Recent structural studies using conventional volumetric MRI yielded conflicting results regarding TS-related regional brain volume alterations. To avoid influences from age, gender, co-morbidities, and medication, we investigated a carefully selected group of TS patients by means of VBM and – for the first time – by MTI analyzed on a voxel-by-voxel basis. Tissue alterations in comparison to controls were detected in the supplementary motor area (SMA), the premotor cortex, the sensorimotor cortex, the prefrontal cortex, the left caudate nucleus, and the ACC. Tic severity was negatively correlated with tissue alterations in the orbitofrontal cortex, parts of the parietal-temporal-occipital association cortex bilaterally, the SMA, and the right ACC.
Current concepts of basal ganglia loops suggest different parallel circuits that connect frontal association areas with basal ganglia. It is thought that these circuits are involved in the selection, programming, initiation, and control of movement . The hypothesis of an involvement of these loops in TS  is further corroborated by our results as we detected changes in several parts of these circuits. In contrast to other volumetric studies, but in agreement with data obtained from functional investigations [16, 31, 32], we detected the most prominent changes not in the basal ganglia but in the prefrontal cortex and the ACC. Our data, therefore, are in line with Peterson's measurements  in adult patients demonstrating decreased frontal brain volumes. However, these results are in contrast to a recent VBM study  in adult TS patients that demonstrated increased regional GMV in the left midbrain and failed to detect alterations in other cortical areas. In this study, however, white matter volumes were not assessed. These discrepancies might be related to differences in the cohorts of patients and methodological factors within the data analysis process . Data from all other aforementioned volumetric MRI studies – although several of them as well demonstrated frontal lobe abnormalities – cannot be directly compared to our results as most data were obtained from children.
Miscellaneous studies are in line with our findings demonstrating most significant changes in frontal areas [16, 31–33]. Therefore, it has been suggested that TS is associated with abnormal pattern-related increases in metabolic activity in different frontal regions that are involved in the execution of movement (e.g., sensorimotor and lateral premotor cortex, SMA) . Thus, anomalous frontal lobe associational and projection fiber bundles may be the basis for basal ganglia abnormalities in TS . This assumption, in turn, is in agreement with the finding that the dorsolateral prefrontal cortex projects primarily to the dorsolateral head of the caudate nucleus .
Only a limited number of volumetric MRI studies is available investigating limbic regions in TS. Disturbances in the anterior cingulate (limbic) cortex, however, are likely to play a critical role in TS pathology [34, 35]. Firstly, stimulation of the ACC is associated with the generation of involuntary vocalizations. Secondly, cingulate epilepsy is combined with complex motor automatisms. Thirdly, the ACC has numerous interconnections with the prefrontal and orbitofrontal cortex, motor systems, other limbic regions, and the striatum – regions that are suggested to be involved in TS pathology. Fourthly, the ACC has an important role not only in movement, but also in the initiation and motivation of goal-directed-behaviour. Fifthly, activation or dysfunction of the ACC can be associated with aberrant social behaviour and psychopathic behaviours such as OCD [36, 37]. From functional neuroimaging studies it is suggested that tic generation is caused not only by basal ganglia and frontal cortex dysregulation but also by alterations in the cingulate gyrus [16, 32, 34]. While an increased ACC activity has been found during tic suppression , an ACC hypoperfusion was detected at rest without tic suppression . A positive correlation between tic frequency and ACC activity has been demonstrated . Using VBM, GMV reductions in the left hippocampal gyrus have been detected in a boys-only group . Our results of reduced gray and white matter cingulate volumes and a negative correlation between tic severity and tissue alterations in the right ACC are in agreement with these data.
Since the ACC seems to be physiologically under control of the prefrontal cortex , it can be speculated that in TS, abnormalities in frontal lobe fiber bundles result in a desinhibition of the ACC. A subsequent autonomous discharge of the ACC might cause motor and vocal tics, OCD, and other behavioural problems. Tics are often stimulated by simply thinking about tics or by surrounding factors. These specific characteristics of tics might be well explained by a dysfunctional ACC, since it is thought that the integration of thought, motivation, and emotion with movement is an important cingulate function .
Regarding basal ganglia volumes our results are in agreement with data from Peterson's large study demonstrating reduced caudate nuclei volumes in adults, although we detected changes only on the left side . Comparable to our findings, only left-sided caudate abnormalities have been reported in several positron emission tomography (PET) and single photon emission computed tomography (SPECT) studies [32, 34, 39]. It can be assumed that basal ganglia volumes are increased in children but decreased in adults . However, it remains unclear whether caudate volumes decrease with age or whether reduced volumes represent a compensatory response to the presence of tics. Since the striatum is involved not only in the initiation and execution of movements but also in behavioural functions  it is conceivable that the complex clinical manifestation of TS – at least in part – is caused by a dysfunctional caudate nucleus.
Our finding of reduced WMV in the corpus callosum is in agreement with one study in adult TS patients  but contrasts to other studies [3, 41]. In adults an increased corpus callosum size may be associated with co-morbidities .
In addition, we found a significantly lower amount of regional GMV in the parietal lobe on the left side. In both, VBM and MTI, YGTSS correlated negatively with parts of the parietal-temporal-occipital association cortex bilaterally. Comparable results has been described before using functional MRI (fMRI) , PET , and SPECT . Reduced metabolic activation and hypoperfusion in the medial and lateral temporal region were found to be related to the severity of tics [30, 32, 34]. Furthermore, functional abnormalities in the superior and inferior parietal lobules and the parietal operculum have been described before and at tic onset, respectively [16, 30, 32, 42]. In the parietal-temporal-occipital association cortex higher perceptual functions relate to somatic sensations, hearing and vision to form complex perceptions from these different sensory modalities. The superior parietal lobule functionally connects to the SMA and the prefrontal cortex and is thought to perform complex integrative functions related to the organization and initiation of movement . Therefore, abnormal connections between the superior parietal lobule, premotor and limbic regions may underlie premonitory sensory urges preceding tics in TS .
Recently, an increase of the regional fractional anisotropy (an indicator of microstructural integrity of brain tissue) has been shown in the white matter underlying the left somatosensory cortex in adult TS patients compared to controls . Additionally, this region showed an inverse linear relationship with tic severity. This is well in line with our finding of increased WMV in the sensorimotor cortex, and is in good agreement with the concept of an involvement of somatosensory pathways in TS. Apart from a recent structural MRI study which demonstrated thinning of the sensorimotor cortex in children with TS , behavioural and neurophysiological studies provide further evidence for abnormal sensory-motor processing in TS patients [45, 46]. These structural alterations may represent an adaptive response of the somatosensory system to abnormal input from fronto-striatal circuits.
Some limitations of this study have to be addressed. One might argue that in TS patients tic severity typically fluctuates over time whereas the tic score on a scale such as the YGTSS relates to symptoms present over the last weeks prior assessment. However, for different reasons we believe that our tic rating indeed reflects disease severity and, therefore, is suitable for a correlation in an MRI morphometric study: (1) the mean age of our sample was quite high (mean age = 30.4) and it is well known that in adults at that age tics do not fluctuate as much as in children and adolescents, (2) since in our sample mean tic severity was quite moderate (28.8 according to YGTSS), patients were able to remain either drug-naïve or drug-free for a longer period of time. This fact further supports the assumption that the patients' tics did not change markedly during the past.
From a methodological point of view, there are potential problems in the use of voxel-by-voxel analysis. First, this analysis was originally intended for use in large samples and requires smoothing of the images, with loss of resolution for small structures. Furthermore, the large number of comparisons requires corrections that greatly reduce the power of the study. This could explain why we did not find significant VBM and MTR differences between TS and controls on a whole brain analysis. Secondly, the small size of our group makes the results vulnerable to type I or type II errors, although in recent VBM and MTI studies [18, 26] a similar number of patients was suitable to detect regional differences compared to normal controls. Thirdly, the location of the MTI and VBM abnormalities overlapped in some regions but was not identical. This apparent inconsistency could be due to the use of a volume of interest (SVC) that, while increasing the power of the analysis by reducing the number of comparisons, may exclude potentially abnormal areas. Furthermore, the segmented VBM images, with a slice thickness of 1 mm, have higher resolution and ability to differentiate white and gray matter abnormalities than the MTI images with a slice thickness of 3 mm. In addition, segmented gray and white matter images have a smaller search volume, which affects sensitivity.
The MTR is largely dependent on the macromolecular density of cell membranes and phospholipids, and gray matter MTR decreases in our TS patients are likely to reflect decreases in the size and number of neurons and dendritic density, while those in the white matter are likely to reflect myelin changes and/or reduced axonal density . The neuropathological counterparts of our VBM findings are less clear. Potential correlates include a simple change in cell size, growth or atrophy of neurons or glia, as well as changes in the intra-cortical axonal architecture (synaptogenesis). Previously, it has been suggested that disturbances in the maturation of fronto-striatal systems contribute to the development, persistence, and severity of tics in adult TS patients . This assumption is supported by the fact that MTR decreases in the prefrontal cortex, which maturates later than other regions , could be indicative for alterations in the degree of myelinization. Furthermore, one might argue that results obtained from adult patients, as in the present study, reflect adaptive mechanisms to compensate for impairments in other brain regions. In line with this, recent studies also suggested that some pathological abnormalities as seen in VBM studies could be due to adaptive neuronal plasticity . However, the problem of separating cause and effect in cross-sectional studies is presently unsolvable.