Nineteen right-handed healthy young adults between 20 and 25 years of age and comprising seven females and 12 males participated in this study. The power calculation for the sample size was performed using G * Power [57] (effect size f = 0.3, α = 0.05, power (1 − β) = 0.9, correlation among repeated measures = 0.5, η p 2 = 0.06). The required sample size with an actual power of 0.9 was of 17. Participants were excluded if they had a history of epileptic seizures, musculoskeletal diseases (including the fatigue of an arm), or cardiovascular diseases. Handedness was assessed using the Edinburgh Handedness Inventory [58]. Eight participants were excluded from the analysis due to noise in the collected data. Thus, a total of eleven participants were included for the final analysis (five females and six males; age, 22.5 ± 1.56 years; height, 165.0 ± 7.59 cm; body weight, 59.5 ± 11.64 kg; Body mass index, 21.7 ± 3.21). The protocol for this study was approved by the Research Ethics Committee of the National Institute of Information and Communications Technology. All the participants signed a written informed consent form.

Surface electromyography (sEMG; EMG Multi Analysis Programe MaP1038L, Nihonsanteku Co., Ltd., Osaka, Japan; sampling frequency: 2048 Hz) was recorded from the right extensor digitorum communis (EDC) and right flexor digitorum superficialis (FDS). Two bipolar Ag-AgCl surface electrodes (BA-U410 m(A)-015, Nihonsanteku Co., Ltd., Osaka, Japan) were placed over the belly of each muscle [59] [60]. The sEMG signals were amplified (gain 1000), band-pass filtered (10 - 1000 Hz), and recorded using the EMG Multi Analysis Programe MaP1038L software.

Single-pulse TMS was provided using a Magstim 200 (Magstim Co., London, UK) with a figure-of-eight shaped coil (7-cm internal diameter). The location of the coil was referred to the 10 - 20 International System and the coil was placed at the C3 of the 10 - 20 System Positions over the hand area of the primary motor cortex on the left hemisphere. The determined location was marked by a reference sticker apposed on the head of the participants wearing a cap for adherence. The coil was tangentially held with the handle pointing backward and laterally approximately 45˚ to the sagittal plane, with an orientation allowing to induce electric current flow in a posterior-anterior direction (Figure 1(a)). The optimal position (hot spot) to elicit a reliable MEP was adjusted by determining the resting motor threshold (RMT) in each participant. The RMT was defined as the minimum intensity of TMS required to evoke MEPs larger than 50 μV peak-to-peak amplitude at least five times out of the 10 trials performed at rest [56].

Vibratory stimulation was provided to the palm of the right hand using a hand intervention device combining vibratory stimulation and visual images (PLANSTAFF CO., LTD., Tokyo, Japan; Figure 1(b)). A vibration frequency of 115 Hz with a low-amplitude of 1 mm was used, considering that previous studies have suggested that a vibratory stimulation with a high-frequency between 80 and 120 Hz and low-amplitude induces strong activation of Ia afferent axons innervating muscle spindles [8] [16] [61]. The vibration device was placed on a table under a small stand holding a video screen (see Figure 1(b)).

The action observation was provided on a video screen (approximately 300 mm wide × 200 mm high) (Figure 1(b)). A 50-sec movie was depicted on the screen showing someone else’s right hand movement. The movie starts with a position of fingers extension, then all fingers are flexed and clenched into the palm for five seconds (i.e., flexion phase), followed by fingers extension toward the start position for five seconds (i.e., extension phase). These movements are repeated five times (Figure 1(c)). The position of the video screen was adjusted to each participant’s hand position, making them feel as much as possible that the hand on the screen was their own hand.

Each participant was comfortably seated in a chair located in front of the hand intervention device and where she/he was able to hold the vibratory device comfortably, so the distance between the chair and table was not fixed for each participant. Also, the participant was asked to wear earplugs in both ears. Electromagnetic wave noise suppression sheets were placed on the table and chair to reduce the noise of the sEMG signal. The tested right forearm of the participant was positioned on the table inside a small stand (Figure 1(b)). The participants were asked to hold the vibratory device softly in the palm of their right hand and relax the rest of the body. Also, the participants were instructed not to move their whole body during the data recording as any body movement could cause noise in the sEMG signals. The participants were allowed to familiarize themselves with the vibratory stimulation before data acquisition.

A custom-written MATLAB program (version 2017a, The MathWorks Inc., Natick, Massachusetts, USA) was used to provide single-pulse TMS pulses which were randomly delivered twice within each flexion phase and extension phase of the movie across all conditions using the same intensity as the RMT (Figure 1(c)). A total of 20 trials of TMS stimulations were provided in a condition, of which 10 trials represented each flexion phase and extension phase of the movie. The TMS system delivered trigger pulses that synchronized with the sEMG system.

A total of six conditions were performed as follows: 1) Vibratory stimulation, imagery, and viewing the movie (V⁺I⁺M⁺) condition where participants were instructed to view the finger movement on the movie and imagine they were moving their fingers while receiving vibratory stimulation, while attempting to move their fingers in accordance with the finger movement displayed on the movie, 2) Vibratory stimulation and viewing the movie (V⁺I⁻M⁺) condition where participants were instructed to view the finger movement on the movie while receiving vibratory stimulation, without imagining the finger movement, when just viewing the movie, 3) Vibratory stimulation and viewing the blank screen (V⁺I⁻M⁻) condition where participants were instructed to view the blank screen while receiving vibratory stimulation, 4) Imagery and viewing the movie (V⁻I⁺M⁺) condition where participants were instructed to view the finger movement on the movie and imagine they were moving their fingers, while attempting to move their fingers in accordance with the finger movement on the movie, 5) Viewing the movie (V⁻I⁻M⁺) condition where participants were instructed to view the finger movement on the movie, without imagining the finger movement, when just viewing the movie, and 6) Viewing the blank screen (V⁻I⁻M⁻) condition where participants were instructed to view the blank screen (Table 1).

In the V⁺I⁻M⁻ and V⁻I⁻M⁻ conditions without the action observation, TMS pulses were delivered in the same manner as in the other conditions. The order of conditions was randomly assigned as determined by computer⁻generated random numbers.

After each condition, the participants were asked to report on their subjective feelings as follows: 1) to what extent they felt that the palm of the right hand was opened during vibratory stimulation (Vib AVR sensation) and 2) to what extent they felt that the palm of the right hand was squeezed during vibratory stimulation (Vib TVR sensation). These questions were scored from 1 (not at all) to 10 (very strong) through verbal expression, and each score in each condition was analyzed to assess differences between the conditions. This questionnaire was conducted as a previous study reported that vibratory stimulation might induce TVR and AVR [9]. In the current study, as vibratory stimulation was applied to the palm (i.e., finger flexors), the TVR and AVR were expected to affect the FDS and EDC, respectively. The mean and standard deviation are presented in Table 2.

The acquired sEMG data were analyzed offline using MATLAB (version 2017a, The MathWorks Inc., Natick, Massachusetts, USA) to calculate the peak-to-peak amplitude of MEP responses from each trial [56]. First, the timing of each TMS trial was detected from the recorded trigger pulses of TMS (Figure 2(a)), and the time window from 10 to 35 ms after the trigger was extracted in each trial (Figure 2(b)). Due to TMS-induced noise on the recorded sEMG signals, a linear trend was subtracted from the time-series data signal within the time window. Also, some trials in which data within the time window displayed only positive (>0) or negative (<0) signals due to noise were eliminated from the analysis. As the result, eight participants were excluded. Further analysis was conducted in the eleven remaining participants. The peak-to-peak amplitudes of MEP in the time window of each trial were calculated in each condition. The MEP amplitudes of the EDC during the extension phase of the movie and MEP amplitudes of the FDS during the flexion phase of the movie were used for further analysis, considering that Yahagi and Kasai [62] reported that the MEP amplitude of the

Notes: Values are mean ± standard deviation. Abbreviations: AVR, antagonist vibratory response; TVR, tendon vibratory response.

wrist palmar flexor muscle is affected by the motor images of palmar flexion. In addition, five trials of response in each condition were shown to be required to reliably assess corticomotor excitability [63] [64]. Thus, the average of the MEPs from five trials in each EDC and FDS during the extension phase and flexion phase of the movie, respectively, were used in each condition for statistical analysis.

A three-way repeated measures analysis of variance (RM ANOVA) was performed using IBM SPSS Statistics version 26 software to determine the effect of three factors on the peak-to-peak amplitudes of MEPs: 1) vibration (V⁺, V⁻; 2); 2) imagery and movie (I⁺M⁺, I⁻M⁺, I⁻M⁻; 3); 3) muscle activity (EDC, FDS; 2). The post-hoc test was adjusted using the Bonferroni test (p < 0.05). To assess the scores of subjective feelings, a two-way RM ANOVA was performed to evaluate the effect of two factors (imagery and movie (I⁺M⁺, I⁻M⁺, I⁻M⁻; 3) and sensation (AVR and TVR; 2)) on the scores of the Vib AVR sensation and Vib TVR sensation.

The findings showed that the peak-to-peak amplitudes of MEPs increased in the conditions with vibratory stimulation compared to the conditions without vibratory stimulation, irrespective of the conditions of imagery or action observation only, and EDC or FDS. These findings are consistent with the findings from previous studies [11] [65] [66] [67] which demonstrated that muscle tendon vibration augmented MEP amplitudes. By contrast, other studies have reported that the increase of MEP amplitudes after vibratory stimulation occurred only in the vibrated antagonistic muscle [13] or only in the vibrated muscle [10]. There are several plausible mechanisms by which vibratory stimulation could augment MEP amplitudes. First, it has been suggested that the modulation of cortico-spinal excitability was induced by proprioceptive inputs generated by the vibratory stimulation. The vibration-induced Ia afferents innervating to muscle spindles caused an activation of the central nervous system, resulting in increased corticomotor excitability. The underlying mechanism may involve a decrease of motor threshold, an increase of intracortical facilitation or decrease of intracortical inhibition of the vibrated muscle [16]. Second, the long loop reflexes displayed as an automatic motor response to somatosensory stimulation might influence the excitability of the motor cortex. The long loop reflexes have suggested that proprioceptive inputs generated by vibratory stimulation can induce descending corticospinal motor response mediated by the primary motor cortex [17] [18] [19] [20]. Third, muscle vibration can elicit TVR, defined as an involuntary contraction of the vibrated muscle, and AVR, referring to the contraction of muscles antagonist to the vibrated muscles, presumably induced by top-down cortical modulation [5] [9]. Therefore, these findings suggest that vibratory stimulation might augment the excitability in the corticomotor and corticospinal circuits mediated by the primary motor cortex.

The remarkable feature of our findings is that the peak-to-peak amplitudes of MEPs were significantly increased in the I⁺M⁺ conditions compared to the I⁻M⁺ and I⁻M⁻ conditions, regardless of the vibratory stimulation and muscle activity conditions. These findings contrast with our hypothesis, which expected stepwise changes between the three conditions, with the highest value to be observed in the I⁺M⁺ conditions and the lowest value in the I⁻M⁻ conditions. Our assumptions were based on previous studies suggesting that imagining voluntary movement can facilitate the excitability of the contralateral primary motor cortex, by conveying signals to the spinal cord level via descending pathway, thus resulting in increased MEP amplitudes due to a modulation of corticospinal excitability [28] [68] [69] [70]. In addition, motor imagery involves multimodal sensation including visual and spatial perspectives, which could have an impact on the central processing [53] [54] [55]. For the influence of action observation on the corticomotor excitability, it was assumed that the visuomotor neurons are activated when observing or perceiving someone else’s action, facilitating corticomotor excitability of the observer through the mirror neuron system [47] - [52]. Additionally, using a visuomotor reaction time task, another study suggested that visual input can elicit the excitability of cortico-cortical connections from the visual cortex to the primary motor cortex, resulting in the modulation of corticospinal excitability [71]. Thus, we expected that motor imagery in addition to action observation would increase more corticomotor excitability compared to action observation alone, and there would be stepwise changes between the three conditions as described above. However, our results instead only showed a significant influence of motor imagery and action observation on the MEP amplitudes. A plausible explanation for these apparently discrepant results could be a lack of corresponding hand movement between the movie and the actual participant’s hand or a lack of feeling that the hand on the movie was as the participant’s own hand, which might have caused insufficient induction of the mirror neuron system. In general, the mirror neurons are activated when viewing someone else performing the same or similar acts or when interacting with the observer’s hand and an object [49] [51]. The task used in the current study was not identical to the hand movement of the movie, which might have resulted in an insufficient induction of corticomotor excitability. Nevertheless, the findings of the current study revealed a significant impact of motor imagery during the action observation on corticomotor excitability. Moreover, based on the participants’ subjective measures, feelings about their palms being opened and closed during the vibratory stimulation appeared more important in the V⁺I⁺M⁺ condition compared to the V⁺I⁻M⁺ condition. These findings emphasize a substantial influence of performing motor imagery while viewing the hand movement of the movie compared to action observation only on the corticomotor excitability. Furthermore, despite the fact that the results of the current study did not clearly show that the V⁺I⁺M⁺ condition had a significantly different influence on MEP amplitudes compared to other conditions, the effect of motor imagery and action observation combined with vibratory stimulation might have a greater impact on the augmentation of corticomotor excitability in order to enhance motor performance.

When comparing MEP amplitudes between EDC and FDS, MEP amplitudes in EDC were found to be greater than those in FDS regardless of the vibration conditions and imagery or action observation conditions. Previous studies have provided inconsistent results. For instance, Suzuki et al., [65] compared how different proprioceptive and visual inputs affect the corticomotor excitability of the wrist dorsi- and palmar-flexors. They demonstrated that the MEP amplitudes of the vibrated wrist dorsiflexors, contrary to the wrist palmarflexors, increased when viewing a video clip compared to viewing a static screen of the video clip, which was proposed to reflect the occurrence of TVR. Likewise, other studies reported identical effects of vibration where the MEP amplitudes increased only in the vibrated muscles [11] [12] [72] [73] [74]. Opposed findings were previously shown by Forner-Cordero et al., [13] who examined whether muscle tendon vibration to wrist palmarflexors induced MEP amplitudes of the wrist dorsiflexor and palmarflexor muscles. They showed that the MEP amplitude of wrist dorsiflexors representing the vibrated antagonistic muscles significantly increased compared to the vibrated wrist palmarflexors. This phenomenon was considered as an AVR which could be caused by a top-down modulation from the primary motor cortex [9] [75]. Overall, the findings on the activation of the agonist and antagonistic muscles from previous studies have remained controversial and seemingly dependent on methodology, including the particular stimuli or tasks used. For the results of the current study, they appeared comparable to the phenomenon of AVR. However, it may be difficult to provide a rational explanation and interpretation for our findings for the MEP amplitudes in the agonist and antagonistic muscles. Thus, a more rigorous methodology is needed to identify the mechanisms underlying the relationship between agonist and antagonist muscles.

In terms of the results of subjective feelings for the Vib AVR and Vib TVR scores, this questionnaire was performed as supplemental analysis since a previous study suggested that vibratory stimulation might induce AVR and TVR [9] and there might be the identical results to the MEP amplitudes in the EDC and FDS (i.e., The vibratory stimulation was applied to the palm (i.e., finger flexors) in this study so the TVR and AVR were anticipated to affect the FDS and EDC, respectively). The results of the current study indicated that the scores in the I⁺M⁺ conditions were higher than those in the I⁻M⁺ conditions, regardless of the EDC or FDS. These results did not show similarity between the MEP amplitudes in the EDC and FDS and the subjective scores in the Vib AVR and Vib TVR scores, respectively. The potential causation of difference between them is that the previous study [9] examined the sEMG signals during vibratory stimulation, not MEP amplitudes. Also, the previous study utilized kinesthetic illusion during vibratory stimulation which was not our aim in this study. The different methodology might cause different results.

The clinical implication from the findings of this study is that motor imagery and action observation may be utilized with vibratory stimulation to facilitate hand function as a therapeutic approach especially in individuals suffering from central nervous system disorders. Our findings together reveal that performing motor imagery of a hand movement while viewing a hand movement in a movie during the vibration had a substantial effect on the corticomotor excitability of finger muscles. In individuals with severe hand paresis or unable to actively perform finger movements, due to neurological diseases such as poststroke, brain injury, or cervical spinal cord injury, using a combination of motor imagery and action observation with vibratory stimulation could help augment corticomotor excitability, which could ultimately contribute to an enhancement of impaired hand function in neurorehabilitation [76].

The sample size was small since some data were excluded from the analysis due to noise. Also, the TMS mapping system allowing to identify the hot spot and center of gravity of the EDC and FDS was not used in this study. In addition, the optimal time for providing the stimulations of vibratory stimulation, motor imagery, and action observation in order to maximize corticomotor excitability in the same conditions was not determined. In addition, this study was not designed to examine the 2 × 2 full design of I⁺M⁺, I⁺M⁻, I⁻M⁺, and I⁻M⁻ conditions so the interaction between the I⁺ and M⁺ conditions could not be evaluated. Therefore, future studies need to be performed to address these issues with the current methodology.

The findings of the current study emphasize that performing motor imagery while viewing a movie of finger flexion and extension movements is more effective at inducing corticomotor excitability compared to action observation alone. In addition, vibratory stimulation significantly increased corticospinal circuit excitability probably via the primary motor cortex. Taken together, these findings suggest that the effect of motor imagery and action observation combined with vibratory stimulation may have a greater impact on the augmentation of corticomotor excitability. Therefore, this approach could be utilized to modulate corticomotor excitability which may contribute to the improvement of finger motor function.