Journal of Psychiatric Research xxx (2014) 1e10
Contents lists available at ScienceDirect
Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires
Review
Diffusion tensor imaging (DTI) studies in patients with obsessivecompulsive disorder (OCD): A review Kathrin Koch a, c, d, *, Tim J. Reeß a, c, e, O. Georgiana Rus a, c, e, Claus Zimmer a, Michael Zaudig b a
Department of Neuroradiology, Klinikum rechts der Isar, Technische Universität München, Ismaningerstrasse 22, 81675 Munich, Germany Windach Institute and Hospital of Neurobehavioural Research and Therapy (WINTR), Psychosomatic Hospital, Windach, Schützenstraße 100, 86949 Windach, Germany c TUM-Neuroimaging Center of Klinikum rechts der Isar (TUM-NIC), Technische Universität München TUM, Ismaninger Strasse 22, 81675 Munich, Germany d Department of Psychiatry and Psychotherapy, Jena University Hospital, Philosophenweg 3, 07743 Jena, Germany e Graduate School of Systemic Neurosciences GSN, Ludwig-Maximilians-Universität, Biocenter, Grobhaderner Strasse 2, 82152 Munich, Germany b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 October 2013 Received in revised form 18 February 2014 Accepted 10 March 2014
This review presents an overview of studies investigating white-matter integrity in patients with obsessive-compulsive disorder (OCD) using diffusion tensor imaging (DTI). There is increasing evidence for white matter alterations in OCD. In adult patients the majority of all studies reported abnormalities in terms of decreased fractional anisotropy (FA) compared to healthy volunteers. Although findings are heterogeneous, the cingulate bundle, the corpus callosum and the anterior limb of the internal capsule are most commonly affected by decreased white matter integrity in adult OCD patients. In pediatric and adolescent patients initial evidence points more towards increased white matter connectivity. Thus, current results suggest alterations in various white matter regions in both pediatric and adult OCD patients. They indicate that alterations may vary as a function of clinical characteristics and may be amenable to pharmacologic treatment. Although the findings have important implications for the neurobiology of OCD they also raise a number of important questions that are discussed in this review and need to be taken into consideration in future studies. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: DTI Obsessive-compulsive OCD White matter Review FA Corpus callosum Internal capsule
1. Objectives of the study and background 1.1. Clinical characteristics of OCD Obsessive-compulsive disorder (OCD) is characterized by recurrent, persistent, and intrusive thoughts or images that cause distress or anxiety (i.e., obsessions), and repetitive behaviors aimed at reducing this feeling of anxiety (i.e., compulsions). Patients may suffer from only obsessions, only compulsion or both obsessions and compulsions (Zaudig, 2011). Compulsions include excessive cleaning or washing, repeated checking, or extreme hoarding; obsessions comprise preoccupation with religious, sexual or violent thoughts, relationship-related obsessions or obsessions related to particular numbers. Importantly, obsessions and compulsions are
* Corresponding author. Department of Neuroradiology, Klinikum rechts der Isar, Technische Universität München, Ismaningerstrasse 22, 81675 Munich, Germany. Tel.: þ49 89 41407972; fax: þ49 89 41404887. E-mail address: [email protected] (K. Koch).
ego-dystonic, i.e., patients recognize that their thoughts are unreasonable and that they represent products of their own mind (APA, 2000). Clinically, patients differ considerably with regard to symptom type, symptom severity, comorbidity and age of onset (Weissman et al., 1994). As discussed below, this clinical heterogeneity may complicate the identification of general and specific brain regions and networks which are involved in the pathogenesis of the disorder. 1.2. Neurobiology of OCD The investigation of the neural circuitry involved in OCD has made substantial progress in the last couple of years. Functional neuroimaging methods have helped to identify altered activity within a predominantly fronto-striato-thalamo-cortical circuitry in patients with OCD (Menzies et al., 2008a). This circuitry originates in prefrontal cortex, has connections to the striatum which connects to the globus pallidus and substantia nigra and from there connects to the thalamus. The thalamus finally provides the link back to the frontal cortex. Amongst the frontal regions, the
http://dx.doi.org/10.1016/j.jpsychires.2014.03.006 0022-3956/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Koch K, et al., Diffusion tensor imaging (DTI) studies in patients with obsessive-compulsive disorder (OCD): A review, Journal of Psychiatric Research (2014), http://dx.doi.org/10.1016/j.jpsychires.2014.03.006
2
K. Koch et al. / Journal of Psychiatric Research xxx (2014) 1e10
dorsolateral prefrontal cortex (DLPFC), the orbitofrontal cortex (OFC) and the anterior cingulate cortex (ACC) seem to be of major psychopathological relevance in OCD (Chamberlain et al., 2008; Evans et al., 2004; Fineberg et al., 2010; Menzies et al., 2008a; O’Neill et al., 2013). According to a frequently discussed model by Saxena and Rauch (2000) excessive activity within frontal subcortical circuits might underlie the repetitive behavior and cognitive inflexibility which characterizes the disorder of OCD (Veale et al., 1996). Here, enhanced activity of the ACC may reflect the increased error-sensitivity and behavioral monitoring in patients with OCD. Thus, individuals with OCD may perform their repetitive, compulsive behaviors as they experience large and persistent error signals (i.e., also called “not just right experience”) that manifest emotionally as feelings of wrongness. Compulsive behaviors thus may represent attempts to reduce the perceived discrepancy between preferred and experienced internal states. However, this “not just right experience” remains relatively unaffected by behavior given continued ACC hyperactivity. As a result, individuals with OCD experience persistent feelings of doubts and anxiety and the urge to engage in behaviors to counteract these emotions. Results of structural studies investigating gray matter structure or volume are partly in line with these models. One meta-analysis identified increased regional gray matter in the basal ganglia and decreased gray matter volumes in the frontal/anterior cingulate gyri as core alterations in OCD (Radua and Mataix-Cols, 2009) while another, very recent meta-analysis, identified, apart from the “classical” fronto-striatal alterations, volume decreases in dorsomedial, ventrolateral and frontopolar prefrontal cortices and connected regions like temporo-parieto-occipital associative areas as well as volume increases in the internal capsule (Piras et al., 2013). Apart from alterations in gray matter volume and structure there is mounting evidence for alterations in white matter fiber tracts in OCD as demonstrated by an increasing number of studies using the method of diffusion tensor imaging (DTI).
single principal diffusion directions (PDD) or mean diffusivity (MD), i.e., the average diffusion along the three major directions calculated as the mean of the three eigenvalues also called apparent diffusion coefficient (ADC). In an attempt to optimize the assessment of white matter microstructure a number of different analysis techniques have been developed. The most frequently used approach is the voxel-wise or voxel-based analysis of diffusion data. With this approach, diffusion data (e.g., FA maps) are coregistered to a high resolution image, spatially normalized and smoothed before statistical tests are performed for each voxel. Consequently, its results are strongly dependent on several parameter settings like, e.g., method of coregistration or size of the smoothing filter (Jones et al., 2005). Another, more recent and increasingly popular method is a tract-based approach as, for instance, implemented in FSL (i.e., tract-based spatial statistics, TBSS) (Smith et al., 2006). Here, individual subjects’ FA data are projected into a common space in a way that does not require perfect nonlinear registration. Strictly speaking, this method is likewise voxel-based, but as only regions containing white matter tracts of the brain are analyzed, the term tract-based analysis prevailed. As no spatial smoothing is necessary this method overcomes a number of drawbacks going along with classical voxelbased methods. Today, most DTI studies in patients with OCD applied voxel- or tract-based methods to investigate potential alterations in white matter microstructure in OCD. The current review intended to provide an overview of DTI studies in patients with OCD. Adolescent and adult patients differ in clinical parameters (such as duration of illness or long-term pharmacological treatment) that are known to have a relevant effect on white matter structure. They also vary with regard to the developmental stage of white matter or myelination which is still developing in adolescent patients. Therefore, results are presented separately for adolescents and adults to highlight potential differences in structural alterations between different patient groups or disease stages.
1.3. The method of DTI to investigate alterations in white matter structure
2. Materials and methods
DTI is a comparatively young imaging method that permits the quantification of the diffusion characteristics of water molecules in vivo. Within cerebral white matter, water molecules diffuse more freely along myelinated tracts than across them. Such directional dependence of diffusivity is called “anisotropy,” and any reduction in white matter anisotropy indicates an alteration in the degree of tissue order or integrity. The reduction may reflect changes to the underlying white matter tracts, including changes in fibers packing, fiber diameter, thickness of the myelin sheaths, and directionality of the fibers. Technically, diffusion is encoded in the MRI signal by magnetic field gradient pulses. Because diffusion can only be detected in the direction of the particular gradient the number of gradient or diffusion encoding directions is a critical parameter in DTI that strongly determines signal strength. There are different parameters that can be assessed to characterize diffusion or, indirectly, fiber tract characteristics. The most commonly used parameter is fractional anisotropy (FA) which describes the fraction of the magnitude of the tensor that can be ascribed to the anisotropic diffusion and is calculated from the three eigenvalues of the diffusion tensor. Axial diffusivity (AD), also called longitudinal or parallel diffusivity, describes the diffusivity along the principal axis whereas radial diffusivity (RD), usually calculated by averaging the diffusivities in the two minor axes, describes the diffusivity perpendicular to the structure. Studies in shiverer mice have shown that AD may indicate axonal loss, while increased RD may reflect a decrease in myelination (Song et al., 2002). In addition, some studies investigated diffusivity along the
We used PubMed to identify original, English-language, studies that investigated white-matter integrity in patients with OCD applying the method of DTI. Studies were retrieved by using the search terms “obsessive-compulsive” and ”diffusion tensor imaging” or “DTI”. We included only studies involving patients with a DSM-IV diagnosis of OCD (i.e., not included were den Braber et al., 2011; Kochunov et al., 2010; Makki et al., 2009; Neuner et al., 2010; Simmons et al., 2012), and excluded studies based on patients only (Koch et al., 2012). Moreover, we excluded ROI analyses (Cannistraro et al., 2007; Chiu et al., 2011; Lochner et al., 2012; Lopez et al., 2013; Oh et al., 2012; Saito et al., 2008) as they investigate specific regions only and would distort the overall picture. The majority of all DTI studies investigated white matter alterations in adult OCD patients using voxel-based methods as described above (although some of them differed in a number of methodological aspects such as, e.g., method of coregistration). Most of these studies included both male and female patients with patients’ mean age ranging between 23 and 39 years and their mean duration of illness ranging between 5 and 19 years (Tables 1 and 2). Up to now, there are only four DTI studies on pediatric or adolescent OCD samples. In these studies patients’ mean age varied between 13 and 17 years and their mean duration of illness ranged between 1 and 5 years (Tables 1and 2). Most of all adolescent and adult patients were medicated with selective serotonin reuptake inhibitors (SSRI). The majority of all studies used voxel-based analysis methods with FA as the main outcome parameter (Table 3).
Please cite this article in press as: Koch K, et al., Diffusion tensor imaging (DTI) studies in patients with obsessive-compulsive disorder (OCD): A review, Journal of Psychiatric Research (2014), http://dx.doi.org/10.1016/j.jpsychires.2014.03.006
K. Koch et al. / Journal of Psychiatric Research xxx (2014) 1e10
3
Table 1 Sample size, age, gender and education (mean and SD) in patients (OCD) and healthy volunteers (HC). Abbreviations: N/A ¼ not available. Author
Sample size
Age
OCD
HC
OCD
Admon et al. (2012) Benedetti et al. (2013) Fan et al. (2012) Fontenelle et al. (2011) Bora et al. (2011) Li et al. (2011) Nakamae et al. (2011) Garibotto et al. (2010) Ha et al. (2009) Menzies et al. (2008)
13 40 27 9 21 23 30 15 25 30
25.5 33.8 25.5 26.2 34.4 27.2 31.6 31.9 23.3 32.2
Nakamae et al. (2008) Yoo et al. (2007) Szeszko et al. (2005)
15 13 15
13 41 23 9 29 23 30 16 25 60 (30 first-degree relatives, 30 unrelated volunteers) 15 13 15
Gruner et al. (2012) Jayarajan et al. (2012) Silk et al. (2013) Zarei et al. (2011)
23 15 16 26
23 15 22 26
Gender HC
OCD
Studies with adult patients (1.0) 27 (0.5) 10 m, 3 f (10.7) 33.8 (13.2) 26 m, 14 f (7.0) 28.8 (7.6) 17 m, 10 f (10.4) 28.0 (10.2) 7 m, 2 f (10.6) 31.4 (8.0) 11 m, 10 f (9.9) 26.7 (9.1) 16 m, 7 f (9.3) 30.8 (8.4) 14 m, 16 f (7.9) 29.7 (6.3) 15 m (5.3) 23.7 (3.9) 25 m 9 m, 21 f (11.1) First-degree relatives: 37.2 (13.3), unrelated volunteers: 33.7 (11.2) 29.7 (6.9) 29.1 (6.0) 9 m, 6 f 27.8 (7.3) 26.9 (7.0) 8 m, 5 f 38.5 (10.9) 38.5 (11.8) 10 m, 5 f Studies with pediatric and adolescent patients 14.3 (2.1) 14.2 (2.2) 13 m, 10 f 14.1 (1.8) 14.3 (2.2) 8 m, 7 f 12.8 (2.8) 11.2 (2.1) 6 m, 10 f 16.6 (1.5) 16.5 (1.4) 14 m, 12 f
3. Results 3.1. Studies reporting decreased white matter connectivity in adult OCD patients So far, most studies investigating adult OCD patients reported decreased FA, mainly in the corpus callosum (CC) and the cingulum (i.e., cingulum bundle). Using tract-based analyses Nakamae et al. (2011) detected decreased FA in the right cingulum, the anterior body of the corpus callosum (CC) and the left anterior limb of the internal capsule (alic) in a sample of unmedicated patients with OCD. Decreased FA in patients in the cingulum was also reported by Szeszko et al. (2005) at a rather liberal threshold, however, Szeszko et al. found decreased FA in patients compared to healthy volunteers in the left and right parietal lobe (i.e., supramarginal gyri) and the left occipital cortex (i.e., lingual gyrus). White matter alterations in the parietal lobe correlated with Yale-Brown Obsessive Compulsive Scale (Y-BOCS) total scores and scores on the obsession subscale. Like Nakamae and colleagues Bora et al. (2011) found decreased FA in the corpus callosum that was underpinned by increased radial diffusivity in that area. As opposed to Nakamae et al., however, in the study by Bora et al. white matter alterations were located in the body of the corpus callosum with no difference between medicated (n ¼ 10) and unmedicated (n ¼ 11) patients. Decreased FA in the corpus callosum was also reported by Garibotto et al. (2010) who investigated a rather homogeneous patient sample containing, according to symptom typology, either checkers (n ¼ 12) or washers (n ¼ 3). They performed voxel-based analyses at a rather liberal threshold and found, apart from decreased FA in the corpus callosum (splenium) and the cingulum, alterations in the superior longitudinal fasciculus, the optic radiation, and the inferior fronto-occipital fasciculus that were significantly correlated with total Y-BOCS scores (i.e., the stronger the FA decreases, the higher the total Y-BOCS score). In addition, they investigated diffusivity along the single principal diffusion directions (PDD) and found PDD differences between patients and healthy volunteers in white matter regions similar to those regions showing altered FA in patients. Fontenelle et al. (2011) examined a sample of chronic OCD patients (i.e., adult patients with an average age at onset of 12 years) with a high degree of symptomatology using tract-based
Education HC
OCD
HC
10 m, 3 f 20 m, 21 f 15 m, 8 f 7 m, 2 f 14 m, 15 f 16 m, 7 f 15 m, 15 f 16 m 25 m First-degree relatives: 9 m, 21 f, unrelated volunteers: 10 m, 20 f 9 m, 6 f 8 m, 5 f 10 m, 5 f
13.5 13.1 14.0 N/A 14.6 13.3 N/A N/A 13.1 N/A
15.2 (2.2) 15.3 (1.8) 14.8 (1.8)
15.7 (1.8) 15.1 (2.9) 15.9 (1.8)
12 m, 11 f 8 m, 7 f 16, 6 14 m, 12 f
N/A 8.1 (1.8) N/A N/A
N/A 8.9 (2.3) N/A N/A
(1) (3.2) (2.9) (2.2) (2.8)
(2.5)
14 (0.5) 15.8 (2.7) 14.6 (3.7) N/A 15.0 (2.3) 13.4 (2.7) N/A N/A 14.0 (1.2) N/A
analyses. Decreased FA in association with increased MD was detected in the corticospinal tract, the internal capsule, the posterior limb of the internal capsule, the corona radiata and the superior longitudinal fasciculus. Regarding clinical characteristics, there was a negative correlation between corpus callosum (genu) FA and total Y-BOCS scores as well as a positive correlation between MD in the corpus callosum (genu), the bilateral anterior limb of internal capsule, the bilateral superior longitudinal fasciculus and total YBOCS scores. In a combined fMRI e DTI study Admon et al. (2012) investigated altered functional and structural connectivity in a sample of 13 patients with OCD and 13 healthy volunteers. On the structural level they found altered white matter structure in terms of decreased FA in clusters within the uncinate and the cingulum bundle in patients applying a voxel-based analysis. These structural alterations correlated with symptom severity (i.e., lower FA was associated with higher total Y-BOCS scores) and with functional connectivity between amygdala and dorsal anterior cingulate cortex (dACC) as well as between nucleus accumbens (Nacc) and OFC (i.e., lower FA was related to lower functional connectivity). In another recent study by Benedetti et al. (2013) tract-based analyses were used to examine alterations in FA, mean diffusivity (MD), radial diffusivity (RD) and axial diffusivity (AD) in a large sample of drug-treated and drug-naïve patients without any comorbidities. Drug-treated patients showed widespread FA decreases and RD increases compared to drug-naïve patients and healthy controls. Benedetti and colleagues concluded from their findings that changes in white matter microstructure could also be a consequence of drug treatment (i.e., SSRI treatment) and might constitute part of the drugs’ mechanism of action. Like in the before mentioned study by Nakamae et al. (2011) one of the structures featuring significant FA decreases in patients was the anterior limb of the internal capsule (alic). The alic contains fibers that connect the cortex with the corpus striatum and receives fibers from the anterior cingulate, the orbitofrontal and medial frontal cortex (Schmahmann and Pandya, 2006). As the fronto-striatal structures are considered to be critically involved in the pathogenesis of OCD the alic constitutes a fiber structure which is not only gaining increasing attention in the field of OCD research but which has also evolved into one of the most common targets of deep brain stimulation in OCD (Bourne et al., 2012).
Please cite this article in press as: Koch K, et al., Diffusion tensor imaging (DTI) studies in patients with obsessive-compulsive disorder (OCD): A review, Journal of Psychiatric Research (2014), http://dx.doi.org/10.1016/j.jpsychires.2014.03.006
4
K. Koch et al. / Journal of Psychiatric Research xxx (2014) 1e10
Table 2 Age of onset, duration of illness, total Y-BOCS score (mean and SD), percentage of medicated patients, and comorbidities. Abbreviations: AD ¼ anxiety disorder, ADHD ¼ attention deficit hyperactivity disorder, BED ¼ binge eating disorder, GAD ¼ generalized anxiety disorder, MDD ¼ major depressive disorder, ODD ¼ oppositional defiant disorder, SSRI ¼ selective serotonin reuptake inhibitor, N/A ¼ not available. Author
Age of onset
Duration of illness (yrs)
Total Y-BOCS score
Admon et al. (2012) Benedetti et al. (2012) Fan et al. (2012)
N/A 14.1 (7.5) 20.4 (7.9)
8 (1) 19.1 (9.9) 4.8 (4.0)
Fontenelle et al. (2011)
11.5 (5.0)
N/A
28.5 (4.8)
Bora et al. (2011) Li et al. (2011) Nakamae et al. (2011)
N/A N/A 25.0 (9.6)
N/A 6.3 (5.4) 6.7 (7.1)
19.2 (5.4) 23.2 (5.1) 23.8 (5.7)
Garibotto et al. (2010) Ha et al. (2009) Menzies et al. (2008)
20.7 (10.9) N/A N/A
10.1 (8.1) N/A N/A
Nakamae et al. (2008) Yoo et al. (2007)
19.6 (8.0) N/A
N/A 7.2 (9.0)
Szeszko et al. (2005)
16.9 (7.7)
N/A
28.2 (5.0) 20.2 (5.4) 22.1 (5.4), firstdegree relatives: 1.4 (3.0) 29.0 (5.3) Pretreatment: 30.2 (4.6) Posttreatment: 16.6 (6.4) 25.9 (4.4)
Author
Age of onset
Gruner et al. (2012) Jayarajan et al. (2012)
N/A 12.7 (1.9)
Studies with pediatric and adolescent patients Duration of illness Total C-YBOCS % Medicated score patients N/A 26.9 (4.5) 52 1.4 (1.0) 21.5 (7.4) 87
Silk et al. (2013)
N/A
N/A
N/A
Zarei et al. (2011)
11.2 (2.8)
5.3 (3.4)
19.5 (7.6)
Studies with adult patients 24.5 (1.5) 31.2 (5.6) 22.0 (4.9)
3.2. Studies reporting increased white matter connectivity in adult OCD patients While the majority of all DTI studies demonstrated FA decreases, there are also a number of studies that found increased FA in adult patients with OCD. Li et al. (2011) performed a voxel-based analysis of FA, AD and RD in a sample of partly medicated patients and identified significantly increased FA in the corpus callosum (genu and body) as well as the white matter of the right superior frontal gyrus in patients compared to healthy volunteers. In addition, they found a positive correlation between FA in left middle temporal gyrus and Y-BOCS total as well as obsessive subscale scores. In an earlier study by Nakamae et al. (2008) patients were found to have increased FA in the bilateral semioval center extending to the subinsular white matter as well as higher apparent diffusion coefficient (ADC) values in the left medial frontal cortex. In a study by Yoo et al. (2007) a group of unmedicated patients underwent 12 weeks of pharmacotherapy with citalopram. At baseline, i.e., in the unmedicated state, patients exhibited increased FA in the corpus callosum, the internal capsule and white matter in the area superolateral to the right caudate. After 12 weeks of pharmacotherapy, these FA increases had largely disappeared. Although care should be taken when interpreting these findings due to some methodological shortcomings such as, for instance, the limited
% Medicated patients
current comorbidity
100 55 0, n ¼ 15 scanned unmedicated and after 12 weeks of SSRI therapy 78
Personality disorder (n ¼ 3) e e
48 57 0 (n ¼ 14 drug naïve) 87 40 70
100 Pretreatment: 0 / 12 week pharmacotherapy with citalopram 80
13 (withdrawn from medication for 24 h before scanning) 62
MDD (n ¼ 2), Tourette syndrome (n ¼ 2), panic disorder (n ¼ 1), GAD (n ¼ 1), BED (n ¼ 1) e N/A e e e Panic disorder (n ¼ 2)
MDD (n ¼ 1), dysthymic disorder (n ¼ 1) MDD (n ¼ 2), dysthymic disorder (n ¼ 1)
MDD (n ¼ 5), panic disorder (n ¼ 1) social phobia (n ¼ 2), eating disorder (n ¼ 1) Current comorbidity MDD (n ¼ 4), AD (n ¼ 4), ADHD (n ¼ 5) Specific phobia (n ¼ 2), MDD (n ¼ 1), social phobia (n ¼ 1), ODD (n ¼ 1), separation anxiety disorder (n ¼ 1) GAD (n ¼ 3)
e
number of diffusion encoding directions, this study is one of the first interventional studies suggesting that white matter alterations in OCD core regions like the striatum or the internal capsule may be partly reversible with pharmacological treatment. Findings by Fan et al. (2012) point into a similar direction. In their study, Fan and colleagues examined FA, AD, MD and RD parameters in a sample of 27 unmedicated OCD patients and 23 healthy controls. In addition, they compared white matter alterations of 15 patients before and after 12 weeks of SSRI treatment. They discovered that patients were characterized by an increased RD in OCD core regions like the striatum, insula and right midbrain, as well as in left medial superior frontal gyrus, temporo-parietal lobe and the occipital lobe. After 12 weeks of SSRI treatment, OCD patients showed a reduction in RD in the left striatum and right midbrain, as well as in MD of the right midbrain. Despite the differences (i.e., Yoo and colleagues identified increased FA in patients whereas Fan and colleagues found increased RD in patients before treatment) both studies illustrate that some alterations in white matter microstructure may be amenable to treatment. 3.3. Studies reporting decreased and increased white matter connectivity in adult OCD patients Apart from studies demonstrating decreased white matter connectivity and reports on increased connectivity, there are also
Please cite this article in press as: Koch K, et al., Diffusion tensor imaging (DTI) studies in patients with obsessive-compulsive disorder (OCD): A review, Journal of Psychiatric Research (2014), http://dx.doi.org/10.1016/j.jpsychires.2014.03.006
K. Koch et al. / Journal of Psychiatric Research xxx (2014) 1e10
5
Table 3 Imaging and analysis method (i.e., MRI field strength, analysis method, statistical threshold), number of diffusion encoding directions, general findings and findings with regard to clinical characteristics. Abbreviations: CBCL-OCS ¼ Child Behavior Checklist e Obsessive-Compulsive Scale, CBCL-AD ¼ Child Behavior Checklist-Anxiety and Depression, acr ¼ anterior corona radiate, alic ¼ anterior limb of internal capsule, atr ¼ anterior thalamic radiation, cc ¼ corpus callosum, cr ¼ corona radiata, cst ¼ corticospinal tract, ic ¼ internal capsule, icp ¼ inferior cerebellar peduncle, ifo ¼ inferior fronto-occipital fasciculus, ilf ¼ inferior longitudinal fasciculus, mcp ¼ middle cerebellar peduncle, mfg ¼ medial frontal gyrus, plic ¼ posterior limb of internal capsule, ptr ¼ posterior thalamic radiation, scp ¼ superior cerebellar peduncle, scr ¼ superior corona radiata, sfg ¼ superior frontal gyrus, slf ¼ superior longitudinal fasciculus, unc ¼ uncinate fasciculus. Author
Method (i.e., field strength, analysis, statistical threshold)
Diffusion encoding directions
Admon et al. (2012)
1.5 T, voxel-based analyses, p < 0.005 corrected, cluster extent 10 voxels 3 T, tract-based analyses (TBSS), p < 0.05 TFCE corrected, cluster extent 100 voxels
6
Benedetti et al. (2013)
35
Measurement
Findings
Studies with adult patients FA Decreased FA in r. unc. and l. cingulum Positive correlation between cingulum/ uncinate FA and functional connectivity between amygdalaedACC/NacceOFC FA, MD, RD, AD Decreased FA in r. cingulum, r. ptr, r. ilf, r. slf, r. alic, l. cc, l. scr, l.þr. ifo, l.þr. unc., l.þr. atr, l.þr. acr. Increased MD in l.þr. ifo, l.þr. forceps minor, l.þr. slf, l. atr, l. unc., l. external capsule, r. ilf. Increased RD in l. slf, l. cingulum, l. forceps minor, l. scr, l.þr. ifo, l.þr. atr. Effects of drug status (see text) FA, AD, RD and MD Increased RD in l. medial sfg, l. temporoparietal lobe, l. occipital lobe, l. striatum, insula and r. midbrain. Decreased RD in l. striatum and r. midbrain after medication. Decreased MD in r. midbrain after medication. FA, MD Decreased FA and increased MD in cst, genu of ic, plic, cr, slf
Fan et al. (2012)
1.5 T, voxel-based analyses, p < 0.001 uncorrected, cluster extent 10 voxels
25
Fontenelle et al. (2011)
1.5 T, tract-based analyses (TBSS), p < 0.05 corrected
6
Bora et al. (2011)
3 T, tract-based analyses (TBSS), p < 0.05 cluster level corrected (t ¼ 2) 3 T, voxel-based analyses, p < 0.05 voxel level corrected, cluster extent 50 voxels 1.5 T, tract-based analyses (TBSS), p < 0.05 cluster level corrected (t ¼ 3 and t ¼ 1.5) 1.5 T, voxel-based analyses, p < 0.005 uncorrected, cluster extent 20 voxels
28
FA, AD, RD
Decreased FA in body of CC
15
FA, AD, RD
Increased FA in genu and body of CC and r. sfg
15
FA
Decreased FA in anterior body of CC (at t ¼ 3). Decreased FA in CC, r. cingulum, and l. alic (at t ¼ 1.5)
35
FA, PDD
Decreased FA in splenium of CC, cingulum, l.þr. slf, l.þr. optic radiation and l.þr. ifo. PDD differences in cc, l.r. slf, l.þr. optic radiation, l.þr. ifo, l.þr. ilf, and l.þr. cerebello-cortical/cortical-cerebellar fiber projections Decreased FA in l. ACC.
Li et al. (2011)
Nakamae et al. (2011)
Garibotto et al. (2010)
Ha et al. (2009)
1.5 T, voxel-based analyses, p < 0.001 uncorrected (p < 0.0001 uncorrected for subgroup analyses)
12
FA
Menzies et al. (2008)
1.5 T, voxel-based analyses, nonparametric testing, p < 0.017 (
Contents lists available at ScienceDirect
Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires
Review
Diffusion tensor imaging (DTI) studies in patients with obsessivecompulsive disorder (OCD): A review Kathrin Koch a, c, d, *, Tim J. Reeß a, c, e, O. Georgiana Rus a, c, e, Claus Zimmer a, Michael Zaudig b a
Department of Neuroradiology, Klinikum rechts der Isar, Technische Universität München, Ismaningerstrasse 22, 81675 Munich, Germany Windach Institute and Hospital of Neurobehavioural Research and Therapy (WINTR), Psychosomatic Hospital, Windach, Schützenstraße 100, 86949 Windach, Germany c TUM-Neuroimaging Center of Klinikum rechts der Isar (TUM-NIC), Technische Universität München TUM, Ismaninger Strasse 22, 81675 Munich, Germany d Department of Psychiatry and Psychotherapy, Jena University Hospital, Philosophenweg 3, 07743 Jena, Germany e Graduate School of Systemic Neurosciences GSN, Ludwig-Maximilians-Universität, Biocenter, Grobhaderner Strasse 2, 82152 Munich, Germany b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 October 2013 Received in revised form 18 February 2014 Accepted 10 March 2014
This review presents an overview of studies investigating white-matter integrity in patients with obsessive-compulsive disorder (OCD) using diffusion tensor imaging (DTI). There is increasing evidence for white matter alterations in OCD. In adult patients the majority of all studies reported abnormalities in terms of decreased fractional anisotropy (FA) compared to healthy volunteers. Although findings are heterogeneous, the cingulate bundle, the corpus callosum and the anterior limb of the internal capsule are most commonly affected by decreased white matter integrity in adult OCD patients. In pediatric and adolescent patients initial evidence points more towards increased white matter connectivity. Thus, current results suggest alterations in various white matter regions in both pediatric and adult OCD patients. They indicate that alterations may vary as a function of clinical characteristics and may be amenable to pharmacologic treatment. Although the findings have important implications for the neurobiology of OCD they also raise a number of important questions that are discussed in this review and need to be taken into consideration in future studies. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: DTI Obsessive-compulsive OCD White matter Review FA Corpus callosum Internal capsule
1. Objectives of the study and background 1.1. Clinical characteristics of OCD Obsessive-compulsive disorder (OCD) is characterized by recurrent, persistent, and intrusive thoughts or images that cause distress or anxiety (i.e., obsessions), and repetitive behaviors aimed at reducing this feeling of anxiety (i.e., compulsions). Patients may suffer from only obsessions, only compulsion or both obsessions and compulsions (Zaudig, 2011). Compulsions include excessive cleaning or washing, repeated checking, or extreme hoarding; obsessions comprise preoccupation with religious, sexual or violent thoughts, relationship-related obsessions or obsessions related to particular numbers. Importantly, obsessions and compulsions are
* Corresponding author. Department of Neuroradiology, Klinikum rechts der Isar, Technische Universität München, Ismaningerstrasse 22, 81675 Munich, Germany. Tel.: þ49 89 41407972; fax: þ49 89 41404887. E-mail address: [email protected] (K. Koch).
ego-dystonic, i.e., patients recognize that their thoughts are unreasonable and that they represent products of their own mind (APA, 2000). Clinically, patients differ considerably with regard to symptom type, symptom severity, comorbidity and age of onset (Weissman et al., 1994). As discussed below, this clinical heterogeneity may complicate the identification of general and specific brain regions and networks which are involved in the pathogenesis of the disorder. 1.2. Neurobiology of OCD The investigation of the neural circuitry involved in OCD has made substantial progress in the last couple of years. Functional neuroimaging methods have helped to identify altered activity within a predominantly fronto-striato-thalamo-cortical circuitry in patients with OCD (Menzies et al., 2008a). This circuitry originates in prefrontal cortex, has connections to the striatum which connects to the globus pallidus and substantia nigra and from there connects to the thalamus. The thalamus finally provides the link back to the frontal cortex. Amongst the frontal regions, the
http://dx.doi.org/10.1016/j.jpsychires.2014.03.006 0022-3956/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Koch K, et al., Diffusion tensor imaging (DTI) studies in patients with obsessive-compulsive disorder (OCD): A review, Journal of Psychiatric Research (2014), http://dx.doi.org/10.1016/j.jpsychires.2014.03.006
2
K. Koch et al. / Journal of Psychiatric Research xxx (2014) 1e10
dorsolateral prefrontal cortex (DLPFC), the orbitofrontal cortex (OFC) and the anterior cingulate cortex (ACC) seem to be of major psychopathological relevance in OCD (Chamberlain et al., 2008; Evans et al., 2004; Fineberg et al., 2010; Menzies et al., 2008a; O’Neill et al., 2013). According to a frequently discussed model by Saxena and Rauch (2000) excessive activity within frontal subcortical circuits might underlie the repetitive behavior and cognitive inflexibility which characterizes the disorder of OCD (Veale et al., 1996). Here, enhanced activity of the ACC may reflect the increased error-sensitivity and behavioral monitoring in patients with OCD. Thus, individuals with OCD may perform their repetitive, compulsive behaviors as they experience large and persistent error signals (i.e., also called “not just right experience”) that manifest emotionally as feelings of wrongness. Compulsive behaviors thus may represent attempts to reduce the perceived discrepancy between preferred and experienced internal states. However, this “not just right experience” remains relatively unaffected by behavior given continued ACC hyperactivity. As a result, individuals with OCD experience persistent feelings of doubts and anxiety and the urge to engage in behaviors to counteract these emotions. Results of structural studies investigating gray matter structure or volume are partly in line with these models. One meta-analysis identified increased regional gray matter in the basal ganglia and decreased gray matter volumes in the frontal/anterior cingulate gyri as core alterations in OCD (Radua and Mataix-Cols, 2009) while another, very recent meta-analysis, identified, apart from the “classical” fronto-striatal alterations, volume decreases in dorsomedial, ventrolateral and frontopolar prefrontal cortices and connected regions like temporo-parieto-occipital associative areas as well as volume increases in the internal capsule (Piras et al., 2013). Apart from alterations in gray matter volume and structure there is mounting evidence for alterations in white matter fiber tracts in OCD as demonstrated by an increasing number of studies using the method of diffusion tensor imaging (DTI).
single principal diffusion directions (PDD) or mean diffusivity (MD), i.e., the average diffusion along the three major directions calculated as the mean of the three eigenvalues also called apparent diffusion coefficient (ADC). In an attempt to optimize the assessment of white matter microstructure a number of different analysis techniques have been developed. The most frequently used approach is the voxel-wise or voxel-based analysis of diffusion data. With this approach, diffusion data (e.g., FA maps) are coregistered to a high resolution image, spatially normalized and smoothed before statistical tests are performed for each voxel. Consequently, its results are strongly dependent on several parameter settings like, e.g., method of coregistration or size of the smoothing filter (Jones et al., 2005). Another, more recent and increasingly popular method is a tract-based approach as, for instance, implemented in FSL (i.e., tract-based spatial statistics, TBSS) (Smith et al., 2006). Here, individual subjects’ FA data are projected into a common space in a way that does not require perfect nonlinear registration. Strictly speaking, this method is likewise voxel-based, but as only regions containing white matter tracts of the brain are analyzed, the term tract-based analysis prevailed. As no spatial smoothing is necessary this method overcomes a number of drawbacks going along with classical voxelbased methods. Today, most DTI studies in patients with OCD applied voxel- or tract-based methods to investigate potential alterations in white matter microstructure in OCD. The current review intended to provide an overview of DTI studies in patients with OCD. Adolescent and adult patients differ in clinical parameters (such as duration of illness or long-term pharmacological treatment) that are known to have a relevant effect on white matter structure. They also vary with regard to the developmental stage of white matter or myelination which is still developing in adolescent patients. Therefore, results are presented separately for adolescents and adults to highlight potential differences in structural alterations between different patient groups or disease stages.
1.3. The method of DTI to investigate alterations in white matter structure
2. Materials and methods
DTI is a comparatively young imaging method that permits the quantification of the diffusion characteristics of water molecules in vivo. Within cerebral white matter, water molecules diffuse more freely along myelinated tracts than across them. Such directional dependence of diffusivity is called “anisotropy,” and any reduction in white matter anisotropy indicates an alteration in the degree of tissue order or integrity. The reduction may reflect changes to the underlying white matter tracts, including changes in fibers packing, fiber diameter, thickness of the myelin sheaths, and directionality of the fibers. Technically, diffusion is encoded in the MRI signal by magnetic field gradient pulses. Because diffusion can only be detected in the direction of the particular gradient the number of gradient or diffusion encoding directions is a critical parameter in DTI that strongly determines signal strength. There are different parameters that can be assessed to characterize diffusion or, indirectly, fiber tract characteristics. The most commonly used parameter is fractional anisotropy (FA) which describes the fraction of the magnitude of the tensor that can be ascribed to the anisotropic diffusion and is calculated from the three eigenvalues of the diffusion tensor. Axial diffusivity (AD), also called longitudinal or parallel diffusivity, describes the diffusivity along the principal axis whereas radial diffusivity (RD), usually calculated by averaging the diffusivities in the two minor axes, describes the diffusivity perpendicular to the structure. Studies in shiverer mice have shown that AD may indicate axonal loss, while increased RD may reflect a decrease in myelination (Song et al., 2002). In addition, some studies investigated diffusivity along the
We used PubMed to identify original, English-language, studies that investigated white-matter integrity in patients with OCD applying the method of DTI. Studies were retrieved by using the search terms “obsessive-compulsive” and ”diffusion tensor imaging” or “DTI”. We included only studies involving patients with a DSM-IV diagnosis of OCD (i.e., not included were den Braber et al., 2011; Kochunov et al., 2010; Makki et al., 2009; Neuner et al., 2010; Simmons et al., 2012), and excluded studies based on patients only (Koch et al., 2012). Moreover, we excluded ROI analyses (Cannistraro et al., 2007; Chiu et al., 2011; Lochner et al., 2012; Lopez et al., 2013; Oh et al., 2012; Saito et al., 2008) as they investigate specific regions only and would distort the overall picture. The majority of all DTI studies investigated white matter alterations in adult OCD patients using voxel-based methods as described above (although some of them differed in a number of methodological aspects such as, e.g., method of coregistration). Most of these studies included both male and female patients with patients’ mean age ranging between 23 and 39 years and their mean duration of illness ranging between 5 and 19 years (Tables 1 and 2). Up to now, there are only four DTI studies on pediatric or adolescent OCD samples. In these studies patients’ mean age varied between 13 and 17 years and their mean duration of illness ranged between 1 and 5 years (Tables 1and 2). Most of all adolescent and adult patients were medicated with selective serotonin reuptake inhibitors (SSRI). The majority of all studies used voxel-based analysis methods with FA as the main outcome parameter (Table 3).
Please cite this article in press as: Koch K, et al., Diffusion tensor imaging (DTI) studies in patients with obsessive-compulsive disorder (OCD): A review, Journal of Psychiatric Research (2014), http://dx.doi.org/10.1016/j.jpsychires.2014.03.006
K. Koch et al. / Journal of Psychiatric Research xxx (2014) 1e10
3
Table 1 Sample size, age, gender and education (mean and SD) in patients (OCD) and healthy volunteers (HC). Abbreviations: N/A ¼ not available. Author
Sample size
Age
OCD
HC
OCD
Admon et al. (2012) Benedetti et al. (2013) Fan et al. (2012) Fontenelle et al. (2011) Bora et al. (2011) Li et al. (2011) Nakamae et al. (2011) Garibotto et al. (2010) Ha et al. (2009) Menzies et al. (2008)
13 40 27 9 21 23 30 15 25 30
25.5 33.8 25.5 26.2 34.4 27.2 31.6 31.9 23.3 32.2
Nakamae et al. (2008) Yoo et al. (2007) Szeszko et al. (2005)
15 13 15
13 41 23 9 29 23 30 16 25 60 (30 first-degree relatives, 30 unrelated volunteers) 15 13 15
Gruner et al. (2012) Jayarajan et al. (2012) Silk et al. (2013) Zarei et al. (2011)
23 15 16 26
23 15 22 26
Gender HC
OCD
Studies with adult patients (1.0) 27 (0.5) 10 m, 3 f (10.7) 33.8 (13.2) 26 m, 14 f (7.0) 28.8 (7.6) 17 m, 10 f (10.4) 28.0 (10.2) 7 m, 2 f (10.6) 31.4 (8.0) 11 m, 10 f (9.9) 26.7 (9.1) 16 m, 7 f (9.3) 30.8 (8.4) 14 m, 16 f (7.9) 29.7 (6.3) 15 m (5.3) 23.7 (3.9) 25 m 9 m, 21 f (11.1) First-degree relatives: 37.2 (13.3), unrelated volunteers: 33.7 (11.2) 29.7 (6.9) 29.1 (6.0) 9 m, 6 f 27.8 (7.3) 26.9 (7.0) 8 m, 5 f 38.5 (10.9) 38.5 (11.8) 10 m, 5 f Studies with pediatric and adolescent patients 14.3 (2.1) 14.2 (2.2) 13 m, 10 f 14.1 (1.8) 14.3 (2.2) 8 m, 7 f 12.8 (2.8) 11.2 (2.1) 6 m, 10 f 16.6 (1.5) 16.5 (1.4) 14 m, 12 f
3. Results 3.1. Studies reporting decreased white matter connectivity in adult OCD patients So far, most studies investigating adult OCD patients reported decreased FA, mainly in the corpus callosum (CC) and the cingulum (i.e., cingulum bundle). Using tract-based analyses Nakamae et al. (2011) detected decreased FA in the right cingulum, the anterior body of the corpus callosum (CC) and the left anterior limb of the internal capsule (alic) in a sample of unmedicated patients with OCD. Decreased FA in patients in the cingulum was also reported by Szeszko et al. (2005) at a rather liberal threshold, however, Szeszko et al. found decreased FA in patients compared to healthy volunteers in the left and right parietal lobe (i.e., supramarginal gyri) and the left occipital cortex (i.e., lingual gyrus). White matter alterations in the parietal lobe correlated with Yale-Brown Obsessive Compulsive Scale (Y-BOCS) total scores and scores on the obsession subscale. Like Nakamae and colleagues Bora et al. (2011) found decreased FA in the corpus callosum that was underpinned by increased radial diffusivity in that area. As opposed to Nakamae et al., however, in the study by Bora et al. white matter alterations were located in the body of the corpus callosum with no difference between medicated (n ¼ 10) and unmedicated (n ¼ 11) patients. Decreased FA in the corpus callosum was also reported by Garibotto et al. (2010) who investigated a rather homogeneous patient sample containing, according to symptom typology, either checkers (n ¼ 12) or washers (n ¼ 3). They performed voxel-based analyses at a rather liberal threshold and found, apart from decreased FA in the corpus callosum (splenium) and the cingulum, alterations in the superior longitudinal fasciculus, the optic radiation, and the inferior fronto-occipital fasciculus that were significantly correlated with total Y-BOCS scores (i.e., the stronger the FA decreases, the higher the total Y-BOCS score). In addition, they investigated diffusivity along the single principal diffusion directions (PDD) and found PDD differences between patients and healthy volunteers in white matter regions similar to those regions showing altered FA in patients. Fontenelle et al. (2011) examined a sample of chronic OCD patients (i.e., adult patients with an average age at onset of 12 years) with a high degree of symptomatology using tract-based
Education HC
OCD
HC
10 m, 3 f 20 m, 21 f 15 m, 8 f 7 m, 2 f 14 m, 15 f 16 m, 7 f 15 m, 15 f 16 m 25 m First-degree relatives: 9 m, 21 f, unrelated volunteers: 10 m, 20 f 9 m, 6 f 8 m, 5 f 10 m, 5 f
13.5 13.1 14.0 N/A 14.6 13.3 N/A N/A 13.1 N/A
15.2 (2.2) 15.3 (1.8) 14.8 (1.8)
15.7 (1.8) 15.1 (2.9) 15.9 (1.8)
12 m, 11 f 8 m, 7 f 16, 6 14 m, 12 f
N/A 8.1 (1.8) N/A N/A
N/A 8.9 (2.3) N/A N/A
(1) (3.2) (2.9) (2.2) (2.8)
(2.5)
14 (0.5) 15.8 (2.7) 14.6 (3.7) N/A 15.0 (2.3) 13.4 (2.7) N/A N/A 14.0 (1.2) N/A
analyses. Decreased FA in association with increased MD was detected in the corticospinal tract, the internal capsule, the posterior limb of the internal capsule, the corona radiata and the superior longitudinal fasciculus. Regarding clinical characteristics, there was a negative correlation between corpus callosum (genu) FA and total Y-BOCS scores as well as a positive correlation between MD in the corpus callosum (genu), the bilateral anterior limb of internal capsule, the bilateral superior longitudinal fasciculus and total YBOCS scores. In a combined fMRI e DTI study Admon et al. (2012) investigated altered functional and structural connectivity in a sample of 13 patients with OCD and 13 healthy volunteers. On the structural level they found altered white matter structure in terms of decreased FA in clusters within the uncinate and the cingulum bundle in patients applying a voxel-based analysis. These structural alterations correlated with symptom severity (i.e., lower FA was associated with higher total Y-BOCS scores) and with functional connectivity between amygdala and dorsal anterior cingulate cortex (dACC) as well as between nucleus accumbens (Nacc) and OFC (i.e., lower FA was related to lower functional connectivity). In another recent study by Benedetti et al. (2013) tract-based analyses were used to examine alterations in FA, mean diffusivity (MD), radial diffusivity (RD) and axial diffusivity (AD) in a large sample of drug-treated and drug-naïve patients without any comorbidities. Drug-treated patients showed widespread FA decreases and RD increases compared to drug-naïve patients and healthy controls. Benedetti and colleagues concluded from their findings that changes in white matter microstructure could also be a consequence of drug treatment (i.e., SSRI treatment) and might constitute part of the drugs’ mechanism of action. Like in the before mentioned study by Nakamae et al. (2011) one of the structures featuring significant FA decreases in patients was the anterior limb of the internal capsule (alic). The alic contains fibers that connect the cortex with the corpus striatum and receives fibers from the anterior cingulate, the orbitofrontal and medial frontal cortex (Schmahmann and Pandya, 2006). As the fronto-striatal structures are considered to be critically involved in the pathogenesis of OCD the alic constitutes a fiber structure which is not only gaining increasing attention in the field of OCD research but which has also evolved into one of the most common targets of deep brain stimulation in OCD (Bourne et al., 2012).
Please cite this article in press as: Koch K, et al., Diffusion tensor imaging (DTI) studies in patients with obsessive-compulsive disorder (OCD): A review, Journal of Psychiatric Research (2014), http://dx.doi.org/10.1016/j.jpsychires.2014.03.006
4
K. Koch et al. / Journal of Psychiatric Research xxx (2014) 1e10
Table 2 Age of onset, duration of illness, total Y-BOCS score (mean and SD), percentage of medicated patients, and comorbidities. Abbreviations: AD ¼ anxiety disorder, ADHD ¼ attention deficit hyperactivity disorder, BED ¼ binge eating disorder, GAD ¼ generalized anxiety disorder, MDD ¼ major depressive disorder, ODD ¼ oppositional defiant disorder, SSRI ¼ selective serotonin reuptake inhibitor, N/A ¼ not available. Author
Age of onset
Duration of illness (yrs)
Total Y-BOCS score
Admon et al. (2012) Benedetti et al. (2012) Fan et al. (2012)
N/A 14.1 (7.5) 20.4 (7.9)
8 (1) 19.1 (9.9) 4.8 (4.0)
Fontenelle et al. (2011)
11.5 (5.0)
N/A
28.5 (4.8)
Bora et al. (2011) Li et al. (2011) Nakamae et al. (2011)
N/A N/A 25.0 (9.6)
N/A 6.3 (5.4) 6.7 (7.1)
19.2 (5.4) 23.2 (5.1) 23.8 (5.7)
Garibotto et al. (2010) Ha et al. (2009) Menzies et al. (2008)
20.7 (10.9) N/A N/A
10.1 (8.1) N/A N/A
Nakamae et al. (2008) Yoo et al. (2007)
19.6 (8.0) N/A
N/A 7.2 (9.0)
Szeszko et al. (2005)
16.9 (7.7)
N/A
28.2 (5.0) 20.2 (5.4) 22.1 (5.4), firstdegree relatives: 1.4 (3.0) 29.0 (5.3) Pretreatment: 30.2 (4.6) Posttreatment: 16.6 (6.4) 25.9 (4.4)
Author
Age of onset
Gruner et al. (2012) Jayarajan et al. (2012)
N/A 12.7 (1.9)
Studies with pediatric and adolescent patients Duration of illness Total C-YBOCS % Medicated score patients N/A 26.9 (4.5) 52 1.4 (1.0) 21.5 (7.4) 87
Silk et al. (2013)
N/A
N/A
N/A
Zarei et al. (2011)
11.2 (2.8)
5.3 (3.4)
19.5 (7.6)
Studies with adult patients 24.5 (1.5) 31.2 (5.6) 22.0 (4.9)
3.2. Studies reporting increased white matter connectivity in adult OCD patients While the majority of all DTI studies demonstrated FA decreases, there are also a number of studies that found increased FA in adult patients with OCD. Li et al. (2011) performed a voxel-based analysis of FA, AD and RD in a sample of partly medicated patients and identified significantly increased FA in the corpus callosum (genu and body) as well as the white matter of the right superior frontal gyrus in patients compared to healthy volunteers. In addition, they found a positive correlation between FA in left middle temporal gyrus and Y-BOCS total as well as obsessive subscale scores. In an earlier study by Nakamae et al. (2008) patients were found to have increased FA in the bilateral semioval center extending to the subinsular white matter as well as higher apparent diffusion coefficient (ADC) values in the left medial frontal cortex. In a study by Yoo et al. (2007) a group of unmedicated patients underwent 12 weeks of pharmacotherapy with citalopram. At baseline, i.e., in the unmedicated state, patients exhibited increased FA in the corpus callosum, the internal capsule and white matter in the area superolateral to the right caudate. After 12 weeks of pharmacotherapy, these FA increases had largely disappeared. Although care should be taken when interpreting these findings due to some methodological shortcomings such as, for instance, the limited
% Medicated patients
current comorbidity
100 55 0, n ¼ 15 scanned unmedicated and after 12 weeks of SSRI therapy 78
Personality disorder (n ¼ 3) e e
48 57 0 (n ¼ 14 drug naïve) 87 40 70
100 Pretreatment: 0 / 12 week pharmacotherapy with citalopram 80
13 (withdrawn from medication for 24 h before scanning) 62
MDD (n ¼ 2), Tourette syndrome (n ¼ 2), panic disorder (n ¼ 1), GAD (n ¼ 1), BED (n ¼ 1) e N/A e e e Panic disorder (n ¼ 2)
MDD (n ¼ 1), dysthymic disorder (n ¼ 1) MDD (n ¼ 2), dysthymic disorder (n ¼ 1)
MDD (n ¼ 5), panic disorder (n ¼ 1) social phobia (n ¼ 2), eating disorder (n ¼ 1) Current comorbidity MDD (n ¼ 4), AD (n ¼ 4), ADHD (n ¼ 5) Specific phobia (n ¼ 2), MDD (n ¼ 1), social phobia (n ¼ 1), ODD (n ¼ 1), separation anxiety disorder (n ¼ 1) GAD (n ¼ 3)
e
number of diffusion encoding directions, this study is one of the first interventional studies suggesting that white matter alterations in OCD core regions like the striatum or the internal capsule may be partly reversible with pharmacological treatment. Findings by Fan et al. (2012) point into a similar direction. In their study, Fan and colleagues examined FA, AD, MD and RD parameters in a sample of 27 unmedicated OCD patients and 23 healthy controls. In addition, they compared white matter alterations of 15 patients before and after 12 weeks of SSRI treatment. They discovered that patients were characterized by an increased RD in OCD core regions like the striatum, insula and right midbrain, as well as in left medial superior frontal gyrus, temporo-parietal lobe and the occipital lobe. After 12 weeks of SSRI treatment, OCD patients showed a reduction in RD in the left striatum and right midbrain, as well as in MD of the right midbrain. Despite the differences (i.e., Yoo and colleagues identified increased FA in patients whereas Fan and colleagues found increased RD in patients before treatment) both studies illustrate that some alterations in white matter microstructure may be amenable to treatment. 3.3. Studies reporting decreased and increased white matter connectivity in adult OCD patients Apart from studies demonstrating decreased white matter connectivity and reports on increased connectivity, there are also
Please cite this article in press as: Koch K, et al., Diffusion tensor imaging (DTI) studies in patients with obsessive-compulsive disorder (OCD): A review, Journal of Psychiatric Research (2014), http://dx.doi.org/10.1016/j.jpsychires.2014.03.006
K. Koch et al. / Journal of Psychiatric Research xxx (2014) 1e10
5
Table 3 Imaging and analysis method (i.e., MRI field strength, analysis method, statistical threshold), number of diffusion encoding directions, general findings and findings with regard to clinical characteristics. Abbreviations: CBCL-OCS ¼ Child Behavior Checklist e Obsessive-Compulsive Scale, CBCL-AD ¼ Child Behavior Checklist-Anxiety and Depression, acr ¼ anterior corona radiate, alic ¼ anterior limb of internal capsule, atr ¼ anterior thalamic radiation, cc ¼ corpus callosum, cr ¼ corona radiata, cst ¼ corticospinal tract, ic ¼ internal capsule, icp ¼ inferior cerebellar peduncle, ifo ¼ inferior fronto-occipital fasciculus, ilf ¼ inferior longitudinal fasciculus, mcp ¼ middle cerebellar peduncle, mfg ¼ medial frontal gyrus, plic ¼ posterior limb of internal capsule, ptr ¼ posterior thalamic radiation, scp ¼ superior cerebellar peduncle, scr ¼ superior corona radiata, sfg ¼ superior frontal gyrus, slf ¼ superior longitudinal fasciculus, unc ¼ uncinate fasciculus. Author
Method (i.e., field strength, analysis, statistical threshold)
Diffusion encoding directions
Admon et al. (2012)
1.5 T, voxel-based analyses, p < 0.005 corrected, cluster extent 10 voxels 3 T, tract-based analyses (TBSS), p < 0.05 TFCE corrected, cluster extent 100 voxels
6
Benedetti et al. (2013)
35
Measurement
Findings
Studies with adult patients FA Decreased FA in r. unc. and l. cingulum Positive correlation between cingulum/ uncinate FA and functional connectivity between amygdalaedACC/NacceOFC FA, MD, RD, AD Decreased FA in r. cingulum, r. ptr, r. ilf, r. slf, r. alic, l. cc, l. scr, l.þr. ifo, l.þr. unc., l.þr. atr, l.þr. acr. Increased MD in l.þr. ifo, l.þr. forceps minor, l.þr. slf, l. atr, l. unc., l. external capsule, r. ilf. Increased RD in l. slf, l. cingulum, l. forceps minor, l. scr, l.þr. ifo, l.þr. atr. Effects of drug status (see text) FA, AD, RD and MD Increased RD in l. medial sfg, l. temporoparietal lobe, l. occipital lobe, l. striatum, insula and r. midbrain. Decreased RD in l. striatum and r. midbrain after medication. Decreased MD in r. midbrain after medication. FA, MD Decreased FA and increased MD in cst, genu of ic, plic, cr, slf
Fan et al. (2012)
1.5 T, voxel-based analyses, p < 0.001 uncorrected, cluster extent 10 voxels
25
Fontenelle et al. (2011)
1.5 T, tract-based analyses (TBSS), p < 0.05 corrected
6
Bora et al. (2011)
3 T, tract-based analyses (TBSS), p < 0.05 cluster level corrected (t ¼ 2) 3 T, voxel-based analyses, p < 0.05 voxel level corrected, cluster extent 50 voxels 1.5 T, tract-based analyses (TBSS), p < 0.05 cluster level corrected (t ¼ 3 and t ¼ 1.5) 1.5 T, voxel-based analyses, p < 0.005 uncorrected, cluster extent 20 voxels
28
FA, AD, RD
Decreased FA in body of CC
15
FA, AD, RD
Increased FA in genu and body of CC and r. sfg
15
FA
Decreased FA in anterior body of CC (at t ¼ 3). Decreased FA in CC, r. cingulum, and l. alic (at t ¼ 1.5)
35
FA, PDD
Decreased FA in splenium of CC, cingulum, l.þr. slf, l.þr. optic radiation and l.þr. ifo. PDD differences in cc, l.r. slf, l.þr. optic radiation, l.þr. ifo, l.þr. ilf, and l.þr. cerebello-cortical/cortical-cerebellar fiber projections Decreased FA in l. ACC.
Li et al. (2011)
Nakamae et al. (2011)
Garibotto et al. (2010)
Ha et al. (2009)
1.5 T, voxel-based analyses, p < 0.001 uncorrected (p < 0.0001 uncorrected for subgroup analyses)
12
FA
Menzies et al. (2008)
1.5 T, voxel-based analyses, nonparametric testing, p < 0.017 (
