- Email: [email protected]
Pediatric OCD structural brain deficits in conflict monitoring circuits: A voxel-based morphometry study
Neuroscience Letters 421 (2007) 218–223
Pediatric OCD structural brain deficits in conflict monitoring circuits: A voxel-based morphometry study Susanna Carmona a,∗ , Nuria Bassas b , Mariana Rovira c , Joan-Domingo Gispert e , Joan-Carles Soliva a , Marisol Prado b , Josep Tomas b , Antoni Bulbena d , Oscar Vilarroya a a
Unitat de Recerca en Neuroci`encia Cognitiva, Departament de Psiquiatria i Medicina Legal, Universitat Aut`onoma Barcelona, Spain b Unitat de Paidopsiquiatria – Servei de Psiquiatria, Hospital Vall d’Hebron, Barcelona, Spain c CRC Corporaci´ o Sanitaria, Barcelona, Spain d Institut d’Atenci´ o Psiqui`atrica, IMAS, Barcelona, Spain e Institut d’Alta Tecnologia – Parc de Recerca Biom` edica de Barcelona, Spain Received 24 January 2007; received in revised form 10 April 2007; accepted 3 May 2007
Abstract The aim of this study is to use a voxel-based morphometry protocol to compare the brains of 18 children with obsessive–compulsive disorder (OCD) with those of a healthy group matched for gender and handedness. Images were acquired with a 1.5-T MRI scanner, spatially normalized, and segmented with an optimized voxel-based morphometry protocol. OCD children presented a 5.93% reduction of gray matter (GM) total volume in comparison with control brains. We identified OCD brain volume reductions in regions that have been extensively related to action monitoring and error signaling processes. Specifically, we found decreased bilateral GM in frontal (significant after Family Wise Error (FEW), multiple comparisons correction) and cingulate regions as well as decreased white matter (WM) in bilateral frontal and right parietal (p < 0.001 uncorrected). Additionally, we found a negative correlation between symptom severity and bilateral hippocampal GM-volume (p < 0.001uncorrected) as well as a positive correlation between age and GM left caudate volume (p = 0.037 FWE small volume corrected) in the OCD group. As a conclusion, our results point to conflict monitoring structural brain regions as primary deficits in pediatric OCD, and to striatal abnormalities as age-related deficits. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Obsessive–compulsive disorder; Child psychiatry; Brain imaging techniques; Conflict monitoring
Current pathophysiological obsessive–compulsive disorder (OCD) models propose a dysregulation of frontal–striatal– thalamic circuits. This system has been involved in monitoring events [1] as well as generating error signals when there is a mismatch between intended and actual performance [15]. Moreover, OCD patients, compared to healthy controls, show larger and longer error signals (as measured by error-related negativity in ERP studies) that correlate with severity symptoms [7]. This exaggerated or false error signaling may be underlying characteristic OCD symptoms, such as the need to perform an action to respond to conflict feelings [18]. Structural neuroimaging studies have partially confirmed fronto-striatal–thalamic models, even if the findings have been heterogeneous. Some studies
∗ Corresponding author at: Unitat de Recerca en Neuroci` encia Cognitiva, Universitat Aut`onoma de Barcelona, Centre Forum (Hospital del Mar), C/Llull 410, Barcelona 08019, Spain. Tel.: +34 661 606 710. E-mail address: [email protected] (S. Carmona).
0304-3940/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2007.05.047
show basal-ganglia (BG) volume reduction [20], others volume enlargement [19], and still others show no differences [3]. Such discordant results are also found in orbitofrontal cortex (OFC) [12,13]. Discrepancies may stem from the sample nature (adults or children), or the volumetric methods employed (manual or automatic). Regarding the sample, the literature seems to point to specific morphometric abnormalities depending on the OCDstage. In this sense, children show reduced [21] or normal [22] striatum (caudate/putamen) while adults seem to have increases in the same areas [19]. Concerning the methods as a source of discrepancies, OCD morphometric studies use principally manual region-of-interest (ROI) analysis.There are just three studies [13,19,26] that used voxel-based morphometry (VBM) methods to examine OCD brain abnormalities, all of them with adults. Regardless of the advantages or disadvantages of manual and automatic methods [8], a pediatric VBM study will complement previous studies and help to improve the understanding of OCD pathophysiology. Our aim is to compare the brains of OCD children with those of a control group using an optimized voxel-
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
based morphometric (OVBM) protocol. In order to do so, we take the following steps: (1) Examine global differences in gray matter (GM), white matter (WM), cerebrospinal fluid (CSF), and total brain volume (TBV). (2) Identify regional differences in GM and WM segments between groups. (3) Pinpoint-specific regions related to OCD clinical severity. The apriori hypotheses specify that OCD children, in comparison to control children, will have a GM volume reduction in regions previously reported in adult VBM studies [13,19,26], such as frontal and cingulate cortex, as well as a GM volume increase in BG and OFC. Subjects included 18 children with OCD, according to DSMIV TR criteria, and 18 control children. Both groups were matched for gender, age and handedness. IQ level was controlled for verbal and performance scores. IQ patients’ scores were obtained using WISC-R full-scale; in the case of control group, Block Design and Vocabulary subtests were used to estimate the measures. OCD patients were recruited from the “Unitat de Paidopsiquiatria, Servei de Psiquiatria” at “Vall d’Hebron” Materno-Infantil Hospital. Eleven of the OCD children had comorbid tic disorder (but did not meet criteria for Tourette’s disorder). Inattention and anxiety symptoms were present in three and six children, respectively. Controls were selected using a convenience sampling from the Traumatology Department. These patients had suffered minor physical trauma (i.e. strain injuries). Exclusion criteria for both groups were IQ below 80, comorbidity of any other psychiatric disorder (according to a semi-structured clinical interview with parents and children based on DSM-IV TR criteria) and evidence of neurological disorders, head trauma or perinatal anoxia. In the patient group, trait anxiety and obsessive/compulsive symptoms severity were also measured using STAI/T-C (State Trait Anxiety Inventory for Children) and CY-BOCS (Child and Adolescent Version of the Yale-Brown Obsessive–Compulsive Scale), respectively. All the children had the parental approval to participate in the study. Demographic and clinical data are summarized in Table 1. Imag-
219
ing was performed by using a 1.5-T MR Unit (Signa Horizon; GE Medical Systems, Milwaukee, I). The scanning protocol was standardized and included axial 3D-T1 weighted images of the entire brain. The imaging parameters were as follows: TR/TE = 13.2/4.2 ms; TI = 500 ms; voxel size = 0.94 × 0.94; 2mm-thick contiguous axial-oblique sections; flip angle 15◦ and matrix size: 256 × 256 × 60. Images were processed using SPM2 (Wellcome Department of Imaging Neuroscience, London, United Kingdom: http://www.fil.ion.ucl.ac.uk/spm/). We performed the analysis following two main stages. First, a study specific template was created for the whole brain and the different tissues (GM, WM and CSF). Second, an OVBM protocol was applied to the 3D original images. Concerning the first step, the templates were based on the images of all OCD and control subjects. The customized templates, based on the subject’s 3D images, were first spatially normalized by non-linearly registering each of them to the standard SPM T1MRI template [2]. Images were segmented into GM, WM and CSF portions using the combined pixel intensity, and finally smoothed with an isotropic Gaussian kernel 8 mm full width at half maximum (FWHM). Regarding the OVBM protocol, we performed the following steps based on previous studies [9,10]: (first) Segmentation of the original 3D images. This step allows the calculation of TBV, GM, WM, and CSF global volumes in millilitres. (second) Normalization of the segmented GM and WM images according to their corresponding tissues template (GM of WM depending on the aim of the specific analysis), thus preventing any contribution of non-tissue voxels and achieving optimal spatial normalization. (Third) Segmentation of the spatially normalized images to reject remaining non-tissue voxels (scalp, skull or venous sinuses). (Fourth) Modulation of the images by the Jacobian determinants derived from their spatial normalization step. This introduces intensity changes in the GM/WM images according to the variation in structural regional volumes that the normalization process
Table 1 Clinical and demographic data
N
Boys Girls
Age
OCD
Controls
13 5
13 5
μ 12.86 (S.D. 2.76) years
μ, 13.03 (S.D. 3.04) years 15 2 1
Laterality
Right-handed Left-handed Cross-dominance
15 2 1
Drug-treatment
SSRIs Sertraline Citalopram Clomipramine Drug-na¨ıve
n, 6; μ, 13.1 (S.D. 12.8) months n, 3; μ, 6.4 (S.D. 3.0) months n, 1; μ, 2.3 months n, 8 Range, 4–98; μ, 60.53 (S.D. 33.7)
STAI/T-C CYBOCS
Obsessions Compulsions Total
μ 10.28 (S.D. 4.28) μ 10.94 (S.D. 4.27) μ 21.39 (S.D. 5.88)
This table summarizes clinical and demographical data of our sample. μ, Mean values; S.D., standard deviation; n, number of patients; STAI/T-C, State Trait Anxiety Inventory for Children.
220
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
Fig. 1. GM and WM reductions in OCD children. Statistically significant regions of GM (A) and WM (B) decrease in OCD children. Images are displayed according to neurological orientation. Colors indicate the t-score. See also Table 2.
produced. The Jacobian-modulated GM step allows making inferences about differences in volumes rather than concentrations. (Fifth) Smooth with a Gaussian kernel of 12-mm FWHM. In order to study global volumetric measures, we performed two-tailed t-test means comparisons for independent samples (threshold at p < 0.05). Regarding regional morphometric analysis, smoothed GM and WM images were analyzed performing two one-tailed t-tests (OCD> controls and controls > OCD) to identify the specific direction of the between-groups differences. Additionally, in the OCD group, we performed two linear correlations between: (a) GM volume and total CY-BOCS and (b) GM volume and STAI/T-C. Results from the regional morphometric analysis and the linear correlations were thresholded at p < 0.001 (uncorrected for multiple comparisons); only clusters larger than 40 voxels were analyzed. Provided that our groups were gender, age and handedness matched, we did not apply TBV covariation since we were interested in finding absolute brain differences between the samples; we considered important to minimize type II rather than type I error, this being the first VBM study in OCD children. As shown below, we did not find significant differences between groups concerning GM volume in the BG (thalamus, caudate, putamen and globus pallidus). Due to this fact, secondary hypotheses were tested in order to see whether age or drug treatment could be modulating the volume in these regions. For this purpose, we lowered our threshold to p < 0.005 uncorrected and performed post-hoc small volume corrected (SVC) correlations between: (a) BG–GM volume and age and (b) BG–GM volume and pharmacological treatment duration. In order to isolate the effect of drug treatment we also performed a correlation between treatment duration and BG, covarying out the possible effect of age. Results were thresholded at p < 0.05 Family Wise Error (FWE)-SVC. Concerning global differences in GM, WM, CSF and TBV, mean comparisons of global volumetric measures showed that OCD children had a GM total volume reduction of 48.75 ml
(p = 0.01), representing 5.93% decrement (95% CI: 1.49– 10.36% or 12.3–85.2 ml). Regional differences in GM and WM segment volumes are shown in Table 2 and Fig. 1. There were no regions of increased GM or WM in the OCD group in comparison with the control group. OCD brains presented GM reductions in frontal lobe (bilateral middle and inferior triangularis areas, right frontal superior and frontal inferior operculum, and left rolandic operculum) and cingulate cortex (bilateral middle cingulate cortex and bilateral precuneus) as well as WM reductions in frontal (bilateral inferior operculum and left frontal middle gyrus) and parietal regions (right parietal superior). The GM reduction located in the right frontal middle cortex shows a p-value of 0.035 corrected for multiple comparisons according to FWE criteria. The FWE-correction prioritizes specificity over sensitivity of the results exerting a strong control over false positive values. Regarding correlation between regional GM volumes and clinical severity, a significant negative correlation was found between symptoms severity (Y-BOCS) and bilateral hippocampi GM volumes (see correlation in Table 2 and Fig. 2). No significant correlations were identified between GM volume and STAI/T-C. Concerning post-hoc BG correlations, we found a significant correlation between age and left caudate volume in the OCD group (p = 0.037 FWE-SVC; cluster size 148 voxels), but no age correlation in the control group. In the OCD group, we also identified positive correlations between drug treatment duration and left putamen volumes (p = 0.005 FWE-SVC; cluster size 375 voxels), right caudate (p = 0.014 FWE-SVC; cluster size 309 voxels) and bilateral thalamus (left, p = 0.049 FWESVC; cluster size = 1170 voxels; right, p = 0.003 FWE-SVC; cluster size = 1386 voxels). None of these correlations survived p-value of 0.05 FWE-SVC once we covariate for age’s effect. Summarizing the main results, our OCD patients showed a 5.93% reduction of total GM. The VBM analysis showed decreased bilateral GM in frontal and cingulate regions as well as decreased WM in bilateral frontal and right parietal.
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
221
Table 2 Voxel-based morphometry results Structure (Brodmann area) GM Frontal lobe R frontal mid. (45) R frontal inf. tri. (45)
R frontal mid. (6) R frontal inf. oper. (44) R frontal mid. (9) R frontal inf. oper. (44) R fontal sp. (8) L fontal mid. (45) L frontal inf. tri. (45)
L frontal inf. tri. (48)
L rolandic oper. (48) Cingulate cortex R cing. (23)
R precuneus
L cing. (23–24)
L precuneus WM Frontal lobe R frontal inf. oper. L frontal mid. L frontal inf. oper. Parietal areas R parietal sup. Correlation (GM/CY-BOCS) Limbic regions R hippocampus L hippocampus
MNIa (x, y, z)
Peakb Z-score
p-value (FWE)-correctedc
Cluster size (voxels)d
40, 39, 21 53, 27, 21 56, 23, 20 50, 31, 19 41, 5, 42 44, 13, 35 43, 13, 31 25, 29, 42 27, 25, 43 46, 7, 22 26, 25, 59 −41, 31, 34 −48, 28, 20 −49, 35, 26 −49, 31, 24 −47, 28, 27 −45, 25, 23 −44, 23, 30 −48, 21, 29 −52, 3, 16
4.39 3.42 3.31 3.30 3.48 3.41 3.36 3.41 3.38 3.41 3.40 3.62 3.33 3.29 3.22 3.19 3.42 3.14 3.10 3.23
0.035 0.594 0.700 0.715 0.526 0.846 0.874 0.601 0.636 0.601 0.611 0.391 0.680 0.720 0.781 0.809 0.591 0.845 0.871 0.771
2104
1, −26, 50 15, −37, 43 1, −31, 47 2, −9, 46 3, −13, 48 1, −18, 46 7, −11, 49 1, 13, 42 16, −43, 42 15, −53, 41 16, −48, 42 −2, −10, 47 −1, −14, 45 −4, −16, 48 −8, −41, 45 −14, −42, 45
4.04 3.91 3.89 3.69 3.69 3.63 3.62 3.43 3.55 3.48 3.39 3.75 3.64 3.58 3.14 3.46
0.119 0.183 0.191 0.329 0.330 0.382 0.391 0.578 0.464 0.527 0.625 0.285 0.376 0.430 0.849 0.552
9399
48, 21, 12 −27, 54, 3 −41, 18, 12
3.42 3.39 3.23
0.359 0.387 0.531
164 76 113
22, −50, 59
3.24
0.523
92
23, −31, −5 −19, −25, −8
3.53 3.77
0.624 0.397
237 626
318
236 292 130 1961
70
Regions of decreased GM and WM in OCD children in comparison with healthy matched controls. The lower part of the table shows GM regions correlating with CY-BOCS scores. L, left; R, right; mid., middle; inf., inferior; sup., superior; cing., cingulate; tri, triangularis; oper., operculum. a Montreal Neurological Institute (MNI) coordinates of the voxel of maximal statistical significance within each region. b Peak Z–score for the voxel of maximal statistical significance within each region. c Statistical significance after multiple comparison family-wise error (FWE), which rejects false positives. d Total number of contiguous voxels in each region that outlive p < 0.001. Given the voxel size (0.94 × 0.94 × 2) each voxel represents 1.77 mm3 .
These regions have been previously involved in action and conflict monitoring [28]. Provided that there are no previous VBM studies on children, our results cannot be compared with other pediatric studies. However, our cingulate results coin-
cide with the three VBM-OCD studies [13,19,26] and the right frontal middle GM reduction (p = 0.035 FWE-corrected) was also reported in two of them [13,19]. Our frontal and cingulate findings may nevertheless have an important bearing on OCD
222
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
Fig. 2. CY-BOCS correlation. Negative correlation (thresholded at p < 0.001) between CY-BOCS scores and GM volume in bilateral hippocampi. Sided scatter plots show the correlation between the GM volume (y-axis) of left (LH) and right (RH) hippocampus and CY-BOCS scores (x-axis). Gray dots represent real values and black dots adjusted values.
pathophysiology. Neuropsychological studies have reported that OCD patients show deficits related to frontal cognitive abilities, such as executive functioning and cognitive-behavioral flexibility [6]. In fMRI studies [28] frontal (BA 9, 44, 45, and 46) and cingulate cortices (BA 23 and precuneus) have been found to be active during conflict monitoring tasks. Moreover, specific OCD fMRI studies [15] have shown a different pattern of activation of these regions in comparison to controls during action monitoring tasks. These results support an initial deficit in the action monitoring system in OCD children. In this sense, the typical OCD feeling of “not just right” could be interpreted as an exaggerated or false error detection [18], and thus partially explained by our findings. Indeed, cingulate cortices have been associated with overactive error monitoring, as measured by the error-related negativity in ERP studies [7,27] as well as in fMRI studies [5,25]. Apart from these frontal and cingulate findings, our study is the first to detect a significant negative correlation between CYBOCS scores and bilateral GM volumes in the parahippocampal region. The negative correlation with CY-BOCS scores points to an association between clinical severity and hippocampal damage at the onset of the illness. Two hypotheses can be advanced to explain this finding. On the one hand, the hippocampal abnormalities could be related to memory OCD deficits. There is strong evidence that OCD is associated with low memory confidence as well as memory biases towards threatening information [17,23,24], specifically when it concerns the ability to recall if a behavior was actually performed or merely imagined [16]. On the other hand, bilateral smaller hippocampi have also been found in other psychiatric conditions, and hence it suggests that hippocampal reduction is more a case of neurodevelopmental damage than an illness-specific abnormality [14]. The main discrepancy of our results with previous OCD structural studies is the lack of BG deficits in our sample. The BG, especially caudate nucleus and putamen, have been extensively associated to routine performance and reinforcement [4,11,29]. This fact has made these regions (together with its OFC connections) suitable candidates to explain OCD pathophysiology. Two possible hypotheses could explain our lack of BG results: (a) the dynamic pattern of the illness and (b) the duration of pharmacological treatment. In relation to the developmental nature
of the illness, behavioral OCD pathological habits, such as the repetition of routines, could specifically affect the normal development of the BG. In fact, the largest structural OCD study to date [19] found a positive correlation between age and striatal abnormalities, which was interpreted as the anatomical expression of enduring striatal dysfunction. Therefore, the absence of striatal alterations [22] – or even the finding of a reduction, rather than an increase of striatal volumes [21] – could be expected in pediatric populations. Regarding the time of pharmacological exposure, Valente’s VBM study [26] found a negative correlation between duration of treatment and bilateral GM volumes of the head of caudate nuclei. To see whether our study supports any of these two hypotheses, we performed age and drug-treatment BG correlations. We found that subjects’ age positively correlated with left caudate volume in the OCD group but not in the control group. Concerning pharmacological treatment duration, we observed positive correlations between drug exposure duration and left putamen, right caudate and bilateral thalamus that do not pass multiple-comparison correction once we covariate for age effect. Thus, we did not find the negative correlation previously reported by Valente [26]. One should take into account that the mean treatment duration in Valente’s study is clearly smaller (mean = 47.6 days, S.D. = 41.9; and range 3–120) than our subjects’ mean treatment duration (10.04 months; S.D. = 10.05; and range 0.2–33.8). This difference in drug-treatment duration, in addition to the young age of our sample, could be partially responsible for the discrepancies concerning the effect of pharmacological time exposure on caudate volumes. However, our drug treatment positive correlation do not support 0.05 FWESVC once we covariate for age affect, which seems to indicate that the duration of pharmacological exposure is, by itself enable to explain BG results. Future studies could clarify this point. As a conclusion, our study pointed to GM and WM early OCD abnormalities associated with conflict monitoring circuits. We also found a significant negative correlation between CY-BOCS scores and bilateral GM volumes in the parahippocampal region. Other deficits previously found in adults (striatum and OFC) are absent in our study. However, according to our results, it seems that the dynamic nature of the illness could be modulating the normal striatum development in OCD children. Finally, given
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
that this is the first VBM study in OCD children our analyses gave priority to sensitivity over specificity, and thus they should be taken with caution waiting for future replications.
[13]
Acknowledgments [14]
We would like to thank all the children and parents who participated in this study, as well as Mercedes Riba for her collaboration, Joseph Hilferty for his editorial review, and Adolf Tobe˜na and Jamil Zaki for their scientific advice. Susanna Carmona was supported by a FPU grant from the “Ministerio de Educaci´on y Cultura”.
[15]
[16]
References [1] B. Aouizerate, D. Guehl, E. Cuny, A. Rougier, B. Bioulac, J. Tignol, P. Burbaud, Pathophysiology of obsessive–compulsive disorder: a necessary link between phenomenology, neuropsychology, imagery and physiology, Prog. Neurobiol. 72 (2004) 195–221. [2] J. Ashburner, K.J. Friston, Voxel-based morphometry—the methods, Neuroimage 11 (2000) 805–821. [3] E.H. Aylward, G.J. Harris, R. Hoehn-Saric, P.E. Barta, S.R. Machlin, G.D. Pearlson, Normal caudate nucleus in obsessive–compulsive disorder assessed by quantitative neuroimaging, Arch. Gen. Psychiatry 53 (1996) 577–584. [4] J.J. Canales, A.M. Graybiel, A measure of striatal function predicts motor stereotypy, Nat. Neurosci. 3 (2000) 377–383. [5] K.D. Fitzgerald, R.C. Welsh, W.J. Gehring, J.L. Abelson, J.A. Himle, I. Liberzon, S.F. Taylor, Error-related hyperactivity of the anterior cingulate cortex in obsessive–compulsive disorder, Biol. Psychiatry 57 (2005) 287–294. [6] L. Friedlander, M. Desrocher, Neuroimaging studies of obsessive– compulsive disorder in adults and children, Clin. Psychol. Rev. 26 (2006) 32–49. [7] W.J. Gehring, J. Himle, L.G. Nisenson, Action-monitoring dysfunction in obsessive–compulsive disorder, Psychol. Sci. 11 (2000) 1–6. [8] N.R. Giuliani, V.D. Calhoun, G.D. Pearlson, A. Francis, R.W. Buchanan, Voxel-based morphometry versus region of interest: a comparison of two methods for analyzing gray matter differences in schizophrenia, Schizophr. Res. 74 (2005) 135–147. [9] C.D. Good, I. Johnsrude, J. Ashburner, R.N. Henson, K.J. Friston, R.S. Frackowiak, Cerebral asymmetry and the effects of sex and handedness on brain structure: a voxel-based morphometric analysis of 465 normal adult human brains, Neuroimage 14 (2001) 685–700. [10] C.D. Good, I.S. Johnsrude, J. Ashburner, R.N. Henson, K.J. Friston, R.S. Frackowiak, A voxel-based morphometric study of ageing in 465 normal adult human brains, Neuroimage 14 (2001) 21–36. [11] M. Haruno, M. Kawato, Heterarchical reinforcement-learning model for integration of multiple cortico-striatal loops: fMRI examination in stimulus-action-reward association learning, Neural. Netw. 19 (2006) 1242–1254. [12] D.H. Kang, J.J. Kim, J.S. Choi, Y.I. Kim, C.W. Kim, T. Youn, M.H. Han, K.H. Chang, J.S. Kwon, Volumetric investigation of the frontal-subcortical
[17] [18] [19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27] [28]
[29]
223
circuitry in patients with obsessive–compulsive disorder, J. Neuropsychiatry Clin. Neurosci. 16 (2004) 342–349. J.J. Kim, M.C. Lee, J. Kim, I.Y. Kim, S.I. Kim, M.H. Han, K.H. Chang, J.S. Kwon, Grey matter abnormalities in obsessive–compulsive disorder: statistical parametric mapping of segmented magnetic resonance images, Br. J. Psychiatry 179 (2001) 330–334. J.S. Kwon, J.J. Kim, D.W. Lee, J.S. Lee, D.S. Lee, M.S. Kim, I.K. Lyoo, M.J. Cho, M.C. Lee, Neural correlates of clinical symptoms and cognitive dysfunctions in obsessive–compulsive disorder, Psychiatry Res. 122 (2003) 37–47. N. Maltby, D.F. Tolin, P. Worhunsky, T.M. O’Keefe, K.A. Kiehl, Dysfunctional action monitoring hyperactivates frontal-striatal circuits in obsessive–compulsive disorder: an event-related fMRI study, Neuroimage 24 (2005) 495–503. R.J. McNally, P.A. Kohlbeck, Reality monitoring in obsessive–compulsive disorder, Behav. Res. Ther. 31 (1993) 249–253. J. Muller, J.E. Roberts, Memory and attention in Obsessive–compulsive Disorder: a review, J. Anxiety Disord. 19 (2005) 1–28. R.K. Pitman, A cybernetic model of obsessive–compulsive psychopathology, Compr. Psychiatry 28 (1987) 334–343. J. Pujol, C. Soriano-Mas, P. Alonso, N. Cardoner, J.M. Menchon, J. Deus, J. Vallejo, Mapping structural brain alterations in obsessive–compulsive disorder, Arch. Gen. Psychiatry 61 (2004) 720–730. D. Robinson, H. Wu, R.A. Munne, M. Ashtari, J.M. Alvir, G. Lerner, A. Koreen, K. Cole, B. Bogerts, Reduced caudate nucleus volume in obsessive–compulsive disorder, Arch. Gen. Psychiatry 52 (1995) 393–398. D.R. Rosenberg, M.S. Keshavan, K.M. O’Hearn, E.L. Dick, W.W. Bagwell, A.B. Seymour, D.M. Montrose, J.N. Pierri, B. Birmaher, Frontostriatal measurement in treatment-naive children with obsessive–compulsive disorder, Arch. Gen. Psychiatry 54 (1997) 824–830. P.R. Szeszko, S. MacMillan, M. McMeniman, S. Chen, K. Baribault, K.O. Lim, J. Ivey, M. Rose, S.P. Banerjee, R. Bhandari, G.J. Moore, D.R. Rosenberg, Brain structural abnormalities in psychotropic drug-naive pediatric patients with obsessive–compulsive disorder, Am. J. Psychiatry 161 (2004) 1049–1056. F. Tallis, The neuropsychology of obsessive–compulsive disorder: a review and consideration of clinical implications, Br. J. Clin. Psychol. 36 (Pt 1) (1997) 3–20. F. Tallis, P. Pratt, N. Jamani, Obsessive–compulsive disorder, checking, and non-verbal memory: a neuropsychological investigation, Behav. Res. Ther. 37 (1999) 161–166. S. Ursu, V.A. Stenger, M.K. Shear, M.R. Jones, C.S. Carter, Overactive action monitoring in obsessive–compulsive disorder: evidence from functional magnetic resonance imaging, Psychol. Sci. 14 (2003) 347–353. A.A. Valente Jr., E.C. Miguel, C.C. Castro, E. Amaro Jr., F.L. Duran, C.A. Buchpiguel, X. Chitnis, P.K. McGuire, G.F. Busatto, Regional gray matter abnormalities in obsessive–compulsive disorder: a voxel-based morphometry study, Biol. Psychiatry 58 (2005) 479–487. V. van Veen, C.S. Carter, The anterior cingulate as a conflict monitor: fMRI and ERP studies, Physiol. Behav. 77 (2002) 477–482. V. van Veen, J.D. Cohen, M.M. Botvinick, V.A. Stenger, C.S. Carter, Anterior cingulate cortex, conflict monitoring, and levels of processing, Neuroimage 14 (2001) 1302–1308. Z.M. Williams, E.N. Eskandar, Selective enhancement of associative learning by microstimulation of the anterior caudate, Nat. Neurosci. 9 (2006) 562–568.
Pediatric OCD structural brain deficits in conflict monitoring circuits: A voxel-based morphometry study Susanna Carmona a,∗ , Nuria Bassas b , Mariana Rovira c , Joan-Domingo Gispert e , Joan-Carles Soliva a , Marisol Prado b , Josep Tomas b , Antoni Bulbena d , Oscar Vilarroya a a
Unitat de Recerca en Neuroci`encia Cognitiva, Departament de Psiquiatria i Medicina Legal, Universitat Aut`onoma Barcelona, Spain b Unitat de Paidopsiquiatria – Servei de Psiquiatria, Hospital Vall d’Hebron, Barcelona, Spain c CRC Corporaci´ o Sanitaria, Barcelona, Spain d Institut d’Atenci´ o Psiqui`atrica, IMAS, Barcelona, Spain e Institut d’Alta Tecnologia – Parc de Recerca Biom` edica de Barcelona, Spain Received 24 January 2007; received in revised form 10 April 2007; accepted 3 May 2007
Abstract The aim of this study is to use a voxel-based morphometry protocol to compare the brains of 18 children with obsessive–compulsive disorder (OCD) with those of a healthy group matched for gender and handedness. Images were acquired with a 1.5-T MRI scanner, spatially normalized, and segmented with an optimized voxel-based morphometry protocol. OCD children presented a 5.93% reduction of gray matter (GM) total volume in comparison with control brains. We identified OCD brain volume reductions in regions that have been extensively related to action monitoring and error signaling processes. Specifically, we found decreased bilateral GM in frontal (significant after Family Wise Error (FEW), multiple comparisons correction) and cingulate regions as well as decreased white matter (WM) in bilateral frontal and right parietal (p < 0.001 uncorrected). Additionally, we found a negative correlation between symptom severity and bilateral hippocampal GM-volume (p < 0.001uncorrected) as well as a positive correlation between age and GM left caudate volume (p = 0.037 FWE small volume corrected) in the OCD group. As a conclusion, our results point to conflict monitoring structural brain regions as primary deficits in pediatric OCD, and to striatal abnormalities as age-related deficits. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Obsessive–compulsive disorder; Child psychiatry; Brain imaging techniques; Conflict monitoring
Current pathophysiological obsessive–compulsive disorder (OCD) models propose a dysregulation of frontal–striatal– thalamic circuits. This system has been involved in monitoring events [1] as well as generating error signals when there is a mismatch between intended and actual performance [15]. Moreover, OCD patients, compared to healthy controls, show larger and longer error signals (as measured by error-related negativity in ERP studies) that correlate with severity symptoms [7]. This exaggerated or false error signaling may be underlying characteristic OCD symptoms, such as the need to perform an action to respond to conflict feelings [18]. Structural neuroimaging studies have partially confirmed fronto-striatal–thalamic models, even if the findings have been heterogeneous. Some studies
∗ Corresponding author at: Unitat de Recerca en Neuroci` encia Cognitiva, Universitat Aut`onoma de Barcelona, Centre Forum (Hospital del Mar), C/Llull 410, Barcelona 08019, Spain. Tel.: +34 661 606 710. E-mail address: [email protected] (S. Carmona).
0304-3940/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2007.05.047
show basal-ganglia (BG) volume reduction [20], others volume enlargement [19], and still others show no differences [3]. Such discordant results are also found in orbitofrontal cortex (OFC) [12,13]. Discrepancies may stem from the sample nature (adults or children), or the volumetric methods employed (manual or automatic). Regarding the sample, the literature seems to point to specific morphometric abnormalities depending on the OCDstage. In this sense, children show reduced [21] or normal [22] striatum (caudate/putamen) while adults seem to have increases in the same areas [19]. Concerning the methods as a source of discrepancies, OCD morphometric studies use principally manual region-of-interest (ROI) analysis.There are just three studies [13,19,26] that used voxel-based morphometry (VBM) methods to examine OCD brain abnormalities, all of them with adults. Regardless of the advantages or disadvantages of manual and automatic methods [8], a pediatric VBM study will complement previous studies and help to improve the understanding of OCD pathophysiology. Our aim is to compare the brains of OCD children with those of a control group using an optimized voxel-
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
based morphometric (OVBM) protocol. In order to do so, we take the following steps: (1) Examine global differences in gray matter (GM), white matter (WM), cerebrospinal fluid (CSF), and total brain volume (TBV). (2) Identify regional differences in GM and WM segments between groups. (3) Pinpoint-specific regions related to OCD clinical severity. The apriori hypotheses specify that OCD children, in comparison to control children, will have a GM volume reduction in regions previously reported in adult VBM studies [13,19,26], such as frontal and cingulate cortex, as well as a GM volume increase in BG and OFC. Subjects included 18 children with OCD, according to DSMIV TR criteria, and 18 control children. Both groups were matched for gender, age and handedness. IQ level was controlled for verbal and performance scores. IQ patients’ scores were obtained using WISC-R full-scale; in the case of control group, Block Design and Vocabulary subtests were used to estimate the measures. OCD patients were recruited from the “Unitat de Paidopsiquiatria, Servei de Psiquiatria” at “Vall d’Hebron” Materno-Infantil Hospital. Eleven of the OCD children had comorbid tic disorder (but did not meet criteria for Tourette’s disorder). Inattention and anxiety symptoms were present in three and six children, respectively. Controls were selected using a convenience sampling from the Traumatology Department. These patients had suffered minor physical trauma (i.e. strain injuries). Exclusion criteria for both groups were IQ below 80, comorbidity of any other psychiatric disorder (according to a semi-structured clinical interview with parents and children based on DSM-IV TR criteria) and evidence of neurological disorders, head trauma or perinatal anoxia. In the patient group, trait anxiety and obsessive/compulsive symptoms severity were also measured using STAI/T-C (State Trait Anxiety Inventory for Children) and CY-BOCS (Child and Adolescent Version of the Yale-Brown Obsessive–Compulsive Scale), respectively. All the children had the parental approval to participate in the study. Demographic and clinical data are summarized in Table 1. Imag-
219
ing was performed by using a 1.5-T MR Unit (Signa Horizon; GE Medical Systems, Milwaukee, I). The scanning protocol was standardized and included axial 3D-T1 weighted images of the entire brain. The imaging parameters were as follows: TR/TE = 13.2/4.2 ms; TI = 500 ms; voxel size = 0.94 × 0.94; 2mm-thick contiguous axial-oblique sections; flip angle 15◦ and matrix size: 256 × 256 × 60. Images were processed using SPM2 (Wellcome Department of Imaging Neuroscience, London, United Kingdom: http://www.fil.ion.ucl.ac.uk/spm/). We performed the analysis following two main stages. First, a study specific template was created for the whole brain and the different tissues (GM, WM and CSF). Second, an OVBM protocol was applied to the 3D original images. Concerning the first step, the templates were based on the images of all OCD and control subjects. The customized templates, based on the subject’s 3D images, were first spatially normalized by non-linearly registering each of them to the standard SPM T1MRI template [2]. Images were segmented into GM, WM and CSF portions using the combined pixel intensity, and finally smoothed with an isotropic Gaussian kernel 8 mm full width at half maximum (FWHM). Regarding the OVBM protocol, we performed the following steps based on previous studies [9,10]: (first) Segmentation of the original 3D images. This step allows the calculation of TBV, GM, WM, and CSF global volumes in millilitres. (second) Normalization of the segmented GM and WM images according to their corresponding tissues template (GM of WM depending on the aim of the specific analysis), thus preventing any contribution of non-tissue voxels and achieving optimal spatial normalization. (Third) Segmentation of the spatially normalized images to reject remaining non-tissue voxels (scalp, skull or venous sinuses). (Fourth) Modulation of the images by the Jacobian determinants derived from their spatial normalization step. This introduces intensity changes in the GM/WM images according to the variation in structural regional volumes that the normalization process
Table 1 Clinical and demographic data
N
Boys Girls
Age
OCD
Controls
13 5
13 5
μ 12.86 (S.D. 2.76) years
μ, 13.03 (S.D. 3.04) years 15 2 1
Laterality
Right-handed Left-handed Cross-dominance
15 2 1
Drug-treatment
SSRIs Sertraline Citalopram Clomipramine Drug-na¨ıve
n, 6; μ, 13.1 (S.D. 12.8) months n, 3; μ, 6.4 (S.D. 3.0) months n, 1; μ, 2.3 months n, 8 Range, 4–98; μ, 60.53 (S.D. 33.7)
STAI/T-C CYBOCS
Obsessions Compulsions Total
μ 10.28 (S.D. 4.28) μ 10.94 (S.D. 4.27) μ 21.39 (S.D. 5.88)
This table summarizes clinical and demographical data of our sample. μ, Mean values; S.D., standard deviation; n, number of patients; STAI/T-C, State Trait Anxiety Inventory for Children.
220
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
Fig. 1. GM and WM reductions in OCD children. Statistically significant regions of GM (A) and WM (B) decrease in OCD children. Images are displayed according to neurological orientation. Colors indicate the t-score. See also Table 2.
produced. The Jacobian-modulated GM step allows making inferences about differences in volumes rather than concentrations. (Fifth) Smooth with a Gaussian kernel of 12-mm FWHM. In order to study global volumetric measures, we performed two-tailed t-test means comparisons for independent samples (threshold at p < 0.05). Regarding regional morphometric analysis, smoothed GM and WM images were analyzed performing two one-tailed t-tests (OCD> controls and controls > OCD) to identify the specific direction of the between-groups differences. Additionally, in the OCD group, we performed two linear correlations between: (a) GM volume and total CY-BOCS and (b) GM volume and STAI/T-C. Results from the regional morphometric analysis and the linear correlations were thresholded at p < 0.001 (uncorrected for multiple comparisons); only clusters larger than 40 voxels were analyzed. Provided that our groups were gender, age and handedness matched, we did not apply TBV covariation since we were interested in finding absolute brain differences between the samples; we considered important to minimize type II rather than type I error, this being the first VBM study in OCD children. As shown below, we did not find significant differences between groups concerning GM volume in the BG (thalamus, caudate, putamen and globus pallidus). Due to this fact, secondary hypotheses were tested in order to see whether age or drug treatment could be modulating the volume in these regions. For this purpose, we lowered our threshold to p < 0.005 uncorrected and performed post-hoc small volume corrected (SVC) correlations between: (a) BG–GM volume and age and (b) BG–GM volume and pharmacological treatment duration. In order to isolate the effect of drug treatment we also performed a correlation between treatment duration and BG, covarying out the possible effect of age. Results were thresholded at p < 0.05 Family Wise Error (FWE)-SVC. Concerning global differences in GM, WM, CSF and TBV, mean comparisons of global volumetric measures showed that OCD children had a GM total volume reduction of 48.75 ml
(p = 0.01), representing 5.93% decrement (95% CI: 1.49– 10.36% or 12.3–85.2 ml). Regional differences in GM and WM segment volumes are shown in Table 2 and Fig. 1. There were no regions of increased GM or WM in the OCD group in comparison with the control group. OCD brains presented GM reductions in frontal lobe (bilateral middle and inferior triangularis areas, right frontal superior and frontal inferior operculum, and left rolandic operculum) and cingulate cortex (bilateral middle cingulate cortex and bilateral precuneus) as well as WM reductions in frontal (bilateral inferior operculum and left frontal middle gyrus) and parietal regions (right parietal superior). The GM reduction located in the right frontal middle cortex shows a p-value of 0.035 corrected for multiple comparisons according to FWE criteria. The FWE-correction prioritizes specificity over sensitivity of the results exerting a strong control over false positive values. Regarding correlation between regional GM volumes and clinical severity, a significant negative correlation was found between symptoms severity (Y-BOCS) and bilateral hippocampi GM volumes (see correlation in Table 2 and Fig. 2). No significant correlations were identified between GM volume and STAI/T-C. Concerning post-hoc BG correlations, we found a significant correlation between age and left caudate volume in the OCD group (p = 0.037 FWE-SVC; cluster size 148 voxels), but no age correlation in the control group. In the OCD group, we also identified positive correlations between drug treatment duration and left putamen volumes (p = 0.005 FWE-SVC; cluster size 375 voxels), right caudate (p = 0.014 FWE-SVC; cluster size 309 voxels) and bilateral thalamus (left, p = 0.049 FWESVC; cluster size = 1170 voxels; right, p = 0.003 FWE-SVC; cluster size = 1386 voxels). None of these correlations survived p-value of 0.05 FWE-SVC once we covariate for age’s effect. Summarizing the main results, our OCD patients showed a 5.93% reduction of total GM. The VBM analysis showed decreased bilateral GM in frontal and cingulate regions as well as decreased WM in bilateral frontal and right parietal.
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
221
Table 2 Voxel-based morphometry results Structure (Brodmann area) GM Frontal lobe R frontal mid. (45) R frontal inf. tri. (45)
R frontal mid. (6) R frontal inf. oper. (44) R frontal mid. (9) R frontal inf. oper. (44) R fontal sp. (8) L fontal mid. (45) L frontal inf. tri. (45)
L frontal inf. tri. (48)
L rolandic oper. (48) Cingulate cortex R cing. (23)
R precuneus
L cing. (23–24)
L precuneus WM Frontal lobe R frontal inf. oper. L frontal mid. L frontal inf. oper. Parietal areas R parietal sup. Correlation (GM/CY-BOCS) Limbic regions R hippocampus L hippocampus
MNIa (x, y, z)
Peakb Z-score
p-value (FWE)-correctedc
Cluster size (voxels)d
40, 39, 21 53, 27, 21 56, 23, 20 50, 31, 19 41, 5, 42 44, 13, 35 43, 13, 31 25, 29, 42 27, 25, 43 46, 7, 22 26, 25, 59 −41, 31, 34 −48, 28, 20 −49, 35, 26 −49, 31, 24 −47, 28, 27 −45, 25, 23 −44, 23, 30 −48, 21, 29 −52, 3, 16
4.39 3.42 3.31 3.30 3.48 3.41 3.36 3.41 3.38 3.41 3.40 3.62 3.33 3.29 3.22 3.19 3.42 3.14 3.10 3.23
0.035 0.594 0.700 0.715 0.526 0.846 0.874 0.601 0.636 0.601 0.611 0.391 0.680 0.720 0.781 0.809 0.591 0.845 0.871 0.771
2104
1, −26, 50 15, −37, 43 1, −31, 47 2, −9, 46 3, −13, 48 1, −18, 46 7, −11, 49 1, 13, 42 16, −43, 42 15, −53, 41 16, −48, 42 −2, −10, 47 −1, −14, 45 −4, −16, 48 −8, −41, 45 −14, −42, 45
4.04 3.91 3.89 3.69 3.69 3.63 3.62 3.43 3.55 3.48 3.39 3.75 3.64 3.58 3.14 3.46
0.119 0.183 0.191 0.329 0.330 0.382 0.391 0.578 0.464 0.527 0.625 0.285 0.376 0.430 0.849 0.552
9399
48, 21, 12 −27, 54, 3 −41, 18, 12
3.42 3.39 3.23
0.359 0.387 0.531
164 76 113
22, −50, 59
3.24
0.523
92
23, −31, −5 −19, −25, −8
3.53 3.77
0.624 0.397
237 626
318
236 292 130 1961
70
Regions of decreased GM and WM in OCD children in comparison with healthy matched controls. The lower part of the table shows GM regions correlating with CY-BOCS scores. L, left; R, right; mid., middle; inf., inferior; sup., superior; cing., cingulate; tri, triangularis; oper., operculum. a Montreal Neurological Institute (MNI) coordinates of the voxel of maximal statistical significance within each region. b Peak Z–score for the voxel of maximal statistical significance within each region. c Statistical significance after multiple comparison family-wise error (FWE), which rejects false positives. d Total number of contiguous voxels in each region that outlive p < 0.001. Given the voxel size (0.94 × 0.94 × 2) each voxel represents 1.77 mm3 .
These regions have been previously involved in action and conflict monitoring [28]. Provided that there are no previous VBM studies on children, our results cannot be compared with other pediatric studies. However, our cingulate results coin-
cide with the three VBM-OCD studies [13,19,26] and the right frontal middle GM reduction (p = 0.035 FWE-corrected) was also reported in two of them [13,19]. Our frontal and cingulate findings may nevertheless have an important bearing on OCD
222
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
Fig. 2. CY-BOCS correlation. Negative correlation (thresholded at p < 0.001) between CY-BOCS scores and GM volume in bilateral hippocampi. Sided scatter plots show the correlation between the GM volume (y-axis) of left (LH) and right (RH) hippocampus and CY-BOCS scores (x-axis). Gray dots represent real values and black dots adjusted values.
pathophysiology. Neuropsychological studies have reported that OCD patients show deficits related to frontal cognitive abilities, such as executive functioning and cognitive-behavioral flexibility [6]. In fMRI studies [28] frontal (BA 9, 44, 45, and 46) and cingulate cortices (BA 23 and precuneus) have been found to be active during conflict monitoring tasks. Moreover, specific OCD fMRI studies [15] have shown a different pattern of activation of these regions in comparison to controls during action monitoring tasks. These results support an initial deficit in the action monitoring system in OCD children. In this sense, the typical OCD feeling of “not just right” could be interpreted as an exaggerated or false error detection [18], and thus partially explained by our findings. Indeed, cingulate cortices have been associated with overactive error monitoring, as measured by the error-related negativity in ERP studies [7,27] as well as in fMRI studies [5,25]. Apart from these frontal and cingulate findings, our study is the first to detect a significant negative correlation between CYBOCS scores and bilateral GM volumes in the parahippocampal region. The negative correlation with CY-BOCS scores points to an association between clinical severity and hippocampal damage at the onset of the illness. Two hypotheses can be advanced to explain this finding. On the one hand, the hippocampal abnormalities could be related to memory OCD deficits. There is strong evidence that OCD is associated with low memory confidence as well as memory biases towards threatening information [17,23,24], specifically when it concerns the ability to recall if a behavior was actually performed or merely imagined [16]. On the other hand, bilateral smaller hippocampi have also been found in other psychiatric conditions, and hence it suggests that hippocampal reduction is more a case of neurodevelopmental damage than an illness-specific abnormality [14]. The main discrepancy of our results with previous OCD structural studies is the lack of BG deficits in our sample. The BG, especially caudate nucleus and putamen, have been extensively associated to routine performance and reinforcement [4,11,29]. This fact has made these regions (together with its OFC connections) suitable candidates to explain OCD pathophysiology. Two possible hypotheses could explain our lack of BG results: (a) the dynamic pattern of the illness and (b) the duration of pharmacological treatment. In relation to the developmental nature
of the illness, behavioral OCD pathological habits, such as the repetition of routines, could specifically affect the normal development of the BG. In fact, the largest structural OCD study to date [19] found a positive correlation between age and striatal abnormalities, which was interpreted as the anatomical expression of enduring striatal dysfunction. Therefore, the absence of striatal alterations [22] – or even the finding of a reduction, rather than an increase of striatal volumes [21] – could be expected in pediatric populations. Regarding the time of pharmacological exposure, Valente’s VBM study [26] found a negative correlation between duration of treatment and bilateral GM volumes of the head of caudate nuclei. To see whether our study supports any of these two hypotheses, we performed age and drug-treatment BG correlations. We found that subjects’ age positively correlated with left caudate volume in the OCD group but not in the control group. Concerning pharmacological treatment duration, we observed positive correlations between drug exposure duration and left putamen, right caudate and bilateral thalamus that do not pass multiple-comparison correction once we covariate for age effect. Thus, we did not find the negative correlation previously reported by Valente [26]. One should take into account that the mean treatment duration in Valente’s study is clearly smaller (mean = 47.6 days, S.D. = 41.9; and range 3–120) than our subjects’ mean treatment duration (10.04 months; S.D. = 10.05; and range 0.2–33.8). This difference in drug-treatment duration, in addition to the young age of our sample, could be partially responsible for the discrepancies concerning the effect of pharmacological time exposure on caudate volumes. However, our drug treatment positive correlation do not support 0.05 FWESVC once we covariate for age affect, which seems to indicate that the duration of pharmacological exposure is, by itself enable to explain BG results. Future studies could clarify this point. As a conclusion, our study pointed to GM and WM early OCD abnormalities associated with conflict monitoring circuits. We also found a significant negative correlation between CY-BOCS scores and bilateral GM volumes in the parahippocampal region. Other deficits previously found in adults (striatum and OFC) are absent in our study. However, according to our results, it seems that the dynamic nature of the illness could be modulating the normal striatum development in OCD children. Finally, given
S. Carmona et al. / Neuroscience Letters 421 (2007) 218–223
that this is the first VBM study in OCD children our analyses gave priority to sensitivity over specificity, and thus they should be taken with caution waiting for future replications.
[13]
Acknowledgments [14]
We would like to thank all the children and parents who participated in this study, as well as Mercedes Riba for her collaboration, Joseph Hilferty for his editorial review, and Adolf Tobe˜na and Jamil Zaki for their scientific advice. Susanna Carmona was supported by a FPU grant from the “Ministerio de Educaci´on y Cultura”.
[15]
[16]
References [1] B. Aouizerate, D. Guehl, E. Cuny, A. Rougier, B. Bioulac, J. Tignol, P. Burbaud, Pathophysiology of obsessive–compulsive disorder: a necessary link between phenomenology, neuropsychology, imagery and physiology, Prog. Neurobiol. 72 (2004) 195–221. [2] J. Ashburner, K.J. Friston, Voxel-based morphometry—the methods, Neuroimage 11 (2000) 805–821. [3] E.H. Aylward, G.J. Harris, R. Hoehn-Saric, P.E. Barta, S.R. Machlin, G.D. Pearlson, Normal caudate nucleus in obsessive–compulsive disorder assessed by quantitative neuroimaging, Arch. Gen. Psychiatry 53 (1996) 577–584. [4] J.J. Canales, A.M. Graybiel, A measure of striatal function predicts motor stereotypy, Nat. Neurosci. 3 (2000) 377–383. [5] K.D. Fitzgerald, R.C. Welsh, W.J. Gehring, J.L. Abelson, J.A. Himle, I. Liberzon, S.F. Taylor, Error-related hyperactivity of the anterior cingulate cortex in obsessive–compulsive disorder, Biol. Psychiatry 57 (2005) 287–294. [6] L. Friedlander, M. Desrocher, Neuroimaging studies of obsessive– compulsive disorder in adults and children, Clin. Psychol. Rev. 26 (2006) 32–49. [7] W.J. Gehring, J. Himle, L.G. Nisenson, Action-monitoring dysfunction in obsessive–compulsive disorder, Psychol. Sci. 11 (2000) 1–6. [8] N.R. Giuliani, V.D. Calhoun, G.D. Pearlson, A. Francis, R.W. Buchanan, Voxel-based morphometry versus region of interest: a comparison of two methods for analyzing gray matter differences in schizophrenia, Schizophr. Res. 74 (2005) 135–147. [9] C.D. Good, I. Johnsrude, J. Ashburner, R.N. Henson, K.J. Friston, R.S. Frackowiak, Cerebral asymmetry and the effects of sex and handedness on brain structure: a voxel-based morphometric analysis of 465 normal adult human brains, Neuroimage 14 (2001) 685–700. [10] C.D. Good, I.S. Johnsrude, J. Ashburner, R.N. Henson, K.J. Friston, R.S. Frackowiak, A voxel-based morphometric study of ageing in 465 normal adult human brains, Neuroimage 14 (2001) 21–36. [11] M. Haruno, M. Kawato, Heterarchical reinforcement-learning model for integration of multiple cortico-striatal loops: fMRI examination in stimulus-action-reward association learning, Neural. Netw. 19 (2006) 1242–1254. [12] D.H. Kang, J.J. Kim, J.S. Choi, Y.I. Kim, C.W. Kim, T. Youn, M.H. Han, K.H. Chang, J.S. Kwon, Volumetric investigation of the frontal-subcortical
[17] [18] [19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27] [28]
[29]
223
circuitry in patients with obsessive–compulsive disorder, J. Neuropsychiatry Clin. Neurosci. 16 (2004) 342–349. J.J. Kim, M.C. Lee, J. Kim, I.Y. Kim, S.I. Kim, M.H. Han, K.H. Chang, J.S. Kwon, Grey matter abnormalities in obsessive–compulsive disorder: statistical parametric mapping of segmented magnetic resonance images, Br. J. Psychiatry 179 (2001) 330–334. J.S. Kwon, J.J. Kim, D.W. Lee, J.S. Lee, D.S. Lee, M.S. Kim, I.K. Lyoo, M.J. Cho, M.C. Lee, Neural correlates of clinical symptoms and cognitive dysfunctions in obsessive–compulsive disorder, Psychiatry Res. 122 (2003) 37–47. N. Maltby, D.F. Tolin, P. Worhunsky, T.M. O’Keefe, K.A. Kiehl, Dysfunctional action monitoring hyperactivates frontal-striatal circuits in obsessive–compulsive disorder: an event-related fMRI study, Neuroimage 24 (2005) 495–503. R.J. McNally, P.A. Kohlbeck, Reality monitoring in obsessive–compulsive disorder, Behav. Res. Ther. 31 (1993) 249–253. J. Muller, J.E. Roberts, Memory and attention in Obsessive–compulsive Disorder: a review, J. Anxiety Disord. 19 (2005) 1–28. R.K. Pitman, A cybernetic model of obsessive–compulsive psychopathology, Compr. Psychiatry 28 (1987) 334–343. J. Pujol, C. Soriano-Mas, P. Alonso, N. Cardoner, J.M. Menchon, J. Deus, J. Vallejo, Mapping structural brain alterations in obsessive–compulsive disorder, Arch. Gen. Psychiatry 61 (2004) 720–730. D. Robinson, H. Wu, R.A. Munne, M. Ashtari, J.M. Alvir, G. Lerner, A. Koreen, K. Cole, B. Bogerts, Reduced caudate nucleus volume in obsessive–compulsive disorder, Arch. Gen. Psychiatry 52 (1995) 393–398. D.R. Rosenberg, M.S. Keshavan, K.M. O’Hearn, E.L. Dick, W.W. Bagwell, A.B. Seymour, D.M. Montrose, J.N. Pierri, B. Birmaher, Frontostriatal measurement in treatment-naive children with obsessive–compulsive disorder, Arch. Gen. Psychiatry 54 (1997) 824–830. P.R. Szeszko, S. MacMillan, M. McMeniman, S. Chen, K. Baribault, K.O. Lim, J. Ivey, M. Rose, S.P. Banerjee, R. Bhandari, G.J. Moore, D.R. Rosenberg, Brain structural abnormalities in psychotropic drug-naive pediatric patients with obsessive–compulsive disorder, Am. J. Psychiatry 161 (2004) 1049–1056. F. Tallis, The neuropsychology of obsessive–compulsive disorder: a review and consideration of clinical implications, Br. J. Clin. Psychol. 36 (Pt 1) (1997) 3–20. F. Tallis, P. Pratt, N. Jamani, Obsessive–compulsive disorder, checking, and non-verbal memory: a neuropsychological investigation, Behav. Res. Ther. 37 (1999) 161–166. S. Ursu, V.A. Stenger, M.K. Shear, M.R. Jones, C.S. Carter, Overactive action monitoring in obsessive–compulsive disorder: evidence from functional magnetic resonance imaging, Psychol. Sci. 14 (2003) 347–353. A.A. Valente Jr., E.C. Miguel, C.C. Castro, E. Amaro Jr., F.L. Duran, C.A. Buchpiguel, X. Chitnis, P.K. McGuire, G.F. Busatto, Regional gray matter abnormalities in obsessive–compulsive disorder: a voxel-based morphometry study, Biol. Psychiatry 58 (2005) 479–487. V. van Veen, C.S. Carter, The anterior cingulate as a conflict monitor: fMRI and ERP studies, Physiol. Behav. 77 (2002) 477–482. V. van Veen, J.D. Cohen, M.M. Botvinick, V.A. Stenger, C.S. Carter, Anterior cingulate cortex, conflict monitoring, and levels of processing, Neuroimage 14 (2001) 1302–1308. Z.M. Williams, E.N. Eskandar, Selective enhancement of associative learning by microstimulation of the anterior caudate, Nat. Neurosci. 9 (2006) 562–568.