Using DTI, we found microstructural changes in the white matter of prefrontal areas, the pars opercularis of the left inferior frontal gyrus, the cingulate gyrus, the medial premotor cortex, the left precentral gyrus, and the left putamen. For the following brain regions there was a positive correlation between tic severity and FA and ADC values, respectively: body of the CC, left thalamus, right superior temporal gyrus, putamen bilaterally, left parahippocampal gyrus, left cingulate gyrus, medial frontal gyrus bilaterally, and left precentral gyrus (Figure
2). In addition, we found a negative correlation between FA values and tic severity in the left superior frontal gyrus, medial frontal gyrus bilaterally, cingulate gyrus bilaterally, and the ventral posterior lateral nucleus of the right thalamus. Thus, most of those brain regions that showed abnormalities in structural organization, also demonstrated a correlation with tic severity. Furthermore, these results obtained by DTI are in broad agreement with our prior findings in the same group of patients using VBM and MTI
. In addition, most of these alterations have been described before in other DTI studies investigating adults with TS including the CC
[9, 10, 13], precentral gyrus
[11, 12], left putamen
, and limbic regions
Comparable to our recent VBM/MTI study in the same group of unmedicated adult patients with TS “only”, we detected the most prominent changes in different frontal areas, particularly in medial and inferior prefrontal areas (reduced regional FA and increased mean diffusivity) including the left pars opercularis (reduced regional FA) and the left precentral gyrus (increased mean diffusivity). These results are in line with several other studies using both functional and structural neuroimaging techniques
[1, 16]. Since our group of patients comprised only patients with TS “only” without comorbidities, alterations in these frontal regions in TS patients have to be related to the presence of tics (and not exclusively to the presence of comorbidities such as attention deficit hyperactivity disorder (ADHD) and obsessive compulsive disorder (OCD)). This assumption is corroborated by the fact that we found not only reduced regional FA and increased mean diffusivity in medial and inferior prefrontal areas, respectively, but also a positive correlation between tic severity and regional ADC values in the medial frontal gyrus bilaterally and the left precentral gyrus. Since, we detected, in addition, a negative correlation between FA values and tic severity in prefrontal areas bilaterally, it can be hypothesized that - at least in adult patients - these frontal regions are involved not only in tic generation but also in tic suppression or compensatory adaptation mechanisms resulting in tic decline with increasing age as suggested earlier
Our data further corroborate an involvement of the CC in the generation of tics in adult TS patients as suggested earlier: in both children and adults, larger CC volumes
[17–19] and reduced FA
[2, 9, 13, 14] have been reported as well as an altered structure-function relationship in the motor CC in adults with TS “only” using a combined TMS-DTI approach
. Since we found a positive correlation between tic severity and FA values in the CC, it can be speculated that CC alterations are correlated to neurodevelopmental abnormalities - resulting in reduced transcallosal inhibition of cortical neurons - rather than adaptive mechanisms to compensate for impairments in other brain regions
. This interpretation is further supported by the fact that even in children a negative correlation has been reported between tic severity and CC volumes
Comparable to our recent study in the same group of patients
, we detected abnormalities in the cingulate gyrus. Using VBM and MTI we found reduced gray and white matter cingulate volumes
; in this study using DTI reduced regional FA and increased mean diffusivity were obvious. We hypothesize that these changes represent secondary compensatory mechanisms, because (1) in both of our studies using three different MRI techniques we found a negative correlation between tic severity and changes in the cingulate gyrus, (2) in functional studies it could be demonstrated that the cingulate is activated during tic inhibition
[21–23], (3) the cingulate gyrus has numerous interconnections with the prefrontal and orbitofrontal cortex, motor systems, and the striatum – regions that have been suggested to be involved in tic generation
, and (4) there is evidence for an involvement of the cingulate gyrus in the initiation and motivation of goal-directed behaviours
. Furthermore, DTI studies in children failed to detect abnormalities in the cingulate gyrus.
In accordance with other DTI studies in adult TS patients, in addition, we found alteration in the thalamus
[11, 12] and the putamen
. These brain regions have been extensively described in association with TS pathophysiology in adults and children not only when using DTI
[4, 5], but also when using other MRI techniques such as VBM
[17, 25]. Accordingly, the thalamus is the most often used target for deep brain stimulation in adult patients suffering from severe treatment resistant TS
. Although some studies in both children with TS “only” (using VBM)
 and neuroleptic-naïve adults with TS or chronic motor tics with and without comorbidities (using large-deformation high dimensional brain mapping)
 failed to detect changes in the basal ganglia and thalamus, today there is little doubt that these brain regions are involved in the pathophysiology of tics.
As described before by Thomalla et al.
 in a group of 15 unmedicated adults with TS “only” and Draganski et al.
 (in a mixed group of TS patients with and without comorbidities and medication), we were able to detect altered FA in the precentral gyrus, a brain region that can be assigned to networks involved in sensory-motor processing
. However, in contrast to our results and those from other studies
[4, 13], Thomalla et al.
 and Draganski et al.
 found an increased (and not reduced) FA with ADC decrease in the somatosensory cortex region bilaterally and in other brain areas such as the thalamus
. These conflicting results could either be related to methodological differences in the different cross-sectional studies or be caused by different groups of patients (e.g. with respect to disease duration). This problem can only be solved by follow-up studies investigating whether the FA changes over time and may reflect compensatory structural changes in patients with TS. In addition, until today it is unclear which histological changes correlate to these alterations detected by DTI. The involvement of the precentral gyrus in the pathophysiology of tics is further supported by the fact that we found not only a reduced FA, but also a positive correlation between the ADC value in the left precentral gyrus (BA 44) and the tic severity.
Our finding of a reduced regional FA in the left pars opercularis should receive special attention. The inferior frontal gyrus consists of three distinct subparts: the pars opercularis, the pars triangularis, and the pars orbicularis. There is substantial evidence that the pars opercularis is not only involved in language and music (for review see
), but also in motor processing. In particular, the pars opercularis seems to be a key component in the human mirror neuron system that is activated during action observation and imitation
. In accordance with this assumption, a significant volume reduction of the pars opercularis could be demonstrated in patients suffering from high-functioning autism spectrum disorders suggesting that this brain region also plays an important role in social (dys-)function
. In addition, the role of the pars opercularis in inhibitory control has been demonstrated using repetitive transcranial magnetic stimulation (rTMS)
. In accordance with these results, structural abnormalities in the pars opercularis have been demonstrated in children suffering from ADHD suggesting that developmental abnormalities of the pars opercularis lead to inhibition difficulties
. In TS, therefore, it is conceivable that the pars opercularis is involved in inhibitory control mechanisms and in particular in the generation of complex motor and vocal tics such as copro- and echophenomena. Accordingly, it has been suggested that echophenomena might be caused by a dysfunction of the mirror neuron system
It should be mentioned that studies using newer techniques (for example TBSS, cortical thickness measurement, probalistic fiber tracking) resulted in additional findings demonstrating alterations in somatosensory pathways
[11, 14], long association fibre tracts such as the inferior fronto-occipitalis fascicle, the superior longitudinal fascicle and the fascicle uncinatus
 and cortical thinning in prefrontal and limbic structures
The following limitations of the study have to be addressed: Firstly, we investigated a carefully selected group of patients including only adult, unmedicated male with TS “only” without comorbidities. Since influences from sex, age, medication, and comorbidities can be excluded, it can be concluded that microstructural abnormalities detected in this study are indeed related to the tic disorder. Since the majority of adult patients with TS, however, suffer not only from tics, but also from one or more psychiatric comorbidities
, it has to remain open whether our group of patients is representative for adults with TS. It is still matter of discussion whether TS is a unitary or heterogeneous condition encompassing different phenotypes (with and without ADHD, aggressive behaviours, OCD, and coprophenomena)
. Secondly, comparable to most other neuroimaging studies in TS, our sample size was relatively small. However, using DTI a sample size of at least 15 subjects is regarded as suitable for reliable analyses. In all but one DTI study performed so far in adults with TS a smaller number of patients has been included. Thirdly, we demonstrated uncorrected data. However, our results were supported by recent findings in the same group of patients using both VBM and MTI
. Fourthly, in this study SPM2 was used to analyze the data. One might argue that a better segmentation and registration tool might provide more reliable findings. To address the issue of compensatory mechanisms in TS in more detail a prospective follow-up study is necessary.