Introduction

Postural control is an essential function of the human body. The main goal of postural control is to keep the body in an upright and stable position. For this reason, even simple daily tasks cannot be performed without efficient postural control (Chiba et al. 2016). To have efficient postural control, a combination of input from the visual, vestibular, and proprioceptive systems is used (Ivanenko and Gurfinkel 2018; Winter et al. 2003). These three systems are connected via direct neural connections in the brainstem area (Edney and Porter 1986) and the central cervical nucleus (Corneil et al. 2002).

As the proprioceptive system uses input from mechanoreceptors, such as muscle spindles (Sjölander et al. 2002), changes in the input from these muscle spindles can change a person’s postural control (Gilman 2002; Riemann and Lephart 2002; Treleaven 2008). When a muscle spindle is stimulated (e.g., using vibration), this induces an illusion of muscle lengthening and joint movement. This manipulation of sensory input engenders subsequent repercussions on the integration of sensory information within the central nervous system, leading to modulations in postural control. As a reaction, this causes a shift in the body’s center of gravity (Cordo et al. 2005). So, probably the response will be different when flexor or extensor muscles are stimulated. Consequently, areas with a high density of muscle spindles, such as the lower limb and cervical spine area (Treleaven 2008), will be important for maintaining balance. The proprioceptive system is extremely developed in the cervical spine due to a high density of mainly the muscle spindles in the deep segmental upper cervical muscles (Bogduk and Mercer 2000). The rectus capitis posterior muscle, for instance, contains 98 spindles per gram of muscle, sternocleidomastoid contains 72.8 spindles per muscle, compared to 55.4 spindles per gram of muscle for the posterior latissimus dorsi muscle (Kulkarni et al. 2001; Ovalle et al. 1999; Radziemski et al. 1991). Afferent proprioceptive information provided by these cervical muscle spindles will be combined with visual and vestibular afference and proprioceptive input from the lower limbs to maintain postural control (Edney and Porter 1986).

Changes in one of the systems providing somatosensory afference results in a change in postural control and related balance (Jamal et al. 2020). In studies investigating the influence of neck dysfunction on balance, neck muscle vibration is frequently used to disturb the cervical proprioceptive input (Jamal et al. 2020). Several of these studies found a significant effect of neck muscle vibration on center of pressure (CoP) displacement (Bove et al. 2006, 2004, 2007, 2001; Courtine et al. 2007; De Nunzio et al. 2018; Ivanenko et al. 1999; Kavounoudias et al. 1999; Morris et al. 2015; Mullie and Duclos 2014; Ribot-Ciscar et al. 2004; Smetanin et al. 2011; Verrel et al. 2011). From these studies, vibration of muscles with higher muscle spindle density has a greater effect on the postural sway (Cordo et al. 2005). If this is the case, the vibration of the dorsal neck muscles should, for instance, have a larger effect on postural sway than the vibration of the sternocleidomastoid muscle.

Additionally, regarding vibration frequencies, an animal study has shown that higher vibration frequencies cause a greater illusion of movement (Zhang et al. 2019) and would therefore, have a bigger effect on the postural sway. The effect of the vibration on the stimulation of the muscle spindles increases linearly with vibration frequencies up to 70–80 Hz, followed by a subharmonic increase at higher frequencies, with sharp falls often observed at frequencies between 150 and 200 Hz (Roll et al. 1980, 1989; Steyvers et al. 2003). Heterogeneity in the methodology used in the different studies and lack of comparison between methodologies makes it currently hard to conclude the ideal vibration location and frequency in humans.

Most studies used a vibration frequency of 80 Hz (Courtine et al. 2007; Ivanenko et al. 2000; Kavounoudias et al. 1999; Mullie and Duclos 2014) or 100 Hz (Bove et al. 2006, 2009; Morris et al. 2015; Smetanin et al. 2011; Verrel et al. 2011) and one study showed the effect of 150 Hz (Seizova-Cajic and Ben Sachtler 2007). Vibration at a low frequency of 20 Hz showed no effect on postural sway, and vibration at 40 Hz and 60 Hz showed smaller postural sway changes than the 80 and 100 Hz vibrations (Pyykkö et al. 1989). Therefore, we decided not to use frequencies below 80 Hz in our study. Currently, no studies comparing the effect of different vibration frequencies are available. Both vibration on the dorsal neck muscles (DNM) (Bove et al. 2006, 2009; Ivanenko et al. 1999, 2000; Kavounoudias et al. 1999; Morris et al. 2015; Mullie and Duclos 2014; Ribot-Ciscar et al. 2004; Smetanin et al. 2011; Verrel et al. 2011) and sternocleidomastoid (SCM) muscles (Bove et al. 2006, Bove et al. 2004, 2007, Bove Diverio 2001, Courtine et al. 2007) has previously shown disturbance of postural sway, but again no studies are available comparing the effect of vibration on both locations. Additionally, several studies have shown that bilateral muscle vibration causes a shift in the center of pressure (CoP), mainly in the anteroposterior direction (Bove et al. 2006, 2009; Ivanenko et al. 1999, 2000; Kavounoudias et al. 1999; Morris et al. 2015; Mullie and Duclos 2014; Ribot-Ciscar et al. 2004; Smetanin et al. 2011). A smaller number of studies also showed the effect of unilateral vibration on the postural sway, showing a shift of the CoP in both anteroposterior and mediolateral direction(Bove et al. 2006, 2004; Courtine et al. 2007; De Nunzio et al. 2005; Dumas et al. 2013; Kavounoudias et al. 1999). Generally, the direction of the shift is expected to be opposite to the muscle function. So, if neck muscles are stimulated bilaterally, the response will be mainly in the anteroposterior direction because the muscles will produce the same illusion of the movement in the sagittal plane (flexion or extension) and an opposite illusion in the frontal plane (right vs left side flexion). In comparison, only one illusion is produced during unilateral vibration, in the frontal (flexion or extension) and sagittal plane (right or left side flexion). For this reason, we can expect a shift in both anteroposterior and mediolateral directions.

Therefore, our study aimed to compare the effect of different vibration locations (unilateral SCM and DNM (semispinalis and splenius capitis muscles)) and frequencies (80, 100, and 150 Hz) on postural sway using unilateral vibration. Our goal was first to confirm that neck muscle vibration disturbs the postural control, second to explore which vibration location results in the largest disturbance of the postural sway, thirdly to explore which vibration frequency causes the largest disturbance on the postural sway, and finally to explore which combination of vibration frequency and location causes the largest disturbance in the postural sway, in order to be able to use this combination to alter cervical somatosensory afference in future functional neuroimaging studies.

Four research hypotheses were investigated: (1) neck muscle vibration increases the postural sway variables compared to the without vibration condition (2) DNM (semispinalis and splenius capitis muscles) vibration shows larger effects on the postural sway variables than SCM vibration, (3) Higher vibration frequencies have a larger effect on postural sway variables than lower frequencies and (4) A 150 Hz DNM vibration causes the largest effect in comparison with other combinations of vibration location and frequency.

Methods

Study design and participants

In this cross-sectional study, postural sway is measured in six different vibration conditions. The six conditions were six different combinations of different vibration locations (right sternocleidomastoid muscle and right dorsal neck muscles [semispinalis and splenius capitis muscles]) and frequencies (80, 100, and 150 Hz). Postural sway measures were compared to a baseline measurement without vibration for each vibration condition. Healthy adult participants were included in the study when they had no history of neck pain or lower limb injury in the previous three months. Participants suffering from diagnosed vestibular dysfunction or recent surgery were excluded from the study, as were participants who suffered from a recent head injury (such as a concussion, less than 6 months before the study) and participants with a Body Mass Index greater than 30.

The measurements took place between January 2021 and March 2021 at the. Before inclusion, all participants were informed about the study by the primary researcher, and a consent form was signed by all participants according to the guidelines provided by the Declaration of Helsinki. The study was approved by the Ethical Committee of the with reference number (B3002020000250).

Sample size calculation (G-Power 3.1) was performed a-priori and was based on literature data from changes in CoP displacement during postural sway tests with and without vibration of the cervical spine (Lekhel et al. 1997). The sample size calculation for this repeated-measures study showed that a minimum of 12 subjects was needed to be able to reject the first null hypothesis that there is no difference in CoP displacement between the vibration of the cervical spine and no vibration and the second null hypothesis that there is no difference in CoP displacement between the six vibration conditions as compared to the no vibration condition, with 80% power and a type I error probability of 0.05.

Vibration

The vibration was applied using a micro-piezo-tactile stimulator (mPTS) system (Dancer Design, Saint Helens, UK), a nonmagnetic system that is magnetic resonance imaging compatible and has high potential for use in future functional neuroimaging studies. In total, six different combinations of different vibration locations (right sternocleidomastoid muscle and right dorsal neck muscles (semispinalis and splenius capitis muscles)) and frequencies (80, 100, and 150 Hz) were tested. The exact placement of the mPTS is specified in Fig. 1.

Fig. 1
figure 1

A Vibrator placement on sternocleidomastoid muscle, in the center of the muscle belly, 4 cm below the insertion on the mastoid bone B Vibrator placement on dorsal neck muscles, below the occiput and 2–5 cm lateral to the cervical spine, red circle: vibrating area

Full size image

The vibration amplitude was 1 mm based on previous studies (Andersson and Magnusson 2002; Gomez et al. 2009). Since neck muscle vibration has at most a 24 h after-effect on cervical sensorimotor performance (Beinert et al. 2018), only one combination was tested on the same day in the same participant. The six different days were evenly spread over a period of 4 weeks. The six vibration conditions were applied in a random sequence to decrease the risk of introducing a sequence bias.

Procedure

The effects of neck muscle vibration were objectified using postural sway tests. The postural sway was measured in a standing position using one embedded force plate (AMTI®, 0.4 0.5 m, 1000 Hz, model OR 6–5-2000, Advanced Medical Technology Inc., Massachusetts, USA) (Fig. 2). During the postural sway test, participants stood barefoot with their eyes closed on the force plate. Both feet were positioned in parallel at a distance of 10 cm (Nunzio et al. 2018). This position was marked on the force plate, and two reflective markers were placed on the heel (middle of the Achilles tendon) as reference points in order to have consistency across trials. The marker position data were captured by eight near-infrared cameras (Vicon T10, 100 Hz., Vicon Motion Systems Ltd., Oxford, UK, 100 fps, resolution 1 Megapixel (1120 × 896)) and the markers were labeled using Vicon Nexus software. During each test session, the postural sway was first tested without vibration and afterward during one of the vibration conditions. Each postural sway measurement consisted of three trials of the 40 s with 30 s rest in between (Fig. 2). The following protocol was used:

  1. 1.

    Postural sway (PS) test without vibration

$$ {4}0 {\text{s PS}}{-}{3}0 {\text{s rest}}{-}{4}0 {\text{s PS}}{-}{3}0 {\text{s rest}}{-}{4}0 {\text{s PS}} $$
Fig. 2
figure 2

The experiment procedure, CoP Centre of Pressure, ap anteroposterior, ml mediolateral, RMS Root Mean Square, Vel Velocity, Cond Condition, Con Control

Full size image
  1. 2.

    Postural sway test with vibration

    $$ {4}0 {\text{s PS}}\, + \,{\text{vibration}}{-}{3}0 {\text{s rest}}{-}{4}0 {\text{s PS}}\, + \,{\text{vibration}}{-}{3}0 {\text{s rest}}{-}{4}0 {\text{s}}\,{\text{PS}}\, + \,{\text{vibration}} $$

The reliability of the protocol was assessed by examining data obtained under conditions without vibration, yielding robust test–retest Interclass Correlation Coefficient (ICC) values (ranging from 0.773 to 0.908) using a two-way mixed effects model with absolute agreement (unpublished data).

Data analysis

The raw data files were exported to MATLAB (R2020a for Windows, The MathWorks, Inc., Natick, Massachusetts, United States), and an existing MATLAB script was used to calculate the mean CoP displacement, the RMS and mean velocity of CoP, all in the anteroposterior and mediolateral direction. Additionally, the mean sway area of CoP was calculated. The CoP positions were calculated from the ground reaction forces (Fx, Fy, and Fz) and moments of the force around the x (Mx) and y (My) axes (Duarte et al. 2010). A second order low pass Butterworth filter with a cut-off frequency of 12.5 Hz was used for filtering the CoP, and a second order low pass Butterworth filter with a cut-off frequency of 6 Hz was used for filtering the marker position data (Duarte and Freitas, 2010). The mean CoP displacement in the anteroposterior and mediolateral direction, the mean position between the two heel markers, was used as a reference position (Fig. 3). The sway area was calculated by fitting an ellipse corresponding to 85% of the surface CoP points (Oliveira et al. 1996) (Fig. 3). The first 10 s of each trial were not used during data analysis, as they are seen as an adaptation period (Mezzarane and Kohn 2007). The average between the three trials was calculated for each postural sway measurement. For mean CoP displacement in the anteroposterior direction, positive values represent a forward shift of the CoP, and negative values a posterior shift. For mean CoP in the mediolateral direction, positive values represent a shift of the CoP to the right side, and negative values a shift to the left. For all the other variables, higher values represent a greater effort from the participants to remain stable.

Fig. 3
figure 3

An example of the CoP variables calculation. The right and left heel marker (red “o”), the average marker position (blue “o”), the CoP line (blue line), the ellipse of the sway area (magenta ellipse), and the mean CoP anteroposterior and mediolateral displacement (turquoise “*”), CoP Centre of Pressure, ap anteroposterior, ml mediolateral, RMS Root Mean Square, MV Mean Velocity

Full size image

Statistical analysis

The normality of the data was tested using a Shapiro–Wilk test. Since some data were not normally distributed (p < 0.05), we used the aligned rank transform tool (ART tool version 0.11.1) for nonparametric factorial ANOVA as described by Wobbrock et al. (2011) and Elkin et al. (2021) using R software 4.0.4. A three-way repeated measures ANOVA was used to identify significant differences in CoP variables.

First, the interaction of the before and during vibration conditions with both frequency and location (conditions*frequency* location) was calculated to find potential differences in the effect of vibration between the combination of different frequencies and locations. This allows us to investigate if a specific combination of vibration frequency and location affects the postural sway more than other combinations. Afterward, interactions between the before and during vibration conditions and vibration frequency (conditions*frequency) were tested to find potential differences in the effect of vibration between the different frequencies. This allows us to investigate if a certain vibration frequency has a larger effect on the postural sway than other frequencies. Additionally, the interaction between the before and during vibration conditions and the vibration location (conditions*location) was tested to find potential differences in the effect of vibration between the different vibration locations. This allows us to identify if one of our vibration locations causes a larger effect on the postural sway. Finally, potential differences between the before and during vibration conditions were investigated to objectify the effect of vibration on the postural sway variables. This analysis was made to be sure that our mPTS vibration causes changes in postural sway. Post hoc pairwise comparisons were not tested due to the lack of significant interaction results. 95% confidence intervals (CI) for the mean differences were calculated. Significance was considered when p < 0.05.

Results

Demographics

Seventeen healthy participants (14 women and three men, mean age: 27.69 years (SD: 3.16), mean height: 1.70 m (SD: 0.06), mean weight: 63.24 kg (SD: 7.81) were included in the study. All tolerated the mPTS vibration well, and no side effects or discomfort was reported during the vibration application.

Effect of vibration on postural sway

We found a statistically significant effect of vibration on all postural sway variables. An overview of all results can be found in Tables 1 and 2 and Fig. 4. An average anterior shift of 6.1 mm (SD: 6.6) (F1,176 = 34.89, p < 0.001) and a right shift of 2.8 mm (SD: 2.6) (F1,176 = 11.40, p < 0.001) of the CoP was observed during vibration compared to the condition without vibration. For the RMS, an increased error of 0.4 mm (SD: 0.3) (F1,176 = 61.15, p < 0.001) and 0.2 mm (SD: 0.2) (F1,176 = 39.28, p < 0.001), was observed during vibration compared to the condition without vibration, in the anteroposterior and mediolateral direction, respectively. The mean velocity was higher during vibration in comparison with the reference condition. The average difference for velocity was 0.5 mm/s (SD: 0.4) in the anteroposterior direction (F1,176 = 26.69, p < 0.001) and 0.2 mm/s (SD: 0.2) in the mediolateral direction (F1,176 = 18.74, p < 0.001). Finally, during vibration, the average sway area was 15.4 mm2 (SD: 15) larger compared with the condition without vibration (F1,176 = 58.38, p < 0.001).

Table 1 Mean center of pressure and root mean square differences between with and without vibration
Full size table
Table 2 Mean velocity and area differences between with and without vibration
Full size table
Fig. 4
figure 4

Postural sway results between without and during vibration. Vibration caused a significant effect in all postural sway variables (p < 0.001). mm = millimeters and s = seconds. Boxplot: upper line: the highest data point in the data set excluding any outliers, lower line: the lowest data point in the data set excluding any outliers, upper box line: third quartile, lower box line: First quartile, mid-box line: median, diamond: average value, black dots: the value of each participant

Full size image

Effect of the vibration frequency and location on the postural sway

We did not find any significant conditions*frequency or conditions*location or conditions*frequency*location interaction in our postural sway variables. (see Tables 1 and 2 and Figs. 5 and 6).

Fig. 5
figure 5

Mean CoP displacement results in anteroposterior direction between without and during vibration in different locations and frequencies. The interaction between vibration*frequency*location was not significant(p > 0.05). mm = millimeters. Boxplot: upper line: the highest data point in the data set excluding any outliers, lower line: the lowest data point in the data set excluding any outliers, upper box line: third quartile, lower box line: First quartile, mid-box line: median, diamond: average value, black dots: the value of each participant

Full size image
Fig. 6
figure 6

Mean CoP displacement results in mediolateral direction between without and during vibration in different locations and frequencies. The interaction between vibration*frequency*location was not significant(p > 0.05). mm = millimeters. Boxplot: upper line: the highest data point in the data set excluding any outliers, lower line: the lowest data point in the data set excluding any outliers, upper box line: third quartile, lower box line: First quartile, mid-box line: median, diamond: average value, black dots: the value of each participant

Full size image

Discussion

Our study aimed to compare the effect of different vibration locations (unilateral SCM and DNM (semispinalis and splenius capitis muscles)) and frequencies (80, 100, and 150 Hz) on postural sway using unilateral vibration. Our goal was first to confirm that neck vibration disturbs the postural control, second to explore which vibration location has the largest influence on the postural sway, thirdly to explore which vibration frequency causes the larger disturbance on the postural sway, and finally to explore which combination of vibration frequency and location causes the largest disturbance of the postural sway.

Our first hypothesis was that neck muscle vibration increases the postural sway variables compared to measurements without vibration. We can confirm this hypothesis, as we found a significant increase in all postural sway variables during neck muscle vibration. Our second hypothesis was that DNM (semispinalis and splenius capitis muscles) vibration affects the postural sway variables more, than SCM vibration. In general, we rejected this hypothesis as we found that both unilateral DNM and SCM vibration has a significant effect in all postural sway variables and that this effect does not significantly differ between the two locations. Based on earlier studies, we expected that DNM (semispinalis and splenius capitis muscles) vibration would result in a larger shift of CoP due to a higher muscle spindle density compared to the SCM (Cordo, Gurfinkel, 2005). Despite that, we did not find a significant difference between the two locations, in five out of seven postural sway variables, the DNM causes a greater increase in these variables. Potentially, the difference in muscle spindle density between the DNM and SCM was too small to result in significant differences in CoP displacement, but also, the minimal detectable change of the method we used is greater than these differences. Another potential reason is that, due to the shape of the vibrator, there was better skin contact on the SCM muscles and that may engage muscles beyond solely the SCM.

For the mean CoP displacement in the anteroposterior direction, our results confirm previous findings from studies investigating vibration applied unilaterally (Courtine et al. 2007, De Nunzio et al. 2018) on the DNM but contradict other studies investigating vibration applied to the SCM, who did not find significant anteroposterior CoP displacement (Bove et al. 2006, Bove et al. 2004, Bove, Diverio, 2001, Courtine et al. 2007). Methodological differences can explain these contradictory results. Specifically, we tried to vibrate the upper part of the SCM muscle. SCM function is flexion of the head relative to the shoulders and extension of the upper cervical spine. In other studies, the vibrator was placed in the middle of the SCM muscle. This might result in a combined illusion of flexion and extension, resulting in very small anteroposterior CoP displacements (Bove et al. 2006, Bove et al. 2004, Bove, Diverio, 2001, Courtine et al. 2007). The vibrator placement in our study’s upper part of the SCM muscle might have resulted in an illusion of upper cervical extension rather than cervical flexion, explaining why we found an anterior CoP displacement during SCM vibration. Additionally, it is plausible that the shape of our vibrator induced an illusory effect on muscles other than SCM. Furthermore, most studies used a vibration amplitude between 0.2 and 0.8 mm (Bove et al. 2006, Bove et al. 2004, Bove, Diverio, 2001, Courtine et al. 2007). A previous study shows that a vibration amplitude between 1 and 2 mm has a larger effect on the muscle spindles (Necking et al. 1996). In our study, we used an amplitude of 1 mm which might also explain why we found a significant increase of the mean CoP displacement in the anteroposterior direction in contradiction to other studies (Bove et al. 2006, Bove et al. 2004, Bove, Diverio, 2001, Courtine et al. 2007). We measured each condition repeatedly with three trials of the 40 s and 30 s rest in between, whereas other studies only measured both conditions once (Bove et al. 2006, Bove et al. 2004, Bove, Diverio, 2001, Courtine et al. 2007). This is the most common method recently used in postural sway evaluation (Duarte and Freitas 2010; Paillard and Noé 2015). Additionally, we calculated the mean CoP displacement from the last 30 s of each measurement and discarded the first 10 s so the participant was allowed to get familiar with the procedure and find the best way to balance (Nunzio et al. 2018; Oliveira et al. 1996; Paillard and Noé, 2015). With this protocol, we calculated the longitudinal effect of the vibration on CoP displacement rather than the immediate effect of the vibration, which happens in the first 1–2 s of the vibration (Andersson and Magnusson 2002, De Nunzio et al. 2018). It is recommended not to use the mean displacement position, especially in case of repeated measures, due to the low reliability and accuracy of this method (Duarte and Freitas 2010; Paillard and Noé 2015). For this reason, we used two heel markers as reference points apart from the line markers in the force plate. In each trial, the mean CoP displacement variables were calculated from the mean position between the two markers. We tested the reliability of the measurement with and without the heel markers and found that this protocol (with heel markers and line markers) is more reliable (ICC = 0.890) and accurate in calculating the mean CoP displacement in comparison to the protocols that were used in the other studies (ICC = 0.575, only line markers without heel markers) (based on unpublished data). In the other studies, line marks in the force plate were used in order to achieve the same foot position between trials (Bove et al. 2006, Bove et al. 2004, Bove, Diverio, 2001, Courtine et al. 2007).

For the mean CoP displacement in the mediolateral direction, we observed a shift of the CoP towards the side of vibration (right side) that does not significantly differ between DNM and SCM vibration. Our findings were in contrast to one other study when vibration was applied unilaterally on the DNM (Nunzio et al. 2018). In case the vibration was applied to the SCM, other studies found a significant displacement to the opposite side of the vibration (Bove et al. 2006, Bove et al. 2004, Bove Diverio 2001, Courtine et al. 2007). The differences in reaction to vibration on DNM and SCM muscle might be explained by the differences in data analysis used to calculate the mean CoP displacement in the mediolateral direction (ICC = 0.908 with line and heel markers vs ICC = 0.374 with only line markers without heel markers) (based on unpublished data). Additionally, the difference in the vibration device, like the shape, dimensions, weight, etc., it is plausible that they induced an illusory effect on muscles other than DNM and SCM and that might have contributed to the differences in mediolateral CoP displacement compared to other studies. As a conclusion, both locations cause a similar shift of the mean CoP displacement towards the side of vibration (right side).

Since mean CoP displacement largely depends on the foot position, small changes in the foot position can have a large effect on the mean CoP displacement. Therefore, we also calculated RMS, the mean velocity, and the sway area (Duarte and Freitas 2010; Paillard and Noé, 2015). The RMS and mean velocity were significantly higher during vibration than the reference condition in both the anteroposterior and mediolateral directions. This increase was similar between the two vibration locations. Moreover, the sway area was higher during vibration than without vibration. This increase was also similar in both vibration locations. These results confirm the findings from one other study, showing that sway area increases during neck muscle vibration (Bove et al. 2007), but this increase is shown to be similar between the two vibration locations.

An animal study has shown that higher vibration frequency causes a greater illusion of movement (Zhang et al. 2019) and, therefore, would have a larger effect on the postural sway. Additionally, the effect of the vibration on the stimulation of the muscle spindles is increased linearly with vibration frequencies up to 70–80 Hz, followed by a subharmonic increase at higher frequencies, with sharp falls often observed at frequencies between 150 and 200 Hz (Roll and Gilhodes 1980; Roll et al. 1989; Steyvers et al 2003). So, our third hypothesis was that the higher frequencies would have a larger effect on postural sway than the lower frequencies. In general, we were able to reject this hypothesis as we found that all vibration frequencies significantly increase the postural sway variables, but we did not find any statistically significant differences between the different frequencies. Despite that, we did not find a significant difference between the three vibration frequencies, in four out of seven postural sway variables, the 150 Hz caused a greater increase in these variables. Potentially, the difference of the effect between different frequencies was too small to result in statistically significant differences in CoP displacement, but also, the minimal detectable change of the method that we used is greater than the observed differences.

The ultimate goal of our study was to define the most disturbing combination of vibration frequency and location for postural sway measurements. Our last hypothesis was that 150 Hz DNM vibration would cause the largest effect on postural sway parameters compared to other combinations of vibration location and frequency. In general, we were able to reject this hypothesis as we found that there is not one combination that results in larger changes in postural sway variables for the investigated locations and frequencies. Although not significant, higher frequency vibration on the SCM does show higher mean velocity figures. This observation confirms previous findings by Pyykkö et al. (1989).

It should be noted that this study included only asymptomatic participants, and the results may not be generalized to symptomatic populations such as neck pain patients. Furthermore, the observed disparity between the percentages of women (14) and men (3) in the sample size could potentially influence the outcomes. However, it appears unlikely that this factor significantly impacts the results. Additionally, the effect sizes observed in this study were smaller than anticipated, potentially impacting the statistical power of the investigation. It is plausible that studies with larger sample sizes could possess greater statistical power to detect potential differences between distinct locations or frequencies of vibration. Moreover, comparative studies exploring various types of vibrators may contribute to establishing the optimal properties of vibrators for neck vibration.

Generally, this procedure can be used to evaluate the cervical somatosensory afference in patients with cervical spine dysfunction caused by several non-specific conditions. One of the targets of cervical spine therapies is to increase the proprioception of the neck muscles. So, neck muscle vibration can also be used as an objective outcome measure in studies evaluating the effect of new therapies for cervical spine conditions.

Conclusion

Neck muscle vibration causes a significant increase in all investigated postural sway variables. More specifically, a forward displacement of the mean CoP was observed in the anteroposterior direction, and a displacement towards the vibration side was observed in the mediolateral direction. An increase in RMS and mean velocity was found in both anteroposterior and mediolateral directions. Additionally, we observed an increase in the sway area. No significant differences in the effect on postural sway variables were observed between the different vibration frequencies (80 vs 100 vs 150 Hz) or between the different vibration locations (unilateral DNM vs SCM). Finally, no significant differences in the effect on postural sway variables were observed between the different vibration locations and frequencies combinations.