Interoception and Premonitory Urges in Children and Adolescents With Tic Disorders

Tics are sudden repetitive movements or vocalizations that occur in up to 21% of children (Cubo et al., 2011; Kurlan et al., 2001). In most cases tics disappear with increasing age and remain stable in only about 1% of people worldwide (Robertson & Cavanna, 2008). Tics regularly are preceded by an unpleasant premonitory urge or sensation (PU). PUs are often perceived as an urge to move, an impulse to move, inner tension or restlessness and mostly occur in the face, neck, shoulders, arms or hands (Kwak et al., 2003).

In psychotherapy, the perception of PUs is both necessary and problematic. On the one hand, in line with habit reversal training, a precise perception of PUs improves the ability to successfully suppress tics (McGuire et al., 2015). On the other hand, PUs illicit tics, negatively reinforce tics and correlate with tic severity (Li et al., 2019).

The capability to perceive bodily signals (‘interoception’) entails several different facets: IAcc is defined as the process of accurately detecting and tracking internal bodily sensations (Garfinkel et al., 2015). Interoceptive sensibility refers to the self-reported attention given to and detection of interoceptive information. Finally, interoceptive awareness refer to the metacognitive correspondence between objective IAcc and self-report of interoceptive information (Garfinkel et al., 2015). In the following, we will only focus on IAcc based on the notion that IAcc may be an underlying dimension in PUs, necessary to perceive interoceptive sensations.

Adult individuals suffering from a TD exhibit a lower IAcc in a heartbeat perception tasks whilst reporting a heightened perception of sensory stimuli (interoceptive sensibility) as compared to individuals without tics. However, this lower IAcc is not evident in children but in adults only (Pile et al., 2018). So far, IAcc in individuals with TDs has exclusively been assessed with a heartbeat-counting task (Schandry, 1981). Arguably, tic symptoms are muscle movements. Thus, assessing IAcc by looking at the ability of an individual to perceive heart activity may not be the best test of the possible involvement of IAcc in TD. Consequently, this study plans to investigate whether IAcc in a muscle tension perception paradigm is associated with PUs in addition to or over the ability to perceive heart activity.

According to a predictive coding account of bodily symptom perception, bodily changes such as heart activity or muscle tension often are weak and imprecise signals. Against this background, a heightened interoceptive sensibility for tic related symptoms may be the result of overly precise interoceptive priors. Specifically, in TDs, an overactive putamen and insula may lead to overly precise predictions at hierarchically higher levels, overriding the actual weak and imprecise sensory inputs. The resulting prediction error may be reduced by performing an ‚involuntary’ tic and/or be the basis for the perception of an unpleasant PU (Rae et al., 2019).

On the behavioral level, PUs may represent a conditioned response to aversive external stimuli such as criticism, offenses or social marginalization as result of tic execution. In the course of a TD, tics may become associated with those negative emotional valences that subsequently constitute the PU. After a tic is executed, the unpleasant PU dissolves and ticking is negatively reinforced, promoting maintenance of tics (O’Connor, 2002). Following the O’Connor model, an attentional focus on PUs and tics may, over time, enhance the overly precise prior even further. As a result, unpleasant PUs are perceived even more, impairing patients’ quality of life. Indeed, adult TD patients with a long history of the experience of tics exhibit lower IAcc (Ganos et al., 2015; Rae et al., 2019). Simultaneously, in adults, physical sensations are self-reported more often (interoceptive sensibility) compared to individuals without tics (Rae et al., 2019). Since a lower IAcc is not evident in children but in adults only (Pile et al., 2018), lower IAcc in adults may reflect a failure to develop, over time, a better IAcc if individuals suffering from TD, which is commonly found in healthy individuals (Murphy et al., 2019). In consequence, given this overly precise prior, the experience of PUs continues and may even be strengthened into adulthood. However, given that IAcc, so far, has been assessed only with a heartbeat-counting task, it is important to additionally assess whether IAcc in children with TDs may be increased with regard to the perception of muscular activity, since muscles are involved in the execution of tics. We opted to assess facial muscle tension given that most TD patients experience at least one tic in the face (McGuire et al., 2016).

Thus, in this study we wanted to test the hypotheses, that children and adolescents with pathological tics exhibit lower IAcc with regard to both heart activity as well as facial muscle tone as compared to children and adolescents without tics. We also test the hypotheses, that variance in PUs is explained by IAcc scores. Furthermore, we compare muscular to cardiac IAcc, using two different paradigms.

Method

Results

Discussion

In the present sample of children and adolescents with and without tics, we found that neither HIAcc nor proprioceptive IAcc scores differed between these groups. These results are in line with recent studies that also did not find any difference in interoceptive accuracy (IAcc) in children with TD (Pile et al., 2018). Our study extends those findings to the perception of muscle tension. Although IAcc scores were numerically lower in participants with TD, these differences did not reach significance (d = -.07 – .43). Whereas lower HIAcc compared to HCs has been established in adult TD patients, we were not able to demonstrate such differences in our sample of children and adolescents. However, such differences in IA may evolve with increasing age of the TD, equal to increasing duration of a childhood-onset TD and may be compared to a model of altered interoception in children with chronic pain: Top-down processes, such as expectations of uncontrollably ticcing, and bottom-up processes, such as a stressed bodily state, may lead to altered interoception over time (Hechler, 2021). It would consequently be highly interesting to assess a sample of adult individuals with TD using our muscle tension paradigm.

Within TD patients, recent studies found IAcc to be positively correlated with PUs in adults (Ganos et al., 2015; Rae et al., 2019). Our findings corroborate this assumption. Adding IAcc scores (MIAcc, CIAcc and HIAcc) to tic severity scores when predicting PUs, a substantial (ΔR² = .22) additional amount of variance in PUs was explained. Note, that neither masseter nor corrugator IA scores were significantly correlated with HIAcc. Consequently, these different measures of interoceptive ability may cover different facets indicating that different body domains may matter when assessing IA. When taking a closer look at the results, the MIAcc (r = .39) correlated more strongly with the PUTS than the CIAcc (r = -.06), which even correlated negatively with the PUTS. In contrast to the corrugator supercilii, the masseter is more frequently contracted deliberately and thus might it be easier to control and it might be easier to estimate its tension. Arguably, this may result in a more reliable measure. However, the corrugator supercilii is linked with emotional expression (Tan et al., 2012) and, arguably, the location of the corrugator supercilii overlaps with locations most frequently affected by tics (McGuire et al., 2016). We therefore recommend that future studies nonetheless continue to assess both facial muscles when looking at IA with regard to muscular tension.

Herbert et al. (2012) compared interoceptive accuracy in eating disorders across two bodily domains – cardiac and gastric –and showed them to be inversely correlated. Similarly, we found cardiac and muscular IAcc to be inversely correlated in TD (r_CIAcc = -.30, r_MIAcc = -.23), although not reaching significance. Following the interpretation by Herbert et al., trying to control muscles that feel uncontrollably at times, might increase activation in the sympathetic nervous system which in turn might increase cardiac IAcc. In line with Flack et al. (2017), activation in the sympathetic nervous system might as well be the result of heightened levels of fear when focusing on muscles associated with unpleasant tic execution.

HIAcc score did marginally significantly correlate with the PUTS in our sample of children and adolescents. Children have smaller hearts and a lower stroke volume, associated with a higher heart rate which influences heartbeat detection positively (Knapp-Kline & Kline, 2005). In light of this it may be concluded that increased HIAcc promotes the perception of PUs in children.

In summary, these results on the association between IA and PU provide an ambiguous picture. On the one hand, IA clearly is associated with PU, but the direction of this association remains unclear. Following our hypotheses, we would have assumed that IA should be negatively associated with PU, as was the case for corrugator perception. In contrast, we find a clear positive association between the perception of cardiac activity as well as tension of the masseter and PU. Obviously, more research is needed here.

Note that the findings linking interoception and PUs rely on the PUTS to represent PUs. However, the PUTS is a self-report measure that more likely represents interoceptive sensibility. Interoceptive sensibility is known to be altered in children and adults with TD (Owens et al., 2011; Rae et al., 2019) and the measure is challenged by the usual problems associated with self-reports. The relative relationship of interoceptive accuracy, interoceptive sensibility und interoceptive awareness constitute PUs’ presence and cognitive, emotional and clinical consequences (Garfinkel et al., 2015).

We find it especially intriguing to interpret our findings within the predictive coding framework (Ainley et al., 2016; Friston, 2010; Khalsa et al., 2018; Rae et al., 2019). In this framework, a difference between sensory input and prior expectation results in a prediction error. Following Bayesian Inference, the prediction error may be resolved by updating the prior expectation or executing a movement to change sensory input. In TD patients, in hierarchical higher brain structures, an over-precise interoceptive prior might predict movements. If the weak bottom-up sensory input does not correspond with this prediction, the insula has to resolve the resulting interoceptive prediction error. The anterior insula is hypothesised to update predictions to reduce prediction errors (Seth, 2013; Seth et al., 2012) and is known to show functional abnormalities in TD patients. Due to the prior’s over-precision and weight, imprecise sensory input may be overcome by the prediction and the prediction error is being ‘explained away’ by the anterior insula as a premonitory sensation (urge to move). In adults, lower IAcc is found which, arguably, indeed indicates weaker priors. However, in line with previous research in adults (Ganos et al., 2015), we found that lower IA (heart activity and tension of the masseter) is correlated with lower PUs. Assuming that good IA, over time, leads to weaker priors, a smaller prediction error may result in less PU. In contrast, IA for the tension of the corrugator, was negatively correlated with PU. Following this finding, one could argue that the worse the perception of actual physical sensations in areas in which tics occur the more top-down predictions will overshadow interoceptive sensations and result in the perception of PUs (perceptual inference).

There are a number of limitations to the current study that offer opportunities for future research. Studies investigating the perception of muscle tension in adult TD patients are yet to be conducted to gain further insights on the development of PUs over the life span. Since the current cross-sectional study allows correlative interpretations only, the longitudinal comparison of chronological changes in HIAcc, MIAcc and CIAcc scores, PUTS and urge thermometers over the lifespan would provide further information on their etiological meanings. In addition, the exploratory findings on IA in child and adolescent TD patients need to be replicated, preferably in larger samples. Our study exclusively focused on IAcc in youth with TD. Future studies may extend our findings to interoceptive sensibility to disentangle the influence of interoceptive sensibility and IAcc on the self-reported perception on premonitory urges. Altogether, our sample size was relatively small, compromising statistical power to some degree.

TD patients usually differ from HCs not only with regard to tics but also with regard to comorbidities such as ADHD or OCD, depression or anxiety. When comparing TD patients to typically developing children, our study cannot account for the impact of comorbidities such as ADHD or OCD due to the relatively small sample size. Therefore, the relationship between interoception and tics and PUs in TD patients with multiple comorbid diagnoses is hard to disentangle. Panic and somatic ratings, for example, were found to correlate with higher HIAcc in adults in children (Eley et al., 2007). Comparing TD patients with a control group exhibiting matching comorbidities could help to disentangle the complex influences of comorbidities on IAcc and PUs. In line with Pile et al. (2018), additional instructions could be added to muscle tension perception tasks to reduce effects of inattention, a frequent comorbid symptom in TD patients, and task-misunderstanding.

The PUTS measures PUs as a whole but does not differentiate between context- and time-dependent urges as a state and the PU as a general trait. It is yet to be examined how context- and time-dependent urges, measured by urge thermometers, vary in relation with changes in PUs as a trait and changes in IAcc.

Similar to the PUTS that measures PUs as a general trait, the SCL-TIC-S and SCL-TIC-P measure self- and parent-reported mean tic severity over the course of a week. It is not clear how accurate self-assessed tic frequency reflects actual tic expression. On the one hand, both child and adult patients underestimate their tic expression (Müller-Vahl et al., 2014; Pappert et al., 2003). On the other hand, parent-reports cannot be accurate either. At least adolescent patients are not observed by their parents most of the day and tic frequency highly depends on context. Another study compared children’s self-reported account of tic frequency to objective video ratings of tic frequency in various experimental situations. Self-reported tic frequency related to objective measures depended on the situation. Interestingly, the higher children scored on the PUTS, the better their self-report predicted objective tic-frequency (Barnea et al., 2016). This implies that self-reported tic frequency depends on PUs and probably IAcc. Still, as our multiple regression analysis showed, IA explains variance in addition to self-reported tic frequency. The moderate correlation with parent-reported tic-frequency (r = .34) further validates self-reported tic-frequency.

The current study holds clinical implications. Interoceptive trainings specifically targeting interoceptive domains that are impaired in TD may be more beneficial than multisystem interventions (Khalsa et al., 2018), so research on the impact of interoceptive trainings aimed at improving heartbeat perception (Schaefer et al., 2014) could be expanded by adding muscle tension biofeedback tasks to improve accuracy of urge perception to foster the ability to control tics via HRT. Observations from intervention studies examining muscle tension biofeedback in TD would provide further insights in causal relationships between IAcc, PUs and tics.