Sensorimotor control depends on the brain's ability to predict what will happen next. In typical motor behavior, the brain uses cues from the environment to prepare actions, update expectations, and correct errors. Tourette syndrome - characterized by sudden, repetitive motor or vocal tics - has long been considered a disorder rooted in disrupted sensorimotor processing. A new open-access study in Brain Communications provides a detailed examination of how these predictive systems differ in Tourette syndrome, combining behavioral testing, EEG recordings, and Bayesian computational modeling to uncover the mechanisms behind impaired motor prediction.
The study included 30 adults with Tourette syndrome and 30 matched control participants. While EEG was recorded, all participants performed a motor cueing version of the Posner task, where cues predicted whether a left- or right-hand response would be required. Crucially, the predictability of these cues varied throughout the experiment - sometimes cues were only 50% valid, other times 70% or 90%. In healthy individuals, the brain uses these probabilities to adapt behavior: when cues are highly reliable, responses to valid targets become faster, and responses to invalid ones slow down, reflecting stronger reliance on predictive information.
This expected interaction between cue validity and cue predictability was clearly present in the control group. However, individuals with Tourette syndrome did not show this adaptive pattern. Their reaction times differed between valid and invalid cues but did not scale with predictability. In other words, while they responded differently to correct versus misleading cues, they did not adjust their behavior in response to changes in how reliable the cues were. This reduced behavioral flexibility indicates an impairment in how the predictive sensorimotor system updates expectations.
To investigate the neural basis for these differences, the researchers applied a Bayesian observer model - the Hierarchical Gaussian Filter - to derive each participant's internal beliefs about cue predictability based on their response speeds. These trial-by-trial estimates were then used in single-trial EEG regressions focusing on two key event-related potentials: the P3a and P3b components, both of which are known to index attention, surprise, and stimulus - response updating.
In control participants, the P3a amplitude increased with higher cue predictability, indicating stronger engagement of attention when the environment became more structured. A similar modulation was present in the Tourette group but significantly reduced in those with more severe tics. Additionally, invalid cues - stimuli requiring behavioral adjustment - elicited larger P3a responses in controls, an effect that was blunted in the Tourette group. These findings suggest that individuals with Tourette syndrome allocate less attentional flexibility to events that signal the need for adaptation.
The P3b component provided even clearer evidence of impaired predictive updating. In typical participants, P3b onset occurred earlier when valid cues were expected and later when invalid cues violated predictions. This timing shift mirrored behavioral adaptation and reflected dynamic updating of stimulus - response associations. P3b amplitudes also decreased in predictable environments - consistent with theories linking lower surprise to reduced need for updating.
However, these patterns were altered in Tourette syndrome. P3b amplitude modulation by cue predictability was significantly diminished, indicating weaker neural differentiation between predictable and unpredictable contexts. Some individuals with Tourette syndrome even showed reversed patterns, with higher P3b amplitudes in predictable trials - suggesting that their internal models of cue - target relationships were not stabilizing as the environment became more regular. Overall, EEG results showed dampened neural signals associated with updating internal models of the environment - a central feature of predictive processing.
These neural findings align with two prominent frameworks for understanding Tourette syndrome. Predictive coding models propose that tics may arise as maladaptive attempts to resolve persistent sensorimotor prediction errors. Meanwhile, the event coding account suggests that individuals with Tourette syndrome exhibit unusually strong binding between sensory cues and motor actions, making it harder to flexibly adjust those bindings when environmental statistics change. Both frameworks are consistent with the reduced sensitivity to cue predictability observed in the study.
The study also examined correlations with clinical severity. Reduced P3a modulation by cue predictability was associated with higher tic severity on the Yale Global Tic Severity Scale, suggesting that the individuals most affected behaviorally also demonstrated the weakest neural adaptation. This link between symptom severity and predictive processing deficits underscores the functional significance of these electrophysiological markers.
Notably, the study did not find group differences in the core Bayesian learning parameters of the computational model. This may indicate heterogeneity within the Tourette group: some individuals may have difficulty forming predictions, while others may form predictions normally but fail to translate them into efficient responses. Larger sample sizes and subgroup analyses may clarify these mechanisms in future research.
From the perspective of Seven Reflections' Dimensional Systems Architecture (DSA), this study illustrates a disruption in the alignment between internal predictive fields and external environmental structure. DSA frames cognitive processes as interactions between structured fields - such as prediction, motor intention, and sensory input - that must remain coherent for adaptive function. In Tourette syndrome, the diminished modulation of P3a and P3b suggests a weakening of the structural interface that updates and synchronizes these fields. Without precise updating, the system may rely on outdated or unstable predictions, leading to the emergence of tics as corrective but maladaptive outputs. Rather than reflecting a failure of will or attention, these results point to a deeper systems-level misalignment in predictive integration.
As research advances, computational and electrophysiological approaches like those used in this study may help refine diagnostic tools and guide interventions that support more stable sensorimotor prediction. The findings provide compelling evidence that Tourette syndrome involves not only motor symptoms but also fundamental differences in how the brain constructs and updates internal models of the environment.
