Interventions Based on Direct Muscle Electrical Stimulation

A variety of electrical stimulation methods have been employed to reduce spasticity including direct muscle stimulation, sometimes also termed patterned electrical stimulation or patterned neuromuscular stimulation, FES and TENS. In the present section, we will examine the effect of interventions based on direct muscle stimulation (or stimulation of the motor nerve over the muscle belly) (i.e. patterned electrical stimulation or patterned neuromuscular stimulation). The objective of direct muscle stimulation is to produce a muscle contraction and related therapies are focused on the beneficial effects of series of muscle contractions. Often this stimulation is cyclical in nature (patterned) so as to simulate natural physiologic conditions such as might be seen in walking or cycling. With FES, the stimulation parameters are set to produce a coordinated contraction of several muscles with the intent of producing purposeful movement. This approach is often used to assist or simulate active exercise paradigms and therefore, the articles addressing FES have been summarized in the previous section on active movement-based approaches. TENS, on the other hand, is focused on stimulating large, low threshold afferent nerves to alter motor-neuron excitability and thereby reduce spasticity. Stimulation intensities are maintained sub threshold for eliciting muscle contraction when stimulating mixed motor and sensory nerves so that only lower threshold sensory nerves are selectively stimulated. For this reason, articles concerning TENS will be included in the next section that is directed towards interventions based on afferent stimulation.

Author Year

Research Design

Total Sample Size

Methods Outcome
Sivaramakrishnan et al. 2018


RCT Crossover



Population: SCI (n=10): Etiology: Trauma=4, Infective=2, Degenerative=2, Tumor=2; Mean age=39±12.98yr; Gender: males=9, females=1; Level of injury: C=6, T=4; Mean time since injury=8.8±9.09mo; AIS scale: A=2, B=0, C=6, D=1, E=1.

Intervention: Participants were randomly allocated to receive either transcutaneous electrical nerve stimulation (TENS) or functional electrical stimulation (FES) first. Interventions were separated by a 24-hr washout period. TENS was administered with biphasic square wave impulses at 100 Hz and a pulse duration of 300 microseconds for 30 min. Stimulation intensity was 15 mA. FES was administered with biphasic rectangular pulses at a 35 Hz pulse rate and 300 microsecond pulse width for 30 min. Stimulation intensity was increased until a visible muscle contraction (motor threshold) was elicited. Outcome measures were assessed at baseline, post intervention, and at 1, 4, and 24-hr follow ups.     

Outcome Measures: Modified Ashworth Scale (MAS) scores for hip adductors, knee extensors, and plantar flexors; Spinal Cord Assessment Tool for Spastic Reflexes (SCATS).

1.     There were no significant between-group differences in MAS and SCAT scores for hip adductors, knee extensors and plantar plexors post intervention, as well as at the 1, 4, and 24-hr follow ups (p>0.05 for all).

2.     TENS and FES showed statistically significant within-group differences for mAS scores of hip adductors and knee extensors. mAS scores were significantly lower post intervention, and at 1 and 4-hr follow ups (p<0.01). However, there were no statistical differences in the 24-hr follow and baseline values.

3.     There were no significant within-group differences in MAS scores for plantar flexors in both FES and TENS groups (p>0.05).

4.     TENS SCAT scores were significantly lower post intervention and at the 1-hr follow up when compared to baseline (p<0.05). There was no statistically significant difference at the 4 and 24-hr follow ups when compared to baseline.

5.     FES SCAT scores were significantly lower post intervention and at the 1 and 4-hr follow ups when compared to baseline (p<0.01). However, there was no statistically significant difference between the 24-hr follow up and baseline.

Gomez-Soriano et al. 2018




Population: SCI with spasticity (SCIS, n=14): Mean age=39.3±15.2yr; Gender: males=10, females=4; Level of injury: C=10, T=3, L=1; Mean time since injury=3.8±2.0mo; AIS scale: C=8, D=6.

SCI without spasticity (SCI, n=14): Mean age=43.5±15.6yr; Gender: males=10, females=4; Level of injury: C=7, T=5, L=2; Mean time since injury=6.5±3.8mo; AIS scale: C=7, D=7.

Non-injured (NI, n=11)

Intervention: Plantar tibialis anterior (Pl-TA) cutaneous reflex (CR) and soleus (SOL) H-reflex were evoked with a controlled current simulator delivering 5 rectangular pulses (with 1 millisecond duration, with a 5-millisecond interval) at 1.2X the required stimulus to induce Pl-TA reflex during a controlled plantarflexion protocol. All participants participated in two randomized session including vibratory and transcutaneous nerve stimulation (TENS). 80 Hz vibratory conditioning was applied to the plantar surface of the foot. TENS consisted of a biphasic asymmetrical current of 100 Hz with a 100-microsecond pulse width using an anode and cathode placed on the dorsum and plantar surfaces of the foot, respectively. Both conditions were applied between the second and third metatarsal joints of the plantar foot for 30 seconds prior to reflex measurement and maintained throughout the testing protocol. Outcome measures were assessed at baseline and during the respective intervention (at rest, and duration ramp and hold phases of the plantarflexion protocol).

Outcome Measures: Pl-TA CR; SOL H-Reflex.

1.     There was a significant decrease in Pl-TA CR during the rest phase and hold phase during the vibrational intervention (p<0.001; p<0.05, respectively). However, there were no significant changes in SOL H reflex for all phases and Pl-TA CR during the ramp phase (p>0.05).

2.     During TENS, Pl-TA CR showed a significant decrease during the hold phase (p<0.05). There were no significant differences for the remaining outcome measures (p>0.05).

van der Salm et al. 2006


Prospective Controlled Trial


Population: Age range: 21-42 yr; Gender: males=8, females=2; Level of injury: C3-T11; Level of severity: AIS A=9, AIS C=1; Time since injury range: 28-275 mo.

Intervention: Electrical motor (agonist or antagonist) or afferent (S1 dermatomal) stimulation of the triceps surae or a placebo (application of electrodes but no current). 1–45 min session of each type of stimulation. Intensity at 3x motor threshold for motor stimulation and 80% of motor threshold for afferent stimulation.

Outcome Measures: Modified Ashworth Scale (MAS), Clonus score, H reflex, and H/M ratios. Measurements were conducted just prior to, immediately after, 1 hr after and 2 hr after the intervention for each of the 4 conditions.

1.     Only the agonist muscle stimulation differed significantly (46% reduction) from the placebo as indicated by reduced MAS (p<0.001).

2.     No significant carry-over effect (over 2 hr) although there was a trend of continued reductions for the MAS (p=0.113).

3.     No significant intervention effect was shown for the clonus score or the H/M ratio.

4.     The reflex-initiating angle showed a significant change for antagonist stimulation (n=8, p=0.015) but the carryover effect was not significant.

Seib et al. 1994


Prospective Controlled Trial


Population: Age range: 19-73 yr; Gender: males=6, females=4; SCI (n=5): Time since injury range: 3-16 yr. TBI (n=5): Time since injury range: 1-6 yr.

Intervention: One, 20 min session of electrical stimulation over the tibialis anterior measured in ipsilateral (intervention) and contralateral (control) leg versus sham stimulation (control).

Outcome Measures: Viscous and elastic ankle stiffness measured by frequency-dependent torque resistance immediately post-stimulation (n=10/9), and 24 hr post-stimulation (n=9/8), and immediate post-stimulation subjective spasticity assessment (n=9).

1.     Spasticity was reduced in 9/10 participants (p<0.05) (p<0.05 for SCI subjects only) and this was sustained for 24 hr over all subjects (p<0.01) but not for SCI subjects only.

2.     Spasticity was not reduced immediately or after 24 hr in the contralateral leg nor with sham stimulation.

3.     Subjective spasticity assessment immediately post-stim only recorded notable reductions for SCI subjects for up to 6 hr post-simulation.

Carty et al, 2013




Population: Gender: males=11, females=3; Injury etiology: SCI=14 Level of severity: AIS A/B=14; Mean time since injury: 10.2 yr

Intervention: Investigate alterations in body composition variables and spasticity following subtetanic neuromuscular electrical stimulation (NMES) training.

Outcome Measures: Lean body mass (LBM), Other body composition variables, Spinal Cord Assessment Tool for Spastic Reflexes (SCATs), Visual analog scale (VAS), Verbal and written feedback for spasticity.

1.     Within subjects, statistically significant decrease in SCATs (p<0.001), showing reduction in measured spasticity.

2.     Within subjects, no significant difference in spasticity VAS scores at any of three test points.

3.     No significant differences between control, testing and pretraining measures.

4.     Significant differences between pretraining and posttraining for lower-limb LBM, total mass and regional body fat; tendency toward reduction in total body fat but no statistical significance.

Tancredo et al. 2013




Population: Mean age: 34.5 yr; Gender: males=11, females=0; Level of severity: AIS A=9, AIS B=1, AIS D=1.

Intervention: Individuals received Neuromuscular electrical stimulation (NMES) on the quadriceps muscles and fibular nerve for 20 min and 15 min respectively. Goniometer and accelerometer sensors for the pendulum test were placed on the right leg with measurements being repeated three times. Assessments were completed at baseline and post- intervention.

Outcome Measures: Modified Ashworth Scale (MAS), Pendulum test, Subjective spasticity scale.

1.     A decrease in spasticity from baseline to post- intervention according to the MAS scores was reported for all but three individuals who remained stable.

2.     The pendulum test revealed a decrease in spasticity with a larger variation between the maximum and minimum peaks of motion from baseline to post- intervention.

3.     When individuals were stratified based on those taking medication (baclofen and clonazepam) and those who were not, both groups still demonstrated improvements on the pendulum test.

4.     Of the nine individuals who completed the subjective spasticity scale, eight reported lower ratings of spasticity after receiving NMES whilst one individual reported stable ratings.

Robinson et al. 1988a




Population: Age range: 21-62 yr; Level of injury: paraplegia=6, tetraplegia=6; Level of severity: complete=6, incomplete=6.

Intervention: 1, 20 min session of electrical stimulation of quadriceps with leg maintained at 60° flexion (isometric exercise).

Outcome Measures: Normalized relaxation index obtained during Pendulum test (R2n) collected prior and immediately after stimulation.

1.     Decrease in spasticity was noted with pendulum test (average R2n increased in most cases) (p<0.005).

2.     The greatest reduction in spasticity after stimulation was noted for individuals who were the most spastic before stimulation.

3.     No carry over effect of stimulation in spasticity measured 24 hr later.

Robinson et al. 1988b





4wk tx N=21,

8wk tx N=8

Population: Level of injury: paraplegia=15, tetraplegia=16; Level of severty: complete, incomplete; Time since injury: <1yr=15, >1yr=16.

Intervention: 4-8 wk quad reconditioning program of twice daily 20 min stimulation sessions of quadriceps, which were at least 4 hr apart, 6x/wk. Stimulus currents were set at 120-160 mA.

Outcome Measures: R2n-index of spasticity (by means of pendulum drop test). Peak isometric quad torque in response to surface electrical stimulation. Biweekly tests of spasticity and torque at baseline, 4 and 8 wk (1-3 evaluations performed each time).

1.     Most participants had increased spasticity after 4 wk of reconditioning but not after 8 wk.

2.     Twice as many legs had R2n decreases (spasticity increases) as had increases or no change.

3.     Baseline versus 4 wk left and right R2n yielded a significant within subject difference between baseline and 4 wk values (p=0.008).


In an RCT (n=10), Sivaramakrishnan et al. (2018) compared the anti-spasticity effects of a single session of FES or TENS electrical stimulation. Changes in spasticity of the hip abductors, knee extensors and ankle plantar flexors were measured using the mAS and SCATS. Upon analysis of mAS scores, TENS and FES significantly reduced spasticity up to 4 hours in the hip abductors and knee extensors (p<0.01). Similarly, SCATS scores showed significant reductions in spasticity 1 hour following TENS (p=0.01) and 4 hours following FES (p=0.01). However, no significant differences were observed between groups at 24 hours. This suggests that both TENS and FES exert similar anti-spasticity effects, which may be effective as temporary therapeutic adjuncts.

Carty et al. (2013) conducted a pre-post study (n=14) to investigate alterations in body composition variables and spasticity following subtetanic neuromuscular electrical stimulation. Subtetanic contractions were elicited bilaterally in the proximal and distal quadriceps and hamstrings muscle groups using a hand-held neuromuscular electrical stimulation device. Measurements were taken before (2x) and after an 8-week neuromuscular electrical stimulation training program for hamstrings and quadriceps. There was a statistically significant within-subject decrease in SCATS (p<0.001), showing reduction in measured spasticity. There was no significant difference in spasticity VAS scores at any of three test points.

Tancredo et al. (2013) conducted a pre-post study (n=11) of the effect of neuromuscular electrical stimulation on spasticity as assessed bythe pendulum test and MAS at baseline and post-treatment. Neuromuscular electrical stimulation was applied within a single session to the quadriceps muscles and fibular nerve for 20 minutes and 15 minutes, respectively. The study revealed a decrease in MAS scores from baseline to post-treatment in eight subjects while spasticity levels remained stable in three subjects. The pendulum test revealed an overall decrease in spasticity with a larger variation between the maximum and minimum peaks of motion from baseline to post-treatment. Even when subjects were analyzed based on those who did or did not take medication (baclofen and clonazepam), both sub-groups still demonstrated improvements on the pendulum test. Of the nine study subjects who completed the subjective spasticity scale, eight reported lower ratings of spasticity after receiving neuromuscular electrical stimulation and one reported stable rating.

Van der Salm et al. (2006), Seib et al. (1994) and Robinson et al. (1988a) tested the effects of a single session of muscle stimulation on spasticity. Each employed slightly different stimulation parameters and a variety of outcome measures. Of note, Van der Salm et al. (2006) and Seib et al. (1994) each employed prospective controlled trials of electrical stimulation and demonstrated immediate effects of reduced spasticity although these effects waned until mostly absent by the next day. In particular, van der Salm et al. (2006) examined three different stimulation methodologies versus a placebo condition and assessed ankle plantar flexor spasticity with the MAS, a clonus score and via EMG responses (i.e., H-reflex and H/M ratio). The various stimulation methods consisted of stimulation over the triceps surae (agonist), the tibialis anterior (antagonist) and the S1 dermatome versus a control placebo condition of electrode application but no current generation. Presumably, subjects were not aware of this because subjects had no sensation in the stimulated areas. Significant spasticity reductions were only obtained with agonist muscle stimulation for the MAS (p<0.001) and not the clonus or EMG responses. This was not sustained for 2 hours post-stimulation although there was still a trend for reduced MAS scores at this time (p=0.113). Spasticity was also reduced (but not statistically significantly) with antagonist muscle stimulation but not for dermatomal or sham (placebo) stimulation.

Interestingly, van der Salm et al. (2006) noted that if they had examined their data by employing t-tests to test for pre-post effects (i.e. univariate analysis) within a specific stimulation method, they also would have demonstrated a reduction in spasticity for antagonist muscle stimulation, thereby illustrating the potential of obtaining false positives in uncontrolled or poorly controlled studies. Robinson et al. (1988a) conducted a pre-post study design without control conditions and Seib et al. (1994) conducted a prospective controlled trial but then inappropriately employed univariate analysis. Regardless, the results of these studies corroborate the finding of an immediate post-stimulation effect (van der Salm et al. 2006). Seib et al. (1994) and Robinson et al. (1988a) employed stimulation of different muscles (tibialis anterior, i.e., ankle dorsiflexion and quadriceps, i.e., knee extension respectively) and each demonstrated short lasting reductions in spasticity. Similar to the findings of van der Salm et al. (2006), Seib et al. (1994) reported that the effect of reduced spasticity waned quickly but was still evident up to 6 hours post-stimulation (mean 4.4 hours) as indicated by subject self-report.

In the only study of the long-term effects of stimulation, Robinson et al. (1988b) employed a similar stimulation protocol for the quadriceps as noted above over a period of four–eight weeks with twice daily 20-minute sessions at least four hours apart, 6 days per week. Although 31 individuals initiated the study and 21 complete 4 weeks of the stimulation program, the study had severe subject retention issues with only eight individuals continuing participation for the intended eight weeks. Study results showed that most subjects actually had increased spasticity at four weeks but for the subjects who remained in the study for eight weeks there was no significant difference. This null result begs further study of the long-term effects of muscle stimulation given the beneficial results obtained with short-term stimulation and in reports involving individuals with other etiologies (Chen et al. 2005; Ozer et al. 2006).

The other aspect of these studies worth noting is the variability in outcome measures. Within these papers there were measures that were clinical, neurophysiological, biomechanical, and subject self-report in nature. Researchers have noted that spasticity is a multi-faceted construct with individual components of spasticity weakly related to each other suggesting that while different tools may measure unique aspects of spasticity the overall construct is best measured with an appropriate battery of tests (Priebe et al. 1996).


There is level 1b evidence (from one RCT: Sivaramakrishnan et al., 2018) that a single session of electrical stimulation with FES or TENS exerts similar anti-spasticity effects, suggesting both TENS and FES may be used as therapeutic adjuncts.

There is level 1b evidence (from one RCT: Gomez-Soriano et al., 2018) that TENS and vibration therapy may reduce plantar flexion spasticity through inhibition of the plantar tibialis anterior cutaneous reflex, rather than the soleus H reflex.

There is level 2 evidence (from two prospective controlled trials and one pre-post study: Van der Salm et al. 2006; Seih et al. 1994; Robinson et al. 1988a) that a single treatment of surface muscle stimulation reduces local muscle spasticity with agonist stimulation more effective than stimulation to the antagonist.

There is conflicting evidence for how long the effects of a single treatment of electrical stimulation on muscle spasticity persist, although they appear to be relatively short lasting (i.e., £ 6 hours).

There is level 4 evidence (from one pre-post study: Robinson et al. 1988b) that a long-term program of muscle stimulation does not reduce muscle spasticity and may even increase local muscle spasticity.

There is conflicting level 4 evidence (from two pre-post studies: Cart et al. 2013; Tancredo et al. 2013) that use of neuromuscular electrical stimulation decreases spasticity.