Tourette syndrome (TS) is a childhood-onset neuropsychiatric disorder which encompasses a spectrum of symptoms inclusive of motor and phonic tics, which may be accompanied by neuropsychiatric comorbidities such as attention-deficit/hyperactivity disorder (ADHD), obsessive compulsive disorder (OCD), impulse control disorder, and other behavioral manifestations.1 According to the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-V) and the International Classification of Diseases, tenth edition criteria, the patients must have 1) a combination of chronic motor and phonic tics, and 2) these must occur several times a day for at least 1 year. 3) The symptoms must have an onset before the age of 18 years and 4) these symptoms must not be explained by other medical/neurological conditions. Alternatively the Tourette Syndrome Classification Group2 published a set of diagnostic criteria with slight differences as compared to the international classifications, with the age of onset prior to 21. The Tourette Syndrome Classification Group embraced the idea that the anatomic location, number, frequency, complexity, type, and severity of the tics could change over time. The latter classification also included the notion that the tics must be witnessed by a reliable examiner, either directly or by videotaping.2 In 2010 the National Institutes for Mental Health proposed a new paradigm for classification of psychiatric disorders, the Research Domain Criteria (RDoC). In contrast to the ICD and DSM, the RDoC identifies constructs and subconstructs based on genetic, structural, and connectivity factors believed to underlie similar behavioral and physiological responses.3,4 TS in this classification fell under the construct of habits in the positive valence systems domain, which is characterized by repetitive, stereotypic, and compulsive behaviors. These behaviors have been thought to be mediated by dopamine and other neurochemical dysfunction in circuits involving the striatum and prefrontal cortex.

The TS motor and phonic tic course typically wanes in most patients as they transition from adolescence into adulthood, however symptoms persist in about half of patients into the adult years.1 Despite several pharmacological approaches, a number of patients remain refractory to therapy and have a persistent and impaired quality of life. There may also be disabling social embarrassment due to both the tics and to psychiatric comorbidities. Severe medication-refractory TS cases have been addressed by surgical therapies since the 1960s. The early therapies focused on ablation of motor and also limbic targets.5 In 1999, Vandewalle et al.6 reported the first case of a patient with TS who was implanted with a deep brain stimulation (DBS) device. The intervention involved bilateral electrode implantation in the centromedian nucleus of the thalamus. At 4 months, the authors reported a significant reduction from 38 to 8 tics per minute (79% improvement) during active stimulation.

The use of TS DBS has generated several case reports and series with stimulation not only of centromedian (CM), but also now inclusive of other targets believed to be part of a dysfunctional neural network ultimately contributing to the underlying pathophysiology thought to be responsible for TS. There has been reported large variability in response to DBS, and large well-powered prospective trials have yet to be conducted. The Tourette Syndrome Association has commissioned a large international database of all implants worldwide.

The pathophysiology of TS remains unknown. Experts have hypothesized the existence of a group of hyperactive dopaminergic neurons that may contribute to a dysfunctional cortico-striato-thalamo-cortical circuit and ultimately lead to decreased cortical inhibition. It has also been pointed out by many experts that this hypothesis cannot explain all of the manifestations of TS.7,8 Additionally, there has been recent data revealing structural changes in the globus pallidus, reduced cyclic adenosine monophosphatea in cortex and striatum, and an increase in dopamine uptake in striatum.1 These largely pathological findings support the notion of enhanced dopaminergic activity. It is also postulated that a developmental defect in migration of gamma-aminobutyric acid (GABA) neurons could play a role in the overall the imbalance of the cortico-striato-thalamic circuit.1,9 The dopaminergic hypothesis has been attractive to many scientists, particularly since dopaminergic blocking drugs result in tic suppression for many TS subjects. Dopamine, however, is likely only one part of the underlying pathophysiological story underpinning TS.

There is growing evidence to support the hypothesis that cortical excitability is abnormally increased in TS. Studies utilizing functional magnetic resonance imaging8,10 and transcranial magnetic stimulation1,7 have shown an increase in excitation of M1 cortical areas that seemed to correlate with the severity of TS.7 Also there appears to be an increase in activation of primary sensorimotor and secondary motor cortices during execution of motor tasks (such as finger tapping),8 as well as coupling of the supplementary motor area and the contralateral M1 motor cortical areas during the movement programming phase (planning to execute a movement) suggesting that this could be an adaptive mechanism that it may also be a marker for TS.8

The execution of motor tasks is known to occur through the activation of a complex circuit involving interaction between the cortex and the basal ganglia. The striatum receives excitatory glutamatergic input from the motor cortex and projects to other basal ganglia structures including the internal portion of the globus pallidus (GPi) and the substantia nigra pars reticulata. These two structures are the main outputs to thalamic nuclei and they function largely through GABAergic projections.11 Non-human primate experiments have revealed that microstimulation of discrete striatal areas can result in stereotyped movements of the extremities.11 It has been suggested that in TS patients, a population of striatal neurons becomes abnormally active, leading to excessive inhibition of basal ganglia output to the thalamus which generates competing motor patterns that could ultimately be the source of tics.11 This network hypothesis, though compelling, may only be part of the story as other factors are likely important to the genesis of tics and complex behavioral manifestations.

Another compelling hypothesis of TS pathogenesis involves dysfunction of basal ganglia dopaminergic structures and the connections to frontocortical circuits.11 Mink11 has proposed a schema where, in a normal person, there is tonically active inhibitory output of the basal ganglia in connections to the cortex and brainstem, suppressing motor pattern generators (MPGs). With the initiation of a purposeful movement, the basal ganglia projections increase their firing rate towards competing MPGs, suppressing undesired motor patterns. During a tic, it is postulated that tics arise from a focal population of striatal neurons that become abnormally hyperactive and result in downstream inhibition into the basal ganglia structures, which in turn may reduce their inhibitory output into cortex and brainstem. Therefore unwanted intrusive MPGs may possibly be translated into repetitive abnormal stereotyped movements.11

The amygdala is involved in the processing of a variety of stimuli inclusive of facial emotional expression, sounds necessary for communication, and also for processing of music.12 The amygdala is also thought to mediate many manifestations of neuro-psychiatric diseases such as depression, anxiety, and antisocial personality disorders.12 The subnuclei of the amygdala can be structurally subdivided into laterobasal, centromedial and superficial groups each associated with specific functional properties as evidenced by recent neuroimaging studies.12

Projections from the amygdala (in particular the superficial nuclei group) not only are important for learning, behavior and emotional distress, but are also thought to play a role in tic suppression in TS patients through close interactions with the frontal cortices.13 Peterson et al.14 observed larger amygdala volumes in TS children when compared to age-matched controls, but smaller amygdala volumes in adult TS patients. These findings suggest that there is a neuromodulatory mechanism involving the amygdala circuits that may be lost in TS patients whose tics persisted into adulthood.13

Neurosurgical interventions for TS have been around since 1955. Early in the history of TS surgery, frontal lobotomies and leucotomies were conducted for cases that were refractory to pharmacological and other therapies.15 There has, since the early Hassler thalamotomy cases, been specific targeting of basal ganglia structures and thalamic nuclei. Early approaches utilized stereotactic coagulation of thalamic nuclei including the intralaminar and centromedian thalamic, as well as the internal portion of ventralis oralis anterior (VOA). Other targets such as GPi, and even isolated reports of globus palidus externa (GPe)5 and subthalamic nucleus,16 have been emerging as potentially successful. The nucleus accumbens and anterior limb of the internal capsule region, a target used successfully for treatment of obsessive compulsive disorder, a common comorbidity in TS, have been used in a handful of patients perhaps with less robust response. In 1999 Vandewalle et al.6 reported the results of the first 3 cases of TS patients who underwent thalamic DBS. These initial patients were implanted to address symptoms unresponsive to medical and behavioral therapies. Since this early report, others have attempted different targets and approaches. A summary of the proposed DBS targets is provided in Fig. 1.

Over the years there has been an increasing knowledge of the pathophysiology of TS and other diseases due to the use of single cell and local field potential (LFP) recordings pre- or post-operatively. There has been a notion that pathological frequencies or oscillations may contribute to the occurrence of symptoms. While excessive beta oscillations (11-30 Hz) have been discovered in hypokinetic conditions such as Parkinson's disease,31 low frequency oscillations (2-7 Hz) may also be encountered in hyperkinetic conditions such as dyskinesias and dystonias.31 There is very little overall known on this topic, however the field is evolving quickly.

In anesthetized TS patients (without ongoing tics), microelectrode recordings from thalamic nuclei have revealed a pattern of low frequency firing in a burst pattern.31 While awake, these patients have LFPs characterized by alpha and low frequency activity (in the absence of beta band oscillations) which seem to correlate with the clinical phenotype of TS, as patients with less tics and more OCD features seem to have fewer low frequency oscillations than patients in whom tics were a predominant feature.31 Similar to thalamus, GPi has shown low frequency oscillations that seem to precede the electromyographic recording of the tic by 50 ms or more, suggesting that in the GPi this activity could reflect the sensations that are premonitory to a motor tic,31 however this remains speculative. These specific frequency oscillations, in particular in thalamus have been targeted and shown to be dramatically altered by application of DBS.32

Anterior limb of internal capsule and NA are structures known to play an important role in the pathophysiology of the reward system and seem to play an important role in patients with OCD, which is a typical comorbidity in TS. These structures have also been used for LFP recordings, and researchers have observed high beta power oscillations in the NA in comparison to VOA/CM-pf thalamus. This may be a physiological marker of the OCD activity that would lead to the lower frequency oscillations found in the thalamic nuclei.31

As research progresses, more efficient therapies may possibly deliver stimulation instead of in a continuous or scheduled fashion, in a more responsive paradigm that triggers off of pathological oscillations.

New stimulation techniques are being developed and some have moved into clinical trials. DBS as a therapeutic technique for movement disorders and neuropsychiatric disorders has been previously delivered in an open-loop paradigm. These paradigms focus on a pre-programmed chronic and continuous stimulation pattern,33 regardless of the internal state of the system, or the environmental factors. TS however, presents in a paroxysmal pattern, characterized by intermittent episodes of motor and/or vocal tics in addition to neuropsychiatric components.1 This presentation in many ways is similar to epilepsy. Newer stimulation delivery devices, can be personalized to the frequency and duration of a behavioral manifestation. These devices, in which stimulation is independent of functional neural feedback have been initially designed to treat seizures (open loop approaches). In this scenario they act to inhibit seizure propagation and to alter seizure thresholds and have been shown to be safe and efficacious.34 The only TS study to date using scheduled rather than constant stimulation, reported clinical outcomes of 5 patients after 6 months using a novel approach.22 Targeting the centromedian brain region, safety and feasibility of intermittent stimulation has been established, however outcomes were less than expected.

Responsive or adaptive DBS (aDBS), so-called closed-loop stimulation devices, rely on functional neural feedback, such as abnormal electrographic discharges or more recently on neurochemical feedback. These closed loop approaches may adjust stimulation parameters and shorten "after-discharges" elicited during functional mapping.34,35 Currently, DBS research has been focusing on how to interpret brain activity and use it as feedback to control delivery of therapeutic stimulation. There has been an emerging development of responsive stimulation in the treatment of Parkinson's disease and a growing discussion about proposing cortical stimulation rather than DBS.36 Recent proof-of-principle studies have shown that by personalizing and optimizing stimulation in real time, efficacy and efficiency of continuous DBS could be improved. LFPs have been reported to provide the most relevant biomarker to close the loop.37 Brain-computer interface (BCI) systems have been tested to control the timing of when stimulation is delivered.38 These BCI systems use the beta activity in the LFP recorded directly from the electrode in order to regulate stimulation. In 8 patients with advanced PD, aDBS was demonstrated to be 30% more effective than conventional DBS therapy while delivering <50% of the stimulation of current DBS strategies. This study suggests that it is possible to track an LFP biomarker from the site of stimulation and that aDBS could be more efficient and efficacious than conventional neuromodulation for PD. Another potential alternative feedback signal is the electrically evoked compound action potential (ECAP). This signal results from activation of an ensemble of neural elements following each DBS pulse. A novel instrumentation that uses commercial amplifiers in a custom serial configuration to suppress the artifact and record short latency ECAPs after each DBS pulse was developed and demonstrated that ECAPs can be recorded with high fidelity during DBS therapy. This novel instrumentation was validated through both in vitro and in vivo testing.39 Recent animal studies have shown that stimulation-induced changes in neurotransmitter release can be associated with the therapeutic benefit of DBS. In a rodent model of DBS, mathematical models were used to describe relationships between stimulation-evoked extracellular dopamine response and DBS parameters and have shown that adjusting stimulation intensity can modulate dopamine concentration.40 Open- and closed-loop configurations for DBS therapy are demonstrated in Fig. 2.

Deep brain stimulation has been shown to be a promising therapy for TS, more so for motor and phonic tics than for associated comorbidities, such as OCD, ADHD, and self-injurious behavior. Overall, a better understanding of the circuitry involved in TS and the mechanisms of brain stimulation will speed development of new techniques and devices. Deciding the best scales and ways to measure outcome will impact the future development of the TS DBS field. Up to 9 targets have been explored individually or in combination, for the treatment of TS with reported improvement in tics. New targets are currently being investigated to treat symptoms less responsive to standard DBS.41

Technological advances in DBS devices or in the systems of stimulation delivery, may enhance clinical outcomes.42 Directional steering through segmented electrodes capable of modeling the electrical field to better target a desired structure or pathway with less side effects, such as a novel DBS electrode with 32 contacts has been shown to be safe, well tolerated, to decrease the thresholds for side effects while improving the therapeutic window of DBS.43 Local field potentials can be also used to close the loop and to identify information regarding high-level sensory processing, perception, and locomotor activity. Whole-brain electropysiological brain activity is measured using far-field sensors located on the scalp by electroencephalogram or directly on the brain surface by electrocorticography (ECoG). A system that combines activity analysis within cortical (ECoG) and subcortical (LFP) networks will likely provide a better depiction of network dynamics.7

Our group at the University of Florida Center for Movement Disorders and Neurorestoration is currently conducting a research study on TS neural network and combining bilateral thalamic stimulation with cortical ECoG strips implanted on the primary and premotor cortices. The goal is to investigate spectral features of tics compared to baseline and volitional movements, and also to investigate the role of phase amplitude coupling in the cortex. Preliminary results have shown that tics are spectrally obvious and can be detected, and strong coupling after application of therapeutic settings can be observed (unpublished data). Further analysis will uncover a more complete understanding on the electrophysiology of tics and how DBS is effective in treating this disorder.

In addition, the development of new neuromodulation techniques such as optogenetics or thermogenetics, may help to better rationalize therapy and to redesign brain stimulation and to make it more efficient. Optogenetics refers to the integration of optics and genetics to achieve gain- or loss-of-function of well-defined events within specific cells of living tissue,44,45,46 by using light to control activity of neurons which have been modified to express light-sensitive proteins, such as channelrhodopsin. Thermogenetics uses temperature to drive neural activity.47 To date, two thermosensitive Transient Receptor Potential (TRP)-based tools have been developed for use. Their thermal sensitivity is such that a neuron expressing a thermoTRP can switch from silent to robustly active in response to temperature shifts as small as 1-2℃. These novel approaches have been widely employed in basic science research. However, with the continual improvements to the molecular toolbox of genetically encoded neuronal proteins, new technologies may support new therapeutic modalities that can be ultimately applied in humans. In the near future, we will likely employ multiple leads to uncover the neural networks and to monitor physiology in real-time in an awake behaving human.