Childhood-onset chronic motor and phonic tics characterise Gilles de la Tourette syndrome (GTS). GTS is commonly complicated by comorbid attention-deficit/hyperactivity disorder (ADHD) or obsessive–compulsive disorder. Since all patients with GTS have tics, abnormalities in the neural circuits relevant to movement control are common to uncomplicated and complicated GTS. In contrast, the function of different neural circuits relevant to the control of behaviour such as obsessions and compulsions, but also the behaviours typical for ADHD, may determine whether GTS patients suffer from comorbidity or not. Hence, we would expect these abnormalities in neural circuits to be more widespread in complicated GTS.

Transcranial magnetic stimulation (TMS) allows the non-invasive study of motor cortex excitability in awake humans. In GTS, TMS studies have demonstrated several alterations of motor cortex excitability, including reduced short interval intracortical inhibition (SICI), reduced short latency sensory afferent inhibition (SAI) and shortened cortical silent periods;1^–3 reduced intracortical inhibition has also been described in patients with OCD4 and ADHD with or without tics.5^–9 In the present study we have therefore examined whether any of the changes of motor cortex excitability differ between uncomplicated adult GTS patients and those complicated by ADHD or OCD.

MATERIAL AND METHODS

Patients and control subjects

Twenty-nine patients (25 men, mean age 31.3 years, range 18–68) with a DSM-IV diagnosis of GTS and 24 control subjects (16 men, mean age 32.5 years, range 21–58) were recruited consecutively from our GTS outpatient clinic. The onset of tics was in childhood (before OCD in those with OCD comorbidity) with a mean duration of GTS of 24 years (range 11–61). The diagnoses of GTS, OCD and ADHD were made according to DSM-IV-TR criteria. This included taking a thorough history of the patient’s childhood behaviour, together with an interview with their life partner, a parent or other close associate. The National Hospital Interview Schedule for Gilles de la Tourette syndrome (NHIS)10 was administered. This semistructured interview schedule allows specific and detailed information to be collected on motor and phonic tics, echophenomena (echolalia and echopraxia), paliphenomena (palilalia and palipraxia), coprophenomena (coprolalia and copropraxia), self-injurious and aggressive behaviour, obsessive–compulsive behaviour and neurodevelopmental comorbid disorders; we complemented the NHIS with the appropriate modules of the structured clinical interview for DSM-IV-TR axis one disorders (SCID-1) for obsessive–compulsive behaviour and depression. On the day of the experiments, the severity of tics was rated using the Yale Global Tic Severity Scale,11 and the Diagnostic Confidence Index assessed the lifetime history of symptoms indicative of GTS (table 1).12 No patient was on medication at the time of the study. All patients except two had not been taking medication for more than a year; one patient with OCD discontinued an SSRI, and one patient with ADHD stopped methylphenidate 4 weeks before the study.

Patients gave informed written consent according to the Declaration of Helsinki, and the Joint Ethics Committee of the Institute of Neurology and the National Hospital for Neurology and Neurosurgery approved the study protocol.

Electromyography recordings

Surface electromyograms (EMG) were recorded from the right first dorsal interosseous (FDI) muscle using silver/silver-chloride disc surface electrodes (1 cm diameter) in a belly tendon montage. The EMG signal was amplified and analogue-filtered (30 Hz to 3 kHz) with a Digitimer D150 amplifier (Digitimer, Welwyn Garden City, UK). Data (sampling rate 4 kHz) were digitised for off-line analysis using Signal software (Cambridge Electronic Devices, Cambridge). The peak-to-peak amplitude of motor evoked potentials (MEP), the area under the curve of the MEP and the silent period duration were measured with in-house software.

Transcranial magnetic stimulation

Patients and controls seated in a comfortable chair were asked to relax as much as possible. Subjects were asked to refrain from caffeine on the day of the experiment. No major irregularities of sleep, mood or other factors could be elicited by direct questioning. Patients had tics throughout the experiments but were asked not to suppress them. Magnetic stimuli were given with a hand-held figure-of-eight coil (outer winding diameter 9 cm) connected to a high-power Magstim 200 stimulator (Magstim Co., Whitland, Dyfed, UK). This stimulator generates a magnetic pulse with monophasic waveform inducing in the brain a current with posterior–anterior flow when the coil handle is positioned at an angle of 45° pointing backwards. The optimal spot for right FDI stimulation was marked with a felt pen.

Motor thresholds

Resting motor threshold (RMT) was defined as the minimum intensity needed to evoke an MEP of >50 μV in five out of 10 consecutive trials in the relaxed FDI. Active motor threshold (AMT) was defined as the minimum intensity needed to evoke an MEP of >200 μV in five out of 10 trials in the tonically active FDI (∼10% of maximal contraction as assessed visually on an oscilloscope). Thresholds were approached from above threshold in steps of 1% stimulator output. Once no MEPs could be elicited, the intensity was increased in steps of 1% stimulator output until a minimal MEP was observed. This intensity was taken as the motor threshold.

Paired pulse paradigm

In each individual, a TMS shock intensity was chosen that elicited an MEP of around 1 mV peak-to-peak amplitude. The conditioning pulse intensity was varied (60, 70, 80 or 90% of AMT) resulting in four different experimental blocks. With each conditioning pulse intensity and in a randomised order, the 2 and 3 ms interstimulus intervals (ISI) and the 12 ms ISI were examined. The former examine short-interval intracortical inhibition (SICI), and the latter intracortical facilitation (ICF). With an interval of 4 s between trials, 10 conditioned MEPs were collected for each ISI, and in each experimental block a total of 20 unconditioned MEPs were recorded. The order of data collection for each conditioning stimulus intensity was randomised between subjects. Trials recorded while the patients contracted the hand muscles or those coinciding with a tic were excluded on-line. No trials were excluded in the off-line analysis. The average of the amplitudes of each conditioned MEP was expressed as a percentage of the average test stimulus MEP amplitude in the same session. SICI thresholds were determined as described previously.13 In brief, in each subject, the %ICI or ICF was plotted against the absolute intensity of the conditioning stimulus, and the data were fitted with a second-order polynomial function. The theoretical threshold was defined as the value where the function crossed the x axis; thus the conditioning stimulus intensity where the net amount of inhibition was zero. This conditioning stimulus intensity was then related to AMT in each individual.

Short-latency afferent inhibition by somatosensory input from the median nerve

Short-latency afferent inhibition of the motor cortex was examined as previously described.14 In brief, a test MEP of ∼1 mV peak-to-peak amplitude was elicited in the FDI by TMS. A paired pulse paradigm examined the influence on MEP size of a supra-threshold electrical stimulus given to the median nerve through bipolar electrodes. The electrical stimulus to the median nerve was delivered at an intensity just above the threshold to elicit a visible contraction in the thenar muscles and preceded the TMS pulse to the FDI hot spot by 14, 18, 20, 22, 24, 26 or 29 ms. Twenty trials of the MEP elicited by TMS alone and 10 trials of conditioned MEPs for each ISI were collected. The amplitude of the MEP in the FDI was measured with inhouse software. The average amplitude of the conditioned MEP was expressed in percent of the average amplitude of the test MEP alone. Trials recorded while the patients contracted the hand muscles or those coinciding with a tic were excluded on-line. No trials were excluded in the off-line analysis.

Data analysis

For baseline data examining SICI, or ICF, we examined whether there was a main effect of “intensity” (60, 70, 80 or 90% AMT) on the amount of SICI, or ICF respectively using analysis of variance (ANOVA). For SICI, ICF or SAI, we tested whether there was a main effect of “ISI” on the size of the conditioned MEP using ANOVA. To test whether controls differed from GTS patients, we also used an ANOVA model examining a main effect of “group” on the size of the conditioned MEP. Possible associations of electrophysiological parameters with YGTSS tic scores were tested using backward stepwise regression analysis with tic ratings as the dependent variable. We only included those electrophysiological parameters that differed significantly between patients and controls. A parameter was removed from the model if the probability of its contribution was less than 0.1.

A statistical difference in the ANOVAs was followed by a post-hoc paired t test analysis without correction for multiple comparisons. Mauchly test was used to test for sphericity in the repeated-measures ANOVAs, and the Greenhouse–Geisser correction applied to the DFs if necessary. Statistical significance levels were set to p = 0.05. All statistical analysis was performed using SPSS 11 for Windows (SPSS, Chicago).

RESULTS

Patients

A total of 29 unmedicated GTS patients participated in the study; according to DSM-IV criteria, 18 had uncomplicated GTS, six had GTS+ADHD, and five had GTS+OCD. Twenty patients were part of a previous study2 where we reported on threshold and SICI data but not on SAI or ICF. The GTS+OCD patients had lower mean scores on the YGTSS motor, vocal and total ratings (table 1); however, this was not significant (ANOVA, p>0.1). Some patients with uncomplicated GTS, or GTS+OCD, also had ADHD symptoms (table 1). The electrophysiological parameters did not predict tic severity (backward stepwise regression analysis with total YGTSS score, motor or vocal YGTSS score as predicted variable). DCI scores were similar in the patient subgroups (table 1).

Motor thresholds

Patients had slightly higher RMT (mean 42.2 (SD 10) vs 38.8 (3.2), F_1,51 = 4.6, p = 0.037, ANOVA) and AMT (31.4 (6.4) vs 28.1 (5.6), F_1,51 = 4.0, p = 0.051) compared with controls. Thresholds in the clinical GTS subgroups, that is uncomplicated GTS, GTS+ADHD, or GTS+OCD, and controls were similar (ANOVA, no main effect of “subgroup,” table 1).

Short intracortical inhibition and facilitation

We extrapolated the theoretical threshold for SICI and ICF, respectively, as described in Material and methods.13 This was not possible in one control subject and two GTS patients for SICI and in four control subjects and two patients for ICF because the data were too variable for a meaningful regression analysis. These subjects were excluded from further statistical analysis. The threshold of both SICI and ICF expressed as percentage of each individual’s AMT was similar in controls and GTS patients (ANOVA, p>0.1). We then examined the effect on SICI, or ICF, of different conditioning stimulus intensities (CSIs) (60, 70, 80, 90% AMT). Increasing the conditioning stimulus intensity resulted in more SICI, or ICF (repeated-measures ANOVA, main effect of “conditioning stimulus intensity,” F_2.55,130.25 = 28.2, p<0.001 for SICI; F_3,153 = 3.6, p = 0.016 for ICF); the amplitudes of the unconditioned MEPs were similar independent of conditioning stimulus intensity or when comparing controls and GTS patients (repeated-measures ANOVA; no main effect of “conditioning stimulus intensity,” no interaction between “conditioning stimulus intensity” and “group,” p>0.1). GTS patients, however, had less SICI, and more ICF, compared with controls (ANOVA, main effect of “group” on data in fig 1A,B; F_1,204 = 5.5, p = 0.02 for SICI; F_1,204 = 7.64, p = 0.006 for ICF).

We then distinguished uncomplicated GTS, GTS+ADHD, and GTS+OCD and compared SICI, or ICF, with controls. SICI was similar in the GTS subgroups (ANOVA, no main effect of “group,” p>0.1).

However, ICF differed significantly between the groups (ANOVA, main effect of “group,” F_3,208 = 5.84, p = 0.001). Post-hoc pairwise comparisons revealed that patients with GTS+ADHD had more ICF than controls (p<0.001), uncomplicated GTS patients (p = 0.002) or GTS+OCD patients (p = 0.024); ICF was similar comparing all other groups (p>0.1) (fig 1C).

Short-latency sensory afferent inhibition

In controls and GTS patients, a suprathreshold electrical stimulus to the median nerve at the wrist before the TMS pulse to the FDI hot-spot reduced the mean MEP amplitude predominantly at ISIs of 20, 22 and 24 ms (fig 2A). Since the early period of inhibition is more likely to have a partly cortical origin than later timings,14 we assessed the maximum amount of afferent inhibition in each individual. GTS patients had significantly less inhibition than controls (ANOVA, main effect of group, F_1,50 = 6.48, p = 0.014, fig 2B).

Maximum SAI differed in the GTS subgroups (ANOVA, main effect of “subgroup,” F_3,48 = 3.86, p = 0.015); post-hoc multiple pairwise comparisons (fig 2C) showed that controls had significantly more SAI than uncomplicated GTS patients (p = 0.006) or GTS+ADHD patients (p = 0.045) but the same SAI as GTS+OCD patients (p>0.5). SAI of GTS+ADHD patients and uncomplicated patients were similar (p>0.5), and there was a trend towards less SAI than GTS+OCD patients (p = 0.07). Uncomplicated patients had less SAI than GTS+OCD (p = 0.041). Combining data of uncomplicated GTS with that of GTS+ADHD, and controls with that of GTS+OCD respectively, the GTS/GTS+ADHD group showed much less SAI compared with the controls/GTS+OCD group (ANOVA, main effect of group, F_1,50 = 11.94, p = 0.001, fig 2D).

DISCUSSION

In this study, we confirm that when we group all GTS patients together, they have less SICI and SAI than controls. New findings are that they have higher resting and active motor thresholds and increased ICF. For SICI and thresholds, the results were the same in uncomplicated and complicated (ADHD or OCD) patients. However, only patients with GTS+ADHD had more ICF than normal; conversely patients with GTS+OCD had normal SAI, while it was reduced in pure GTS and GTS+ADHD. Thus, there was a “gradient” of abnormalities in our subgroups: the group of GTS+ADHD patients showed abnormalities of all our measures of motor cortex excitability, patients with pure GTS had abnormal results in all measures apart from ICF, and patients with GTS+OCD had abnormal SICI and thresholds but normal levels of ICF and SAI.

In previous reports, others and we had found that thresholds were the same in GTS and controls.1–3 5 Here, we report a small but significant increase. We think that the smaller numbers of patients in previous studies may account for this difference and had made it difficult to observe this subtle difference. TMS to the motor cortex activates axons of cortical neurons, which synaptically excite pyramidal tract neurons that in turn synapse on alpha motoneurons in the spinal cord. Threshold thus depends on the excitability of axon membranes at the site of stimulation and the membrane potential of postsynaptic neurons in motor cortex and spinal cord. It is not possible to say from our data which is more important for the increase in thresholds we observed, although any changes are likely to have been small. We also found that GTS patients as a whole have increased ICF, whereas others had not described any significant effects.3 5 This could simply be because of the well-recognised variability in measurements of ICF.13 In addition, only one intensity of the conditioning stimulus around the active motor threshold had been used in previous studies. Alternatively, the difference may, like the threshold changes we observed, be due to the larger number of patients studied here in comparison with other studies. A final reason could relate to the presence of six patients in the present series with ADHD comorbidity (see below). GTS is often complicated by comorbidity, in particular ADHD (about 60%) and OCD (about 30%).

We examined whether uncomplicated GTS patients differed electrophysiologically from patients with comorbidity. Previous studies in GTS, OCD and ADHD patients all found reduced SICI compared with normal, so it is perhaps not surprising that there was no difference in SICI between our pure and cormorbid GTS patients. This implies abnormal GABAergic neurotransmission common to all patients. One previous study reported that patients with OCD have lower thresholds than normal,2 particularly in patients with coexistent tics, whereas there was no difference between thresholds in our subgroups. This may relate to the relatively small numbers of GTS+OCD patients in the present study, or to the fact that they were unmedicated. Finally, a previous study of ADHD in children3 but not in adults4 found increased ICF similar to our data in the GTS+ADHD group. Others, however, have commented on the range of variation between patients with ADHD, so that further work is needed to substantiate these conclusions.5^–7

There are no previous data on SAI in pure OCD or ADHD. In the present study, uncomplicated patients, and those with GTS+ADHD, had less SAI than in the GTS+OCD group, whose data looked similar to controls. SAI is known to be modulated by cholinergic inputs,15 so its preservation in GTS+OCD may relate to the primary involvement of the serotonergic system in OCD and indicate relative sparing of cholinergic function.

Tic severity may have accounted for the differences we observed between the GTS subgroups. However, the GTS+ADHD subgroup’s tics were similar to the uncomplicated group’s, and, even if the means of the YGTSS scores were lower in the GTS+OCD group, they were not statistically different. Further studies with larger numbers of patients in the subgroups could help resolve this issue. However, in the present study, we do not think that tic severity accounted for the differences between subgroups because none of our electrophysiological measures correlated with any of the tic ratings. The amount of SICI, ICF and SAI may not be directly related to the severity of the tic phenotype.2 5 In contrast, the amount of SICI correlated inversely with ADHD severity, in particular hyperactivity.5 Thus, reduced cortical excitability may have a more direct influence on motor hyperactivity as part of ADHD. Taken together, the data suggest that patients with ADHD comorbidity have the most widespread abnormalities in our measures of cortical excitability, while patients with GTS+OCD had the fewest. The difference between the GTS and the GTS+OCD group in our measure of sensory motor integration suggests that the circuits mediating tics may differ slightly in the two groups. However, in uncomplicated GTS and OCD, there is evidence to implicate striatal processing.16^–18 Conversely, imaging, pharmacological and neuropsychological data support the hypothesis that ADHD pathophysiology predominantly relates to fronto-cortical (mainly prefrontal) dysfunction, in combination with an impairment of the mesocortical dopaminergic system.19 20 Patients with both GTS and ADHD might therefore suffer in equal measures from striatal abnormalities and abnormal frontal lobe function, and therefore show more extensive abnormalities of cortical excitability. Supporting this hypothesis, GTS patients who are also affected by ADHD exhibit a higher burden of executive dysfunction than patients with GTS only;21 22 they are more prone to oculomotor abnormalities, related to frontal eye field dysfunction,23 a more severe decrease in GABA-mediated intracortical inhibition24 and more definite volumetric alterations of the frontal lobes.25

A clear limitation of our study is the small number of patients in the comorbidity subgroups. Differences between patient subgroups, for example regarding tic severity, may have been lost because of high variability between patients. Hence, our findings have to be reproduced with larger patient groups. In addition, a comparison of patients with GTS, and comorbid OCD or ADHD, and patients with OCD, or ADHD, but without GTS, might further our understanding of the electrophysiology of these disorders. This should include more quantitative measures of the severity of OCD and ADHD.

In summary, we provide more evidence to suggest that in GTS, changes in motor cortex excitability are more widespread than previously thought. In addition, we demonstrate that ADHD comorbidity is associated with more extensive changes in the excitability of motor cortex circuits than uncomplicated GTS or OCD comorbidity. This provides further evidence for the notion that the extent to which various different neuronal circuits are affected is relevant for the phenotype of Tourette spectrum disorders.