A systematic review of the literature was performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [42]. PUBMED, MEDLINE, CINAHL, and Cochrane Library databases were searched up to September 6, 2025. The searches included the following Medical Subject Headings (MeSH) terms: (“Dystonia” OR “blepharospasm” OR “cervical dystonia” OR “cranial dystonia” OR “laryngeal dystonia” OR “meige syndrome” OR “musician’s dystonia” OR “oromandibular dystonia” OR “spasmodic dysphonia” OR “spasmodic torticollis” OR “task-specific dystonia” OR “torticollis” OR “writer’s cramp”) AND (“Vibration” OR “vibration therapy” OR “vibratory training” OR “vibrotactile stimulation” OR “focal vibration” OR “local vibration” OR “muscle vibration”). The search was limited to English language and human subjects. A complete list of combinations of search terms used in each database can be found in Supplementary Table S1.

Inclusion criteria were: 1) Inclusion of participants who presented with a form of focal dystonia (FD), 2) application of superficial VTS adjacent or directly to dystonic muscles or regions, 3) reporting of at least one biomechanical and/or electrophysiological outcome measure reflecting possible VTS effects.

Exclusion criteria were: 1) Application of VTS for assessing an unrelated experimental effect, 2) not reporting selected parameters of VTS such as frequency, amplitude or duration, 3) not constituting an original, empirical investigation (e.g., review, comment or opinion article), 4) not written in English.

Search results were exported to Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia). After duplicate studies were removed, two reviewers (CO, SR) independently screened titles, abstracts and then the full-texts against the inclusion-exclusion criteria. The third reviewer (JK) was consulted to resolve any disagreements where required.

Two reviewers (CO, SR) independently assessed the quality of the included articles according to the Joanna Briggs Institute (JBI) critical appraisal checklists [43]. The case-control studies, case reports, quasi-experimental studies, and crossover studies were assessed using the 10-item case-control study checklist, 8-item case report checklist, 9-item quasi-experimental study checklist, and 13-item randomized controlled trials checklist, respectively [44]. The overall risk of study bias ranked into one of the four levels (High, Moderate, Low, Very Low) in line with the recommendation from the JBI reviewer manual [43]. Any disparities in the quality assessment were resolved through consensus with a third reviewer (JK) if necessary. No studies were excluded based on the results of the quality assessment. The methodological quality of the selected studies is presented in Supplementary Table S2.

Two reviewers (CO, SA) extracted data and entered these into a standard data extraction table (Table 1). Extracted data related to 1) study details (author names, year of publication, and study design), 2) participant characteristics (i.e., population, age, sex, disease duration, Botulinum neurotoxin history, controls), 3) VTS protocol information (i.e., type of vibration device, frequency, amplitude, vibrated muscles or sites, duration), 4) outcomes measures and results.

The heterogeneity of dystonia types and outcome measures across articles meant that a meta-analysis was not possible. Therefore, a narrative synthesis of the data was conducted, detailing the findings, the VTS protocol, and the measured outcomes for each study. For each included study, the intervention effect was determined by extracting reported outcome data. Specifically, we focused on the primary biomechanical, neurophysiological or electrophysiological outcome measures. Where available, we included secondary outcomes or subgroup analyses to provide a broader understanding of the intervention’s effect across different dystonia populations. The intervention effect was inferred from pre- and post-VTS comparisons and statistical significance. In case reports, the intervention effect was determined based on the pre- to post-VTS percentage change in the reported outcome measure [53, 68]. For case-control studies that focused on comparing pre- to post-VTS changes between people with FD and healthy individuals, we considered the statistical significance of the difference in the intervention effect between groups [49, 50, 57, 64].

A Joanna Briggs Institute (JBI) bias risk analysis was performed [69]. The risk of bias was categorized as high when a study had up to 49% of “yes” responses, moderate when it ranged from 50% to 69%, and low when it exceeded 70% [69]. Based on these criteria, 6 studies were classified as having a high risk of bias, 7 as moderate, and 11 as low. None of the studies were excluded based on the quality assessment results. A summary of the applied risk analysis can be found in Supplementary Table S2.

A search of the PUBMED (n = 73), MEDLINE (n = 68), CINAHL (n = 28), and Cochrane Library databases (n = 16) identified a total of 185 studies. After removing 93 duplicate studies, 92 studies were reviewed through title-abstract assessment. A total of 66 records were excluded, because they did not meet the inclusion criteria. Of the 26 records sought for retrieval, one full text of one study was unavailable, another study was not published in English. Thus, 24 studies were included in this review. The process of study selection for this review is presented in Figure 1.

Thirteen studies (54%) applied a case-control study design [47–50, 52, 57, 58, 60, 61, 63–66] focusing on elucidating possible neurophysiological mechanisms of VTS as a form of neuromodulation that influences motor output. These studies were not designed as clinical trials that attempted to determine if VTS can reduce dystonic symptoms. The remaining 11 studies (46%) investigated VTS as a possible treatment for focal dystonia [45, 46, 51, 53–56, 59, 62, 67, 68]. Two of these studies (8%) were case reports [53, 68], 7 (29%) were quasi-experimental studies [45, 46, 51, 54, 56, 59, 62], and 2 (8%) were randomized controlled trials [55, 67].

The included studies reported a total of 587 participants (sample size: min = 1; max = 67; median = 7.5), among whom 235 presented with adult-onset, isolated focal cervical dystonia (CD), 100 with adult-onset, isolated laryngeal dystonia (LD), 10 with adult-onset, non-task-specific upper limb dystonia, 48 with adult-onset, task-specific focal hand dystonia (30 = writer’s cramp, 18 = musician dystonia), 2 with adult-onset, task specific dystonia of the face, jaw or mouth (musician embouchure dystonia), 7 with adult-onset, isolated focal cranial dystonia affecting the eyelids (blepharospasm), 10 with adult-onset segmental dystonia involving cervical and upper limb (n = 7) or the lower limb (n = 3) and 175 participants served as neurologically normal controls.

The mean age was 50.9 (SD ± 11.8) years for the dystonia participants and 43.1 (SD ± 12.0) years for the healthy controls. Of the 23 studies that reported sex, 61% of the participants were females [45–59, 61–68]. Nineteen studies reported disease duration (mean: 10.1 years; range: 2–21 years) [45–50, 52–57, 59–62, 64, 65, 67]. See Table 1 for an overview of all reviewed studies.

The majority of studies used a variety of commercially available vibratory motors [47–50, 52–58, 60–68]. Considering all vibratory motors, the applied frequencies of VTS ranged from 70–150 Hz (mean: 95 Hz, median: 90 Hz) with a vibration amplitude between 0.2–3 mm (mean: 0.85 mm, median: 0.5 mm). VTS was applied for a minimum of 5 s [53] up to a maximum of 54 min [55] and longitudinally for up to 11 weeks [55]. The application sites were generally on or near the clinically identified dystonic muscles. Across the included studies, the VTS targets varied by dystonia subtype. In blepharospasm, stimulation was applied to the region around the mouth [52]. In cervical dystonia, the application sites were the sternocleidomastoid in 5 studies [46–48, 67, 68], trapezius in 6 studies [46, 49, 50, 57, 67], right extensor carpi radialis in one study [64], splenius in one study [53], and broader regions, such as the forehead, occiput, cervical spine, paraspinal and shoulder girdle muscles, masseter, suprahyoid muscles, neck flexors and extensors in one study [56]. For laryngeal dystonia, all studies consistently targeted the lateral aspect of the thyroid cartilage bilaterally [45, 51, 54, 55, 59]. In musician’s dystonia, writer’s cramp and upper-limb dystonia studies, stimulation was applied to the finger muscle [58, 60–62, 65, 66].

Because a possible link between somatosensory stimulation and the motor manifestations of dystonia have long been recognized, several studies examined the effects of vibration in focal dystonia to elucidate the underlying pathomechanisms of the disease and to understand how the nervous system responds to vibration in people with focal dystonia. These studies applied VTS in short bursts between 4 and 50 s.

Abnormalities in somatosensory function affecting the central processing of tactile, proprioceptive and nociceptive afferent signals have long been recognized in patients with focal dystonias [17]. In addition, many people affected by FD apply some form of sensory trick by touching the skin over dystonic body regions to alleviate symptoms indicating that the underlying physiology is susceptible to the manipulation of somatosensory inputs [70, 71]. The pioneering early work reviewed here focused on using VTS as a form of controlled somatosensory input and mostly applied short bursts of VTS. This research obtained a range of neurophysiological parameters to determine VTS response differences of people with FD compared to healthy controls or between different phenotypes of FD.

Using positron emission tomography to measure cerebral blood flow showed that VTS produced a consistently localized and robust peak response in primary sensorimotor cortex contralateral to hand vibration in healthy participants, but this response was reduced in patients with unilateral dystonia [65], writer’s cramp [66] and blepharospasm [52]. Moreover, regardless of whether the vibration was applied to the affected or unaffected side, the studies showed a reduction in cerebral blood flow in both ipsilateral and contralateral primary sensorimotor cortex. This research established for the first time that cortical processing of mechanical vibration signals is affected in FD.

Subsequently, TMS-based studies reported that MEP-derived markers indicating intracortical inhibition and/or facilitation (ICF) are altered in FD. For example, in healthy musicians and non-musicians, VTS decreased short-interval cortical inhibition (SICI) in the vibrated muscle increased in the non-vibrated hand muscles. In contrast, VTS strongly reduced SICI and ICF in all hand muscles in people with MD [60, 61] but had little effect on cortical excitability in task-specific upper limb dystonia [61, 62]. Applying VTS to non-dystonic forearm extensor muscles in people with CD significantly increased ICF compared to healthy controls [64] (see Figure 3). This body of research corroborated previous evidence by showing that people with FD respond to VTS, but the MEPs of the target muscles, their synergists or antagonists is reduced or altered when compared to the MEPs of healthy controls. This is important, because it means that the cortical and myoelectric response to VTS is not restricted to the vibrated dystonic or non-dystonic muscle, but that the underlying functional muscle synergies are modulated by VTS. It further suggests that VTS-induced changes in cortical function extend beyond the projection areas of the clinically affected dystonic muscles.

Another complementary set of studies investigating the effects of VTS on postural and force control demonstrated that force control of dystonic muscles is impaired, but that VTS can modulate force control in people with FD [47–50, 57, 63] (see Figure 2). Importantly, its effect depends on which muscle groups are stimulated (e.g., non-dystonic vs. dystonic), again underlining the notion that VTS affects synergistic muscle control. In addition, the fact that VTS of dystonic neck muscles has no or only minor effects on whole body postural control indicates that VTS effects are regional and not generalized.

Given that the application of sensory tricks to alleviate symptoms and the increasing empirical evidence that FD is associated with deficits in somatosensory processing, there has been an interest in understanding whether people with a sensory trick react differently to VTS than those without it. One study reported that for CD patients who could apply a successful sensory trick, neck muscle VTS vibration led to an initial forward sway of the body that slowly increased during prolonged vibration [49]. In contrast, for patients without a sensory trick, the initial sagittal sway was significantly reduced and the slow increase in forward sway was absent [49]. In addition, neck vibration improved backward head tilt in CD patients, particularly those with sensory tricks (see Figure 3). However, two larger sample clinical trials found that the use of an effective sensory trick did not clearly differentiate between responders and non-responders to VTS and the effect on perceived pain and head righting in CD [46, 67]. Thus, current empirical evidence shows that the presence of an effective sensory trick is associated with a higher susceptibility to somatosensory stimulation (e.g., as a signal to regulate body posture), but such susceptibility does not automatically translate to larger or more effective dystonic symptom relief. Yet, one also needs to recognize that the available evidence is limited. No clinical trials have systematically investigated the efficacy of VTS as a function of the presence or absence of a sensory trick.

A set of phase 1 and phase 2 clinical trials examined the effectiveness and response rates to VTS in a combined sample of 194 participants with CD and LD [45, 46, 54, 55, 59, 67]. These studies applied VTS over longer periods (up 45 min per session) in a single session with two trials examining VTS effects longitudinally for up to 2 months [45, 55]. Depending on the outcome measure, between 57% and 85% of participants responded to VTS and exhibited improvements in voice and speech quality, in head posture and/or reductions in pain. In approximately 40% of CD participants, 3D head posture and/or pain levels acutely improved by 50% or higher [46, 67]. The studies on people with LD demonstrated that typical voice and speech symptoms such as voice breaks, prolonged speech or reduced loudness can be ameliorated by VTS [54, 55, 59] (see Figure 4). However, a recent pilot study reported that a short 7-min application of VTS did not induce reductions in voice and speech symptoms in LD participants [51], indicating that short-duration VTS may be not or less effective than longer applications exceeding 20 min or more.

A cortical correlate to the observed symptoms in CD and LD is the excessive synchronization of motor cortical neurons [54, 72]. The empirical evidence reviewed here indicates that the effect of VTS on cortical oscillatory activity over somatosensory and premotor cortex is an almost immediate suppression of low-frequency oscillations, which reduces the excessive somatosensory-motor cortical activity and corresponds to the observed fast occurring behavioral changes such as increased loudness of voice, reduced number of voice breaks and sentence durations [54, 59]. Invasive neuromodulation techniques, such as deep brain stimulation, attempt to normalize the irregular neuronal discharge patterns by applying high-frequency impulses to targeted subcortical nuclei [73] with the aim to restore the activity of upstream motor cortical networks. The results from the reviewed studies as well recent recordings from basal ganglia output nuclei showing that neck muscle VTS normalized firing of neck-sensitive neurons in globus pallidus internus [74] indicate that non-invasive high-frequency peripheral stimulation via VTS may modulate the discharge patterns of neurons in the somatosensory-motor network in a similar way.

There is inconclusive evidence on the long-term effects of VTS on dystonic symptoms. In two longitudinal studies, participants with LD applied VTS at home for up to 2 months [45, 55]. They confirmed that the repeated use of laryngeal VTS was safe and feasible at home. The majority of responders to VTS showed a consistent, repeatable response over weeks to a daily 20-minute application of VTS with approximately 20% of those reporting that overall speech effort decreased slowly over the 2-month period [45]. From a clinical and patient perspective, it is important to know for how long a VTS will last. Based on the available research reviewed here, the therapeutic effects after 20–45 min of VTS vary widely from less than 30 min to several days.

Several important limitations should be considered when interpreting the findings of this review. First, the applied VTS protocols and the reported outcome measures varied largely across studies. This limited the interpretation of the evidence, because it was not possible to perform a rigorous meta-analysis. Second, most studies had relatively small sample sizes and besides a single study [55], no randomized clinical trials are yet available. Third, few studies considered the influence of confounding variables in clinical characteristics (e.g., disease duration, BoNT injection history) on VTS effects. Fourth, no fully sham-controlled randomized clinical trials have been conducted. While the available neurophysiological evidence on the afferent pathways and the brain areas activated by VTS is inconsistent with an unspecific placebo response, one cannot exclude the possibility that, at least in a subset of participants, the observed symptom reductions were due to a placebo effect. In addition, the evidence of VTS-induced symptom relief was also based on objective markers (e.g., head righting in CD, CPPS in LD), which is also inconsistent with a placebo response. Lastly, most studies examined the short-term effects of VTS, and the long-term results of VTS in patients with dystonia are unknown and an effective protocol cannot yet be recommended. At present, two longitudinal studies [45, 55] showed that people with LD may receive consistent, repeatable symptom relief over time with some showing an increased magnitude of symptom relief over time.

This review documents a long line of research spanning over nearly four decades that applied VTS to various presentations of FD. What have we learnt from this research and how can it be applied to clinical practice? In summary, there is converging evidence that VTS to dystonic regions can provide temporary symptom relief of varying duration for people affected with FD. There is no evidence that one form of FD is more susceptible to VTS compared to others–VTS has been applied to numerous subtypes of FD and they mostly showed a measurable, repeatable response based on the reviewed neurophysiological and behavioral data (see Figures 2, 3). While the reported response rates are high, indicating that well over half of FD patients may benefit from VTS, there is no clear understanding what determines the susceptibility to VTS in FD. Moreover, most studies applied VTS in a single session and, depending on the reported outcome measures, approximately 15%–42% of FD participants did not respond to a single dose of VTS. Yet, there is initial evidence that continued VTS use over weeks reduces overall symptoms severity in people with LD [45]. Nevertheless, there is still a paucity of longitudinal data on the effects of VTS on dystonic symptoms.

From a clinical perspective one needs to ask whether VTS is a practical strategy to manage patients with dystonia. Why is VTS still not an established treatment for FD? Part of the reason why VTS was not employed earlier stems from the fact that earlier vibration technology was too heavy, too cumbersome and not wearable. Today small vibrators with enough vibration amplitude and a variable range of frequencies are available to build wearable VTS devices for clinical and in-home use. From a practical perspective, there is initial evidence available from two longitudinal studies [45, 55] that one significant advantage of VTS is that it can be easily and safely applied at home. Adverse events due to VTS appear to be rare, with only one study reporting that VTS induced dystonic cramps in patients with unilateral dystonia [65]. Research has already documented that direct skin contact of the vibrators is not necessary to be effective [45]. Thus, wearable VTS devices in the form of collars (see Figure 7B) or arm braces that are rechargeable and wirelessly controlled are a realistic scenario for future use.

What is the path forward? Besides advancing wearable vibration technology, the next necessary steps are the following: First, to delineate the VTS-induced changes in somatosensory-motor networks affected by FD, such as determining short-term and long-term plasticity due to VTS in cortical as well as subcortical regions of the network. This work would further solidify the neurophysiological basis for any future VTS therapy and address the criticism that VTS is mainly a placebo treatment. Second, to conduct rigorous randomized clinical trials to establish treatment efficacy against various sham conditions or other non-invasive treatments (e.g., acupuncture, transcranial magnetic stimulation). These studies would further address the issue of VTS as a placebo treatment. Third, to conduct the necessary clinical trials to determine the optimal VTS frequency, amplitude, duration, number of sessions, and relevant body parts to be stimulated for the various FD subtypes. This work will establish guidelines for the standardization of VTS therapy for treating FD.

In closing, the current empirical knowledge on VTS as a possible non-invasive neuromodulation treatment for FD is incomplete but the present results document that VTS has the potential to reduce dystonic symptoms in people with FD. Given the current limited treatment options for people with FD, VTS could enlarge the available therapeutic arsenal.