Is there symmetry in motor imagery? Exploring different versions of the mental chronometry paradigm

Motor imagery and motor execution share similar processes. However, only some factors that affect motor execution affect motor imagery in the same way. We investigated whether bimanual coordination constraints (parallel movements are performed slower than symmetric movements) are observed in motor imagery and whether the way of implementing the mental chronometry paradigm, which is used to investigate motor imagery, influences the results. Participants imagined and executed repetitive symmetric and parallel bimanual movements in three different tasks. Participants performed a certain number of movement repetitions (number task), repeated movements for a fixed duration (duration task), and performed movements in synchrony with pacing sounds (synchronization task). In both, imagination and execution, inter-response intervals were longer with parallel movements than with symmetric movements (number task and duration task), and the percentage of correct movements was lower with parallel than with symmetric movements (synchronization task). Performance of imagined and executed movements was correlated in all tasks. However, imagination took longer or was rated as less accurate than execution, and in the synchronization task the coordination constraint affected accuracy more in execution than in imagination. Thus, motor imagery and overt execution involve shared and unique processes. The synchronization task offers a promising alternative to investigate motor imagery, because the speed-accuracy trade-off is taken into account, different tempi can be used, and psychometric functions can be calculated. Electronic supplementary material The online version of this article (doi:10.3758/s13414-016-1112-9) contains supplementary material, which is available to authorized users.


Number task
. Means and standard errors of inter response intervals (IRIs) in imagination (IMA), execution based on reported data (EXE), and execution based on actual performance (EXE actual) for symmetric and parallel movements, depending on requested number, in the number task Duration task Figure SM2. Means and standard errors of inter response intervals (IRIs) in imagination (IMA), execution based on reported data (EXE), and execution based on actual performance (EXE actual) for symmetric and parallel movements, depending on requested duration, in the duration task.

Correlations between tasks
Pearson correlations between the tasks, separately for the different measures (imagination, reported execution and actual execution) and the coordination patterns (symmetric and parallel) can be seen in Table SM2. Note that whereas correlations between the number and the duration task are positive (IRIs are correlated), they are negative between those tasks and the synchronization task (higher accuracy in the synchronization task coincides with shorter IRIs).

Simultaneousness of hands (inter onset intervals, IOIs)
For all tasks we calculated the inter onset interval (IOI), i.e. the absolute value of the interval between button presses of the left and the right hand of valid bimanual presses as a measure of simultaneousness. Means and standard errors of IOIs can be seen in Figure SM3. A repeated measures ANOVA with the factors pattern (symmetric, parallel) and task (number task, duration task, synchronization task) was performed on IOIs. A significant main effect of pattern, F(1, 20)=72.7, p<.001, η² p =.78, indicated shorter IOIs in symmetric (M=13 ms) than in parallel movements (M=19 ms). The significant interaction between pattern and task, F(2, 40)=3.7, p=.034, η² p =.16, indicated that the difference between IOIs of parallel and symmetric movements was significantly smaller in the number task (4.6 ms) than in the duration task (7.7 ms; p=.023, d=.63). The difference in the synchronization task (7.7 ms) did not differ significantly from the differences in the other tasks (p min =.079). The main effect task was not significant, F<1. Thus, in all three tasks, IOS were lower in symmetric movements (higher simultaneousness) than in parallel movements. This further supports the observation that symmetric movements are easier to perform than parallel movements. Figure SM3. Means and standard errors of inter onset intervals (IOIs) for symmetric and parallel movements in the number task, the duration task, and the synchronization task.

Strength of representation
In addition to performance, we investigated the strength of kinesthetic/tactile and visual representation during imagination and execution using performance reports.
This was of interest, because even when imagination and execution durations are similar, movements might still be represented in a slightly different way. Given the equivalence assumption (Jeannerod, 1995) In line with previous results (Dahm & Rieger, 2016) execution and imagination did not differ significantly from each other and stronger representation of kinesthesis/touch than vision was observed. Importantly, these effects did not significantly differ between tasks and the the strength of representation did not significantly differ between imagination and execution. Figure SM4. Means and standard errors of the strength of kinesthetic/tactile and visual representations in execution (EXE) and imagination (IMA) in the number task, the duration task, and the synchronization task.

Order effects
For the number and duration task we calculated a repeated measures ANOVA with the between factor order (EXE first, IMA first) and the within-factors task (number, duration), action (EXE, IMA), and pattern (symmetric, parallel) on IRI. For the synchronization task we calculated a repeated measures ANOVA with the between factor order (EXE first, IMA first) and the within-factors action (EXE, IMA), pattern (symmetric, parallel), and ISI (167,182,200,222,250,286,333,400, and 500 ms) on COR%. Neither the main effects of action order nor any of the interactions with order became significant in both analyses. The results of the ANOVAs can be seen in Table SM7.