Combining motor imagery with action observation training does not lead to a greater autonomic nervous system response than motor imagery alone during simple and functional movements: a randomized controlled trial

Study design

The present study was a randomized controlled trial. The study was planned and performed in accordance with the requirements of the CONSORT (Consolidated Standards of Reporting Trials) statement (Schulz et al., 2010).

Recruitment of participants

A sample of asymptomatic subjects was obtained from La Salle University and from the Community of Madrid through media and social networks, posters, brochures, and emails. The subjects were recruited between March and June of 2017. The inclusion criteria were as follows: healthy and with no pain subjects, and age between 18 and 60. The exclusion criteria included the following: (a) subjects who presented systemic, cardio-respiratory, central nervous system or rheumatic diseases, or those who presented any musculoskeletal pathology with a source of symptoms at the time of the study; (b) underage subjects; (c) subjects with pain at the time of the study; and (d) subjects who were not in full use of their mental faculties and thus were not able to perform the intervention of the study.

Informed written consent was obtained from all subjects prior to inclusion. All participants were given an explanation of the study procedures, which were planned under the ethical standards of the Helsinki Declaration, and which had been approved by the ethics committee of the La Salle University Center for Higher Education (CSEULS-PI-008). The trial was registered with the United States National Institutes of Health Clinical Trials Registry, with the registration number NCT03232879.

Randomization

Randomization was performed using a computer generated random sequence table with a balanced three-block design (GraphPad Software, Inc., CA, USA). A statistician generated the randomization list, and once the initial assessment and inclusion of the participants were completed, the included participants were randomly assigned to any of the three groups using the random sequence list (MI + AO, only MI or the control group) (Fig. 1).

Intervention

MI in isolation

All the subjects in the MI group were informed of the procedure at the beginning of the intervention, which consisted of the following: first, the subjects were supine for 5 min in order to achieve a baseline condition. After that, they sat for two more minutes. Afterwards, two consecutive 30 s visual MI tasks were performed, both based on two movements that are recorded in the revised movement imagery questionnaire in third-person perspective, which is based on the representation of somebody performing a movement.

The first movement chosen for mental practice, starting from a standing up position, consisted of raising the right knee as high as possible and then returning to the starting position. The second movement, starting from a standing up position, and with arms extended along the head, consisted of bending the trunk while trying to touch their toes with their hands and then returning to the initial position. Finally, at the end of the mental training the subjects were at rest for 3 more minutes. Hence, the total time of the intervention was 11 min (Fig. 2).

Control group

The subjects included in the control group (CG) had their neurovegetative variables recorded without having undergone any intervention, and they followed the same measurement times as the two previous groups.

Procedures

After giving their consent to partake in the study, all the participants received a set of questionnaires prior to the intervention. These included a sociodemographic assessment, as well as an evaluation of their physical activity and of their ability to generate kinesthetic and visual MI. The questionnaires given were the following: the Spanish validated version of the International Questionnaire of Physical Activity (IPAQ) and the Spanish validated version of the Revised Movement Imagery Questionnaire (MIQ-R). The latter was given to the participants after the intervention, where the time spent for each task of imagination was also recorded. A Biopac MP150 system, a device to measure autonomic activity, was used for the recordings and processing of the autonomic variables.

Five measurements were taken for each participant: the first of them was taken after lying down for 5 min; the second, after sitting for 2 min; the third, after finishing the first 30 s task of motor imagination; the fourth, after finishing the second 30 s task of motor imagination; and, finally, the fifth measurement was taken 3 min after finishing the mental exercise.

The Biopac MP150 system has an internal microprocessor to control the data acquisition and communication with the computer. There are 16 analog input channels, 2 analog output channels, 16 digital channels that can be used for either input or output, and an external trigger input. The digital lines can be programmed as either inputs or outputs and function in 8 channel blocks. Block 1 (I/O lines 0 through 7) can be programmed as either all inputs or all outputs, independently of block 2 (I/O lines 8 through 15).

Each condition involved a series of inter-sampling intervals (ISIs). In each condition, the first two ISIs (Pre and Pre-Pre’, 420 s) was the baseline (rest) period; the next two ISIs (Pre’- ^1stI and ^1stI- ^2ndI, 30 s each) was the intervention period. Finally, the last ISI (^2ndI-Post, 180 s) was the return to calm phase. Autonomic parameters were continuously monitored and were sampled every 30 s. In the first two ISIs, the Pre interval lasted 300 s. Ten measurements were taken, and the device calculated a mean of the 10 measurements. The Pre-Pre’ interval was measured in the same way (120 s, 4 measurements). In this way, two basal measurements were obtained, the first with the participant lying down and second with the participant sitting. Then, regarding the two subsequent ISIs corresponding to the intervention period, the device averaged the values over intervals of 30 s where 5 values per second were registered during the total intervention time and an average value was automatically obtained after each 30-second interval. Thus, two measurements were taken, one every 30 s (the first relative to the first mental task and the second relative to the second). Finally, the last ISI (^2ndI-Post, 180 s) value was obtained in the same way as the baseline data. The procedure was similar to that performed by Decety et al. (1993).

Previous research has shown how after MI intervention, the LF/HF ratio as an index of ANS activity returned to the resting level. Therefore, the device was programmed to record the ANS activity values concurrent with the MI session; the measurements are the mean of these values (Bunno, Suzuki & Iwatsuki, 2015).

All measurements were taken in a sitting position except for the first one that was taken in a lying position. The measured variables were: (1) heart rate (HR); (2) respiratory rate (RR); and (3) the electrodermal activity (EDA).

Primary outcomes

Secondary outcomes

Mental chronometry

Mental chronometry evaluation was also used to measure the subject’s motor imagery ability. Using a stopwatch, the time spent for performing each MIQ-R task was recorded. Time recorded corresponds to the interval between the command to start the task, given by the evaluator, and the verbal response of conclusion of the task, given by the subject. Mental chronometry is a reliable behavioral task that has previously been employed to collect an objective measure of MI ability (Guillot & Collet, 2005b; Malouin et al., 2008; Williams et al., 2015).

The level of physical activity

The level of physical activity was objectified through the IPAQ questionnaire, which allows the subjects to be divided into three groups according to their level of activity, which can be high, moderate, and low or inactive (Roman-Viñas et al., 2010). This questionnaire has shown an acceptable validity to measure total physical activity. Therefore, the psychometric properties of the questionnaire were accepted for use in studies that required the measurement of physical activity.

Sample size calculation

The sample size was estimated with the program G*Power 3.1.7 for Windows (G*Power© from the University of Dusseldorf, Germany) (Faul et al., 2007). The sample size calculation was considered as a power calculation to detect between-group differences in the primary outcome measures (skin conductance). We considered three groups and four measurements for primary outcomes to obtain 95% statistical power (1-β error probability) with an α error level probability of .05 using analysis of variance (ANOVA) of repeated measures, within-between interaction, and an effect size of ${\eta }_{\mathrm{p}}^2=0.25$ obtained from a pilot study conducted with eight participants (four per group). This generated a sample size of 15 participants per group (a total of 45 participants).

Data analysis

The data analysis was performed using the Statistics Package for Social Science (SPSS 22.00, IBM Inc., Armonk, NY, USA).

For data analysis, we used a confidence interval of 95%, considering all those values that had a p value of less than .05 to be statistically significant.

Descriptive statistics that were used to summarize data for continuous variables are presented as mean ± standard deviation and the 95% confidence interval, while categorical variables are presented as an absolute number or relative frequency percentage. A chi-square test with residual analysis was used to compare categorical variables. The normal distribution of all primary and secondary measures data was assessed using the Shapiro–Wilk test (p > .05). In order to compare the means of the ANS measures, the repeated measures ANOVA test was used. The effect size (Cohen’s d) was calculated for the main autonomic variables. According to Cohen method, the effect was considered as small (0.20 to 0.49), medium (0.50 to 0.79), and large (>0.80) (Cohen, 1988). To conclude, the association between the autonomic response and the ability to generate motor imagery, the mental chronometry, and the level of physical activity was examined using the Pearson’s correlation coefficient. A Pearson’s correlation coefficient greater than 0.60 is considered to have a strong association, one between 0.30 and 0.60 indicates a moderate association, while less than 0.30 indicates a poor association (Hinkle, Wiersma & Jurs, 1990).

Primary outcomes

Secondary outcomes

In the correlation analysis, no association was found between the ANS response and the ability to generate motor images (p > .05). However, regarding mental chronometry, it was found that there is a moderate-strong positive association in the MI group with respect to HR increase (p < .05, r = .643), RR and the time spent for performing each MIQ-R task (p < .05, r = .575). Regarding the level of physical activity, based on the data obtained, there appears to be a moderate-strong association between a higher IPAQ score and a higher HR in the MI (p < .05, r = .559) and MI + AO (p < .05, r = .621) groups, even though no correlation was found in the other variables.

Complexity of movement

The movements chosen to carry out the mental practice in the present study were movements that were easy to perform, comfortable, daily, and at a neurophysiological level, they were movements with a broad subcortical participation, minimizing the requirements of cortical activity, which included raising the right knee as high as possible and bending the trunk while trying to touch the toes with the hands. The data revealed that all the groups started from medium-high levels of physical activity. Several studies have suggested that higher levels of physical activity are associated with a greater ability to generate mental motor images (Robin et al., 2007; Di Corrado, Guarnera & Quartiroli, 2014). The hypotheses to be able to explain this finding reside in relation to the differences in perception and somatosensory afferent integration of the body as well as the processing and control of the voluntary movement, neurophysiological mechanisms widely related to the ability to generate mental motor images. Therefore, the authors of the present study hypothesize that the movements chosen for mental practice did not present any complexity for the participants and they were widely automated. In relation to this, Kuhtz-Buschbeck et al. (2003) used functional magnetic resonance imaging (fMRI) to observe cortical activity during MI while performing simple and high-complexity tasks. They found that at the neurophysiological level, premotor, posterior parietal and cerebellar regions were significantly more active during motor imagery of complex movements than during mental rehearsal of the simple task.

Demougeot et al. (2009) obtained that imagined wrist cyclic movements, which are easy to visualise and have a very low intensity, did not provoke any variation in the HR compared to the control group, however, during mental simulation of greater complexity, such as trunk or leg movements against gravity, physiological parameters (e.g., AP and HR) significantly increased. In this study, digital motor gestures are performed even less intensely than those used in our study, which may be the explanation of the results obtained. Moreover, it has been demonstrated that MI combined with AO training is able to achieve a motor learning greater than the two interventions in an isolated way (Smith & Holmes, 2004; Wright & Smith, 2009; Eaves et al., 2016). Holmes & Calmels (2008) argued that in tasks of high complexity, providing a visual input through the AO facilitates their subsequent motor imagination, as it eliminates the need for visual imagination, allowing, therefore, the attentional focus of the imagery to concentrate on the kinesthetic subdomain.

A possible answer to this finding could be that imagined movements, the visual input that is obtained after AO training in a widely integrated and automated motor gesture, does not provide relevant visual information that facilitates the later imagination of the gesture, and therefore both the combined group and the group with the MI training in isolation undergo the same changes concerning the vegetative response. Thus, in more complex motor tasks, AO training in combination with MI would cause a greater autonomic response than MI in isolation. Therefore, movement complexity is a limited factor in the present study; more complex movements should have been used in the present study to verify this finding.

Mental effort and intensity

Decety et al. (1991) found that the self-representation of walking activated the SNS, and that both HR and RR varied with the degree of mental effort of the representation as well as with the intensity of the action during the mental simulation of locomotion. In addition, Paccalin & Jeannerod (2000) obtained that the RR in subjects that were visualizing people running on a treadmill increased as the intensity of the observed exertion did. Therefore, the intensity of the changes caused in the ANS is closely related to the observed effort, which is greater when the action requires a greater effort.

Bolliet, Collet & Dittmar (2005) found that participants, when they observed squats at both 50% and 90% of the personal best mark, had ANS responses that varied as a function of movement intensity: autonomic responses recorded during movement observation at 90% were significantly greater and longer than those recorded during movement observation at 50%. They concluded that autonomic responses were strongly linked to the amount of observed effort. These data could be correlated with those obtained in the present study, although it would have been necessary to include exercises that require greater effort to obtain stronger conclusions. Finally, regarding mental effort, a recent study performed by Wriessnegger et al. (2017) examined, at a cortical level, force-related hemodynamic changes during the performance of a motor execution and MI task by means of multichannel functional near-infrared spectroscopy. It was found that the most difficult mental task, which needed greater mental effort to perform, led to higher oxygen-hemoglobin concentration changes.

Electrodermal activity: unique innervation

Electrodermal activity is one of the most studied and relevant measures when assessing the activation of the SNS. The reason for this assertion is that the variations in EDA arise from the activity of the sweat glands, which have an exclusively sympathetic innervation. This is an exception, since usually the sympathetic responses can be modulated by a parasympathetic response, as it occurs with the HR or the RR. However, as for the EDA, this double innervation does not occur, and the decrease in the activity of the sweat glands is produced by the cessation of sympathetic innervation (Shields et al., 1987). Consequently, it is suspected that the increase in the EDA supposes a physiological change in SNS prior to the accomplishment of a movement, and that increase might be preceded by a cortical level processing (Vissing, Scherrer & Victor, 1991; Vissing & Hjortsø, 1996; Critchley, 2002). The results here in obtained for the EDA follow in agreement with the findings described in the scientific literature since Di Rienzo et al. (2015) and Oishi, Kasai & Maeshima (2000) reported increases in the electrodermal response after the completion of a protocol of motor imagery.

It should be noted, however, that no significant differences were found between the two intervention groups in the results of the EDA. This may be due to the fact that, as Tremayne & Barry (2001) found, the increases in EDA depend on the cortical regulation of the ANS and, hence, on the planning of the energy expenditure to be performed depending on the needs of the task being executed. The reason for this difference in the values obtained could be due to the fact that the movements included in the protocol of our study are of low intensity and widely integrated in subcortical structures, and said tasks would not require any energy expenditure or any great previous electrodermic preparation.

In addition, Guillot et al. (2008) based on simple movements, concluded that subjects who had not been trained in a motor imagery process had lower values in the EDA compared to subjects who had been trained prior to the imaging process. This result suggests the importance not only of the type of task, but also of the mental effort and the ability to generate motor imagery of the subjects.

Ability to imagine and level of physical activity

The secondary objective of this research was to observe if there were any relationships between the autonomic response and the ability to generate MI, the mental chronometry, and the level of physical activity.

The scores of both the MIQ-R and the mental chronometry were used with the aim of assessing the ability to generate both kinesthetic and visual MI of the subjects of study. The scores found on the MIQ-R questionnaire were high in all groups, and there were no differences between them, hence showing that the subjects had a great ability to generate MI. The results obtained in the study conducted by Peixoto Pinto et al. (2017) and Guillot et al. (2008) were similar to those obtained in the present study. Nonetheless, no correlation was found between the autonomic response and the ability to generate motor imagery. Concerning mental chronometry, the study of correlations showed that the longer the time spent in the practice, the greater the autonomic response obtained. Thus, we can agree that mental effort and attention seem to be very closely related to autonomic response.

Regarding the level of physical activity, the data revealed that all the groups started from medium-high levels of physical activity, with no differences between them. A high level of physical activity correlates with a higher ANS response, as reported by Oishi, Kasai & Maeshima (2000). Their study compared professional and novice skaters and it was shown that there was a higher increase in HR and RR in higher levels of physical activity. In addition, Di Corrado, Guarnera & Quartiroli (2014) found that athletes had higher mean scores on imagery ability than the non-athlete group, suggesting the relationship between physical activity and the ability to generate motor imagery. A moderate-strong association was obtained in both intervention groups, between the HR and the level of physical activity, but neither in the RR nor in EDA, probably because of the characteristics of the motor gestures used for the mental motor task.

Clinical implications

The results of this research must be interpreted carefully because the study was conducted with healthy participants. It is not possible to completely extrapolate the results to patients who have pain or functional disorders. Despite this, the movements used in the present study, lifting the leg and bending the trunk, are widely used in a clinical setting. Thus, if the clinician wants to perform a mental practice of a simple movement, it is not necessary to perform AO training in addition to MI training, thus optimizing the treatment time.

Currently, it is unknown whether MI and AO might play a role as a cognitive tool in rehabilitation. It is therefore relevant to the theoretical explanation of MI and AO to increase knowledge about ANS activity. These findings could be useful in the therapeutic use of MI and AO as cognitive tools in rehabilitation, providing further clues about movement complexity and intensity for clinical use.

ANS activation during mental practice could provide useful indications to clinicians for the level of effort that patients develop during rehabilitation by means of mental training. This can help to optimize mental training and to minimize discomfort or fatigue for patients, as Demougeot et al. (2009), argued.

MI therefore enables the practice of movements without the need to physically perform them, and has been widely used in the training of technical skills to athletes and musicians, as well as in neurorehabilitation (Calmels et al., 2006). In the field of sports fitness, numerous research studies have suggested that MI accelerates and improves sports performance and the learning of motor skills. Mental practice has also been shown to improve (at the psychosocial level) athletes’ motivation and confidence, thereby reducing the anxiety inherent in a competitive event (Guillot & Collet, 2008; Ridderinkhof & Brass, 2015). It could be used clinically to improve the psychosocial sphere of patients through a mental practice adapted to them.

Finally, MI could be considered a complement to physical training because the combination of the two has been widely shown to be more effective in terms of performance than isolated physical training (Feltz & Landers, 1983). Thus, MI could have an impact on rehabilitation to improve performance in patients seeking to achieve physical goals.

Limitations

One limitation of our study was that a movement of greater complexity and intensity could have been included to be able to compare the results between groups and to verify their influence in relation to the response of the ANS. Movement complexity should be considered as a dependent variable to be tested in future research. We also consider that it is necessary to evaluate the ANS activity evoked by mental practice according to the different levels of previous physical activity.

Moreover, our results do not allow us to isolate the effects of only AO in ANS activity. It could be interesting in future research to perform an isolated AO intervention to test how it influences ANS activity. Another limitation was the impossibility of performing masking in the present study. Finally, a placebo intervention was not performed for the CG. It would have been interesting to record the changes in a placebo intervention for the CG.