Secondary Logo

Journal Logo

Original Research

Higher Quadriceps Roller Massage Forces Do Not Amplify Range-of-Motion Increases nor Impair Strength and Jump Performance

Grabow, Lena1; Young, James D.2; Alcock, Lynsey R.2; Quigley, Patrick J.2; Byrne, Jeannette M.2; Granacher, Urs1; Škarabot, Jakob3; Behm, David G.2

Author Information
Journal of Strength and Conditioning Research: November 2018 - Volume 32 - Issue 11 - p 3059-3069
doi: 10.1519/JSC.0000000000001906
  • Free



Self-myofascial release, self-massage therapy, and neuromuscular rolling are terms that describe the use of a tool to massage muscles and connective tissues with a rolling motion (4). Applicable tools can be a foam roller (17,25,34), roller massage (RM) stick (17,21,26), or tennis ball (16). Neuromuscular rolling has been shown to increase range of motion (ROM) immediately after the intervention (8,17,21,25,33,34) with changes present for as long as 20 minutes (min) (22,23,27). The degree of increase in ROM reported in the literature is highly variable, ranging from 2.8% (33) to 23.4% (16). Variability may be explained by the type of tool used (15,28), the target muscle group, the task instructions (17), the overall volume of neuromuscular rolling (14), and the applied rolling force (14,31,33).

A limited number of studies have suggested that the rolling forces applied to the target muscle may influence changes in ROM. Bradbury-Squires et al. (8) applied 25% of the body mass, which was equivalent to 205.9 N of RM force, to the anterior thigh. Rolling for 5 sets of 20 and 60 seconds (s) increased ROM by 10% and 16%, respectively. Another study (34) used 127.5 N of RM force applied to the hamstrings and reported an increase of 4.3% after only 1–2 sets of 5–10 seconds. Recently, 68% of the body mass (mean force of 266.7 N) was used on a foam roller, and a small but significant 2.4° increase in the hip flexors and quadriceps muscles was found (30). Although a greater increase in ROM reported by Bradbury-Squries et al. (8) in comparison with that of Sullivan et al. (34) could be the result of longer intervention, the possible effect of higher force application that would be accompanied with greater discomfort or pain cannot be excluded. Although one attempt to explain this variance showed that smaller contact area and more rigid roller design would lead to greater pressure (15), the impact of rolling force that is associated with differing perception of pain (17) remains unclear. Neuromuscular rolling has been shown to increase pain threshold associated with muscle tender spots, acute electrically evoked pain, and delayed onset muscle soreness (1,10,24). This change is also observed on the contralateral, nonrolled limb, suggesting the involvement of a central pain modulatory system (1,10). Rolling-induced improvements in ROM could be related to an increased pain or stretch tolerance (17). However, this relationship has not been previously examined.

Based on recent studies, neuromuscular rolling exerts global effects (1,10,23). For example, Monteiro et al. (29) showed improvements in overhead deep squat performance regardless of the area under treatment, i.e., lateral thigh, plantar surface of the foot, and latissimus dorsi. This finding suggests a degree of likelihood that findings from a specific muscle (e.g., quadriceps in this study) can be extrapolated to others (e.g., hamstrings).

Although neuromuscular rolling is reported to increase flexibility, it does not seem to attenuate athletic performance (4). Several studies have shown that muscle strength, power, or balance performance remained unaffected by the self-applied treatment (5,17,18,25,26,34). To the authors' knowledge, no study has examined varied RM forces on athletic performance. Whether a high or a low intensity roll has a different impact on strength and jump parameters is of direct relevance with athletic activities that includes maximal strength and power performances.

There is a practical need to identify the optimal rolling force to achieve the greatest ROM without attenuating muscular performance. The aim of this study was to compare the effects of low (RMlow), moderate (RMmod), and high (RMhigh) RM forces, calculated relative to the individual's perception of pain applied to the anterior thigh, on ROM, strength, and jump performance. With reference to the relevant literature (4,5,8,17,25,34), it was hypothesized that all interventions would enhance ROM without causing a subsequent decrease in performance. It was assumed that higher RM forces produce greater ROM improvements.


Experimental Approach to the Problem

A randomized ( within-subject design was used to investigate the effects of 3 conditions: RMlow, RMmod, and RMhigh forces applied to the anterior thigh on active and passive knee flexion ROM, single-leg drop jump (DJ) performance, and maximum voluntary isometric contraction (MVIC) measures (Figure 1). At the beginning of each session, the subject's maximum point of discomfort was reevaluated to control for daily variations of the individual's rating of perceived pain (RPP). After the warm-up (5 minutes warm-up on a Monark cycle ergometer at 60–70 revolutions per minute and 1 kilopond of resistance), dependent measures were tested, which included active and passive knee flexion ROM, 2 single-leg DJs, 2 knee extension, and 2 knee flexion MVICs. After the pretests, subjects sat quietly for 10 minutes followed by another set of dependent variable measures. After the pretests, the intervention consisted of three 60-second RM bouts either at RMlow, RMmod, or RMhigh rolling intensities. To monitor effects of repeated bouts, and thus determine possible effects of the RM volume, knee flexion ROM and single-leg DJ performance were measured after each RM set. Immediately after (INTpost) and 10 minutes after the last bout of RM (INTpost10), all measures (ROM, DJ, and MVIC) were performed again.

Figure 1.:
Flow chart of the experimental design. ROM = range of motion; DJ = drop jump; MVIC = maximum voluntary isometric contraction.


A statistical power analysis was calculated based on related previous publications (8,17,24,25,34), which determined that 16 participants would be needed to achieve an alpha of 0.05 and a power of 0.8. Thus, 16 young, healthy individuals (8 men, 27 ± 4 years, range = 22–37 years, 178 ± 5 cm, 87 ± 9 kg and 8 women, 26 ± 2 years, 170 ± 4 cm and 69 ± 8 kg) were recruited to participate in this study. All participants were either resistance or aerobically physically trained (minimum 3 sessions × 20 min/wk) and reported no previous experience with RM. Exclusion criteria included any history of neurological or musculoskeletal injuries in the past year. Participants were instructed to refrain from vigorous physical activity and to abstain from alcoholic beverages 24 hours before testing.

All subjects completed the Physical Activity Readiness Questionnaire (Canadian Society for Exercise Physiology 2011) form and signed a written letter of consent before testing. A brief explanation of the study was given during the familiarization. In addition, participants were accustomed to the RM device. This orientation allowed participants to experience the force of RM necessary to elicit their maximum point of discomfort before the first testing session. The maximum tolerable pain was considered equivalent to a 10/10 on a visual analog scale, reaching from 0 (no pain or discomfort) to 10 (maximum tolerable pain) as perceived by the participant. Participants were encouraged to practice single-leg DJs several times from a platform set at a height corresponding to 50% of the length of the tibial tuberosity. This study was approved by the Newfoundland and Labrador Human Research Ethics Board (reference # 15.226). All procedures were in accordance with the Declaration of Helsinki (2013).



The Theraband RM (the Hygenic Corporation, Akron, OH, USA) composed of dense foam wrapping around a solid plastic cylinder was used for this study. The ridged design is supposed to allow for both superficial and deep tissue massage when performing RM on the muscle (8,34). The RM was placed in a specially designed constant pressure roller apparatus (custom-designed by Technical Services, Memorial University of Newfoundland, St John's, Newfoundland and Labrador, Canada: Figure 2), which was previously used in this laboratory (8,34). This device allowed for consistent force application and frequency of rolling, thereby eliminating variations that would be typical if each individual applied the roller action to their own limb (Figure 3). A pilot study on experienced individuals was conducted beforehand. It revealed that the average force that a person would exert while rolling a muscle was strongly dependent on the day and the individual. Weighted plates were added to the vertical poles until the load of the apparatus for 1 full cycle of rolling reached the individual's 10/10 RPP on the specific testing day. The evaluation of the highest point of discomfort on each day was conducted before the warm-up. The maximum weight put on the apparatus was calculated relatively for 50% (RMlow: 116.7 ± 27.5 N; 15% of body mass), 70% (RMmod, 160.6 ± 29.4 N; 21% of body mass), and 90% (RMhigh, 205.9 ± 34.3 N; 27% of body mass) of the participants' maximum (10/10) RPP. These relative loads were chosen because most of our previous publications (1,8,10,17,24,25) used rolling pressures at 7/10 on a pain scale. While 70% (7/10) provides moderate discomfort, the choice of 90% and 50% would provide a spectrum of either extreme discomfort or minimal discomfort while still providing varying pressure on cutaneous, fascial, and muscle sensory receptors.

Figure 2.:
Custom-designed constant rolling pressure apparatus.
Figure 3.:
Roller massage procedure.

One full cycle (distal to proximal and return) was completed by the researcher in 4 seconds, as conducted in previous studies (8,30). The pace was monitored with a metronome. Each roll commenced at the distal end of the quadriceps superior to the patellar tendon and continued to the proximal end of the quadriceps (hip crease) and was then reversed. After each of the 60-second bouts of rolling, the participant performed active and passive knee flexion ROM as well as 2 single-leg DJs. The between-RM-bout measures were conducted to control for possible changes dependent on RM volume. The between-set-interval was 2 minutes. The RM was performed for 3 sets of 60 seconds irrespective of the RM force. To determine the individuals' RPP during each RM bout, participants were asked to mark 3 separate, blank 10-cm lines that represented the visual analog scale after 5, 30, and 60 seconds of RM. After 60 seconds of rolling, RM caused low (3.9/10 ± 0.64 RPP), moderate (6.2/10 ± 0.64 RPP), and intense (8.2/10 ± 0.44 RPP) pain in RMlow, RMmod, and RMhigh, respectively. The individual pain scores at 60 seconds were further used to investigate possible pain-related ROM increases.

Dependent Variables

Two pretests were analyzed to determine possible effects of repeated measures (2). The second set of measures before the intervention (INTpre) was used as a baseline for comparison with the intervention. To monitor effects of repeated bouts, and thus determine possible effects of the RM volume, knee flexion ROM, and single-leg DJ performance were measured after each RM set. Immediately after (INTpost) and 10 minutes after the last bout of RM (INTpost10), all measures (ROM, DJ, and MVIC) were performed again. The 3 experimental sessions were conducted at similar times during the day to minimize diurnal effects. A minimum of 48 hours and maximum of 4 days was scheduled between each session.

Knee flexion ROM was assessed using a slight modification of the kneeling lunge position published previously (24,25). With their torso upright in an erect position, participants were asked to position their malleolus of the nondominant leg over a horizontal line on the floor with their tibia perpendicular to the floor. A rectangular frame over the horizontal line served to visualize the position of the nondominant limb and support the erect position. The dominant knee was then moved back until the participant felt a maximal tolerable stretch in the dominant hip flexor without deviating from the initial position. The position of the dominant knee was marked and kept for all subsequent ROM measurements for each session. All measures were performed barefoot and on the dominant leg as identified by the lateral preference inventory (13). Knee joint ROM was assessed by the same researcher with an analog goniometer placed in accordance with MacDonald at al. (24,25) at the following landmarks: the lateral malleolus, the lateral epicondyle, and the center of the vastus lateralis (VL). The participants were instructed to engage their abdominal muscles to maintain trunk posture. By actively contracting the hamstrings, active knee flexion ROM was then measured. For the passive knee flexion ROM assessment, one researcher passively flexed the individual's knee until the participant reached the maximal point of discomfort (Figure 4).

Figure 4.:
Kneeling lunge position for measurement of knee flexion range of motion (ROM).

All DJ trials were performed on a force platform (400 × 600 ×83 mm, model BP400600 HF-2000; AMTI, Watertown, NY, USA) and amplified at a gain of 1,000. Vertical ground reaction forces (GRFs) were measured. Participants stood on a platform set at a height corresponding to 50% of the length of their tibial tuberosity, and with their hands placed akimbo. They were instructed to drop onto the force plate with their dominant leg and vertically jump at the maximal effort with as little ground contact time as possible and land on their dominant leg. A trial was considered valid when the foot was placed on the force plate and when participants maintained a stable landing position for 2 seconds. The GRF was used to determine contact time and jump height. Jump height was calculated using the formula: jump height = 0.5 × g × t2 in which g refers to the acceleration due to gravity, and t is the flight time. In addition, given that ground contact time is a relevant parameter for DJ performance, a performance index was calculated according to the following formula: performance index = jump height/contact time (35).

Participants performed unilateral MVICs by flexing or extending the knee joint against a strap attached to the ankle while sitting on a padded table. To prevent hip extension, the subjects were fastened to the table at the proximal part of the thigh while the upper body was fastened to a back rest. The ankle strap was secured by a high-tension wire to a Wheatstone bridge configuration strain gauge (Omega Engineering Inc., Don Mills, ON, Canada), which was connected to a metal plate on the opposing wall for knee flexion and to the table for knee extension. Differential voltage from the strain gauge was amplified (DA 100; Biopac Systems Inc., Holliston, MA, USA), A/D converted (MP100WSW, DA100 amplifier; Biopac Systems Inc., Holliston, MA, USA), and monitored on a personal computer at a sampling rate of 2,500 Hz. The subjects were shown how to perform an MVIC with their arms crossed and then instructed to flex or extend their knee as fast and as strong as possible for hamstring or quadriceps MVIC. Verbal encouragement was given during the 4-second trials to ensure maximal force production. Two trials per movement were separated by a 1-minute rest. The order of knee flexion or extension was randomized. The maximal force level and F200 were taken into consideration for further analysis. F200 was considered an indicator of how rapidly force can be produced (20).

Statistical Analyses

SPSS software (Version 22.0, IBM) was used to analyze the data. The Shapiro-Wilk's test confirmed a normal distribution. To determine the effects of the RM intervention on MVIC force and F200, a 3 (RMlow, RMmod, RMhigh) × 3 (INTpre, INTpost, INTpost10) analysis of variance (ANOVA) was used. Effects on ROM and DJ parameters were calculated in a 3 × 5 ANOVA because measures were also taken between RM bouts. Differences were considered significant at p ≤ 0.05. When the condition × time interaction was significant, post hoc paired t tests were used to identify the statistically relevant comparisons. Moreover, effect sizes were assessed to determine the pertinence of differences by computing Cohen's d. Classifications of the effect sizes were in accordance with the literature (12) (small: d < 0.5; medium: 0.5 ≤ d < 0.8; large: d ≥ 0.8). In addition, Spearman's correlation coefficients between pain and ROM increases were determined. Intersession reliability was calculated using an intraclass correlation coefficient Cronbach's alpha. Data were presented as mean and SD.


Intraclass correlation coefficient reliability scores were classified as acceptable to high for active (0.70) and passive (0.71) ROM, DJ height (0.70), contact time (0.76), performance index (0.74), knee extension MVIC forces (0.97), and F200 (0.94) as well as knee flexion MVIC forces (0.89) and F200 (0.91).

Active Range of Motion

A significant main effect (p < 0.001, d = 2.54) for testing time was found for active ROM. Both INTpost (∆ 7.0%, p = 0.029, d = 2.25) and INTpost10 (∆ 6.9%, p = 0.026, d = 2.38) measures showed significantly more active ROM than INTpre measures in all intervention groups (Figure 5). No significant main effects for different intensities (load) of rolling force or interactions for rolling force with testing time were found.

Figure 5.:
Changes in active knee flexion ROM based on 2-way repeated measures ANOVA (main effect of time); significance level set at p ≤ 0.05; significant (*) findings of 5 and 6 are relative to the second pretest (2), significant difference of 2 is relative to 1. ROM = range of motion; RM = roller massage; mod = moderate; 1 = first pretest; 2 = second pretest (INTpre); 3 = after the first RM bout; 4 = after the second RM bout; 5 = after the intervention (INTpost); 6 = 10 minutes after the intervention (INTpost10).

Weak but significant correlations were found between the pain of each bout of RM and the ROM changes from INTpre to measures after the first (r = −0.29, p = 0.04), second (r = −0.308, p = 0.03), and third (INTpost) (r = −0.321, p = 0.02) RM application. Regarding the separate rolling forces, the changes in active ROM from INTpre to INTpost correlated with the recorded pain values during the first (r = 0.565, p = 0.023) and second (r = 0.633, p = 0.008) RMlow force application. There were no correlations with RMmod. Significant correlations with RMhigh were evident between pain during second and third RM bout and the changes in active ROM from INTpre to testing after the second (r = 0.500, p = 0.048) and third RM bout (INTpost) (r = 0.620, p = 0.010), respectively (Table 1).

Table 1.:
Effects of 3 RM forces on active and passive range of motion (ROM).*

Passive Range of Motion

There was a significant main effect for testing time (p < 0.001, d = 3.22) with significantly greater passive ROM after the second RM treatment (∆ 9.3%, p = 0.007, d = 2.40), for INTpost (∆ 15.4%, p = 0.000, d = 3.73), and INTpost10 (∆ 11.9%, p = 0.000, d = 2.90) measures in comparison with INTpre measures. The increases in passive ROM from one bout to another was significant from the first to the second bout (∆ 4.6%, p = 0.049, d = 1.12) and from the second RM bout to INTpost (∆ 7.3%, p = 0.029, d = 1.57) measures. In addition, there was a 7.4% increase in ROM from the first pretest to the INTpretest (p = 0.000, d = 1.96) (Figure 6). There was no significant effect for the intervention or the interaction of rolling force application and testing time. There was a small effect size between the greater ROM increases from INTpre to INTpost performance of RMsham and RMmod (d = 0.43) and a moderate effect size of RMlow compared with RMhigh measures (d = 0.55) (Table 1).

Figure 6.:
Changes in passive knee flexion ROM based on 2-way repeated measures ANOVA (main effect of time); significance level set at p ≤0.05; significant (*) findings of 5 and 6 are relative to the second pretest (2), significant difference of 2 is relative to 1. ROM = range of motion; RM = roller massage; mod = moderate; 1 = first pretest; 2 = second pretest (INTpre); 3 = after the first RM bout; 4 = after the second RM bout; 5 = after the intervention (INTpost); 6 = 10 minutes after the intervention (INTpost10).

There were small but significant correlations between the pain of each bout of rolling and the changes in knee flexion passive ROM between the second and INTpost measures (pain in the first: r = 0.407, p = 0.004; second: r = 0.419, p = 0.003; and third RM bout: r = 0.427, p = 0.002). The pain of the third 60-second RMsham bout correlated with the INTpost increases in passive ROM (r = 0.713, p = 0.002). There were no correlations for RMmod and RMhigh.

Drop Jump Performance

There were no significant effects on DJ height, DJ contact time, and DJ performance index (Table 2).

Table 2.:
Effects of 3 RM forces on drop jump (DJ) measures.*

Knee Flexion and Extension MVIC Force and F200

Significant main effects for testing time (p < 0.001, d = 1.53) showed 6% higher forces in the first pretest compared with INTpre performance in knee flexion MVIC (p = 0.038, d = 1.74). Main effects for testing time in knee flexion MVIC F200 (p = 0.029, d = 0.94) showed that forces achieved in the first pretest were 11.8% higher than INTpost force (p = 0.048, d = 1.5) (Table 3). Knee extension MVIC force showed main effects for testing time (p = 0.009, d = 1.08). Participants produced 4.8% more force in the initial measures than in INTpre trials (p = 0.003, d = 0.75). No further significant differences were present (Table 4).

Table 3.:
Effects of 3 RM forces on MVIC knee flexion measures.*
Table 4.:
Effects of 3 RM forces on MVIC knee extension measures.*


The most important findings of this study were that, besides the overall increase in active and passive ROM, there was no significant effect of varied rolling forces. Although higher forces elicited greater pain, participants could roll without substantial discomfort and still increase ROM. Secondly, strength (MVIC) and power (DJ) parameters remained unaffected regardless of the rolling forces. Finally, the second pretest (INTpre) was significantly different than the initial pretest for knee flexion passive ROM, knee flexion, and knee extension MVIC force, emphasizing the impact of initial pretests on subsequent measures.

Without any effect of the varied force application, RM treatment induced 7.0% and 15.4% greater active and passive ROM, respectively. The significant increases persisted for 10 minutes. Previous literature reported highly varying significant greater ROM after neuromuscular rolling (4,31). No significant findings were reported by Mikesky et al. (26) who did not control for any parameter, possibly influencing RM treatment. Vigotsky et al. (37) did not find increases in knee flexion ROM (modified Thomas test) after two 60-second bouts of foam rolling. The Thomas test only uses the weight of the leg and force of gravity to determine any length changes. On the contrary, the present kneeling lunge test was dependent on the biceps femoris (BF) strength for active ROM and on the researcher's force for passive ROM assessment. Therefore, the different outcomes might be due to different ROM assessment. Vigotsky et al. (37) considered neither rolling intensity nor pain. Both variables were taken into consideration for explaining varying ROM increases caused by neuromuscular rolling theoretically in previous studies (8,15,24,25,34) and practically (direct measurement) in this study.

Overall, differing levels of pain associated with RMlow (3.9/10), RMmod (6.2/10), and RMhigh (8.2/10) had similar effects on knee flexion ROM. According to these findings, individuals can roll without substantial discomfort or pain but still achieve significant ROM increases. However, small overall correlations revealed that increasing pain and active knee flexion ROM could be related (0.29 < r < 0.321). Conversely, RMmod-dependent active ROM changes and pain did not significantly correlate. Differing correlations between pain and ROM increases might be due to a very individual response to both the perception of pain and neuromuscular rolling, which was also elucidated previously (37). Small effect sizes between each INTpost result of active ROM, a small effect size between RMlow and RMmod, and a moderate (d = 0.55) effect size between INTpost passive ROM outcomes of RMlow and RMhigh indicate that there are practical relevant differences that need to be further investigated.

Even though this study was the first to focus on the impact of varied force application, few studies mentioned possible force-related mechanisms. Curran et al. (15) compared a multilevel rigid roller with a biofoam roller and strongly encouraged further research in pressure-related neuromuscular rolling mechanisms. The authors hypothesized that higher forces and less cutaneous contact time and consequent higher pressure might be beneficial for facilitating movement. Their theory was not supported in this study. Bradbury-Squires et al. (8) suggested that rolling force and duration could possibly amend viscoelasticity and thixotropic properties of fascia (32). However, this theory has been rejected because forces outside human physiological range would be required to induce mechanical deformation in firm tissues, including fascia (11). If fascial ground substance was altered to a more gel-like constitution, it would more likely be a long-term effect (3,36). Noticeably, Bradbury-Squires et al. (8) found 7% M. VL and 8% M. BF activation, relative to the maximum activation during MVIC, at 20 and 60 seconds during one 60-second RM bout, regardless of the increasing pain. They indicated that the cocontraction while rolling would protect the musculature from RM forces. Secondly, heat might be generated, which would result in reduced muscle and connective tissue viscoelasticity, further leading to greater ROM. An observation in this study was that participants began to sweat during RM application, irrespective of the force. Muscular cocontractions cannot be excluded as a possible mechanism contributing to ROM increases.

Another possible neuromuscular mechanism involved might be related to a greater stretch tolerance, as the extrafusal and intrafusal (muscle spindle) muscle length alters when muscles contract, similarly to the contract-relax proprioceptive neuromuscular facilitation (19). An increased pain pressure threshold (decreased pain sensitivity) over tender spots in the plantar flexors after ipsilateral and contralateral heavy RM and massage treatments was reported (1), thus supporting neural mechanisms. Increased stretch tolerance (increased resilience against the pain or discomfort of stretching) might be attributed to the diffuse noxious inhibitory control mechanism (9). Pain transmission would be inhibited because of monoamine transmitters when nociceptive stimuli are evoked and ascend to the brain. The highly variable individual RPP indicate that further research with greater populations is needed to determine whether force-induced pain affects ROM outcomes of neuromuscular rolling.

That varied forces did not cause impairments in muscle strength or jump performance parameters is in alignment with previous research (17,25,34). These findings put emphasis on different working mechanisms from traditional static stretching. Possible static stretching mechanisms include a reduction in active or passive stiffness in musculotendinous unit or a reduced crossbridge overlap (6). Considering that static stretching has been reported to lead to decrements in performance (6), RM, even when applied with high forces, may be an alternative treatment to increase ROM in an athletic field that involves maximum performances.

Finally, this study showed significantly different results between the 2 pretests. These findings put emphasis on the impact of the initial pretest on subsequent measures. While Atha and Wheatley (2) reported mobilizing effects of repeated measures, Bergh and Ekblom (7) found that higher muscle temperature enhances muscle strength and power output. Bradbury-Squires et al. (8) suggested that VL and BF cocontractions might generate additional intramuscular heat; however, an impact of one 60-second bout of quadriceps foam rolling on muscle temperature has not been found in a different study (30). Because MVIC force in this study only substantially changed from the initial pretest to all subsequent measures, it is likely that the warm-up resulted in more muscle hyperthermia than possible RM-induced cocontractions. Therefore, multiple pretests or warm-ups involving MVICs, and ROM measures should be performed to prevent testing effects.

The most important limitation is that this study did incorporate a low intensity RM (RMlow) rather than a control group. The RMlow condition provided a similar environmental condition with negligible force application. However, while the RMlow condition could have activated cutaneous receptors, the lack of preintervention to postintervention force-dependent RM changes suggests that the RMlow condition was a suitable control replacement. Another limitation of this study may concern whether an individual's upper body strength might not produce sufficiently high forces. However, the present findings indicate that intense forces do not need to be reached to substantially increase ROM. Finally, participants did not have previous experience with neuromuscular rolling. The results of this study may not necessarily be extrapolated to a population that uses RM regularly.

In conclusion, the increase in ROM was not dependent on the intensity (pressure or load) of the RM. Fortunately for rolling enthusiasts, higher levels of pain or discomfort are not necessary to achieve an increased ROM. Furthermore, the intensity of rolling did not have differential effects on strength or power measures. Moreover, the second pretest (INTpre) was significantly different than the initial pretest for knee flexion passive ROM, knee flexion, and knee extension MVIC force, emphasizing the impact of initial pretests on subsequent measures.

Practical Applications

Previous literature suggested that the forces at which neuromuscular rolling is performed might have an impact on rolling-induced ROM increases (4,15). The present results suggest that the intensity of rolling forces (50–90% of maximum discomfort) do not differentially affect strength and jump performance, nor do they substantially amplify ROM. Pain with rolling is not necessary for increasing ROM. While previous studies have shown increased ROM with as little as 5–10 seconds of RM (34), the research tends to show higher ROM with longer durations and thus 2–3 sets of 30–60 seconds of rolling per muscle group (1,8,10,17,24,25) below a level of significant pain or discomfort would be suggested. These findings are of clinical relevance, as practitioners do not need to roll to the point of discomfort or pain to achieve improvements in flexibility.


The authors thank Dr. Thamir Alkani for his technical assistance. The results of this study do not constitute endorsement of the product by the authors or the National Strength and Conditioning Association. There are no conflicts of interests for the authors of this paper.


1. Aboodarda SJ, Spence AJ, Button DC. Pain pressure threshold of a muscle tender spot increases following local and non-local rolling massage. BMC Musculoskelet Disord 16: 265, 2015.
2. Atha J, Wheatley DW. The mobilising effects of repeated measurement on hip flexion. Br J Sports Med 10: 22–25, 1976.
3. Barnes MF. The basic science of myofascial release: Morphologic change in connective tissue 231–238. J Bodyw Mov Ther 1: 231–238, 1997.
4. Beardsley C, Škarabot J. Effects of self-myofascial release: A systematic review. J Bodyw Mov Ther 19: 747–758, 2015.
5. Behara B, Jacobson BH. Acute effects of deep tissue foam rolling and dynamic stretching on muscular strength, power, and flexibility in Division I linemen. J Strength Cond Res 31: 888–892, 2017.
6. Behm DG, Chaouachi A. A review of the acute effects of static and dynamic stretching on performance. Eur J Appl Physiol 111: 2633–2651, 2011.
7. Bergh U, Ekblom B. Influence of muscle temperature on maximal muscle strength and power output in human skeletal muscles. Acta Physiol Scand 107: 33–37, 1979.
8. Bradbury-Squires DJ, Noftall JC, Sullivan KM, Behm DG, Power KE, Button DC. Roller-massager application to the quadriceps and knee-joint range of motion and neuromuscular efficiency during a lunge. J Athl Train 49, 2014.
9. Broucker TH, Cesaro P, Willer JC, Le Bars D. Diffuse noxious inhibitory controls in man. Brain 113: 1223–1234, 1990.
10. Cavanaugh MT, Doweling A, Young JD, Quigley PJ, Hodgson DD, Whitten JHD, Reid JC, Aboodarda SJ, Behm DG. An acute session of roller massage prolongs voluntary torque development and diminishes evoked pain. Eur J Appl Physiol 117: 109–117, 2017.
11. Chaudhry H, Schleip R, Ji Z, Bukiet B, Maney M, Findley T. Three-dimensional mathematical model for deformation of human fasciae in manual therapy. J Am Osteopath Assoc 108: 379–390, 2008.
12. Cohen J. Statistical Power Analysis for the Behavioral Sciences. (2nd ed.). Hillsdale, NJ: Erlbaum, 1988. pp. 8–17.
13. Coren S. The lateral preference inventory for measurement of handedness, footedness, eyedness, and earedness: Norms for young adults. Bull Psychon Soc 31: 1–3, 1993.
14. Couture G, Karlik D, Glass SC, Hatzel BM. The effect of foam rolling duration on hamstring range of motion. Open Orthop J 9: 450–455, 2015.
15. Curran PF, Fiore RD, Crisco JJ. A comparison of the pressure exerted on soft tissue by 2 myofascial rollers. J Sport Rehabil 17: 432–442, 2008.
16. Grieve R, Goodwin F, Alfaki M, Bourton A-J, Jeffries C, Scott H. The immediate effect of bilateral self myofascial release on the plantar surface of the feet on hamstring and lumbar spine flexibility: A pilot randomised controlled trial. J Bodyw Mov Ther 19: 544–552, 2015.
17. Halperin I, Aboodarda SJ, Button DC, Andersen LL, Behm DG. Roller massager improves range of motion of plantar flexor muscles without subsequent decreases in force parameters. Int J Sports Phys Ther 9: 92–102, 2014.
18. Healey KC, Hatfield DL, Blanpied P, Dorfman LR, Riebe D. The effects of myofascial release with foam rolling on performance. J Strength Cond Res 28: 61–68, 2014.
19. Hindle KB, Whitcomb TJ, Briggs WO, Hong J. Proprioceptive neuromuscular facilitation (PNF): Its mechanisms and effects on range of motion and muscular function. J Hum Kinet 31: 105–113, 2012.
20. Janes WCI, Snow BBG, Watkins CE, Noseworthy EAL, Reid JC, Behm DG. Effect of participants' static stretching knowledge or deception on the responses to prolonged stretching. Appl Physiol Nutr Metab 41: 1052–1056, 2016.
21. Jay K, Sundstrup E, Sondergaard SD, Behm D, Brandt M, Saervoll CA, Jakobsen MD, Andersen LL. Specific and cross over effects of massage for muscle soreness: Randomized controlled trial. Int J Sports Phys Ther 9: 82–91, 2014.
22. Junker DH, Stoggl TL. The foam roll as a tool to improve hamstring flexibility. J Strength Cond Res 29: 3480–3485, 2015.
23. Kelly S, Beardsley C. Specific and cross-over effects of foam rolling on ankle dorsiflexion range of motion. Int J Sports Phys Ther 11: 544–551, 2016.
24. MacDonald GZ, Button DC, Drinkwater EJ, Behm DG. Foam rolling as a recovery tool after an intense bout of physical activity. Med Sci Sports Exerc 46: 131–142, 2014.
25. MacDonald GZ, Penney MDH, Mullaley ME, Cuconato AL, Drake CDJ, Behm DG, Button DC. An acute bout of self-myofascial release increases range of motion without a subsequent decrease in muscle activation or force. J Strength Cond Res 27: 812–821, 2013.
26. Mikesky AE, Bahamonde RE, Stanton KA, Fitton T. Acute effects of the stick on strength, power, and flexibility. J Strength Cond Res 16: 446–450, 2002.
27. Mohr AR, Long BC, Goad CL. Effect of foam rolling and static stretching on passive hip-flexion range of motion. J Sport Rehabil 23: 296–299, 2014.
28. Monteiro ER, Cavanaugh MT, Frost DM, Novaes JDS. Is self-massage an effective joint range-of-motion strategy? A pilot study. J Bodyw Mov Ther 21: 223–226, 2017.
29. Monteiro ER, Škarabot J, Vigotsky AD, Brown AF, Gomes TM, da Silva Novaes J. Acute effects of different self-massage volumes on the fms overhead deep squat performance. Int J Sports Phys Ther 12: 1–11, 2017.
30. Murray AM, Jones TW, Horobeanu C, Turner AP, Sproule J. Sixty Seconds of foam rolling does not affect functional flexibility or change muscle temperature in adolescent athletes. Int J Sports Phys Ther 11: 765–776, 2016.
31. Paolini J. Review of myofascial release as an effective massage therapy technique. Athl Ther Today 14: 30–34, 2009.
32. Schleip R. Fascial plasticity—a new neurobiological explanation: Part 1. J Bodyw Mov Ther 7: 11–19, 2003.
33. Škarabot J, Beardsley C, Štirn I. Comparing the effects of self-myofascial release with static stretching on ankle range-of-motion in adolescent athletes. Int J Sports Phys Ther 10: 203–212, 2015.
34. Sullivan KM, Silvey DB, Button DC, Behm DG. Roller-massager application to the hamstrings increases sit-and-reach range of motion within five to ten seconds without performance impairments. Int J Sports Phys Ther 8: 228–236, 2013.
35. Taube W, Leukel C, Lauber B, Gollhofer A. The drop height determines neuromuscular adaptations and changes in jump performance in stretch-shortening cycle training. Scand J Med Sci Sports 22: 671–683, 2012.
36. Twomey L, Taylor J. Flexion creep deformation and hysteresis in the lumbar vertebral column. Spine J 7: 116–122, 1982.
37. Vigotsky AD, Lehman GJ, Contreras B, Beardsley C, Chung B, Feser EH. Acute effects of anterior thigh foam rolling on hip angle, knee angle, and rectus femoris length in the modified Thomas test. PeerJ 3: e1281, 2015.

self-massage therapy; neuromuscular rolling; pressure; self-myofascial release

© 2017 National Strength and Conditioning Association