The Severity and Neural Correlates of Premonitory Urge in Tourette Syndrome: A Systematic Review and Meta-Analysis

Introduction : Premonitory urge (PU) is an aversive bodily sensation that signals the onset of tic disorder.To our knowledge, PU typically precedes the appearance of tic symptoms, and both age and tic severity are correlated with PU. However, inconsistent findings have also been reported. Hence, we conducted a meta-analysis to examine the relationship among premonitory symptoms, patient age and the severity of tic symptoms, as well as to summarize the research on the neural underpinnings of PU in Tourette syndrome (TS). Methods : We conducted a literature search of relevant studies published between December 2005 and April 2022 using databases such as PubMed, Elsevier, PsycINFO, and Web of Science. Our analysis was carried out using R software with the assistance of the “meta” and “metafor” packages. Results : Our meta-analysis included 22 studies with a total of 1236 tic disorder patients. The mean Premonitory Urge for Tics Scale (PUTS) score was 20.17, with a 95% confidence interval of [18.14, 21.68]. Through meta-regression, we found that age and tic severity play important moderating roles in PU severity ( p < 0.0001). Neuroimaging studies suggest that PU is related to the insula, prefrontal cortex (PFC), anterior cingulate cortex (ACC), and supplementary motor area (SMA), regardless of the structural or functional level. Conclusions : Our meta-analysis confirmed the positive relationship between the severity of tics and PU and identified age as a significant factor influencing PU. The neural mechanisms underlying PU remain largely unknown, but evidence suggests that the insula, PFC, ACC, and SMA are related regions.


Introduction
Tic disorder (TD) is a prevalent neurodevelopmental condition, with Tourette syndrome (TS) being its most severe form.Many patients with chronic TDs and TS report experiencing pressure-like or tickling sensations prior to the occurrence of tics [1,2].These sensations are well-known symptoms accompanying tics [3].The term "premonitory urge (PU)" was first used by Leckman et al. [4].The majority of patients describe these pressure, itch, and sore sensations as aversive or unpleasant before the onset of a tic.A prior study found that PU was present in nearly 90% of patients with TS [5].The most common sites of PU are in the head and neck and are separate from the anatomic region where the tics occur [6].
PU plays a crucial role in the etiology of TD.Many studies have reported that patients can temporarily relieve the sensations associated with PU by producing tic symptoms.In other words, tics are often seen as a motor response aimed at reducing the intensity of PU [7].For instance, the recognition of an itchy tongue as a PU could lead to tongue biting as a tic symptom, which in turn can alleviate the itchy sensation [1].Our previous meta-analysis showed a mild to moderate correlation between PU strength and tic severity [8].Additionally, PU has been shown to play a significant role in behavioral therapy for TS [9].Awareness of PU before the onset of tics is the first step in habit reversal training, a widely used behavioral psychotherapeutic approach for individuals with TD.This approach highlights the importance of increased insight into PU in helping patients improve their ability to suppress tics.Pertinently, PU is a fundamental component in the formation of tics [10,11].Despite this, the mechanism underlying PU and its relationship with tic symptoms remains unclear.
The onset of PUs typically occurs around the age of 10 years and may be linked to the patient's increasing awareness and cognitive development [4,[12][13][14].However, some studies have reported an earlier onset of PU [15].This may be due to a lack of introspection and the inability of younger children to describe feelings [15,16].The severity of PU is generally mild in children and adolescents, but it can become more pronounced in adults [7,10,17].These findings suggest that age is an important factor influencing PU, although the relationship between patient age and PU severity still requires further investigation.
Moreover, the underlying neurobiological mechanism of PU remains unclear.However, several studies have implicated the anterior insula as a potential source of PU [18,19].One study using functional magnetic resonance imaging (fMRI) found that a region in the right insular cortex was associated with PU scores [19].Other neuroimaging studies, including functional and structural magnetic resonance imaging (MRI), have suggested that PU may be related to activity in the prefrontal cortex (PFC), anterior cingulate cortex (ACC), and supplementary motor area (SMA) [13,20,21].These findings suggest that the generation of PU may involve multiple brain networks.Further research is needed to explore the relationship between PU and tics, particularly from a network perspective.
The aim of this meta-analysis was to examine the predictors of PU, particularly patient age and tic symptom severity.We conducted a meta-regression analysis to achieve this goal.This study also examined the relationship between PU and tic symptom severity.Additionally, to gain a deeper understanding of PU, we focused on magnetic resonance imaging (MRI) studies and aimed to present an updated summary of the evidence-based neural correlates of PU.

Protocol and Registration
This systematic review and meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [22] and was registered with the International Prospective Register of Systematic Reviews (PROSPERO Register CRD42022379230).PRISMA checklist is shown in Supplementary material.

Search Strategy
We conducted a comprehensive literature search for relevant studies published between December 2005 and April 2022 in PubMed, Elsevier, PsycINFO, and Web of Science.Our main search terms included "premonitory urge", "sensory phenomena", and various variants of "tic disorder" or "Tourette syndrome".We also manually checked the reference lists of the included studies to ensure completeness.
To be eligible for inclusion in our meta-analysis, studies had to meet the following criteria: assessment of PU using the Premonitory Urge for Tics Scale (PUTS), assessment of tic severity using the Yale Global Tic Severity Scale (YGTSS), and inclusion of both child and adult subjects.Studies were excluded if they failed to meet any of the following criteria: no mention of PU score, lack of focus on TDs, not original research, and sample size less than 10.Additionally, we performed a separate search using the terms "premonitory urge" and "MRI" or "imaging" to identify studies examining the neural correlates of PU.

Data Extraction
The following data were extracted from each study: first author, publication date, sex and age of participants, sample size, mean value and standard deviation (SD) of PU as assessed by the PUTS, mean value and SD of tic severity as assessed by the YGTSS.In addition, the main findings from included MRI studies examining the neural correlates of PU were documented.

Assessment of Quality for Included Studies
The quality of each study was evaluated using a modified version of the critical appraisal skills program (CASP) scale [23,24].This widely used tool assesses the strengths and limitations of qualitative research methodologies and includes 11 items, such as "Item 1: Did the study address a clearly defined issue?" and "Item 4: Were the controls chosen appropriately?" Two authors independently assessed each included trial and reached a consensus.Studies were excluded if they had fewer than 5 'yes' responses on the CASP scale.

Statistical Methods
We employed the use of a pooled mean with 95% confidence intervals (CIs) to quantify PUs and conducted a meta-regression analysis to examine the predictors of PUs, specifically patient age and the severity of tic symptoms.To assess the heterogeneity of each effect size, we utilized the I 2 statistic, with a value greater than 50% indicating substantial heterogeneity.We sought to identify the source of heterogeneity, including differences within groups, between groups, and in statistical methods, and employed a random effects model in the meta-analysis.Our analysis was performed using R software (version 3.5.3,The University of Auckland, Auckland, New Zealand) with the "meta" and "metaphor" packages for the meta-regression analysis.To detect publication bias, we utilized a funnel plot.

Mean PUTS Scores and Publication Bias
The results of our meta-analysis showed that the average PUTS score was 20.17, with a 95% CI of [18.76, 21.69] and I 2 of 98%.Significant heterogeneity was confirmed by a Q test, with a value of 934.86 and 21 degrees of freedom (df = 21, p < 0.0001, as shown in Fig. 2).There was no evidence of publication bias, as indicated by the funnel plot (as shown in Supplementary Fig. 1 in the supplementary material).

Subgroup Analysis of the Pooled Mean PUTS Scores
The source of heterogeneity was explored through subgroup analysis, which showed a significant difference between adult and child participants.The mean PUTS score in the adult group was 22.98 (95% CI [21.58, 24.47],I 2 = 77%), and in the child group, it was 18.15 (95% CI [16.67, 19.76],I 2 = 97%).The heterogeneity between the adult and child groups was determined to be significant by the Q test (Q = 19.12,df = 1, p < 0.0001) (More details see Fig. 3).

Neural Correlates of PU
Both MRI and magnetic resonance spectroscopy (MRS) were employed to investigate the neural basis of PU at both functional and structural levels.Our analysis identified 12 relevant studies (refer to Fig. 2).In terms of functional aspects, the results showed that PU severity as measured by PUTS scores and the inhibition of PU were correlated with decreased connectivity between the cortex and insula seeds, as revealed by resting-state functional MRI [19,34].On the other hand, functional connectivity between the right dorsal anterior insula and bilateral SMA was found to positively correlate with PU severity [18].At the structural level, structural MRI showed an inverse association between PU and gray matter volume in the insula and sensorimotor cortex [13].Additionally, increased activity in other regions, such as the ACC and right temporoparietal junction, was observed during the inhibition of PU [43,44].Furthermore, gray matter volume in the cerebellar lobule was implicated in sensorimotor processing, which is related to PU [41,45].Finally, Gamma-aminobutyric acid (GABAergic) processing was found to be crucial in sensorimotor processing [46].Gamma-aminobutyric acid (GABA)-edited MRS studies revealed lower levels of GABA in the primary sensorimo-  tor cortex (SM1) in children with TS and a correlation between more severe and frequent PUs with lower levels of SMA GABA+ [47].A summary of these findings can be found in Table 2 (Ref.[13,18,21,34,43,[48][49][50][51][52][53][54][55]).

Discussion
To our knowledge, this was the first study to conduct a meta-analysis investigating the relationship between PU and patient age.Our findings indicated that younger individuals tended to have a lower level of PU, while older individuals tended to experience a higher level.This could be because with age, children develop greater self-awareness and the ability to describe their sensory experiences [12,28].As a result, the reporting of PU increases [4,12].By age 10, children are better able to describe the sensations they experience and report more awareness of PU [4,5,56].On the  other hand, while the tic symptoms of some patients start to subside at approximately 17 years of age, approximately 23% of severe tic cases become intractable [57][58][59].Additionally, the PUTS tool for PU assessment is self-reported, and its diagnostic validity in young patients is limited, making it challenging to use in this population [26].Therefore, developing a structured interview to assess PU in young patients may be a valuable direction for future research.Traditionally, PUs have often been observed to present three years after the onset of tics [60].In addition, a recent study found that associations between reported urge levels and instantaneous tic intensity tended to be less pronounced in children and adolescents than in adult TS patients, which also supports the age effect in PU [61].Furthermore, tics can be voluntarily suppressed for a brief period [62].Thus, PU is a valuable cue for tic suppression.Behavioral treatments, such as habit reversal therapy (HRT), are considered a promising therapy for tics [8,63].HRT, which is widely used to treat TS, has been confirmed by numerous studies to be effective [64][65][66].For instance, a meta-analysis of HRT showed that PU is closely related to tic symptoms [8,67].Detecting PU is the first step in HRT [68].The therapy aims to increase awareness of urges and to help patients develop coping behaviors to replace tic symptoms [69].This implies that greater awareness of PU leads to improved HRT efficacy.Furthermore, several studies have reported that PUs may play a crucial role in maintaining TD and have confirmed that PUs have a strong association with tic symptoms [11,60].
The widely accepted neurobiological basis for TS is dysfunction of the cortico-striatal-thalamo-cortical (CSTC) circuit, as supported by numerous neuroimaging studies [13,39,45,70].These studies have shown that aberrant activity in the striatum may be linked to PUs.Structural MRI data, such as gray matter volume, cortical thickness, and connectivity matrices obtained from resting/task-state fMRI data, provide valuable morphological insights into PUs.Decreased gray matter thickness in the insula and sensorimotor cortex has been linked to PUs in young people with TS [13].Whole-brain analysis of cortical (gray matter) thickness demonstrated that in young adults with TS, a reduction in cortical thickness within the insula, sensorimotor cortex, anterior cingulate, and insula was associated with higher ratings of PU severity and was significantly lower than that in a group of typically developing controls [13].
At the neurochemical level, dopaminergic modulation of glutamatergic corticostriatal afferents has been linked to the expression of diverse motor and nonmotor symptoms, including those seen in TS [71].This anatomical substrate is reflected in the lower functional connectivity between the orbitofrontal cortex and sensorimotor cortex, between the inferior frontal gyrus and the putamen, and between the inferior frontal gyrus and insula as PU severity increases [34,43].The SMA shows a broader profile of crosscorrelation with the motor cortex during tics compared to that during intentional movements, with its proportion varying with premonitory sensations [18].Increased PU frequency was found to be associated with greater connectivity between the primary motor cortex (M1) and the caudate nuclei and increased transfer of information between M1 and the insula [72].Moreover, greater PU severity was associated with greater activity in the hippocampus, middle temporal gyrus and subcortex [55].
The insula has been indicated to play a critical role in PU, suggesting that it is associated with the urge-to-tic [19].The brain network involving the insula, ACC, and SMA is associated not only with the neural mechanism of PU but also with tic symptoms [73,74].Further imaging studies are needed to better understand the relationship between PU and tic symptoms in the future.
Overall, the current neuroimaging research suggests that PUs may be an early component of a network that includes the insula, ACC, and SMA.The insula appears to play a key role in the generation of PUs, connecting the sensory and emotional aspects of PUs.Moreover, the ACC and SMA may be involved in the urge-to-tic.These areas are recognized as important nodes of the salience network, suggesting that this network may play a role in both urge generation and the urge-to-tic.Further studies are required to fully understand PUs at the network level.
In addition to the generation of PUs, PU inhibition is also an important aspect to consider.Evidence from multiple studies has indicated that TS patients exhibit dysfunction in the inhibition of PUs [75][76][77].This evidence highlights the need to investigate the role of inhibition in the development of PUs.Of note, impaired automatic inhibition has been reported [78].We speculate that the generation and inhibition of PUs appear to be interrelated processes, and an imbalance between them may contribute to an increased urge-to-tic.Whether PUs translate into tics may depend on the balance between these two processes.
To advance PU research, a reliable and validated measure of sensory symptom severity is crucial.One widely used instrument is the PUTS, which was developed by Woods and his team [14].The PUTS was optimized to nine items and has been widely used in clinical practice [78][79][80].Multiple studies have provided evidence for the reliability and validity of PUTS [21,26,81].However, new assessment tools have also been developed, including the Individualized Premonitory Urge for Tics Scale (I-PUTS) [30] and the Rumination and Awareness Scale for Tic-associated Sensations (RASTS) [40].While these tools provide new insights into PUs, they also have limitations, and further development is necessary.The inclusion of multiple dimensions, such as location, severity, frequency, and intensity, is important for future advancements in PU assessment.
This meta-analysis supports the correlation between PUs and tic symptoms, reinforcing the significance of PUs in the treatment of TDs.The findings indicated that patient age is a crucial moderator affecting PU severity, im-plying that the impact of age should be considered during PU assessments.Considering these results, future research should be directed toward further understanding the nature of PUs.However, several challenges need to be addressed in future studies.First, current PU assessment tools are subjective and lack objectivity.New tools such as the I-PUTS and the RASTS require further validation across different countries and languages.Second, neuroimaging studies face difficulties in separating the processing of PUs and tic symptoms.Thus, future studies should aim to explore PUs at the neural circuitry level.Third, while computational psychiatry approach holds potential in understanding PUs, it requires further development to align with clinical practices.Finally, considering dopaminergic hyperinnervation in TS, investigating PUs based on this aspect may also be worthwhile.
There were two limitations to this study.First, the sample size of the included studies was small, which may have affected the accuracy of the results.Second, most of the included studies were from clinical samples with different inclusion and exclusion criteria, which might increase the heterogeneity of the meta-analysis.

Conclusions
The present meta-regression study confirmed the positive correlation between the severity of tics and PUs.Moreover, patient age appears to be a significant influencing factor for PU, as evidenced by the increasing PUTS score with age and tic severity.Neural correlates of PUs were associated with the insula, ACC and SMA, as well as the network associated with these regions.To further support these findings, future studies with larger sample sizes are necessary.