Increased scale-free and aperiodic neural activity during sensorimotor integration-a novel facet in Tourette syndrome

Abstract

Tourette syndrome is a common neurodevelopmental disorder defined by multiple motor and phonic tics. Tics in Tourette syndrome resemble spontaneously occurring movements in healthy controls and are therefore sometimes difficult to distinguish from these. Tics may in fact be mis-interpreted as a meaningful action, i.e. a signal with social content, whereas they lack such information and could be conceived a surplus of action or 'motor noise'. These and other considerations have led to a 'neural noise account' of Tourette syndrome suggesting that the processing of neural noise and adaptation of the signal-to-noise ratio during information processing is relevant for the understanding of Tourette syndrome. So far, there is no direct evidence for this. Here, we tested the 'neural noise account' examining 1/f noise, also called scale-free neural activity as well as aperiodic activity, in n = 74 children, adolescents and adults with Tourette syndrome and n = 74 healthy controls during task performance using EEG data recorded during a sensorimotor integration task. In keeping with results of a previous study in adults with Tourette syndrome, behavioural data confirmed that sensorimotor integration was also stronger in this larger Tourette syndrome cohort underscoring the relevance of perceptual-action processes in this disorder. More importantly, we show that 1/f noise and aperiodic activity during sensorimotor processing is increased in patients with Tourette syndrome supporting the 'neural noise account'. This implies that asynchronous/aperiodic neural activity during sensorimotor integration is stronger in patients with Tourette syndrome compared to healthy controls, which is probably related to abnormalities of GABAergic and dopaminergic transmission in these patients. Differences in 1/f noise and aperiodic activity between patients with Tourette syndrome and healthy controls were driven by high-frequency oscillations and not lower-frequency activity currently discussed to be important in the pathophysiology of tics. This and the fact that Bayesian statistics showed that there is evidence for the absence of a correlation between neural noise and clinical measures of tics, suggest that increased 1/f noise and aperiodic activity are not directly related to tics but rather represents a novel facet of Tourette syndrome.

Keywords: Tourette syndrome; cognition; dopamine; signal-to-noise ratio.

Graphical Abstract

Figure 1

Event file coding paradigm. Schematic illustration of the event file coding paradigm used in this study. For further details please see the main body of the manuscript.

Figure 2

Behavioural data in patients with Tourette syndrome and healthy controls. Full depiction of the behavioural data (mean response times and hit rates) in patients with Tourette syndrome and healthy controls as a function of task conditions (feature overlap and response). Each plot shows the individual data as a scatterplot (y axis jitter added for discernibility) in addition to boxplot information and a density-based distribution. Illustrations were accomplished using the Raincloud plot toolbox. Outliers are indicated by diamonds. For the hit rates, the data underlying the significant interaction ‘group’בfeature overlap’בresponse’ [F(1,146) = 4.227; P = 0.042; η_p2 = 0.028] are shown. This interaction was not significant for the reaction time data shown [F(1,146) ≤ 1.286; P ≥ 0.259].

Figure 3

Results of 1/f analyses. (A) Log–log-transformed PSD data of SR task conditions averaged across electrodes in the different groups (Tourette patients and healthy controls). Neural noise is defined as the respective log₁₀-transformed PSD slope parameter with more positive values indicating higher neural noise. Scalp topographies are illustrated for each group and condition (Tourette patients: upper right corner; healthy control group: lower left corner). (B) Top: electrode sites with significant differences of neural noise between groups. Significant values underwent false discovery rate correction using the Benjamini–Hochberg method; (q<0.05). Bottom: PSD plots averaged over this electrode set and task conditions. (C) Bayes factors obtained for the correlations between neural noise and clinical scales as a function of sample size. All Bayes factors (BF₁₀<1; detailed values ranged between 0.15 and 0.67 depending on clinical parameter examined as shown in Table  2) indicate moderate evidence that no relationship between neural noise and the respective clinical scale exists in this sample. Note that for clarity Bayes factors in this plot are based on a bivariate model, which is why they diverge from values in Table  2. GTS = Gilles de la Tourette syndrome; Rush total = Rush video-based rating, total score; Rush mot/min = Rush video-based rating, motor tics per minute; Y-BOCS = Yale-Brown Obsessive Compulsive Scale; YGTSS = Yale Global Tic Severity Scale (total score).

Figure 4

Results of the FOOOF analyses. (A) Log–log-transformed PSD data of SR task conditions averaged across electrodes in the different groups (Tourette patients and healthy controls). Neural noise is defined as the respective log₁₀-transformed PSD slope parameter with more positive values indicating higher neural noise. Scalp topographies are illustrated for each group and condition (Tourette patients: upper right corner; healthy control group: lower left corner). (B) Top: electrode sites with significant differences of neural noise between groups. Significant values underwent false discovery rate correction using the Benjamini–Hochberg method; (q<0.05). Bottom: PSD plots averaged over this electrode set and task conditions. (C) Bayes Factors obtained for the correlations between neural noise and clinical scales as a function of sample size. All Bayes factors (BF₁₀<1; detailed values ranged between 0.20 and 0.86 depending on clinical parameter examined as shown in Table  2) indicate moderate evidence that no relationship between neural noise and the respective clinical scale exists in this sample. Note that for clarity Bayes factors in this plot are based on a bivariate model, which is why they diverge from values in Table  2. GTS = Gilles de la Tourette syndrome; Rush total = Rush video-based rating, total score; Rush mot/min = Rush video-based rating, motor tics per minute; Y-BOCS = Yale-Brown Obsessive Compulsive Scale; YGTSS = Yale Global Tic Severity Scale (total score).