Muscle activity and lower body kinematics change when performing motor imagery of gait

Abstract

Motor imagery (MI) is a mental simulation of a movement without its actual execution. Our study aimed to assess how MI of two modalities of gait (normal gait and much more posturally challenging slackline gait) affects muscle activity and lower body kinematics. Electromyography (biceps femoris, gastrocnemius medialis, rectus femoris and tibialis anterior muscles) as well as acceleration and angular velocity (shank, thigh and pelvis segments) data were collected in three tasks for both MI modalities of gait (rest, gait imagery before and after the real execution of gait) in quiet bipedal stance in 26 healthy young adults. No significant change was observed in electromyography activity and lower body kinematics when comparing MI tasks of normal gait. A significantly higher acceleration for the lower limb segments in the vertical direction and for the pelvis in the mediolateral and anteroposterior direction and angular velocity for pelvis in the frontal plane were found during MI of slackline gait after its real execution compared to rest. The results show that MI of normal gait does not lead to any significant changes, while MI of slackline gait affects lower body kinematics parameters.

Introduction

Motor imagery (MI) is a dynamic cognitive state in which movement is rehearsed internally, without being performed and without overt muscular contraction^1,2. Imagination of performing the movement from first-person perspective (kinesthetic imagery) is reportedly similar to the real movement performed when considering neural activation and spatiotemporal parameters (mainly speed and cadence)^3–5. MI is a part of movement planning and neuroimaging studies have reported similar activation of the cortical and subcortical centres during real and imagined movements^4,6–9. Neural activity during MI may be further facilitated by previous practice of motion imagination¹⁰. MI of lower-limb movements, including gait, rely mainly on the supplementary motor area, premotor and parietal cortex, cingulate motor area, basal ganglia, thalamus and cerebellum^4,6,7. The activity of cortical areas related to MI of gait may facilitate the excitability of spinal motoneurons with a subthreshold level¹¹, which might influence even muscle activity. As the demands of locomotor tasks require increased cognitive and sensory information processing, higher brain centres become increasingly involved in MI of gait^4,12.

MI has been reported to improve movement execution in individuals with intact brain regions and in those who have the ability to imagine movements according to movement imagery questionnaires, especially when combined with actual movement practice^13–15. The Movement Imagery Questionnaire-Revised (MIQ-R) has been used primarily in healthy participants for demanding movement tasks and has shown good internal reliability and validity¹⁶. MI can influence movement execution by optimising the regulation of autonomic arousal, reducing anxiety, enhancing motivation (i.e., self-efficacy and confidence), and modulating the inhibition and excitation in the brain, muscles, and autonomic nervous system¹⁷. The neurophysiological similarities of movement imagination and execution support the usefulness of MI as a training technique in professional sport^3,18 and in clinical practice, where MI as a therapeutic approach has been shown to enhance the performance of real movement, including improved range of motion, muscle strength, flexibility, agility, and speed^19,20. Neuromuscular coupling during MI has been used effectively e.g. in stroke survivors or in patients with spinal cord injury²¹ to stimulate function of the paretic/plegic muscles^21,22 and to induce brain plasticity by a motor learning process^13,21,23.

MI of gait is being effectively used as a therapeutic technique to improve real gait in individuals with Parkinson’s disease or poststroke^24,25. However, the effect of gait imagery alone is usually masked by the complexity of other aspects which might occur during therapy as concurrent physical activity or external and internal condition changes. Firstly, it is important to objectify changes in MI of gait in healthy individuals without gait impairment and unmasked follow-up MI therapy.

Previous studies have proved that electromyographic (EMG) signals from muscles can be detected during imagined movements, which are based on the incomplete inhibition of corticospinal activity during the mental process^18,26. In accordance with previous studies, it seems that MI of non-gait lower limb tasks does not affect the EMG activity of the involved body segments^27,28. However, MI of gait has previously been shown to have the potential to modulate activity of muscles engaged in the gait execution^29–31. The main finding of these studies was no difference or decrease in lower limb muscle activity, with the inhibitory effect of gait imagery mainly present for the tibialis anterior^29–31 or also the proximal lower limb muscles³¹ in normal gait after its real execution compared with at rest. MI can also modulate postural sway^27,28 in a task-dependent manner, especially in participants with good imagery skills²⁷. Previous studies have also shown that EMG and kinematic patterns during MI are related to corticospinal excitability, Hoffman reflex and cortical activity^5,32.

Overall, it is evident, that training of motion in imagination influences the final mechanical output, which should be reflected by changes in both muscle activity and kinematics of engaged segments. However, studies concerning the effect of gait imagery on muscle activity or on changes in postural sway are still lacking. Therefore, this study aimed to investigate how the muscle activity and lower body kinematics of healthy young adults change when imagining performing a normal gait (GI) and a more posturally challenging gait modality on a slackline (SLI). As MI of gait in rugged terrain or with obstacles can lead to increased activation of cortical and subcortical brain regions than gait imagery alone^4,33,34, even the effect on lower body kinematics might be more evident. We hypothesised that imagining performing SLI would produce greater muscle activation in selected dominant lower limb muscles and greater acceleration and angular velocity in dominant lower limb segments and pelvis than imaging performing GI. Compared to the normal gait, gait on a slackline is characterized by placing the feet parallel to the sagittal plane, and therefore increases the demands on mediolateral stability of the body’s centre of gravity over a narrow base of support³⁵.

In addition, we hypothesised that MI of gait after its real execution would change lower limb muscle activity and lower body kinematics compared to resting task and MI of gait before its real execution. This assumption is supported by previous studies demonstrating that MI of movements immediately following real performance allows learning of motor skills³⁶ and more lively imagination of movement which could have a facilitative effect on the motor periphery (in our case, detected muscle activity and kinematics of the dominant lower limb segments and pelvis).

Methods

Participants

Participants were recruited from the student population of Palacký University Olomouc in 2019. The inclusion criteria were the absence of congenital or acquired neuromuscular deficits, absence of cognitive deficits, a good level of MI according to the MIQ-R¹⁶, no previous experience with MI, no previous experience with gait on a slackline, and an absence of pain at the time of measurement. All participants signed an informed consent form before data collection, in accordance with the Declaration of Helsinki, and the study was approved by the Ethics Committee of the Faculty of Health Sciences of Palacký University Olomouc (No. UPOL-28609/1040-2019).

Experimental tools

Movement imagery questionnaire-revised

The MIQ-R consists of four test items for visual MI (which involve perceiving the execution of movement from a first- or third-person perspective in which individuals imagine themselves seeing performing a movement) and four items for kinesthetic MI (which involve perceiving the execution of movement from a first-person perspective and the individual imagines the feeling of the movement, including muscle contractions and sensation during a movement)^1,2,37. These included lifting the flexed knee upward, bending forward and then jumping upward, extending the arm straight out to the side and moving it forward, and bending forward¹⁶. Each participant rated the quality of MI of the above movements on a seven-point scale, with higher scores (≥ 4 points) indicating better ability of MI¹⁶.

Electromyographic and segment accelerations and angular velocities data

EMG activity, segment acceleration, and angular velocity data were collected using Trigno™ IM Sensors (Delsys, Inc., Natick, MA, USA). Each wearable EMG sensors contained a built-in triaxial accelerometer (± 16 g) and gyroscopes (± 2,000 deg/s). The EMG data were sampled at 1, 111 Hz at a 16-bit resolution using a 20–450 Hz analogue bandwidth filter, and the segment acceleration and angular velocity data were sampled at 148 Hz at a 16-bit resolution. Each inertial sensor measured 3.7 cm × 2.6 cm × 1.5 cm and weighed 14.7 g.

The EMG data were obtained from the biceps femoris long head (BF), gastrocnemius medial head (GM), rectus femoris (RF), tibialis anterior (TA) muscles of the dominant lower limb^29,31. Before sensor application, each muscle was palpated during isometric contraction, the skin surface overlying the centre of the muscle belly was cleaned with alcohol, and the hair was removed. Sensors were attached to the skin using hypoallergenic double-sided tape. The sensor locations were determined according to the Surface Electromyography for Non-Invasive Assessment of Muscles guidelines³⁸. Lower limb dominance was determined by asking all participants, "Which lower limb would you kick a ball with?"³⁹. Sensors for collecting segment acceleration and angular velocity data were placed on the skin overlying the pelvis (centre of the sacrum), bilaterally on the thighs (lateral side of the thigh 15 cm above the femoral epicondyle), and on the shanks (lateral side of the shank 15 cm above the lateral malleolus)^40,41.

Data collection

Participants took part in a single experimental session and were asked to imagine the GI in similar manner as tested previously^29–31 and SLI modalities (Fig. 1a,b). The GI modality consisted of rest, GI before and after the real execution of normal gait tasks; the SLI modality consisted of rest, SLI before and after the real execution of gait on a slackline tasks. The order of the imagined gait modalities was randomised; however, the order of the tasks was not randomized and was in the indicated order. The characteristics of each task and instructions for participants are listed below.

1. Rest consisted of a non-MI task in which participants were asked to sing “Happy Birthday” in their minds, which is commonly used for MI studies^29–31,42. The participant’s initial position was an upright bipedal stance with feet placed pelvic-width apart, arms held loosely at the sides of the body, eyes open, and facing forward (Fig. 1c). The rest conditions were recorded before both gait imagery before its real execution tasks.

2. GI before the real execution of normal gait and SLI before the real execution of gait on a slackline were tasks using only the imagination, in which the participants only observed the environment in which they subsequently imagined walking. The participant initially viewed a 3 m section of the corridor for GI^29–31 or a slackline on a self-supporting structure placed in the centre of the corridor. The slackline tape (Gibbon Slacklines®, Germany) was 3 m long, 4 cm wide, and 30 cm from the ground (Fig. 1b). The participant then stood in the initial position and kinesthetically imagined walking after receiving verbal instructions. For GI, they were told, "Imagine yourself walking down the corridor you just viewed”^29–31. For SLI, they were told, "Imagine yourself walking on the slackline you just viewed”.

3. GI after the real execution of normal gait and SLI after the real execution of gait on a slackline were imagined tasks that followed the real execution of the gait tasks. The verbal instructions were, "Imagine yourself walking down the corridor you just walked"^29–31 and "Imagine yourself walking on the slackline you just walked,” respectively.

Fig. 1.

Experimental tasks of (a) normal gait imagery, (b) slackline gait imagery, and (c) initial position for all tasks. Informed consent was obtained from the participant to publish the photo in an online open access publication. Photo source – own.

All tasks were performed once for 30 s^30,31. This period was selected to maintain attention and prevent fatigue²³. The participants were asked about the ease or difficulty of imagining the MI of gait before and after the real execution of gait tasks using a 5-point scale (5 points [easy] to 1 point [difficult]) [similarly as in Ref.^30,31] (Table 1).

Table 1.

Demographics and outcome scores (mean ± SD) of the questionnaires.

Characteristics
Age (years)	23.5 ± 1.69
Height (cm)	176.2 ± 9.08
Weight (kg)	68.5 ± 8.89
Sex (Female:Male)	13:13
LL dominance (Left:Right)	0:26
MIQ-R
TS	5.97 ± 0.18
KMI	6.03 ± 0.16
VMI	5.9 ± 0.17
Evaluation of gait imagery
GI before	3.35 ± 0.78
GI after	4.31 ± 0.87
SLI before	3.31 ± 0.91
SLI after	4.19 ± 0.88

LL lower limb, MI motor imagery, MIQ-R Movement Imagery Questionnaire-Revised, TS total score, KMI kinesthetic motor imagery, VMI visual motor imagery, GI before normal gait imagery before the real execution of normal gait, GI after normal gait imagery after the real execution of normal gait, SLI before gait on slackline imagery before the real execution of gait on slackline, SLI after gait on slackline imagery after the real execution of gait on slackline.

Data analysis and statistical evaluation

EMGworks (Delsys, Inc., Natick, MA, USA) was used to process the EMG, segment acceleration, and angular velocity data. The mean of the middle 20 s of each record was evaluated for the dominant lower limb segments and pelvis⁴³. The middle period was chosen because of the high consistency of data between tasks. The EMG data were processed using Visual 3D software (C-Motion Inc., Germantown, MD, USA) with the remove mean function and root mean square with 50 window frames³¹. EMG data from the BF, GM, RF, and TA were normalised to the maximum observed signal from all tasks prior to statistical analysis^31,43–45. Segment acceleration and angular velocity data from the shank and thigh of the dominant lower limb, and pelvis were recorded in three directions (anteroposterior, mediolateral and vertical) and planes (sagittal, frontal and transverse).

All data were analysed using Statistica software version 14 (TIBCO Software Inc., Palo Alto, CA, USA). The normal data distribution was verified using the Shapiro–Wilk test. A multivariate analysis of variance (MANOVA) with two factors (gait imagery modalities – GI and SLI and tasks for both MI modalities of gait – rest, gait imagery before and after its real execution) was used to assess the effect of gait imagery modality. The interaction between these two factors was also assessed. Twenty-two variables were used for the analysis, i.e. four lower limb muscles (BF, GM, RF and TA), three segments (shank, thigh and pelvis) for acceleration (all three directions) and angular velocity (all three planes) in dominant lower limb. A Bonferroni post-hoc test was used for pairwise comparison. Subsequently, MANOVAs were also used for GI and SLI modalities separately for one factor (task) for all twenty-two variables. A Bonferroni post-hoc test was used for pairwise comparison between rest, gait imagery before and after its real execution. The level of statistical significance was set at p < 0.05. Effect sizes were reported using partial eta-squared (ηp²), as small (˃ 0.01), medium (˃ 0.06), or large (˃ 0.14)⁴⁶.

Results

Participants

Overall, 26 healthy young participants were selected according to the inclusion criteria and completed the study. The demographic data and questionnaire outcome scores are summarised in Table 1.

Electromyographic and segment accelerations and angular velocities data

MANOVA (Table 2) showed a significant main effect of gait imagery modality and no significant main effect of task. Significant interactions were not observed between gait imagery and task.

Table 2.

Multivariate analysis of variance results for all gait imagery data.

Main Effect	Wilks value	F	p	ηp2
Gait imagery (Normal, Slackline)	0.720	2.177	0.004	0.2803
Task (Rest, Before, After)	0.729	0.959	0.550	0.1464
Interaction
Gait imagery × task	0.669	1.247	0.151	0.1824

Rest – non-imagery task; Before – gait imagery before the real execution of gait; After – gait imagery after the real execution of gait. Statistically significant differences are in bold.

Post-hoc analysis showed a significantly higher muscle activity for BF when comparing GI rest and SLI before the real execution (p = 0.018). Data analysis did not show statistical significance for segment acceleration. Significantly higher angular velocity was observed for pelvis in frontal plane when comparing SLI after the real execution and GI before the real execution (p = 0.002) and when comparing SLI after the real execution and GI after the real execution (p = 0.01).

MANOVAs for GI (Wilks value 0.573, F = 0.745, p = 0.863, ηp² = 0.243) and SLI (Wilks value 0.399, F = 1.351, p = 0.109, ηp² = 0.368) modalities separately showed no significant main effect of task.

Post-hoc analysis for SLI modality did not show statistical significance for EMG activity (Fig. 2). Analysis showed a significantly higher acceleration for shank (p = 0.026) and thigh (p = 0.034) in vertical direction, pelvis in anteroposterior direction (p = 0.037), mediolateral direction (p = 0.024) after the real execution compared with at resting task (Fig. 3). Additionally, there was a significant increase for pelvis in frontal plane (0.017) after the real execution compared with at resting task (Fig. 4).

Fig. 2.

The means and standard deviations of EMG data. Rest – non-imagery task; Before – gait imagery before the real execution of gait; After – gait imagery after the real execution of gait; *p < 0.05.

Fig. 3.

The means and standard deviations of segment acceleration data. Rest – non-imagery task; Before – gait imagery before the real execution of gait; After – gait imagery after the real execution of gait; *p < 0.05.

Fig. 4.

The means and standard deviations of angular velocity data. Rest – non-imagery task; Before – gait imagery before the real execution of gait; After – gait imagery after the real execution of gait; *p < 0.05.

Discussion

This study aimed to investigate how muscle activity and lower body kinematics change when imagining two different complex movements (normal gait and much more posturally challenging slackline gait) in healthy young adults.

The main result of this study was that we found a significant difference between gait imagery modalities. It was detected higher muscle activity for BF when comparing GI rest and SLI before the real execution and higher angular velocity for pelvis in frontal plane when comparing SLI after the real execution and GI before the real execution and when comparing SLI after the real execution and GI after the real execution. The difference could be due to lower means in tasks for most muscles in SLI than in GI, respectively higher means in SLI than in GI for angular velocity. In our study, we can exclude the influence of the modality (i.e. GI and SLI) on each other due to the pre-selected randomisation. Therefore, it was surprising that the resulting muscle activity for the SLI task was lower than for the GI-related task. Our EMG results are in agreement with our previous study in stroke survivors, where MI of a more demanding modality (tandem gait on the line) resulted in lower EMG than normal gait imagery³¹. The significant result of previous study was the decrease in muscle activity of the non-paretic BF during tandem gait on the line imagery after its real execution compared to GI after its real execution and for non-paretic RF during tandem gait on the line imagery before its real execution compared to GI before its real execution³¹. The current and previous EMG data contrast with the subjective assessment of the participants’ ability to generate MI of both gait modalities before and after its execution. Imagining a more demanding option of gait before and after its real execution showed lower EMG, whereas MI generation according to the chosen assessment was more difficult during the more demanding movement³¹. Whereas for lower body kinematics it is the opposite.

The second main result of this study was that we found a significant difference between tasks in SLI. It was detected higher acceleration for shank and thigh in vertical direction, pelvis in anteroposterior and mediolateral direction after the real execution compared with at rest. There was also a significant increase for pelvis in frontal plane after the real execution compared with at rest. In this study, all muscles consistently exhibited a trend towards a decrease in EMG activity across the observed tasks during GI, which was in our study a posturally simpler movement. Normal gait during internal MI is a familiar movement that does not induce the need for excessive EMG activity, which was non-significantly reduced compared to that in the “rest” task in our study. This claim can be supported by the motor learning process during an actual movement (in our case, normal gait), where the control system selects the most optimal brain and neuromuscular activation among all possible options based on the degree of movement familiarity⁴⁷. This was reflected in the EMG pattern, where we expected a decrease in activity rather than an increase^29–31. Through motor learning, an internal model of movement (i. e. specific imagery of the movement) and a sensorimotor map are created and used to predict the degree of movement dynamics and generate anticipatory forces⁴⁸. Thus, in this study, GI might have potential to optimise sensorimotor model of movement and induce lower muscle activity than the rest task. Rest task in our study was associated with the internal imagery of non-motor task (specifically singing a song), which should not lead to activation of similar sensorimotor centres as MI and participants did not perform singing a song as often as normal gait. In our study, we can exclude the influence of the modality on each other due to the pre-selected randomisation, and we can also exclude the influence between tasks, as each imagined movement had its own preceding rest tasks. Our study is in agreement with previous concerning the effect of gait imagery on muscle activity in healthy participants or stroke survivors^29–31, where no difference or decrease in lower limb muscle activity was found, with the inhibitory effect of gait imagery mainly present for the tibialis anterior^29–31 and for the proximal lower limb muscles (BF and RF)³¹ in normal gait imagery after its real execution compared with at rest^29–31 or in normal gait imagery before its real execution compared with at rest³¹. Simultaneously, our present and previous results suggest an alteration in lower-limb muscle activity when the selected movements are imagined, most likely due to the incomplete inhibition in the frontoparietal regions of the brain, basal ganglia, thalamus, and cerebellum during their internal representation^3,7,14,49. During posturally more demanding movement (SLI), the EMG activity tended to increase consistently across the observed tasks, however these changes were not significant. We assumed that there would be the increase mainly in RF activity may be related to its important role in the real execution of gait on a slackline, where the RF or the entire quadriceps femoris is crucial for stabilising the knee joint^50,51. Its role in stabilising the knee and ankle joints during gait on a slackline has been described⁵¹. It was proved that higher brain centres showed an increased activation when demands of imagined locomotor tasks require more attention or visuo-spatial abilities which are more demanding on sensory information processing^12,34. Our assumption of increased muscle activity during MI of posturally more demanding gait was also based on previous studies demonstrating increased activation in the brain regions involved in MI^4,12, which might potentially influence the corticospinal excitability in muscle-specific manner^5,52. In a previous MI study of more demanding option of gait in stroke survivors, the EMG activity consistently decreased across the observed tasks. The lowest activity was observed during tandem gait on the line imagery after its real execution compared with at rest task. The result was probably related to the level of gait achieved (without the need for support and assistance from another person) and thus postural stability of the selected stroke survivors for this study. We concluded that although gait on a slackline is a posturally more challenging task compared to normal gait, a quality movement plan is not created when imagining the task without real experience of the movement^10,36,53, and sufficient muscle activation cannot occur. This is because muscle activity during imagined movement reflects activity during real movement, although at a lower intensity¹⁸. This was supported by the results of asking participants about the ease or difficulty of generating MI after each imagined task, where MI was more difficult to generate during GI before the real execution of normal gait or SLI before the real execution of gait on a slackline than during either GI and SLI after its real execution (Table 1).

In terms of lower body kinematics, the segment acceleration data confirmed that the SLI modality had significantly higher acceleration for the shank, thigh and pelvis, while the GI modality led to a slight decrease, but without statistical significance. In SLI, the segment acceleration data showed a trend towards consistent increases in all segments across the observed tasks. Compared with traditional balance training (solo stance on a soft mat or balance beam), gait on a slackline is a challenge to the postural system because of the constant interaction between the participant and tape dynamicity, which translates into the perception of the task and thus the response of the postural system (i. e. also to the lower body kinematics)⁵⁴. The base of the support is extremely small, movable mainly mediolaterally, unfixed, and characterised by elastic properties⁵¹. Furthermore, gait on a slackline is also challenging in terms of attention and awareness of the contact of the foot against the slackline (information from exteroceptors and proprioceptors)⁵⁴, which may translate into an increase in acceleration and angular velocity of the lower body segments. Considering that the participant is moving on such a small base of support when walking on the tape with one foot placed in front of the other, in addition to the effort required to thoroughly perceive oneself during this movement. Montull et al. (2020) found that any deviation from the optimal alignment during walking must be stabilised by compensatory mechanisms that are ascribed to postural stability and lower body kinematics. This can also explain our results, where the participant stands in an initial position and is asked to purposefully not move but is simultaneously asked to imagine that they are walking. Moreover, in the case of gait on a slackline, the participant does not maintain the base of support that they are used to in normal gait but narrows it. If we consider MI as an essential part of the planning and preparation for a purposeful movement with the definition of its parameters and its effects on one’s own body and environment³, then the aforementioned perceptions must be “interwoven”.

Specifically, for the pelvis, the segment closest to the centre of gravity of the body in bipedal standing, an increase in segment acceleration data was observed in the mediolateral and anteroposterior direction. However, this significant change occurred only during the imagining that followed the real experience with movement, as in previous MI studies of normal gait^29–31. Pure imagery (MI without real-movement experience) did not produce a significant change. An explanation may be sought for the intensity of perception obtained after experiencing the movement. In particular, we assumed that the participants had no previous experience with gait on a slackline. Focusing on the changes that occur across the lower limb segments, the answer can be sought in the actual movement. The pelvis in our study, as the initiator of movement (i.e., hip and trunk muscles), shows demonstrable changes in mediolateral and anteroposterior direction of movement but only during SLI after the real execution of gait on slackline. MI uses a forward internal model that mimics the causal process of real movement by predicting the future sensorimotor state of the body based on an efferent copy of the motor command and current state^55,56. The more distal the segment (i.e., the thigh and shank), the more evident the change in terms of a significant increase in segment acceleration data in the vertical directions. The resulting segmental acceleration is typical for the slackline gait, particularly the significant activity of the thigh and foot muscles, that directly influence the maintenance of segmental stability in our selected lower limb segments⁵¹.

Lower body kinematics was also detected using angular velocity data, which, to the best of our knowledge, has not yet been applied to gait imagery. We confirmed a significant increase in the angular velocity data for pelvis in frontal plane; however, the change was significant only for SLI. This significant change occurred during MI following the real experience and when comparing with rest and not during pure imagination, which is consistent with the segment acceleration data presented above. All changes in the angular velocity data were likely related to the attempt to reproduce the slackline gait in detail in the internal MI. The assumption based on our results is that the kinematics of the lower limb, but not the electromyographic activity, mimics to some extent the real slackline gait after experiencing it through performance. According to our results, it seems that the imagery of a challenging slackline gait task compared to simple gait not only leads to change in kinematics of body but is also reflected in activity of cortical and subcortical networks³⁴. It might be speculated here that the greater the activation of neural structures, the greater the influence on body kinematics. Normal gait, as previously hinted, can be perceived as an automated movement which does not require any extra attention³⁴ and to which the participant is adapted (performing it frequently throughout the day); therefore, there is no significant projection of lower body kinematics in either type of MI but only in SLI⁵⁷. These findings might help for better understanding of the mechanical and neuromuscular consequences during gait imagery, even with respect to the cognitive processing necessary during more demanding gait modalities.

A trend of a slight decrease or no change was observed between the measured tasks during the GI modality, which may be related not only to the lower difficulty of the movement but also to the nature of the gait itself, i.e., tandem versus normal. According to these findings, the initial position or the position of the participants’ feet may also play a role. For both modalities, a standing stance with feet placed at the pelvic width was chosen rather than tandem foot placement, as might be the case for gait on a slackline. According to Holmes and Collins (2001)⁵⁸, this may not have been entirely appropriate for the individual and may have caused a greater tendency to approach movement in the imagined modality. More distally on the thigh and the shank, the change occurred in vertical plane. Based on the changes, we considered the dominant movement of the hip and knee joints during slackline gait and the projection of the elasticity of the tape into the kinematics of the movement.

Overall, even small changes in lower body kinematics levels during gait imagery (Figs. 3, 4) reflect to some extent the effect of imagination on movement execution level. Nevertheless, increasing segment acceleration and angular velocity data might be useful in segmental stability training in imagination, for example, in athletes or patients with musculoskeletal or neurological deficits. For example, MI can be applied in physiotherapy of the lower limb joints, where standing on tape on one leg is recommended to increase proprioception⁵⁹. For conditions requiring immobilisation, mental training is then performed to activate the muscles. A combination of MI and real movement practice is more suitable and increases the effectiveness of training more than physical training alone⁶⁰.

A limitation of this study was the small sample size of only healthy participants without movement limitations, so in a future study we would like to focus on determining the effects of gait imagery in a larger sample of patients with a specific diagnosis (e.g. stroke survivors). As a further limitation, we see the inability to randomize between the GI and SLI tasks due to the nature of our study design and thus the potential learning effect from the first task of imagery (gait imagery before the real execution of gait).

Conclusions

When imagining a posturally less demanding movement (normal gait), muscle activity and lower body kinematics did not change significantly. Imagining a posturally more demanding movement (gait on a slackline) led to a change in lower body kinematics corresponding to the direction of the deviations when performing the real movement. A significantly higher acceleration and angular velocity for the lower limb segments and pelvis occurred during MI of slackline gait after its real execution compared to rest. These results indicate that the difficulty of the chosen movement during imagining is reflected in lower body kinematics parameters. A better understanding of the mechanical and neuromuscular consequences during gait imagery may lead to a better understanding of gait control mechanisms which might be further utilized in rehabilitation practice.

Author contributions

H.H. carried out the data collection, participated in the protocol design, performed the data analysis, participated in the results implementation and wrote the manuscript; M.J. participated in the results implementation and helped to draft the manuscript; B.K. participated in the protocol design, participated in the results implementation and helped to draft the manuscript. All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication.

Data availability

The data sets used and/or analysed during this study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.