Distinct role of central predictive mechanisms in tactile suppression

Summary Tactile sensitivity on a limb is reduced during movement. This tactile suppression results presumably from central predictive mechanisms that downregulate sensations caused during voluntary action. Suppression also occurs during passive movements, indicating a role for peripheral mechanisms, questioning the predictive nature of suppression. Yet, predictions existing beyond the motor domain (non-motor predictions) can also modulate tactile suppression. This study aimed to disentangle central motor predictive and peripheral feedback mechanisms while accounting for non-motor predictions. Participants detected tactile stimuli on their limb shortly before it moved in an active or passive manner. Passive movements were either fully (100%) or partially (50%) predictable. We found tactile suppression during both active and passive movements irrespective of whether the passive movements were predictable. Importantly, tactile suppression was stronger in active than passive movements highlighting the specific role of central predictive mechanisms.

Table S1.Fixed effects of interest from the LMM on stimulation relative to movement onset related to Figure 3C.Statistics are reported for each item with sum of squares (SS), mean squares (MS), degrees of freedom (dfs), F-value (F) and p-value (p).In addition, we calculated the median stimulation amplitude in each bin in order to account for potential influence of overall vibration intensity on detection judgments. 1To determine if time course of tactile sensitivity differed as a function of movement and time, we conducted a LMM with interaction between movement and time bin as fixed factors of interest along with movement, time bin, cue and their interaction as fixed factors, participant as random factor and amplitude (median amplitude in each bin) as covariate.The LMM showed an interaction between movement and time bin; F(4, 586.17) = 5.51, p < .001.For the interaction effect, planned post-hoc comparisons looking at detection rate in each bin as a function of movement revealed significant differences in the active compared to the passive condition in time bins 1 to 3, showing decreased detection rates within the 200-50ms time bin for active as opposed to passive movements.Significant differences revealed by the LMM analysis are indicated by bold asterisks (n = 35 in cue-and n = 32 in cue+).

Figure S1 .
Figure S1.Individual psychometric functions depicting tactile detection as a function of vibrotactile probe intensity in the cue-condition related to Figure 1.The size of the symbols indicates the number of presented trials for each probe intensity.Horizontal lines indicate lapse (upper boundary) and guess (lower boundary) rates.

Figure S2 .
Figure S2.Individual psychometric functions depicting tactile detection as a function of vibrotactile probe intensity in the cue+ condition related to Figure 1.The size of the symbols indicates the number of presented trials for each probe intensity.Horizontal lines indicate lapse (upper boundary) and guess (lower boundary) rates.

Figure S3 .
Figure S3.Group-level detection thresholds as a function of movement and cue related to Figure 1.Dots represent individual data points (n = 35 in cue-and n = 33 in cue+).

Figure S4 .
Figure S4.Correlation between detection measures of d' and c, and stimulus presentation relative to movement onset related to Figure 2 and Figure 3. Dots represent individual data points (n = 35 in cue-and n = 33 in cue+).Line and shading represent the best-fitting regression line and its 95% confidence interval, respectively.All p values > .05after correction for multiple comparisons.

Figure S5 .
Figure S5.Correlation between detection measures of d' and c, and movement duration related to Figure 2 and Figure 3. Dots represent individual data points from all participants (n = 35 in cue-and n =33 in cue+).Line and shading represent the best-fitting regression line and its 95% confidence interval, respectively.All p values > .05after correction for multiple comparisons.

Figure S6 .
Figure S6.Distribution of stimulation onsets relative to movement onsets in each condition collapsed across all participants related to Figure 1.Negative values indicate stimulation occurring before movement.

Figure S7 .
Figure S7.Time course of detection as a function of movement and cue related to Figure2.In order to test if time course of tactile suppression differed between active and passive movements, we conducted an analysis on detection judgments, i.e. to trials in which a vibrotactile stimulus was present and the stimulation occurred max.200ms before movement onset (-200ms) and max.50ms (50ms) after movement onset based on the available data in each condition of interest (active, passive cued and uncued) (see also FigureS6for a distribution of all trials in each condition of interest).For each participant and condition, we then binned the detection responses as a function of stimulation time relative to movement onset into five bins (-200:50:50ms) and calculated the proportion of detected trials in each bin.Data from one participant in the active cue+ condition was excluded due to 0 detection responses in all bins.In addition, we calculated the median stimulation amplitude in each bin in order to account for potential influence of overall vibration intensity on detection judgments.1To determine if time course of tactile sensitivity differed as a function of movement and time, we conducted a LMM with interaction between movement and time bin as fixed factors of interest along with movement, time bin, cue and their interaction as fixed factors, participant as random factor and amplitude (median amplitude in each bin) as covariate.The LMM showed an interaction between movement and time bin; F(4, 586.17) = 5.51, p < .001.For the interaction effect, planned post-hoc comparisons looking at detection rate in each bin as a function of movement revealed significant differences in the active compared to the passive condition in time bins 1 to 3, showing decreased detection rates within the 200-50ms time bin for active as opposed to passive movements.Significant differences revealed by the LMM analysis are indicated by bold asterisks (n = 35 in cue-and n = 32 in cue+).