Subjects

21 TS patients (age at imaging: 7.90±1.95 yrs; range: 5~11 yrs; 1 female) were recruited from outpatient clinics in our hospital from November 2009 to April 2010. All met DSM-IV-TR (Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, text revision) criteria for TS. 20 age and gender matched health controls (age at imaging: 8.05±2.30 yrs; range: 4-10 yrs; 3 females) were included in this study. All controls were healthy, with no history of tic disorder, OCD, or ADHD. All participants were right-handed. Eight patients who had previously taken medications were drug-free for at least 1 month prior to entering the study [18]. The mean score of disease severity for all patients was 41.71±12.46 (range, 17.5-76.1) as documented with a Chinese translation of the Yale Global Tic Severity Scale (YGTSS) [28]. Clinical interview and the Children’s Yale-Brown Obsessive Compulsive Scale (CY-BOCS) [29] were used to diagnose OCD. The German short version of Wender Utah rating scale (WURS-k, translated to Chinese) [30] was used to diagnose ADHD. One TS patient had a history of uterine-incision delivery. The duration of TS ranged from 1 month to 5 years ([mean±STD]: 1.84±0.56 yrs). For those who had course less than 1 year, TS diagnosis was made by follow-up call. After the study was approved by Beijing Children’s Hospital review board, written informed consent was obtained from all the parents/guardians according to the Declaration of Helsinki. Details of the patients are shown in Table 1.

Data acquisition

Magnetic resonance imaging was acquired using a 1.5T MR scanner (Gyroscan Interna Nova, Philips, Netherland). Head positioning was standardized using canthomeatal landmarks. After acquisition of a T2 localizer scan (TR = 1400ms, TE = 100ms, NEX3), axial three-dimensional T1-weighted image (3D T1WI) and diffusion tensor imaging (DTI) were acquired from all the subjects. 3D T1-weighted imaging were performed with axial three-dimensional-Fast Field Echo (3D FFE) sequence with the following parameters: repetition time (TR) = 25 ms, echo time (TE) = 4.6 ms, flip angle = 30°, reconstructed image matrix = 256×256, field of view (FOV) = 23×23 cm, slice thickness = 1 mm. DTI was performed using the following protocol: spin-echo diffusion-weighted echo-planar imaging sequence, 5 mm slice thickness, no inter-slice gap, repetition time = 4268 ms, echo time = 93 ms, field of view (FOV) = 23×23 cm, reconstructed image matrix = 192×192. Diffusion MRI images were obtained from 15 non-collinear directions with a b value of 1000 s/mm².

Quantification of local volumetric changes

After the data acquisition, 3D T1-weighted Images processing was performed by statistical parametric mapping (SPM8, http://www.fil.ion.ucl.ac.uk/spm, Wellcome Department of Cognitive Neurology, London, UK, 2008) and executed in Matlab 7.9 (MathWorks, Natick, MA, USA). All of the 3D T1-weighted images were reoriented with the origin set close to the anterior commissure (AC) and then were segmented into GM, WM and CSF in native space with unified segmentation [31]. Afterwards, all the segmented GM and WM images were rigidly transformed to produce a series of aligned GM and WM images. The study-specific GM/WM templates were then created with the aligned serial images from all the subjects and during the template creation processing, all aligned images were warped to the template yielding a series of flow fields, which parameterized the deformation. Data normalization was done, which was followed by data modulation to correct volume changes. Since the previous processing was performed in native space, it was necessary to transform the modulated data into MNI space. After the transformation, all the images were smoothed with a 6-mm full-width at half-maximum (FWHM) Gaussian filter. Two-sample t-test was then applied to evaluate the abnormalities of GM/WM between groups, using SPM8. Two contrasts were defined to examine both decreased and increased regions in the patients. Resulting statistical parametric maps of VBM were derived at a significance level of p < 0.001, uncorrected with an extent threshold of 10 voxels. Then small volume corrections (SVC) [32] limited to the volume of the regions we were interested in were performed using a sphere of 10mm radius.

Quantification of microstructural changes

DTI images were transformed to nifti format in a workstation and processed off-line using FMRIB’s Diffusion Toolbox (FDT 2.0) within FSL 4.1 (http://www.fmrib.ox.ac.uk/fsl). For each subject, fifteen DTI volumes with b value of 1000 s/mm² were first affine registered to the b0 volume for correction of eddy current distortion and simple head motion. Non-brain voxels were removed using Brain Extraction Tool (BET) [33] of FSL; a fractional intensity threshold of 0.3 was selected, resulting in a brain-extracted 4D image and a binary brain mask for each subject. The 4D image and corresponding brain mask created by BET were then used for fitting diffusion tensor model at each voxel using FDT. Eigenvalues of diffusion tensor matrix (λ₁, λ₂, λ₃) were obtained and maps of axial diffusivity (AD, or λ₁), mean diffusivity (MD = (λ₁+λ₂+λ₃)/3), and fractional anisotropy (FA) were generated. Radial diffusivity (perpendicular eigenvalue, λ₂₃ = (λ₂+λ₃)/2) was calculated by averaging the maps of λ₂ and λ₃. The standard TBSS [23,34] procedure was then applied to the data. First, all individual FA images were nonlinearly registered to the pre-defined FMRIB58_FA standard-space image provided by FSL, and affine-aligned into 1×1×1 mm MNI152 standard space. All subsequent processing was carried using this space and resolution. Then the mean of all FA images were created and fed into FA skeletonisation program to generate a FA skeleton, which represented the center of all tracts common to the subjects; a threshold of 0.2 was selected to define the borders of WM and GM. Each subject’s local maximum FA intensity along the perpendicular direction of white matter tract was projected to the mean FA skeleton to carry out the voxel-wise statistics across subjects. The same projection was applied to the MD, AD and RD images. Voxel-wise group comparisons of patients versus normal controls on the skeleton image were carried using FSL’s randomise tool. Randomise uses a permutation based statistical inference that does not rely on a Gaussian distribution [35]. Random Monte Carlo simulated samples of 10,000 permutations were used as null distribution. A statistical threshold p < 0.05, corrected for multiple comparisons with threshold-free cluster enhancement (TFCE) [36] method was used. TFCE can identify cluster-like structures without definition of an initial cluster-forming threshold or carrying out a large amount of data smoothing. The Johns Hopkins University JHU–ICBM-DTI-81 white-matter tractography atlas was used to identify the abnormal white matter tracts reveled by TBSS. TBSS results of MD, RD and AD were analyzed in the same manner.

Link changes to clinical measures

To further determine the relation between the alterations and clinical scores and duration of tics, Pearson’s correlation was calculated using SPSS to relate corresponding values of that ROI (abnormal regions found by VBM) to the YGTSS scores, controlling for age, gender, and the whole brain volume. For DTI data, the identified tracks showed abnormal AD values were selected as ROI. The mean AD values in each ROI were correlated with the clinical data (YGTSS scores and duration of illness) using Pearson’s correlation, controlling for age and gender. The same operations were applied to the MD and FA results.

Alterations in motor-sensory areas, CSTC and limbic structures

We found increased GMV and decreased WMV in the motor-sensory gyrus. This indicates that the main motor pathway—the corticospinal tract—is affected by TS. This, in turn, may have some relation to the conclusion that prefrontal cortical regions, sensory-motor pathways, and limbic structures take part in the modulation of tics [37]. AD values in right anterior thalamic radiation and cingulum bundle projecting to the cingulate gyrus were positively correlated with YGTSS scores. That is, more severe tics are corresponding to higher AD values or larger AD increase. The AD increase starts from the corona radiata and extends to the anterior thalamic radiation, conforming the results of WMV decrease in these areas. The CSTC circuits are believed to be connected directly to the ventral striatum where motivational and incentive behaviors are regulated [38]. The anterior thalamic radiation connects the medial thalamic nuclei and the prefrontal lobe and the cingulum bundle, which in turn, is of key importance in the tics’ pathophysiologic mechanism [39-41]. The AD increase in anterior thalamic radiation indicates the micro-structural changes in the CSTC. Forceps minor of corpus callosum takes part in the modulation of attention by transferring information between the hemispheres [42] and inhibition of motor cortical activities [43]. The AD and MD increase in the forceps minor may result from reduced cell density [9,44,45], indicating reduced interhemispheric connectivity via the corpus callosum, hence reduced interhemispheric inhibition or less long-range interaction control of the motor system.

Alterations suggesting regulation of tics symptoms

Tics are the hallmark of TS. It is known that interactions of the frontal cortex and striatum guide and regulate motor behavior [46,47]. Considerable evidence suggests that the cortical portion of these connections is involved in regulating the severity of tic symptoms. Regional volumes of the dorsal prefrontal and parietal cortex were found to be significantly larger in children and adults with TS compared with healthy controls [12]. In our research, the decrease of WMV in right superior frontal gyrus was inversely associated with YGTSS scores (r = -0.594, p = 0.009). TS patients with the smaller WMV displayed more tics. Previous brain morphological study results obtained in TS children are different from those obtained in adults [11,48]. Regional volumes of the dorsal prefrontal and parietal cortex were found to be larger in TS children than the controls, whereas it was significantly smaller in TS adults compared to the controls. The volume enlargement was seen as sign of compensatory function. Morphometry study results of other areas like hippocampus, amygdale, and corpus callosum were inconsistent, even when the subjects were all children [48]. This may have resulted from comorbid diseases and/or secondary impairments, which we were able to avoid. We reduced the probability of confounds by studying short disease duration TS children.

Cross-validation of morphological and microstructural changes

Using DTI we found that the white matter was wildly influenced in our TS children. WMV decrease cluster size as measured by VBM study was about ten times greater than the GMV decrease, and more WM than GM areas showed volume decrease. Furthermore, the only correlation of volume change with tic severity score (YGTSS) was WMV decrease in right frontal pole. Our results showed decreased WMV in frontal lobe of TS children which was consistent with Kates et al [19], but conflicted with Frederickson et al [18]. This discrepancy may have resulted from the difference in segmentation method and patients’ tic duration.

Our comprehensive analysis of brain white matter alterations in TS children demonstrates significant WM volume loss in different structures such as somatosensory cortex (BA3), primary motor cortex (BA4), frontal pole and fusiform gyrus, which contains corticospinal tract, left cingulum bundle and forceps minor projections from the frontal-parietal-temporal areas. Forceps minor is a WM pathway that connects the lateral and medial areas of the prefrontal cortex. The WM volume reduction along these fiber pathways was much greater than the grey matter loss in our study. When viewed in conjunction with the adult studies which reported mainly loss of grey matter [14,25], this suggests a difference in change pattern between TS children and adults. This conclusion is consistent with previous reports showing increased AD in the corticospinal tract and post- and pre-central gyrus found in TS adults [15,17]. No significant FA decrease was found. This may indicate AD and MD are more sensitive to microstructural changes in TS children than other DTI indices.

Pathological reasons and brain development of TS children

Evidences suggested that Tourette syndrome was caused by genetic variations [49,50]. Some children thus do not produce enough myelin, or they have metabolic problems that may cause white matter degeneration. Because cerebral WM plays a fundamental role in transmitting electrical signals between neurons both in terms of short and long-range connections in the brain, it may be that such white matter problems become only apparent (symptomatic) when the TS child matures. The lack of activation input that follows white matter loss or the morphological loss of axons may lead to both anterograde and retrograde degeneration of brain (“use it or lose it”), which also prevents normal GM function. The associated social isolation and the suppression of the symptoms by the TS child may make the tics even more severe, which would then be a non-developmental, secondary problem. Brain development during childhood is a complex and dynamic process for which the pathological mechanism of TS can be quite complicated.

Limitations

Despite these interesting findings, our study had some limitations. The number of our participants was too small to allow generalization of the radiological findings to other patients with TS. The TS severity score has a large range and so does the standard deviation of the correlations. Also, to save the scanning time, we applied 5mm slice thickness in DTI, which is not the preferred method. In addition, we did not expand our study to follow through to patients’ adulthood, making it difficult to answer why many TS patients improves as they enter adulthood.