Focal Muscle Vibration for Stroke Rehabilitation: A Review of Vibration Parameters and Protocols

Featured Application: This review summarized the focal muscle vibration devices, parameters, and protocols that have been used for stroke rehabilitation. The review discussed the application of wearable focal muscle vibration technology for sustainable stroke rehabilitation and suggested future research on the relationship between vibration frequency, amplitude, treatment protocol, and outcome measures for optimized and individualized intervention in stroke. Abstract: In this review, we present a narrative synthesis of studies on the use of focal muscle vibration (FMV) in stroke rehabilitation with a focus on vibration device, parameters, and protocols. A search was conducted via PubMed, SCOPUS, PEDro, REHABDATA, and Web of Science using the keywords “stroke and focal vibration” or “focal muscle vibration”. Inclusion and exclusion criteria to select the articles were determined. Twenty-two articles involving FMV and stroke were included in this review. Eight di ﬀ erent vibration devices were used in the 19 articles that reported the vibration apparatuses. The vibration frequencies ranged from 30 Hz to 300 Hz with amplitudes ranging from 0.01 mm to 2 mm. The vibration treatment frequency ranged from a single treatment to 5 days / week. The session duration ranged from 14 s to 60 min / session with a duration of a single treatment to eight weeks. Twenty di ﬀ erent muscles were targeted with 37 di ﬀ erent outcome measures used to assess the e ﬀ ects of FMV. The clinical applications of FMV were not conﬁrmed based on available evidence. More research is needed to improve the FMV technology, guide the selection of vibration parameters, optimize the vibration dosage, and develop standardized protocols for FMV therapy in patients with stroke.


Introduction
Stroke is the second leading cause of death and one of the most common causes of adult-onset disability worldwide. According to a recent study, 26% of individuals with stroke have a disability in activities of daily living (ADL), and 50% have reduced mobility due to hemiparesis [1]. The post-stroke disruption of the sensory system plays an important role in motor dysfunction of the hemiparetic limb [2]. Loss of proprioception impairing corrections to movement errors and loss of tactile sensation are common consequences of stroke and affect control of limb motion [3]. Another major problem affecting nearly 20-30% of stroke survivors is gait disorders [4]. These disorders further increase the risk of falls and loss of balance, reducing patients' social participation. According to National Stroke Association's post-stroke recovery guidelines, only 10% of stroke survivors recover almost completely; 25% recover with minor impairments; 40% experience moderate to severe impairments requiring special care; 10% require care in a nursing home or other long-term care facility; and 15% die shortly after the stroke [5]. While numerous therapies have been developed over the last 10 years to treat acute ischemic stroke, the stark reality remains that 95% of these patients continue intervention in the chronic stage and go on to live with significant disability for many years.
Because of the complexity of a stroke, various approaches to chronic stroke rehabilitation, such as facilitation technique [6,7], functional electric stimulation (FES) [8,9], transcutaneous electrical stimulation (TENS) [10,11], electromyography (EMG) biofeedback [12], exercise [13,14], physical and occupational therapy [15,16], robotics [17,18], and virtual reality [19,20], have been studied to help functional recovery from hemiplegia due to brain damage. The limitations of the aforementioned intervention strategies are their sustainability due to one or more of these challenges: the requirement of trained and licensed professionals to administer the right dose to ensure safety; lack of precision and accuracy of intervention; lack of consensus among the findings; lack of sufficient evidence to establish the effectiveness of the intervention strategies; awareness of and access to existing intervention strategies; the cost of administration; and other similar disparities [21][22][23][24]. One intervention strategy that has the potential for sustainable stroke rehabilitation is the use of mechanical vibration as a therapeutic intervention known as vibration therapy (VT) [25][26][27][28][29]. According to Murillo et al. (2014), VT as an intervention in rehabilitation can be dated back to 1969, when Hagbarth and Eklund observed tonic vibration reflex (TVR) in which the application of vibratory stimulus resulted in agonist muscle contraction and antagonist relaxation. Hagbarth and Eklund then used this observation as a basis to use vibration to decrease muscle spasticity in individuals with stroke [28,29].
There are two types of VT: whole body vibration (WBV), in which mechanical vibrations are transmitted from the feet to the rest of the body using a vibrating platform, and focal muscle vibration (FMV), where mechanical vibrations are applied to a localized point in muscles, generally the muscle belly or the tendon on the affected/paretic side. The potential mechanism behind using vibration as an intervention in the treatment of motor disorders in patients is that vibration stimulates the primary muscle spindle endings, causing Ia afferent impulses to be conducted to alpha motor neurons and Ia inhibitory interneurons in the spinal cord. This afferent pathway produces involuntary contraction in the vibrated muscle (that is, a tonic vibration reflex, TVR) and inhibits the antagonist muscle [30,31]. The effect of VT on the human body depends on the characteristics of the vibration applied, such as type of vibration (vertical, horizontal, or multidirectional), frequency, amplitude, and the protocol [27]. The effects also depend on the characteristics of the person, such as age, gender, and health condition [26,27].
WBV has been widely studied, and the evidence agrees on the pros and the cons of its application in patients with stroke [32][33][34][35][36][37][38][39][40]. The application of FMV for patients with stroke has been less widely studied [28,29]. Only one review specifically focused on FMV in stroke [28]. The authors summarized eight studies and concluded that FMV showed some evidence in reducing hemiplegic upper extremity spasticity in patients with stroke, and additional randomized controlled trials were needed to study the effects on FMV on spasticity in individuals with stroke [28].
Multiple studies have been conducted on the use of FMV for stroke rehabilitation in upper and lower limb impairments. These studies showed some improvements in functionality and reduction of muscle spasticity. However, there is a lack of consensus regarding its clinical application. The other gaps include the lack of protocol (frequency and amplitude of vibration, number of days and duration of intervention and overall study, etc.), standardized outcome measures, and recommended vibration devices. The purpose of this review was to focus on the current FMV devices in use, the vibration parameters applied, and protocols of FMV therapy and outcome measurements in post-stroke rehabilitation.

Search Strategy
A search was conducted in the following electronic databases: PubMed, SCOPUS, PEDro, REHABDATA, and Web of Science. The key search terms were (focal muscle vibration OR fmv OR vibration OR focal vibration OR focal-muscle-vibration OR segmental muscle vibration OR localized mechanical vibration) AND (neurological OR central nervous system OR nervous system OR diseases OR disorders OR spinal cord OR brain OR cerebral OR neurological manifestations) AND (motor OR motor impairments OR physical OR impairment OR activity OR disability OR function OR movement). No time restraint was applied to the literature search that was finished in July 2020.

Study Selection
Studies were included if they were written in English, treated patients who were diagnosed with stroke, and used focal muscle vibration as the primary intervention for rehabilitation. Studies were excluded if they did not use focal vibration as the main intervention, treated multiple diagnoses, did not have at least one motor outcome, or did not report any parameters for the application of vibration.

Data Extraction
All authors (H.W., R.C., J.R., M.G.) searched the different databases for relevant publications using the aforementioned keywords. The searched articles were screened by authors H.W. and R.C. for relevance, followed by a title and abstract review through a discussion between H.W. and R.C. Then, a full text review by all four authors was performed based on a pre-developed data extraction form. Two of the authors (R.C. and M.G.) developed an Excel document with the following data extracted: participant characteristics (number, gender, mean age), vibration devices, vibration parameters (frequency and amplitude), protocols (dosage and duration), region of application of vibration, outcome measures, and results. All four authors made the final decision on articles to be included in this review and discussed the studies with a focus on vibration parameters and protocols.

Vibration Frequency and Amplitude
The frequencies and the amplitudes used in the included studies are represented graphically in Figure 2. The most frequently used combination was 100 Hz and 0.2-0.5 mm [47,50,58,62], followed by 120 Hz and 0.01 mm [42,51,53] and 99.5 Hz and 1 mm [45,49,59]. The FMV frequency ranged from 30 Hz to 300 Hz and the amplitude ranged from 0.01 mm to 2 mm. Articles did not report amplitude were plotted as 0 mm.

Vibration Protocols
In terms of the muscles where the vibration was applied, 20 different muscles were targeted, with five muscles for lower extremities and 15 muscles for upper extremities (Figure 3). Biceps brachii was targeted the most, followed by triceps brachii and flexor carpi radialis. Articles did not report amplitude were plotted as 0 mm.
Thirty-seven different outcome measures were used across the 22 studies. The most common outcome measure was assessment of the spasticity via the modified Ashworth scale, which was used in 11 studies, followed by grip strength assessed in five studies, and Fugl-Meyer scale and box and block test evaluated in four studies each. Figure 4 shows all outcome measures and how commonly they were used across studies.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 13 of 21 Thirty-seven different outcome measures were used across the 22 studies. The most common outcome measure was assessment of the spasticity via the modified Ashworth scale, which was used in 11 studies, followed by grip strength assessed in five studies, and Fugl-Meyer scale and box and block test evaluated in four studies each. Figure 4 shows all outcome measures and how commonly they were used across studies.

Discussion
To our knowledge, this is the first review aimed to investigate FMV devices, parameters, protocols, and outcome measures in post-stroke rehabilitation. The only other review of FMV for stroke rehabilitation was a systematic review focused on the effectiveness of FMV on hemiplegic upper extremity spasticity in individuals with stroke [28]. Our review agreed with the findings that FMV therapy may reduce spasticity in both upper and lower extremities and improve function in individuals with stroke [28,29]. The positive effect of FMV in inhibiting hemiplegic upper and lower extremity spasticity in patients with strokes was confirmed with other reviews [25,28,29].
Included studies were primarily quasi-experimental design and RCTs. Most studies did not justify the choice of target muscles for vibration or provide the rationale behind the vibration protocols. Blinding of participants and therapists was poor, although the assignment of control and experimental groups was randomized. There was overall a lack of follow-up post FMV intervention to determine how long the improvements would last. Marconi et al. (2011) and Jung Sang-mi et al.
(2017) examined the effects of FMV therapy after two weeks of intervention and reported that, even though the changes on the main outcome measures were less than observed immediately post interventions, patients were still better than baseline [47,59]. This finding indicates that the benefit of FMV therapy might last for two weeks. Caliandro et al. (2012) and Calarbo et al. (2017) checked the participants one month after the FMV therapy, and there were no significant differences on the outcome measures [50,57]. In addition, the included studies did not compare FMV therapy with other interventions, except for traditional physical therapy (PT). There was agreement with our review and others that a wide variety of FMV devices with different vibration frequencies, amplitude, targeted muscles, vibration protocol, and outcome measures were used [25,28,29].
Seven different vibration devices were used in the 19 studies that reported the vibration devices. The technical details of those devices were ambitiously described, but their availability for clinical

Discussion
To our knowledge, this is the first review aimed to investigate FMV devices, parameters, protocols, and outcome measures in post-stroke rehabilitation. The only other review of FMV for stroke rehabilitation was a systematic review focused on the effectiveness of FMV on hemiplegic upper extremity spasticity in individuals with stroke [28]. Our review agreed with the findings that FMV therapy may reduce spasticity in both upper and lower extremities and improve function in individuals with stroke [28,29]. The positive effect of FMV in inhibiting hemiplegic upper and lower extremity spasticity in patients with strokes was confirmed with other reviews [25,28,29].
Included studies were primarily quasi-experimental design and RCTs. Most studies did not justify the choice of target muscles for vibration or provide the rationale behind the vibration protocols. Blinding of participants and therapists was poor, although the assignment of control and experimental groups was randomized. There was overall a lack of follow-up post FMV intervention to determine how long the improvements would last. Marconi et al. (2011) and Jung Sang-mi et al. (2017) examined the effects of FMV therapy after two weeks of intervention and reported that, even though the changes on the main outcome measures were less than observed immediately post interventions, patients were still better than baseline [47,59]. This finding indicates that the benefit of FMV therapy might last for two weeks. Caliandro et al. (2012) and Calarbo et al. (2017) checked the participants one month after the FMV therapy, and there were no significant differences on the outcome measures [50,57]. In addition, the included studies did not compare FMV therapy with other interventions, except for traditional physical therapy (PT). There was agreement with our review and others that a wide variety of FMV devices with different vibration frequencies, amplitude, targeted muscles, vibration protocol, and outcome measures were used [25,28,29].
Seven different vibration devices were used in the 19 studies that reported the vibration devices. The technical details of those devices were ambitiously described, but their availability for clinical and home use were not clear. In the 22 included studies, participants visited the clinics for the vibration interventions, which could lead to poor compliance for sustainable usage of the FMV therapy.
Recently, newer wearable FMV technologies were developed, including the Equistasi ® (Equistasi S.R.L. Via C.Porta, 16 20064 Gorgonzola, Italy), VibraCool ® (Pain Care Labs, 195 Arizona Ave LW08, Atlanta, GA 30307, USA), and Myovolt (Myovolt Limited 146a Litchfield Street, Christchurch 8011, New Zealand). Equistasi ® uses nanotechnology fibers to deliver frequency as high as 9000 Hz with very low amplitude less than 0.002 mm. It has been used to treat Parkinson's disease [63][64][65], multiple sclerosis [66], and ataxia [67,68]. However, due to the much higher frequency and the lower amplitude, the mechanism of Equistasi ® might not be the same as that of the FMV discussed in this study. In addition, Equistasi ® has not been used for patients with stroke to our knowledge. VibraCool ® uses proprietary high-speed vibration frequencies and intense cold for pain relief and to treat muscle tension and myofascial trigger points. Research evidence on VibraCool ® appears unavailable, and its technical specifications were not reported on their website. Myovolt combines therapeutic vibration together with a gentle warming effect to massage and relieve muscle soreness and stiffness. Studies conducted using Myovolt reported improvement in muscular power performance [69] and alleviation of muscle soreness in healthy adults [70] and improved muscle function in patients with peripheral artery disease [71]. All of these wearable FMV technologies showed promise but with limited application or evidence in stroke rehabilitation. Future studies are warranted to explore their benefits with individuals with stroke.
More than half of the studies used vibration frequencies from 85 to 120 Hz and vibration amplitudes of 0.01-2 mm. A reduction in spasticity was observed with various frequency ranges. Due to the variations in amplitude and treatment frequency and duration, and contradictory to what is stated in the recent review [28], we speculate that vibration frequency cannot be disregarded as a discriminative factor in FMV intervention. We believe that studies with rigorous design controlling for vibration amplitude and treatment protocol will be needed to investigate the impact of vibration frequency in FMV intervention. The vibration amplitude for stroke rehabilitation ranging from 0.01 mm to 2 mm was considered comparable to the vibration amplitude ranging from 0.005 mm to 10 mm in studies using FMV intervention for patients with spinal cord injury, multiple sclerosis, and other movement disorders [25,29,72]. About one third of articles on FMV did not report the amplitude delivered. The improvements observed in the outcome measurement scores were better in studies using FMV with amplitudes greater than or equal to 1 mm and frequencies in the range of 91-108 Hz or greater [45,49,52,54,55,[59][60][61], unless FMV was paired with other forms of intervention such as robotic assistive device [57] or progressive modular re-balancing (RMP) [58]. FMV alone with lower frequencies of less than 90 Hz and lower amplitudes of less than 1 mm seemed to have a lesser change in the outcome measures. Given that frequency range 75-120 Hz was particularly effective on the central nervous network underlying motor control [73], and amplitude of 1-2 mm was sufficient to drive Ia spindle afferents while remaining safe for the tonic vibration reflex [74] and avoiding muscle fiber injury [75], these frequency and amplitude combinations could be recommended for future studies. The duration of intervention did not seem to have much effect on the total improvement, although the change scores were slightly greater in studies with longer durations of intervention, which could be because of the long-lasting effects on cortical excitability. In addition to exploring the impact of vibration frequency, it is necessary to conduct basic science research on how muscle spindles, neurons, and human tissues respond to the different amplitudes delivered by the vibration motor to understand the individual and the combined impact of the vibration parameters as well as to optimize vibration parameters for individual patients.
A single session of vibration while walking for 14 s was reported to improve the walking speed of patients with stroke [41]. Further, a single session lasting 5 min inhibited spasticity and improved muscle performance, as measured by EMG [45,46,49]. Although these studies were of high quality with larger sample sizes, the results were insufficient for generalization. These findings may implicate the acute effect of FMV in stroke rehabilitation, as also observed in professional athletes [69]. With more frequent and longer duration (5 min FMV + 30 min PT × 3/week × 8 week) and more FMV (30 min FMV × 3/week × 6 week), small to moderate effect sizes (0.11-0.52) were observed in studies with relatively high methodological quality [55,61]. Other studies with less FMV (30 min FMV + 60 min PT × 5/week × 2 week) also reported significant reduction in spasticity in the experimental group compared with the control group [51][52][53]59]. More FMV might lead to better outcomes, but there is a lack of evidence regarding the best vibration dosage and duration. Thus, future studies to investigate and standardize the protocol for FMV interventions are warranted. The overall lack of follow-up after the FMV interventions made it difficult to determine the long-term effects, even though some studies stated that FMV intervention effects could last as long as two weeks [47] and even four months after the intervention in elder adults [76].
A variety of muscles were targeted for FMV therapy. For upper extremity rehabilitation, triceps brachii and biceps brachii were targeted the most. Shorter fascicle lengths have been reported for the brachialis muscle on the affected side, as shown by ultrasound [77]. The flexor carpi radialis was also frequently targeted because shortening of wrist flexor muscles is associated with poor recovery after stroke [78]. For the lower extremity rehabilitation application of FMV, three out of the five studies targeted the tibialis anterior muscle. This focus could be due to the importance of tibialis anterior in gait and that the affected lower limb exhibited significantly longer delays in initiation and termination of tibialis anterior contraction relative to the unaffected limb in individuals with stroke [79,80]. Overall, there was a lack of justification for the choice of target muscles and discussions on the clinical rationales and applications of the findings based on the muscles that received vibration. Six studies reported electrophysiology [42,45,[47][48][49]53]. There was no clear mechanism through which FMV acts on the sensorimotor system. All six studies hypothesized the mechanism of increasing Ia afferent fiber discharges because of the activation of muscle spindles via FMV. The modulation of Ia inputs altered the excitability of the corticospinal pathway as well as the activation of cortical motor regions. However, excitability remained unchanged in other cortical motor representations, indicating that the increased neuronal excitability was specific to the vibrated muscle's movement representation [42,53]. For FMV applied to low extremity muscles, such as quadriceps during the stance phase of walking, group Ia afferent discharges also contributed to the triggering of locomotor phase transitions [42]. It was also noted that FMV affects not only the contralateral but also the ipsilateral hemisphere, thereby modulating the relationship between the two hemispheres [42]. For FMV applied to upper extremity muscles, it was hypothesized that FMV applied to the forearm improved regulation of reflex excitability and improved cortical control of the movement [48]. In addition, the effects of FMV might depend on the inhibitory/excitatory state within the motor system reflecting the site of lesion for stroke patients [47]. The underlying mechanism of motor and function recovery due to vibration at different muscles could be more complex than simply the activation of Ia afferents. Measurements to detect activation of sensorimotor cortex network, nitric oxide production, and blood flow could be included in the future studies to better understand how vibration at different muscles impact the outcomes. This could be partly due to the overall short durations of those studies, which could make investigating changes in participation and quality of life difficult. This finding could also be attributed to the infancy of FMV therapy for stroke rehabilitation and the lack of accessible and sustainable FMV devices for researchers, clinicians, and individuals with stroke.
This review was limited to articles published in English. Regarding the locations where the research was conducted, only one was conducted in the United States; one was in Germany; one in France; three in Japan; four in Korea; and twelve in Italy. This can lead to bias, as studies published in languages other than English were not included. In addition, it is known that significant results have greater likelihood of publication than do studies that do not have significant results. Effect sizes were not reported for most outcome measures because of the inconsistency of outcome measures and insufficient data. We did not conduct meta-analysis because the focus of this review was vibration technology and protocol, because of the heterogeneity of treatment protocols, dosages, and assessment, and because of the inability to contact the authors of many articles.

Conclusions
In conclusion, FMV may reduce spasticity and improve function in individuals with stroke when it is applied to the antagonist muscles. However, the effects of FMV on stroke rehabilitation are not fully understood. The accessibility and the sustainability of existing FMV technology, effectiveness of treatment protocol, and dosage remain unclear. Furthermore, the included studies did not report details on the vibration devices, with highly varied muscles vibrated, vibration frequency and amplitude, treatment protocol, and outcome measures. These variations make it difficult to recommend the clinical application of FMV therapy. These findings illustrate the need for more research to understand the mechanisms of FMV in stroke rehabilitation, and the impact of characteristics of the vibration device on outcome measures. Further high-quality studies with large sample sizes are warranted.

Acknowledgments:
The authors would like to thank Sarah Bulloch for her contribution to the edits and literature search. The authors would like to thank Bethany Block for her assistance with categorizing the outcome measures based on ICF model.

Conflicts of Interest:
The authors declare no conflict of interest.

Abbreviations
Acronyms used in "Study Design and Participant" column, "Outcome Measures" column, and "Results" column: