vitalis et al., 2002). Tight links between the neurotrophin system and 5-HT have previously been shown (Castrén and Rantamäki, 2010) demonstrating the role of 5-HT in regulating neuronal morphology and circuitry (Daubert and Condron, 2010). Selective 5-HT reuptake inhibitors represent the first line medication for depression (Bauer et al., 2007), anxiety disorder, OCD, posttraumatic stress disorder and eating disorders (Aigner et al., 2011; Bandelow et al., 2008). At the serotonergic synapse, SSRIs bind to a binding site at the SERT and block reuptake of 5-HT (Kasper et al., 2009; Stahl, 1998). Treatment with SSRIs results in rapid 20-fold increase in 5-HT levels within the midbrain raphe nuclei (Tao et al., 2000) that increases 5-HT binding at 5-HT1A autoreceptors there, which subsequently alters neuronal firing rates and promotes desensitization of 5-HT1A receptors (Stahl, 1998; Zimmer et al., 2004). The resulting lack of autoinhibition triggers 5-HT release at axon terminals (Gibbons et al., 2012). Most of the existing studies using phMRI applied functional MRI and investigated task related blood oxygen dependent level (BOLD)

C. Kraus et al. / NeuroImage 84 (2014) 236–244

responses. Evidence from phMRI and VBM, which addresses the impact of pharmaceuticals on gray matter structure, is scarce. Yet, this technique is a powerful tool that is able to detect morphological alterations in vivo at high resolution. Recent work has confirmed gray matter alterations detected by MRI with ex vivo MRI scans as well as post-mortem volumetric analysis (Vernon et al., 2011), and structural MRI exhibits an excellent test–retest reliability (Wonderlick et al., 2009). In vivo gray matter changes found by MRI and VBM were validated by postmortem findings (Hornberger et al., 2012). Taken together, a convergent line of evidence demonstrates that 5-HT is involved in development and regulation of gray matter morphology through a series of mechanisms associated with neuroplasticity. Treatment with SSRIs might thus trigger gray matter changes, yet confirmation in healthy subjects is missing and the impact of regionally altered gray matter on neuronal functionality is hardly known. Hence, the aim of this study was to (1) investigate the influence of 5-HT on gray matter and (2) to elucidate the associated functional neuronal network changes. This was accomplished by administration of SSRIs to healthy subjects followed by structural and functional phMRI with quantification of gray matter changes through VBM and assessment of neuronal networks through resting functional connectivity (rFC) analyses.

Materials and methods Subjects A longitudinal, crossover, double-blind, placebo-controlled study design was used. The study sample is part of a previously published fMRI study (Windischberger et al., 2010), yet all analyses were previously not considered. Twenty-four healthy adult subjects were recruited by advertisement at community boards at the General Hospital in Vienna, four subjects did not meet inclusion criteria or refused to participate, 20 subjects were randomized, two subjects dropped out (not related illness, non-compliance) and structural data from one subject was not available for all three points. Hence, structural MRI datasets were available from 17 healthy Caucasian subjects (6 female, 11 male 26.5 ± 6.1 years, mean ± SD, see Table 1). All subjects provided written informed consent and received reimbursement after participation. All subjects underwent a medical examination at the screening visit that included medical history, electrocardiogram and routine blood tests. Exclusion criteria were history of severe disease, any psychiatric (according to assessment by Structured Clinical Interview for DSM-IV Axis I + II Disorders, SCID I + II) or neurological disorder, drug abuse including anabolic steroids, psychiatric medication, use of hormonal contraceptives for the past 6 months, and a positive urine pregnancy test. All subjects were naïve to SSRIs and psychotropic medication. No particular menstrual phase for scanning of female subjects was defined. The interventions ended with a final check-up visit for each participant. All study related procedures were approved by the Ethics Committee of the Medical University of Vienna.

Table 1 Demographic data of the study sample. Data are given as means ± SD. Alcohol units per week = alcohol consumption (liter) × alcohol by volume ratio. BMI = body mass index. tGMD = total gray matter density (placebo condition). P compares males and females with independent sample t-test or Mann–Whitney U test (+) where normal distribution was not obtained by Levene's test.

N Age (years) BMI (kg/m2) Cigarettes/day Alcohol/week tGMD

All subjects

Males

Females

P

17 26.5 ± 6.1 22.3 ± 2.3 1.8 ± 4 5.7 ± 6.5 919.3

11 28 ± 7.1 22 ± 2.7 0.8 ± 2.2 3.9 ± 4.9 918.6 ± 19.9

6 23.8 ± 2.1 23 ± 1.3 3.6 ± 5.7 9 ± 8.2 920.6 ± 14.8

0.185 0.382 0.533+ 0.128 0.746

237

Study design and medication All study subjects received 10 mg escitalopram (S-citalopram), an equivalent dosage of 20 mg citalopram (the 1:1 racemic mixture of R-citalopram and S-citalopram) and placebo, in randomized order respectively, for 10 days prior to MRI scanning. This period of medication intake was chosen to reach the plasma steady-state condition (Kasper et al., 2009; Klein et al., 2007). Study subjects consecutively underwent three MRI scanning sessions (one after citalopram, escitalopram and placebo) with an average interval of 21.8 ± 13.0 (mean, SD) between screening visit and MRT1 (no wash-out period), 33.8 ± 6.5 days between MRT1 and MRT2 and 33.1 ± 4.9 between MRT2 and MRT3. According to the half-lives of citalopram and escitalopram (BezchlibnykButler et al., 2000; Rao, 2007), visit intervals have provided enough time to ensure previous drug/placebo washout. Treatment adherence was ascertained by announcing control of medication intake through measurements of plasma-levels at any given time point during study duration. Color-matched dextrose tablets were used as placebo. In order to blind all study personnel and participants to medication group assignment, independent pharmacists prepared the medication in accordance with a computer generated randomization list and each blister was encoded with a unique number to prevent inferences on treatment type and subject. For quantification of plasma levels, blood samples were taken from each subject approximately 10 min before each fMRI session. Plasma was frozen at −20 °C and shipped for analysis (Quintiles Analytical Services, Sweden). MRI measurements and image analyses Structural MRI measurements were performed on 3 Tesla (T) whole-body MEDSPEC S300 MR-scanner (Bruker BioSpin, Ettlingen, Germany) using a standard quadrature single-loop transmit/receive birdcage head coil at the MR Center of Excellence at the Medical University of Vienna, Austria. The imaging protocol comprised a magnetizationprepared rapid gradient echo (MPRAGE, T1-weighted) sequence yielding 128 slices, a 256 × 256 matrix, at a slice thickness of 1.56 mm and a voxel size of 0.78 × 0.86 mm. Additionally, all study subjects underwent fMRI with a facial expression task (described below), of which the fMRI results were published previously (Hahn et al., 2009). In the same MRI session a single-shot gradient-recalled echo planar imaging (GR-EPI) sequence was applied, optimized for imaging blood oxygen dependent (BOLD) contrast. This EPI-sequence was done at a TE = 31 ms, TR = 1000 ms and a matrix size = 128 × 91, which resulted in a total slab width of 34.5 mm with 10 axial slices of 3 mm thickness aligned to the AC–PC line (0.5 mm slice gap). Voxel-based morphometry In order to test the main hypothesis of this study, which was to detect alterations of gray matter after 5-HT reuptake inhibition, we used VBM for structural brain assessment. All analyses of images were performed with statistical parametric mapping (SPM8, Wellcome Trust Centre for Neuroimaging, Institute of Neurology, University College London, London, United Kingdom, http://www.fil.ion.ucl.ac.uk/spm/software/ spm8/) and MATLAB 7.10 (MathWorks, Natick, MA). An optimized VBM protocol was used, applying the DARTEL (Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra) algorithm (Ashburner, 2007). The images were segmented into gray matter, white matter, and cerebrospinal fluid (CSF) compartments and successfully passed visually checking for major artifacts. Subsequently, the gray matter maps obtained by this procedure were separately normalized to a gray matter template representing the stereotactic standardized Montreal Neurological Institute (MNI) space at a voxel size of 1.5 × 1.5 × 1.5 mm. Based on deformation fields calculated during segmentation, a template was generated by the DARTEL algorithm. The Jacobian determinants derived

238

C. Kraus et al. / NeuroImage 84 (2014) 236–244

from the spatial normalization were used during nonlinear spatial transformations, which modulated the regional differences of partitioned gray matter on images from relative to absolute amounts (volume). Unmodulated images constitute gray matter density (GMD), while modulated images constitute gray matter volume (GMV). Gray matter images were smoothed with a Gaussian filter of 8 × 8 × 8 mm full width at half maximum (FWHM). Preprocessing resting functional connectivity Resting functional connectivity was determined using a method to separate network-specific low frequency BOLD (LF-BOLD) signals from task-evoked BOLD responses (Fair et al., 2007). Intrinsic connectivity networks are present during task performance (Fox et al., 2006) so that the obtained resting activity mimics real-life setting, where the brain is usually continuously engaged in task processing. So gathered intrinsic network activity at rest offers a reliable alternative to restingstate fMRI at the advantage of controlling for inter- and intraindividual variations in thought processes and mentation (Shih et al., 2011). Functional MRI data preprocessing comprised slice-timing correction, realignment, spatial normalization and spatial smoothing as implemented in SPM8. An in-house, scanner-specific EPI template was created from previous data by spatial normalization of each individual scan to the SPM template which was followed by averaging across individuals. The advantage of using an in-house template is that potential local field inhomogeneities of individual scans match those of the template, which in turn improves the spatial normalization. This template was used for normalization to MNI-space and a Gaussian smoothing kernel of 8 × 8 × 8 mm FWHM for spatial smoothing. Resting functional connectivity was extracted from a blocked facial fMRI paradigm (Hahn et al., 2009), consisting of 20 s baseline blocks, during which a black screen with a white fixation cross at the center was presented. Baseline blocks were alternated by 20 s presentation of faces as active task blocks, whereby the facial expression and attractiveness had to be rated. The entire paradigm consisted of 5 alternating baseline and active paradigm blocks leading to total paradigm duration of 200 s. In accordance with Fair et al. (2007), the active task periods were removed by cutting out the five 20 s baseline periods. In consideration of the delay in the hemodynamic response function, cuts were set 5 s after active and resting blocks, respectively, leading to a time shift of 5 s. This number was used to maximize the number of frames within steady-state data. Resting functional connectivity Because both SSRIs and brain changes underlying VBM findings have previously been demonstrated to alter brain connectivity (Reetz et al., 2012; van Marle et al., 2011), we were particularly interested in differences in brain rFC associated with a principal VBM finding between the SSRI and placebo group. Hence, seed voxel correlation analysis was performed with resting data extracted from the FEDT fMRI paradigm as described above. Resting data were corrected as published in an earlier study (Weissenbacher et al., 2009). Linear regression was applied to correct for changes in white-matter, ventricular and global signals. Then, data were band-pass filtered by a 12-term finite impulse response (FIR) filter to 0.009–0.08 Hz with MATLAB. Connectivity maps were calculated by cross-correlation between the BOLD time course from the seed region obtained by VBM (see the Results section) and the time course from the remaining voxels of the entire brain. For group comparisons, correlation maps were converted to z-values using Fisher's r-to-z transformation. Resting functional connectivity density The brains' resting activity has recently been shown to be subdivided into functional “hubs” representing network nodes with high functional

connectivity (Bullmore and Sporns, 2009), which differ between their resting activity and anatomical topology (Tomasi and Volkow, 2011). We hypothesized that the brain's functional hubs are changed themselves by 5-HT reuptake inhibition. To investigate this question we applied resting functional connectivity density (rFCD) mapping. This recently developed functional connectivity analysis to define functional network hubs is a fast voxel-wise data-driven approach sensitive to the number of local functional connections in brain regions (Tomasi and Volkow, 2010, 2011). Technically, functional hubs represent network nodes with a large number of edges as defined by graph theory. This approach yields FCD, in other words the node degree, and resembles the cross-correlation function of every voxel. Functional connectivity density has a “scale-free” distribution in the brain (Tomasi and Volkow, 2010, 2011), with few hubs and numerous weakly connected nodes, consistent with the emergence of scaling in neural networks (He et al., 2010). To reduce spatial dimensionality for more efficient computation of the BOLD time series, rFCD maps were downsampled by spline interpolation to a resolution of 4 × 4 × 4 mm. Only gray matter voxels were further processed using a gray matter mask. White-matter, ventricular and global signals were regressed-out and the time series were detrended by a 4th order polynomial function, which was chosen because of optimal results after visually inspecting the plots of the time-series. For FCD, a connectivity matrix with Pearson correlation coefficients was generated and a threshold was set at the 1% strongest correlations in each graph, the remaining voxels were set to zero (Fornito et al., 2012). Functional connectivity density maps were generated by summing the number of connections within the connectivity matrix for each voxel separately, corresponding to global FCD as defined previously (Tomasi and Volkow, 2010). Statistical analyses To decrease the variance in the verum condition, we combined citalopram and escitalopram groups by averaging the effects within repeated-measure analysis of variance (rmANOVA) design (i.e., the average of the two verum conditions was equally weighted as the placebo condition). Three main analyses were performed: one for VBM (1), rFC (2) and rFCD (3), respectively. (1) In the main VBM statistical analysis a contrast for gray matter density (GMD) differences between structural MRI scans after SSRIs (pooled citalopram and escitalopram scans) and those after placebo was defined by rmANOVA as implemented in SPM8. In this analysis we controlled for total GMD (tGMD) and sex as two additional factors. Results were corrected applying the voxel-level family-wise error (FWE) rate at a significance level of Pcorr b 0.05. Only results reaching a cluster-size above 100 voxels are reported. An identical analysis was performed with gray matter volume (GMV), stronger results are reported. (2) Proceeding from the major VBM finding, the cluster coordinates were used as seed in functional connectivity statistics. Seed voxel rmANOVA with rFC maps was performed in SPM8 and contrasted between SSRI intake (pooled) and placebo condition. Total GMD and sex were controlled as two additional factors. Results were corrected for false positives using FWE-correction at a significance level of Pcorr b 0.05. (3) Finally, group differences in rFCD maps from scans after SSRI intake (pooled) and placebo intake were calculated with rmANOVA in SPM8 and again controlled for sex and tGMD. Results were corrected for false positives using FWE-correction at a significance level of Pcorr b 0.05. Follow-up Pearson correlations were calculated between principal voxel-based morphometry (VBM) finding and selective serotonin reuptake inhibitor (SSRI) plasma levels, to test the relationship between plasma levels and gray matter changes. Furthermore, to investigate a

C. Kraus et al. / NeuroImage 84 (2014) 236–244

more global effect of SSRI plasma levels on VBM-results, we performed an identical rmANOVA controlling for SSRI plasma levels, only. Sex and gray matter density (GMD) were not added here, because the number of variables roughly should not exceed N/10 to avoid overfitting. To test the validity of the extracted resting data and subsequent rFCD analysis, we aimed to depict major FCD hubs as previously demonstrated (Tomasi and Volkow, 2011). Therefore, in an intermediate step before SSRI vs. placebo analysis we calculated a one-sample t-test over rFCD maps of subjects receiving placebo only. Results were corrected with voxel-level FWE-correction at a significance level of Pcorr b 0.05. Additional statistical tests were used: to rule out cortical atrophy as source of variance and to investigate global gray matter differences between SSRIs and placebo. Subject gender differences were calculated with independent sample t-test or Mann–Whitney U test where appropriate. We aimed to exclude cortical atrophy due to aging as a potential source of variance. Since constant covariates across the scans are not part of the general linear model in SPM and could therefore not be included in the main rmANOVA analyses, the linear effect was computed ([citalopram + escitalopram] / 2-placebo) between the different scanning sessions resulting in a separate map for every subject. The subsequent one-sample t-test across these maps is equal to the contrast in the rmANOVA reported above. Hence, the linear effect maps could be used in a regression analysis with age as independent variable to investigate the effects of age on VBM. Age effects were corrected using voxellevel FWE-correction at a significance level of Pcorr b 0.05. Additionally, we aimed to compare global gray matter across treatment with SSRIs or placebo. Therefore, we performed a repeated-measure analysis of variance (rmANOVA) in SPSS 19.0 comparing total GMD after SSRIs (tGMD = citalopram + escitalopram / 2) or placebo intake using sex as between-subjects variable and age as covariate. A statistical level of significance was accepted at Puncorr b 0.001.

239

found in the posterior cingulate cortex (PCC), the ventral precuneus (peak T-values = 16.6), the fusiform gyrus, the insula and the medial superior frontal cortex (all regions bilaterally, 8.1 b T b 16.6, all FWE Pcorr b 0.05; for all regions see Fig. 1 and Table 2). There was no significant correlation between GMD and SSRI plasma levels (r = 0.27, P = 0.278) in the PCC cluster. Topologically similar results were obtained after correcting for SSRI plasma levels (data not shown). However, in the PCC and precuneus, T-values were attenuated and cluster size decreased (T = 6.3, FWE Pcorr = 0.012). Furthermore, significant decreases of GMD after SSRI intake were observed bilaterally in the precentral gyrus, the cerebellum and the cingulate cortex (8.4 b T b 13.4, all FWE Pcorr b 0.05, see Table 3). Gray matter density and GMV results were topologically comparable (data not shown). Seed-voxel rFC differences between SSRIs and placebo The PCC is a known hotspot of neuronal activity and exhibited gray matter increases in the VBM results. Hence, the PCC cluster was chosen for further analysis and subsequent discussion. The PCC result from VBM served as seed region (VBM peak MNI: x, y, z = 8, −45, 38; 3261 voxels) for calculating differences in rFC between the SSRI and placebo group. In the voxel-wise rFC analysis, enhanced connectivity from the PCC seed region was observed after SSRI intake compared to placebo in the bilateral PCC spreading to the ventral precuneus (T = 5.7, FWE Pcorr b 0.05, Fig. 2A). The resulting rFC cluster in the PCC/ventral precuneus was located overlapping and caudally adjacent to the VBM seed region (Fig. 2C). No other regions exhibiting significant positive or negative rFC alterations. Applying less stringent correction, increased rFC was obtained in the cuneus (Puncorr b 0.001, T = 3.6, 40 voxels).

Results

FCD connectivity hub changes under SSRI administration

Age, body mass index, total brain gray matter, and alcohol and cigarette consumption were not significantly different between males and females (all P N 0.1, Table 1). No significant effect of age was observed between the linear-effect maps of SSRI vs. placebo condition and age (all FWE Pcorr N 0.05). Comparison of total gray matter revealed no significant effect of SSRI treatment on total brain GMD (Puncorr = 0.471).

The extraction of resting activity and subsequent rFCD analysis at baseline levels of block design fMRI data in subjects after intake of placebo yielded typical rFCD hubs as reported previously (Tomasi and Volkow, 2011), specifically, in the brain's visual- and default mode network (8.4 b T b 15.2, all FWE Pcorr b 0.05, see Fig. 3). The enhanced VBM and rFC signals in the PCC and adjacent ventral precuneus are located in and around the significant rFCD hub observed in scans after placebo intake (Fig. 3). Compared to placebo, whole-brain rFCD was significantly increased after SSRI-intake in the ventral precuneus spreading to the PCC (T = 4.3, FWE Pcorr b 0.05, Fig. 2B). We observed no other regions exhibiting significant FCD alterations.

Voxel-based morphometry differences between SSRIs and placebo The main VBM analysis revealed significant increases of GMD after oral intake of SSRIs compared to placebo. Gray mater increases were

Fig. 1. Increases of gray matter density in 17 healthy subjects after 10 days of SSRI intake compared to placebo in the bilateral posterior cingulate cortex and adjacent ventral precuneus as measured with structural MRI and voxel-based morphometry. Color bar represents T-values displayed at P b 0.05 (FWE-corrected). Numbers represent coordinates in MNI standard space at the location of the crosshair, warm color tones represent increases.

240

C. Kraus et al. / NeuroImage 84 (2014) 236–244

Table 2 Regions exhibiting increases of gray matter density (GMD) after SSRI intake vs. placebo intake. Voxel-wise whole brain repeated measurements ANOVA using GMD MRIs with treatment modality (SSRIs or placebo) as factors controlling for whole brain GMD and sex. Stereotactic coordinates (x, y, z) represent cluster peaks in the standard Montreal Institute of Neurology (MNI) space. FWE = family-wise error. L = left, R = right. Region

Peak x

Fusiform (see Fig. 2C, coronal plane) Posterior cingulate ventral precuneus (Fig. 1) Insula L Insula R Medial superior frontal L Supramarginal L Supramarginal R Medial superior frontal R Superior occipital L Medial temporal pole R Cerebellar crus R Precentral L Postcentral L Inferior frontal pars triangularis L Medial temporal L Precentral R Posterior hippocampus L Cerebellar crus L Medial frontal L

P-FWE y

z

T

28 −59 −19 15.3 Inf 8 −45 −33 33 −10 −52 38 21 −21 34 50 −62 −44 −39 −45 33 −26 −16 −27

Voxels

z

38 16.6 Inf

b0.001 7758 b0.001 3261

−14 5 10.8 6.9 b0.001 −20 11 11.8 7.2 b0.001 48 51 9.4 6.4 b0.001 −39 26 9.6 6.5 b0.001 −32 42 11.7 7.2 b0.001 45 54 9.2 6.4 b0.001 −69 33 15.0 Inf b0.001 12 −33 8.3 6.0 b0.001 −74 −37 10.8 6.9 b0.001 1 38 10.4 6.8 b0.001 −20 38 9.0 6.3 b0.001 34 12 8.1 5.9 b0.001 −54 20 10.1 6.7 b0.001 −6 48 10 6.7 b0.001 −33 −3 8.8 6.2 b0.001 −87 −37 8.8 6.2 b0.001 −12 51 9.5 6.5 b0.001

748 739 635 581 548 409 235 223 218 191 176 174 155 155 119 116 102

Discussion Our main finding is the localization of dynamic gray matter changes in healthy subjects after 10 days of SSRI intake. These results were independent of VBM gray matter modality (GMD or GMV), controlled for sex and total gray matter and not associated with age. Increased gray matter did not correlate with SSRI plasma levels, statistical significance was, however attenuated upon correcting for plasma levels. Total gray matter was not significantly different between treatment groups. Furthermore, we observed increased functional connectivity after SSRI intake associated with increased gray matter in the PCC within the PCC and

Table 3 Regions exhibiting decreases of gray matter density (GMD) after SSRI intake vs. placebo intake. Voxel-wise whole brain repeated measurements ANOVA using GMD MRIs with treatment modality (SSRIs or placebo) as factors and whole brain GMD and sex as controlling factors. Stereotactic coordinates (x, y, z) represent cluster peaks in the standard Montreal Institute of Neurology (MNI) space. FWE = family-wise error. L = left, R = right. Region

Precentral/postcentral L/R Cerebellum R Cerebellum R Posterior cingulate R Cerebellum L Medial occipital/temporal R Inferior frontal trigonum R Cerebellum R Superior temporal L Anterior cingulate L Medial cingulate R Cuneus L Cerebellum L Superior temporal R Precentral R Medial temporal R Rectus R Precuneus L Medial cingulate R Calcarine R

Peak x

y

z

T

z

37 36 18 −4 −9 50 60 −21 −36 −10 15 −8 −27 34 48 48 2 −10 −12 12

−8 −32 −60 −35 −62 −83 31 −27 −41 33 −8 −78 −99 −38 −2 −33 40 −60 −3 −78

68 −49 −28 17 −25 22 12 −39 6 6 45 18 −21 12 27 −7 −39 36 45 18

13.4 11.4 10.0 12.4 10.3 8.7 11.0 10.9 10.1 10.6 10.1 9.4 8.4 9.2 8.6 8.5 14.0 9.8 8.2 8.4

7.6 7.1 6.6 7.4 6.7 6.2 7.0 7.0 6.7 6.9 6.7 6.4 6.0 6.4 6.1 6.1 7.8 6.6 5.9 6.0

P-FWE

Voxels

b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001

6259 1086 845 807 806 582 578 566 542 373 334 271 257 228 218 146 144 136 122 119

to adjacent gray matter in the ventral precuneus. Connectivity density, a measure for the brain's functional network hubs, was increased after SSRI intake in a topologically identical region in the PCC/ventral precuneus. Taken together, our results provide consistent evidence that SSRI intake is associated with changes of gray matter and neuronal functionality as measured by phMRI. Our findings are striking for two apparent reasons. First, although differently calculated, both increased rFCD and rFC spatially overlap in the PCC/ventral precuneus. Enhanced functional connectivity seeded from increased PCC gray matter was validated by whole brain connectivity density, which resembles whole brain correlational strength of LF-BOLD signals without the need of seeds. Second, these brain structures exhibit large neuronal densities and are among the brain's metabolically most active (Gusnard and Raichle, 2001). Peak values in blood flow, metabolic activity and oxygen turnover are located there, which all might impact changes of structural and functional MRI signals. We observed several regions with significant gray matter increases, such as the fusiform gyrus, the insula, and the frontal cortex, and the posterior cingulate cortex where serotonergic neurotransmission might impact on structural plasticity as measured with MRI. Furthermore, several decreases were found, such as in the pre- and postcentral gyri. The principal neurobiological mechanisms affecting volumetric changes in the living brain as reported by numerous MRI studies are currently being investigated with large effort. The findings revealed by this study appear very interesting because they indicate a role of 5-HT in changing gray matter and functional neuronal activity. Decreased metabolism was discovered at one week and reversal towards increased metabolism after 6 weeks of treatment in the posterior cingulate cortex, hippocampus, insula, putamen as well as in temporal and prefrontal cortices (Mayberg et al., 2000). This was explained by the authors to be caused by receptor downregulation or changes in second messenger systems and further undermines the notion that SSRIs impact on energy metabolism (Webhofer et al., 2011). Furthermore, response dependent differences in regional blood flow were demonstrated in depressive patients in the PCC (Joe et al., 2006). Beyond this, previous evidence from animal research enables a brief discussion on 5-HT mediated neuroplasticity. Hypothetically, SSRI-dependent altered plasticity might occur due to binding at serotonergic receptors, which are linked to second messenger systems effecting restructuration. Candidate molecules potentially mediating such remodeling in the PCC and ventral precuneus are 5-HT1A, 5-HT1B and 5-HT2A receptors (Savli et al., 2012) and crosslink to the neurotrophin system with targets such as BDNF and CREB. Unfortunately, MRI data at the current spatial resolution do not allow inference on the kind of neuronal reactivity or the type of cellular remodeling that cause gray matter signal changes. A number of factors such as altered neuronal or glial reactivity (Lesch and Waider, 2012), liquor circulation, blood flow and angiogenesis (Seevinck et al., 2010) might likewise account for brain volume changes and subsequently altered neuronal functionality. Our data show that changes of regional gray matter MRI signals are also associated with altered functional activity. Higher amount of gray matter in the PCC was associated with increased resting activity in the PCC itself and the adjacent ventral precuneus. This region is a key node of the default mode network and differentially active during learning, memory, reward and task engagement (for review see Pearson et al., 2011). We demonstrated that the brain's functional connectivity in our dataset is compartmentalized in resting network hubs at locations in line with earlier studies (Tomasi and Volkow, 2010, 2011). The anatomical localization and functional activation of these hubs is consistently reported (Gusnard and Raichle, 2001; Tomasi and Volkow, 2010, 2011), though, according to our study, SSRIs might alter the activity of such hubs. In consideration of the fact that multiple imaging modalities detected increases in gray matter, which altered functional connectivity and hub activity in the PCC/ventral precuneus, this study provides robust evidence that SSRIs interfere with a broad cascade of mechanisms affecting several physiological brain systems.

C. Kraus et al. / NeuroImage 84 (2014) 236–244

241

Fig. 2. Increased resting functional connectivity (rFC) and resting functional connectivity density (rFCD) associated with increased gray matter (VBM) in the PCC after SSRI intake compared to placebo. (A) Increased seed-voxel rFC associated with increased gray matter in the PCC (VBM peak MNI: x, y, z = 8, −45, −19; 3261 voxels) to gray matter within small parts of the PCC and caudally adjacent in gray matter of the ventral precuneus. (B) Increased rFCD in subjects after SSRI intake compared to placebo in the PCC and ventral precuneus. (C) Triple overlay indicating close proximity of increased gray matter, rFC and rFCD in the PCC and ventral precuneus, known to represent one of the brain's metabolically most active region (Gusnard and Raichle, 2001). Blue color bars represent rFC T-values at P b 0.05 (FWE-corrected), displayed at Puncorr b 0.001. Purple color bars represent rFCD T-values at P b 0.05 (FWE-corrected), displayed at Puncorr b 0.001. Spectral color bar represents VBM T-values displayed at P b 0.05 (FWE-corrected). Numbers represent coordinates in MNI standard space at the location of the crosshair.

Moreover, these data are indicative that functional changes occur in concert with structural alterations, even after short treatment periods. However, our study design does not allow inference on the length of time the observed network changes last. Previous reports indicate that structural modifications usually take more time than changes in functionality, yet demand further research on the structure–function relationship of brain networks (Bullmore and Sporns, 2009), which emphasizes the relevance of our findings. We are not aware of a similar study in healthy subjects investigating the impacts of SSRIs on neuronal structure and function by combining structural and functional MRI. Increases of gray matter have been recently published in studies using phMRI in psychiatric patients undergoing SSRI therapy. A coinciding longitudinal study using the DARTEL algorithm shows increases of gray matter in the dorsolateral prefrontal cortex after intake of sertraline (Smith et al., 2012). Yet herein the authors investigated depressed patients over a time course of 12 weeks.

Moreover, gray matter increases were reported for the left putamen after 12 week intake of fluoxetine in patients with obsessive–compulsive disorder (Hoexter et al., 2012). Another coinciding study investigated patients with post-traumatic stress disorder and found hippocampal volume increases after 9–12 months of continuous treatment with paroxetine (Vermetten et al., 2003). Gray matter increases after SSRI intake are further supported by a study in first-episode drug-naïve depressive disorder and panic disorder, detailing moderate increases in subcortical structures such as the nucleus accumbens, the putamen and the hippocampus (Lai and Wu, 2011). On the other side, a cross-sectional study with patients suffering from depression and anxiety disorders did not reveal a protecting effect of stable SSRI treatment on gray matter reductions observed in these disorders (van Tol et al., 2010). In consideration of neuroplastic effects of 5-HT, the observed decreases of gray matter in our study are challenging to interpret, but potentially the same mechanisms mediating gray matter increases could be involved. Only one

242

C. Kraus et al. / NeuroImage 84 (2014) 236–244

Fig. 3. Resting functional connectivity density (rFCD) hubs in 17 subjects receiving placebo. Resting functional connectivity was extracted from baseline blocks of a previously published block design fMRI study (Windischberger et al., 2010) and further analyzed by rFCD mapping. The topological pattern matches previously published large datasets (Tomasi and Volkow, 2011) and demonstrates activation of typical resting-state networks in the cuneus/precuneus, cingulate cortex and medial frontal cortex. Color bar represents T-values displayed at P b 0.05 (FWE-corrected) and numbers represent coordinates in MNI standard space at the location of the crosshair.

study reported gray matter decreases, located in the superior temporal cortex in patients with social anxiety disorder after a 12 week-long intake of 20 mg escitalopram (Cassimjee et al., 2010). Taken together, different patient collective drug types and duration in relation to our study impair comparability. Until now, rFC alterations upon SSRI challenge were only scarcely investigated. A study in healthy subjects identified decreased functional coupling between the amygdala and the ventral medial prefrontal cortex (McCabe and Mishor, 2011). Enhanced functional coupling was reported between the amygdala and the anterior insula after duloxetine intake (van Marle et al., 2011), here 19 healthy subjects were analyzed. VBM-based seed regions were not considered in either of these studies. Notably, rFC and rFCD changes after SSRI intake were regionally restricted to intrinsic areas around increased gray matter in the PCC. Cognitive load adherent in blocked design task activation was previously located at the PCC (Newton et al., 2011). Furthermore, transient BOLD responses at block transitions occur at the PCC and many other regions (Fox et al., 2005), yet inserted intervals accounting for delay in the hemodynamic response function limit these transitions. Variance from cognitive load and transitions might theoretically spill into the analyzed baseline blocks, which should be taken into account upon comparison of our data with traditionally obtained resting-state data. Hence, it remains intriguing that whole-brain connectivity density analysis, which is not related to seed-based connectivity, identifies the same region as connectivity from increased gray matter signal. The mechanisms that SSRIs interfere with in this region are therefore likely to be associated with several factors altering substrates detectable both in T1-weighted as well as EPI MRI sequences. The existing gap between underlying molecular mechanisms and alterations of voxel-intensity values is vigorously debated (for critical review and comments see Draganski and Kherif, 2013; Erickson, 2013; Fields, 2013; Thomas and Baker, 2013), so that strong gray matter changes as shown in this study emphasize the need for more translational work on molecular players mediating in vivo structural and functional changes of neuronal networks as measured by MRI. The following study limitations must be reported. Though we analyzed a rather low subject number, sample sizes of active groups within previous studies have been even lower than in our study (Anand et al., 2005; McCabe and Mishor, 2011). Therefore, this factor, though indeed a limitation, remains a common feature in many pharmacologic neuroimaging studies. In addition, subjects were not balanced according to sex, this issue was however addressed by including sex as nuisance variable.

In summary, we found that study subjects after SSRI intake exhibited significant gray matter changes. Moreover, almost identical locations of increased resting functional connectivity and connectivity density associated with gray matter increases in the PCC provide evidence for the involvement of SSRIs in multiple mechanisms changing brain structure and functionality. When taken together, these results point towards plastic changes of brain structure and function as neuronal substrate of effects associated with SSRI intake and offer a paradigm for further exploration of these mechanisms in psychiatric patients. Acknowledgments Data have been measured within a project funded by an investigatorinitiated and unrestricted research grant from H. Lundbeck A/S, Denmark to S. Kasper. The sponsors and funders did not participate in the design and conduct of the study and were not involved in the preparation, review, or approval of the manuscript. The study protocol has been planned by the authors who retained full academic control. In the study presented here we applied new data analysis approaches in structural and functional magnetic resonance imaging recently available beyond the scope of a study already published (Windischberger et al., 2010). The work of C. Kraus has been funded by an intramural grant of the research cluster between the Medical University of Vienna and the University of Vienna (FA103FC001) to R. Lanzenberger and C. Lamm. A. Hahn was funded by a DOC fellowship of the Austrian Academy of Sciences at the Department of Psychiatry and Psychotherapy. The authors are grateful to C. Spindelegger, U. Moser, P. Stein, M. Fink, L. Pezawas, A. Erfurth, and M. Willeit for their medical support, and to A. Holik, S. Friedreich, F. Gerstl, and E. Moser for technical support. We thank M. Spies for native English editing. The study is part of C. Kraus' thesis “Serotonin and Neuroplasticity” supervised by R. Lanzenberger in the Clinical Neurosciences PhD program at the Medical University of Vienna, Austria. Parts of this study have been or will be presented by P. Baldinger at the 19th European Congress of Psychiatry (EPA), March 12–15, 2011, Vienna, Austria, by M. Savli at the 24th European College of Neuropsychopharmacology (ECNP) Congress, September 3–7, 2011, Paris, France, and by C. Kraus at the 11th World Congress of Biological Psychiatry (WFSBP), June 23–27, 2013, Kyoto, Japan. Conflict of interest Without any relevance to this work, S. Kasper declares that he has received grant/research support from Eli Lilly, Lundbeck A/S, Bristol-

C. Kraus et al. / NeuroImage 84 (2014) 236–244

Myers Squibb, Servier, Sepracor, GlaxoSmithKline, and Organon, and has served as a consultant or on advisory boards for AstraZeneca, Austrian Sick Found, Bristol-Myers Squibb, GlaxoSmithKline, Eli Lily, Lundbeck A/S, Pfizer, Organon, Sepracor, Janssen, and Novartis, and has served on speakers' bureaus for AstraZeneca, Eli Lilly, Lundbeck A/ S, Servier, Sepracor and Janssen. R. Lanzenberger received travel grants and conference speaker honoraria from AstraZeneca and Lundbeck A/S. References Aigner, M., Treasure, J., Kaye, W., Kasper, S., Disorders, W.T.F.O.E., 2011. World Federation of Societies of Biological Psychiatry (WFSBP) guidelines for the pharmacological treatment of eating disorders. World J. Biol. Psychiatry 12, 400–443. Anand, A., Li, Y., Wang, Y., Wu, J., Gao, S., Bukhari, L., Mathews, V.P., Kalnin, A., Lowe, M.J., 2005. Antidepressant effect on connectivity of the mood-regulating circuit: an FMRI study. Neuropsychopharmacology 30, 1334–1344. Ashburner, J., 2007. A fast diffeomorphic image registration algorithm. NeuroImage 38, 95–113. Bandelow, B., Zohar, J., Hollander, E., Kasper, S., Moller, H.J., WFSBP Task Force on Treatment Guidelines for Anxiety, O.-C., Post-Traumatic Stress, D., Allgulander, C., Ayuso-Gutierrez, J., Baldwin, D.S., Buenvicius, R., Cassano, G., Fineberg, N., Gabriels, L., Hindmarch, I., Kaiya, H., Klein, D.F., Lader, M., Lecrubier, Y., Lepine, J.P., Liebowitz, M.R., Lopez-Ibor, J.J., Marazziti, D., Miguel, E.C., Oh, K.S., Preter, M., Rupprecht, R., Sato, M., Starcevic, V., Stein, D.J., van Ameringen, M., Vega, J., 2008. World Federation of Societies of Biological Psychiatry (WFSBP) guidelines for the pharmacological treatment of anxiety, obsessive–compulsive and post-traumatic stress disorders — first revision. World J. Biol. Psychiatry 9, 248–312. Bauer, M., Bschor, T., Pfennig, A., Whybrow, P.C., Angst, J., Versiani, M., Möller, H.-J., Disorders, W.T.F.o.U.D., 2007. World Federation of Societies of Biological Psychiatry (WFSBP) guidelines for biological treatment of unipolar depressive disorders in primary care. World J. Biol. Psychiatry 8, 67–104. Benninghoff, J., van der Ven, A., Schloesser, R.J., Moessner, R., Moller, H.J., Rujescu, D., 2012. The complex role of the serotonin transporter in adult neurogenesis and neuroplasticity. A critical review. World J. Biol. Psychiatry 13, 240–247. Bezchlibnyk-Butler, K., Aleksic, I., Kennedy, S.H., 2000. Citalopram—a review of pharmacological and clinical effects. J. Psychiatry Neurosci. 25, 241–254. Bullmore, E., Sporns, O., 2009. Complex brain networks: graph theoretical analysis of structural and functional systems. Nat. Rev. Neurosci. 10, 186–198. Cassimjee, N., Fouche, J.-P., Burnett, M., Lochner, C., Warwick, J., Dupont, P., Stein, D.J., Cloete, K.J., Carey, P.D., 2010. Changes in regional brain volumes in social anxiety disorder following 12 weeks of treatment with escitalopram. Metab. Brain Dis. 25, 369–374. Castrén, E., Rantamäki, T., 2010. The role of BDNF and its receptors in depression and antidepressant drug action: reactivation of developmental plasticity. Dev. Neurobiol. 70, 289–297. Daubert, E.A., Condron, B.G., 2010. Serotonin: a regulator of neuronal morphology and circuitry. Trends Neurosci. 33, 424–434. Draganski, B., Kherif, F., 2013. In vivo assessment of use-dependent brain plasticity— beyond the “one trick pony” imaging strategy. NeuroImage 73, 255–259. Erickson, K.I., 2013. Evidence for structural plasticity in humans: comment on Thomas and Baker (2012). NeuroImage 73, 237–238. Fair, D.A., Schlaggar, B.L., Cohen, A.L., Miezin, F.M., Dosenbach, N.U., Wenger, K.K., Fox, M.D., Snyder, A.Z., Raichle, M.E., Petersen, S.E., 2007. A method for using blocked and event-related fMRI data to study “resting state” functional connectivity. NeuroImage 35, 396–405. Fields, R.D., 2013. Changes in brain structure during learning: fact or artifact? Reply to Thomas and Baker. NeuroImage 73, 260–264. Fornito, A., Zalesky, A., Pantelis, C., Bullmore, E.T., 2012. Schizophrenia, neuroimaging and connectomics. NeuroImage 62, 2296–2314. Fox, M.D., Snyder, A.Z., Barch, D.M., Gusnard, D.A., Raichle, M.E., 2005. Transient BOLD responses at block transitions. NeuroImage 28, 956–966. Fox, M.D., Snyder, A.Z., Zacks, J.M., Raichle, M.E., 2006. Coherent spontaneous activity accounts for trial-to-trial variability in human evoked brain responses. Nat. Neurosci. 9, 23–25. Frodl, T., Koutsouleris, N., Bottlender, R., Born, C., Jäger, M., Mörgenthaler, M., Scheuerecker, J., Zill, P., Baghai, T., Schüle, C., Rupprecht, R., Bondy, B., Reiser, M., Möller, H.-J., Meisenzahl, E.M., 2008. Reduced gray matter brain volumes are associated with variants of the serotonin transporter gene in major depression. Mol. Psychiatry 13, 1093–1101. Gaspar, P., Cases, O., Maroteaux, L., 2003. The developmental role of serotonin: news from mouse molecular genetics. Nat. Rev. Neurosci. 4, 1002–1012. Gibbons, R.D., Hur, K., Brown, C.H., Davis, J.M., Mann, J.J., 2012. Benefits from antidepressants: synthesis of 6-week patient-level outcomes from double-blind placebocontrolled randomized trials of fluoxetine and venlafaxine. Arch. Gen. Psychiatry 69, 572–579. Gould, E., 1999. Serotonin and hippocampal neurogenesis. Neuropsychopharmacology 21, 46S–51S. Gusnard, D.A., Raichle, M.E., 2001. Searching for a baseline: functional imaging and the resting human brain. Nat. Rev. Neurosci. 2, 685–694. Hahn, A., Holik, A., Gerstl, F., Savli, M., Stein, P., Akimova, E., Angleitner, P., Windischberger, C., Kasper, S., Lanzenberger, R., 2009. Anxiety scores are related to amygdala activity induced by facial attractiveness and emotional expressions. NeuroImage 47, 48S.

243

He, B.J., Zempel, J.M., Snyder, A.Z., Raichle, M.E., 2010. The temporal structures and functional significance of scale-free brain activity. Neuron 66, 353–369. Hoexter, M.Q., de Souza Duran, F.L., D'Alcante, C.C., Dougherty, D.D., Shavitt, R.G., Lopes, A.C., Diniz, J.B., Deckersbach, T., Batistuzzo, M.C., Bressan, R.A., Miguel, E.C., Busatto, G.F., 2012. Gray matter volumes in obsessive–compulsive disorder before and after fluoxetine or cognitive-behavior therapy: a randomized clinical trial. Neuropsychopharmacology 37, 734–745. Hornberger, M., Wong, S., Tan, R., Irish, M., Piguet, O., Kril, J., Hodges, J.R., Halliday, G., 2012. In vivo and post-mortem memory circuit integrity in frontotemporal dementia and Alzheimer's disease. Brain 135, 3015–3025. Joe, A.Y., Tielmann, T., Bucerius, J., Reinhardt, M.J., Palmedo, H., Maier, W., Biersack, H.J., Zobel, A., 2006. Response-dependent differences in regional cerebral blood flow changes with citalopram in treatment of major depression. J. Nucl. Med. 47, 1319–1325. Kasper, S., Sacher, J., Klein, N., Mossaheb, N., Attarbaschi-Steiner, T., Lanzenberger, R., Spindelegger, C., Asenbaum, S., Holik, A., Dudczak, R., 2009. Differences in the dynamics of serotonin reuptake transporter occupancy may explain superior clinical efficacy of escitalopram versus citalopram. Int. Clin. Psychopharmacol. 24, 119–125. Klein, N., Sacher, J., Geiss-Granadia, T., Mossaheb, N., Attarbaschi, T., Lanzenberger, R., Spindelegger, C., Holik, A., Asenbaum, S., Dudczak, R., Tauscher, J., Kasper, S., 2007. Higher serotonin transporter occupancy after multiple dose administration of escitalopram compared to citalopram: an [123I]ADAM SPECT study. Psychopharmacology (Berl.) 191, 333–339. Lai, C.H., Wu, Y.T., 2011. Duloxetine's modest short-term influences in subcortical structures of first episode drug-naive patients with major depressive disorder and panic disorder. Psychiatry Res. 194, 157–162. Lesch, K.P., Waider, J., 2012. Serotonin in the modulation of neural plasticity and networks: implications for neurodevelopmental disorders. Neuron 76, 175–191. Mayberg, H.S., Brannan, S.K., Tekell, J.L., Silva, J.A., Mahurin, R.K., McGinnis, S., Jerabek, P.A., 2000. Regional metabolic effects of fluoxetine in major depression: serial changes and relationship to clinical response. Biol. Psychiatry 48, 830–843. McCabe, C., Mishor, Z., 2011. Antidepressant medications reduce subcortical–cortical resting-state functional connectivity in healthy volunteers. NeuroImage 57, 1317–1323. Mogha, A., Guariglia, S.R., Debata, P.R., Wen, G.Y., Banerjee, P., 2012. Serotonin 1A receptor-mediated signaling through ERK and PKCalpha is essential for normal synaptogenesis in neonatal mouse hippocampus. Transl. Psychiatry 2, e66. Newton, A.T., Morgan, V.L., Rogers, B.P., Gore, J.C., 2011. Modulation of steady state functional connectivity in the default mode and working memory networks by cognitive load. Hum. Brain Mapp. 32, 1649–1659. Pearson, J.M., Heilbronner, S.R., Barack, D.L., Hayden, B.Y., Platt, M.L., 2011. Posterior cingulate cortex: adapting behavior to a changing world. Trends Cogn. Sci. 15, 143–151. Rao, N., 2007. The clinical pharmacokinetics of escitalopram. Clin. Pharmacokinet. 46, 281–290. Reetz, K., Dogan, I., Rolfs, A., Binkofski, F., Schulz, J.B., Laird, A.R., Fox, P.T., Eickhoff, S.B., 2012. Investigating function and connectivity of morphometric findings—exemplified on cerebellar atrophy in spinocerebellar ataxia 17 (SCA17). NeuroImage 62, 1354–1366. Savli, M., Bauer, A., Mitterhauser, M., Ding, Y.S., Hahn, A., Kroll, T., Neumeister, A., Haeusler, D., Ungersboeck, J., Henry, S., Isfahani, S.A., Rattay, F., Wadsak, W., Kasper, S., Lanzenberger, R., 2012. Normative database of the serotonergic system in healthy subjects using multi-tracer PET. NeuroImage 63, 447–459. Seevinck, P.R., Deddens, L.H., Dijkhuizen, R.M., 2010. Magnetic resonance imaging of brain angiogenesis after stroke. Angiogenesis 13, 101–111. Shih, P., Keehn, B., Oram, J.K., Leyden, K.M., Keown, C.L., Muller, R.A., 2011. Functional differentiation of posterior superior temporal sulcus in autism: a functional connectivity magnetic resonance imaging study. Biol. Psychiatry 70, 270–277. Smith, R., Chen, K., Baxter, L., Fort, C., Lane, R.D., 2012. Antidepressant effects of sertraline associated with volume increases in dorsolateral prefrontal cortex. J. Affect. Disord. Stahl, S.M., 1998. Mechanism of action of serotonin selective reuptake inhibitors. Serotonin receptors and pathways mediate therapeutic effects and side effects. J. Affect. Disord. 51, 215–235. Tao, R., Ma, Z., Auerbach, S.B., 2000. Differential effect of local infusion of serotonin reuptake inhibitors in the raphe versus forebrain and the role of depolarizationinduced release in increased extracellular serotonin. J. Pharmacol. Exp. Ther. 294, 571–579. Thomas, C., Baker, C.I., 2013. Teaching an adult brain new tricks: a critical review of evidence for training-dependent structural plasticity in humans. NeuroImage 73, 225–236. Tomasi, D., Volkow, N.D., 2010. Functional connectivity density mapping. Proc. Natl. Acad. Sci. U. S. A. 107, 9885–9890. Tomasi, D., Volkow, N.D., 2011. Association between functional connectivity hubs and brain networks. Cereb. Cortex 1–11. van Marle, H.J., Tendolkar, I., Urner, M., Verkes, R.J., Fernandez, G., van Wingen, G., 2011. Subchronic duloxetine administration alters the extended amygdala circuitry in healthy individuals. NeuroImage 55, 825–831. van Tol, M.-J., van der Wee, N.J.A., van den Heuvel, O.A., Nielen, M.M.A., Demenescu, L.R., Aleman, A., Renken, R., van Buchem, M.A., Zitman, F.G., Veltman, D.J., 2010. Regional brain volume in depression and anxiety disorders. Arch. Gen. Psychiatry 67, 1002–1011. Vermetten, E., Vythilingam, M., Southwick, S.M., Charney, D.S., Bremner, J.D., 2003. Long-term treatment with paroxetine increases verbal declarative memory and hippocampal volume in posttraumatic stress disorder. Biol. Psychiatry 54, 693–702. Vernon, A.C., Natesan, S., Modo, M., Kapur, S., 2011. Effect of chronic antipsychotic treatment on brain structure: a serial magnetic resonance imaging study with ex vivo and postmortem confirmation. Biol. Psychiatry 69, 936–944.

244

C. Kraus et al. / NeuroImage 84 (2014) 236–244

Vitalis, T., Cases, O., Gillies, K., Hanoun, N., Hamon, M., Seif, I., Gaspar, P., Kind, P., Price, D.J., 2002. Interactions between TrkB signaling and serotonin excess in the developing murine somatosensory cortex: a role in tangential and radial organization of thalamocortical axons. J. Neurosci. 22, 4987–5000. Webhofer, C., Gormanns, P., Tolstikov, V., Zieglgansberger, W., Sillaber, I., Holsboer, F., Turck, C.W., 2011. Metabolite profiling of antidepressant drug action reveals novel drug targets beyond monoamine elevation. Transl. Psychiatry 1, e58. Weissenbacher, A., Kasess, C., Gerstl, F., Lanzenberger, R., Moser, E., Windischberger, C., 2009. Correlations and anticorrelations in resting-state functional connectivity MRI: a quantitative comparison of preprocessing strategies. NeuroImage 47, 1408–1416.

Windischberger, C., Lanzenberger, R., Holik, A., Spindelegger, C., Stein, P., Moser, U., Gerstl, F., Fink, M., Moser, E., Kasper, S., 2010. Area-specific modulation of neural activation comparing escitalopram and citalopram revealed by pharmaco-fMRI: a randomized cross-over study. NeuroImage 49, 1161–1170. Wonderlick, J.S., Ziegler, D.A., Hosseini-Varnamkhasti, P., Locascio, J.J., Bakkour, A., van der Kouwe, A., Triantafyllou, C., Corkin, S., Dickerson, B.C., 2009. Reliability of MRI-derived cortical and subcortical morphometric measures: effects of pulse sequence, voxel geometry, and parallel imaging. NeuroImage 44, 1324–1333. Zimmer, L., Riad, M., Rbah, L., Belkacem-Kahlouli, A., Le Bars, D., Renaud, B., Descarries, L., 2004. Toward brain imaging of serotonin 5-HT1A autoreceptor internalization. NeuroImage 22, 1421–1426.