Research Report

Effects of Massage on Alpha Motoneuron Excitability

The purpose of this study was to investigate the specijicity of the effects of massage @etrisscage)o n spinal motonarron excitability as measured by changes in the peak-to,peak amplitude of H-reflex recordings. H-repexes (and M-responses) were recordedfrom the distal aspects of the @ht triceps surae muscle of 8 men and 8 women, aged 2 0 to 3 7 years, with no neummuscular impairments of the lower extremities. The H-reflexes were recorded during Jive control and four e x p a a m tal coruiitions (20 trials at each condition). The control conditions (CI-C5)preceded and followed each experimental condition, prouidiqg a measure of the stability of the H-reflex. Each experimental condition comsted of a 4-minute period of mussage of the ipsilateral and contralateral triceps surae and hamstring muscle groups (ITS, CTS, IHS and CHS, respectively). The mean peak-to-peak amplitude of the H-reflex wasfound to be stable (range=I.91-1.95 mV) across theJive control conditions. H-reflex amplitudes recorded during the experimental conditions indicate that massage of the ITS resulted i n a reduction of the H-repex (0.83 mV) in comparison with the pretest c o n m l condition (CI) and the remaining experimental conditions (range=I . 77-2.23 mV). This d t j i m c e was signijicant, and subsequent Newman-Keuls tests indicated a specijicity of the effects of massage on the muscle group being massaged. /Sullivan SJ, WilliamsIRT, Seaborne DE, Morelli M. Effects of massage o n alpha motoneumn excitability. Pbys Thm 1991;71:555-560.1 Key Words: H-reflex, Massage, Motoneumn excitability, Triceps surae muscle.

S John Sulllvan Leslle RT Willlams Derek E Seaborne Moreno Morelll

Massage is one of the oldest therapeutic modalities available to physical therapists.1~2 Although therapists frequently report empirical changes in the refltx activity of muscles resulting from m:assage,there is a lack of scien-

tifically documented evidence to support the use of massage as a thempeutic intervention in the alteration of the reflex activity of muscles. Some authorities3.4 claim that petrissage, o r massage in which the muscles are

kneaded, can exert an inhibitory effect on the motoneurons of the muscle being massaged. There is obviously a need to substantiate such claims. Two recent studies5g6 conducted in our laboratory have investigated the question of whether local massage can bring about a change in the level of spinal motoneuron activity in neurologically nonimpaired subjects. In these studies, we investigated the effects of petrissagez on spinal motoneuron activity by measuring changes in H-reflex amplitude, an indirect measure of motoneuron excitability.' Short periods (3 and 6 minutes) of massage were applied to the triceps surae muscle. H-reflex recordings were obtained from the same muscle before, during, and following

SJ Sullivan, PhD, is Associate Professor, Department of Exercise Science, Concordia University, 7141 Sherbrooke St W, Montreal, Quebec, Canada H4B 1R6, and is affiliated with the Centre de Recherche, Institut de Readaptation de Montreal, MontrCal, Quebec, Canada H3S 2J4, and L'Ecole de Readaptation, Facult6 de Medecine, Universite de Montreal, Montreal, Quebec, Canada H3C 3J7. Address all correspondence to Dr Sullivan at the first address.
LRT Williams, PhD, is Director, Motor Control and Learning Laboratory, University of Otago, Dunedin, New Zealand.

DE Seaborne, MSc, is Associate Professor, Division of Physiotherapy, L'Ecole de Readaptation, FacultC de Medecine, Universite de Montreal. M Morelli, MSc, is enrolled in the Doctor of Medicine program, Faculte de Medecine, Universite de Montreal.
This study was approved by the Human Ethics Committee of Concordia University

This article was submitted March 20, 1990, and was accepted March 27, 1991

Physicall TherapyNolume 71, Number 8/August 1991

555 / 11

massage. The H-reflex amplitude was found to be markedly reduced during the period of massage in comparison with that obtained before or following massage. This finding was interpreted as indicating a reduction in (inhibition of) motoneuron activity. Massage may be compared with other therapeutic techniques such as tendon pressure8~9 muscle tapping,lo and which have recently been investigated as to their effects on alpha motoneuron excitability. These two techniques, like massage, result in a reduction in H-reflex amplitude, indicating reduced motoneuron excitability. The studies of Kukulka and c0lleagues8~9 have documented changes specific only to the muscle group under investigation. Bdanger and colleagues10 also reported a reduction in H-reflex amplitude recorded from the triceps surae muscle, resulting from the application of muscle tapping to other sites on the ipsilateral limb, in neurologically nonimpaired subjects. This generality of response led to the speculation that massage may also result in a generalized neurophysiological response, the effects of which may extend beyond the muscle being massaged. Such a finding would provide both practical clinical information and further insights into the neurophysiological mechanisms subserving the previously reported inhibitory effects of massage. The purpose of this investigation was to document the effects of massage, applied to two ipsilateral and two contralateral sites of the lower limbs, on motoneuron excitability of the right triceps surae muscle in neurologically nonimpaired subjects. We expected that the inhibitory effects (a decreased H-reflex amplitude) of massage would extend beyond the

muscle being massaged. A secondary aspect of the study was to confirm our previous finding6 that both female and male subjects responded in a similar manner to the massage, which was applied by a male therapist.

pass filtered (3 dB down at 20 and 2,000 Hz) and sampled by an IBM P C - A P compute6 at 5,000 Hz. Following the location of the ideal stimulating site (ie, one that resulted in an H-reflex devoid of artifacts) by a hand-held probe, the cathode (10-mm diameter) was positioned over the posterior tibial nerve in the region of the popliteal fossa. The anode (20 cm2) was placed across the thigh at the superior border of the patella. Individual square-wave pulses (1-millisecond duration) were delivered from a Grass S88 stimulators and related stimulus isolation (Grass ~ 1 ~ 5 and constant current (Grass ') CCUI~) units at 10-second intervals. The stimulator was controlled by a software management package specifically designed for H-reflex data acquisition and processing.13 In this investigation, the peak-to-peak amplitudes (in millivolts) of the H-reflex and M-response were recorded and analyzed. The peak-to-peak amplitude is easily quantifiable and a widely used characteristic6.8J4 in describing the H-reflex.

Subjects
Sixteen subjects (8 men, 8 women), recruited from a population of undergraduate students and university staff, volunteered to participate in the study. The subjects, who ranged in age from 20 to 37 years @=22.3, SD=3.8), had no known neurologic or orthopedic impairments. Each subject signed an institutionally approved informed consent form before participating in the study. Subjects were asked to refrain from ingesting substances containing caffeine or alcohol on the day of the data collection. These substances have been demonstrated to alter the H-reflex amplitude.11

H-reflex Acquisition
Details of the H-reflex recording procedures have been reported in detail elsewhere12 and are summarized as follows. Following preparation of the recording site, two miniature (4-mm diameter) Beckman surface recording electrodes* were positioned (parallel to the direction of the muscle fibers) along the midline of the posterior surface of the triceps surae muscle, 1116th of the distance between the flare of the medial malleolus and the distal crease of the popliteal fossa of the right leg. The reference electrode was positioned over the external malleolus of the same leg. The electrical activity of the stimulated muscle (Hreflex and M-response) was recorded by a DISA differential amplifiert (common mode rejection ratio= >90 dB, input resistance=250 Ma) and band-

Massage Technique
The form of massage used in this study and in previous investigations596 consisted of a one-handed, rhythmicpressure grasping and releasing of the muscle tissue in the region of the muscle belly. This deep-massage technique, also referred to as petrissage,l was administered by the same person (q,experienced in the apwho is plication of this technique in both clinical and research situations. The pressure applied during the massage was not experimentally controlled, but we believe it corresponded to that used in clinical practice and in previous studies. The rate of application was standardized at 0.5 Hz, based on the rhythm developed during the previous studies and verified during this experiment. Talcum powder was applied to the massage site before beginning the massage. The procedure used closely followed the standard procedure for a clinical massage. The massage was applied to the triceps surae and hamstring muscles of each leg.

'Beckman Instruments Inc. 110 Technology Park, Norcross, GA 30071. +Dantech, 100 Dynamic Dr, #103, Scarborough, Ontario, Canada N1V 5C4. *~nternational Business Machines Corp, PO Box 13283, Boca Raton, FL 33432. 'Grass Instrument Co, 101 Old Colony Ave, Quincy, MA 02169.

Physical TherapyNolume 71, Number 8/August 1991

Table 1. Sumnary of Main Effects for Three-Way Analysis of Variance of M-response Amplitudesfor Pretest Control (CI) and Experimental Conditiom
Source
Between subjects Gender Error Within subjects Conditions Error Trials Error 4 56 0.39 0.88 0.01 0.02 0.62 ,6524 1 14 2.38 75.44 2.38 5.39 0.44 ,5176 df

Procedure

SS

MS

Following a 5-minute period of lowlevel stirnulation, which allowed the subject to accommodate to the electrical stimulus, a brief H-reflex recruitment pmfile was generated in order to establish the voltage necessary to generate a c:ontrol H-reflex of approximately 50% of the maximum H-reflex. This control signal amplitude, which allowed for the monitoring of both facilitatory and inhibitory effects, was retained for the duration of the investigation. In most subjects, this stimulation voltage also resulted in the recording of a small M-response. We believe that, if the amplitude of the M-response remains stable throughout the experimental (and control) conditions, then it can be assumed that no changes in the stimulation conditions have occurred. Thus, any changes in H-reflex amplitude can be interpreted with confidence. Twenty trials were recorded during each of five control conditions (ClC5) and four experimental conditions. The subjects were instructed to remain still and relaxed throughout the testing procedure. The experimental conditions corresponded to 4-minute periods of massage applied to (1) the ipsilateral (right) triceps surae muscle (ITS), (2) the ipsilateral hamstring muscles8(IHS), (3) the contralateral

triceps surae muscle (CTS), and (4) the contralateral hamstring muscles (CHS). Each experimental condition was preceded and followed by a control condition. For example, the order of presentation for subject 1 was as follows: C1, CTS, C2, IHS, C3, CHS, C4, ITS, and C5. The first H-reflex of the massage condition was delivered immediately following the beginning of the massage. The order of administration of the four experimental conditions was controlled so that each subject received a unique order of presentation and so that there were no systematic trends in the presentation of conditions. We believe this procedure minimized any carry-over effects between conditions. During the 45-minute testing session, the subject lay prone with the head turned to the right to control for possible influences resulting from the tonic neck reflex.15 The lower legs were supported in a specially fabricated mold that ensured a constant angle of both the knee (17" of flexion) and the ankle (15" of plantar flexion). Ambient noise and light were reduced in order to minimize competing stimuli and provide a relaxing atmosphere.

female subjects to the massage, (2) to establish whether there was stability in the five control conditions over time, and (3) to determine whether there were differences between the pretest control condition (Cl, or the control condition that preceded the first experimental condition) and the four experimental conditions. The basic experimental design for the analysisof-variance (ANOVA) procedure can be expressed as a three-factor (gender x conditionsx trials) experimental design with repeated measures on the latter two factors. Gender was included as a between-subjects factor because of the possibility of a differential effect of male and female subjects receiving massage from a male investigator. Trials were included in order to investigate whether there were changes in the pattern of response within the various conditions over the 20 trials. A Newman-Keuls post hoc procedure was used to detect significant differences in the results of the ANOVA for main effects.An a priori 5% level of statistical significance was adopted for all statistical tests. In addition, the ensemble average of selected conditions was determined for illustrative purposes. The ensemble average is the mathematical average of the sampled data points recorded for the 20 trials in each condition. Essentially, it represents the mean recorded waveform.

_*

Results

Data Analysls
Statistical analyses were performed (1) to determine whether there was a difference in the response of male and

The mean recorded values for the M-response ranged from 0.25 to 0.26 mV across all conditions. As expected, the ANOVA did not reveal any significant differences among the five control conditions for gender (F=0.44; df= 1,14;P>.05), conditions (F=0.62; df=4,56; P>.05), or trials (F=0.55; df= 19,266;P > .05). None of the related interactions approached signficance. A similar result was established when the four experimental conditions were combined with the pretest control condition (C 1). The corresponding F ratios for the three main effects were all less than 1.0 (P>.05).

Physical TherapyNolume 71, Number 8/August 1991

Figure 1. Mean H-reflex amplitudes and standard errors for the four experimental conditions (ie, massage of ipsilateral and contralateral triceps surae and hamstring muscles iffS, CTS, IHS, and CHS, respectively])and the pretest control condition (CI). H-reflexes were obtainedwm the f l S (right lower extremity).
These data indicate that there were no changes in the recording conditions during the testing session. Table 1 contains the corresponding ANOVA summary data. P>.05), conditions (F=0.14; df=4,16; P>.05), or trials (F= 1.81;df= 19,266; P>.05) or for the related interactions, thus indicating the stability of the control conditions over time. These findings also justify the selection of C1 (as being representative of all control conditions) for inclusion in the subsequent analysis of the experimental conditions. Of primary importance was the analysis of the four experimental condi-

tions and the pretest control condition. A summary of the obtained amplitudes of H-reflexes recorded from the ITS during massage is presented in Figure 1. The ANOVA revealed a significant difference for the conditions factor (Tab. 2). Subsequent post hoc procedures revealed that the amplitude of the H-reflex recorded during massage of the ITS (0.83 mV) was significantly reduced in comparison with those in all other conditions (C1= 1.94, IHS=1.77, CTS=2.23, and CHS=2.03). This response was observed for all subjects. In addition, the amplitude of the H-reflex obtained during massage of the CTS was elevated in comparison with that obtained during massage of the IHS. The ensemble averages (20 trials) for the peak-to-peak H-reflexes for each of the four expenmend conditions and the pretest control condition for two subjects are presented in Figure 2.

Mean peak-to-peak H-reflexes recorded during the five control conditions ranged from 1.91 to 1.95 mV. No statistically significant differences were located for gender (F=0.46; df=1,14;

Table 2 Sumrnaty of Main Efects for Three-WayAnalysis of Variance of H-rejlex . Amplitudes for Pretest Control (CI) and Bperimental Conditions
Source

Between subjects Gender Error Within subjects Conditions Error Trials Error 4 56 384.27 209.96 96.07 3.75 25.62 .0001 1 14 51.45 2245.85 51.45 160.42 0.32 ,5800

The principal finding of this study was that there was a reduction in the amplitude of the H-reflex recorded from the ITS while it was being manually massaged. This finding confirms those of our earlier reports.5j6 The reduction in H-reflex amplitude during massage was clearly observed in all subjects evaluated. The recorded decrease was observed in relation to the control condition and the other three massage sites (IHS, CHS, and CTS), thus confirming the specificity of the effects of massage. This reduction in H-reflex amplitude may be interpreted as a reduction in alpha motoneuron excitability of the triceps surae muscle motoneuron pool. The amplitude of the H-reflex provides an indirect measure of spinal motoneuron excitability7 and therefore should be interpreted with caution. It must be noted that a reduction in H-reflex amplitude could also result from presynaptic inhibition6 being exerted on the afferent terminals. Despite this limitation, the use of H-reflex techniques provides an acceptable measure of motoneuron excitability.6*7J4 The finding that the H-reflex amplitude decreased during the period of

14 1558

Physical Therapyllrolume 71, Number 81August 1991

C1

54

C1

514

ITS

ITS

IHS

IHS

CTS CTS

undoubtedly activates the same receptors as does intermittent pressure. The role of muscle spindles and Golgi tendon organs in producing this observed inhibition awaits confirmation. The fact that inhibition was observed only during massage of the ITS confirms the specificity of the response and suggests that the stimulus and receptors activated during massage are specific to the muscle being massaged. A facilitation, albeit nonsignificant, was observed when the CTS was massaged. The fact that both male and female subjects responded in an identical manner to the massage further reinforces the notion that the observed effects are physiological and not psychological in nature.
Clinical Implications in Neurologically Nonlmpalred Persons

-L JCHS
CHS

... ..........

F ~ ~ U2. O T Ensemble average (20 triak)for two typical subjects (S4 and S14) obtained for the H-reflex recordings for the pretest contra1 condition (Cl) and the four acperimetttal conditiom (ITS, O S , IHS, and CHS /see Fig. 1 legend for definitions]).A 6@millisecondtime frame is presented. The ordinate value of 100% represents 5 mV Tbe peak-topeak amplitude of the H-rejkx is cotzsiderably reducedfor ITS in reference to CI for both subjects.
massage of the ITS is enhanced by the stability of the control conditions. The nonsignificant differences in H-reflex amplitude over the duration of the testing assion permitted the choice of C1 as being representative of all control conditions for the principal analysis. Within the same context, changes in M-response amplitude reflect changes in the stimulating conditions and thus limit the interpretation of the resulting H-reflex changes. The data reported in this article are statist i d y stable and provide confidence in the interpretation of the results observed for H-reflex changes. The reduction in motoneuron excitability (as evidenced by a decrease in H-reflex amplitude) observed during massage provides further evidence for the inhibitory effects of a range of peripheral stimuli. Both pressure8 and muscle tappinglo have been shown to reduce the level of motoneuron excitability in neurologically nonimpaired subjects. The receptors and origins of the observed inhibition, however, are still spe~ulative.~aJO Massage probably provides a multidimensional sensory input capable of exciting a number of different receptors. Petrissage, the form of massage used in this study, involves an active grasping and release of the triceps sume muscle, thereby activating both cutaneous and deep muscle receptors. The possible contribution of cutaneous receptors has been discussed in relation to the inhibition subserving pressures and muscle tapping.1 This effect, however, is probably of a limited magnitude6 during massage. The active manipulation of the muscle during massage

Massage is often used as a therapeutic intervention to increase circulation and reduce muscle reflex activity following an injury. Although anecdotal evidence supports the use of massage for reducing muscle reflex activity, solid evidence as to its therapeutic effectiveness is lacking. The results of this study indicate that massage not only reduces the level of motoneuron excitability, and consequently of muscle reflex activity, but does so in a specific manner. The inhibitory effects are specific to the muscle group being massaged.
As these results were obtained with

neurologically nonimpaired subjects, the results cannot be extrapolated to neurologically impaired subjects. The implications for the use of massage in subjects manifesting hyperactive muscle reflexes, however, should not be ignored. Further studies into the effects of massage in neurologically impaired subjects are currently being undertaken.

Massage has been used for the treatment of a variety of pathological conditions. This study investigated the effects of massage, administered at four different muscle sites, on the ex-

Physical TherapyNolume 71, Number B/Augus t 1991

citability of the ITS of 16 neurologically nonirnpaired individuals. A significant decrease was observed in the amplitude of the H-reflex, recorded during massage of the ITS. This decrease was interpreted as representing a reduction in the excitability of the triceps surae muscle motoneuron pool. This finding confirms the specificity of the neuromuscular response and suggests that the stimulus and receptors activated during massage are specific to the muscle being massaged.
Acknowledgments

Readaptation de Montreal, for their development of the data-acquisition and data-analysis software and their support.
References
1 Beard G, Wood EC. Massage: Principles and Techniques. Philadelphia, Pa: WB Saunders Co; 1964. 2 Tappan FM. Healing Massage Techniques:A Study of Eastern and Western Methods Reston, Va: Reston Publishing Company Inc; 1978. 3 Hollis M. Massage for Therapists Oxford, England: Blackwell Scientific Publications Ltd; 1987. 4 Tappan FM. Healing Massage Techniques: Holktic, Classic and E m q i n g Methods. 2nd ed. East Norwalk, Conn: Appleton & Lange; 1988. 5 Morelli M, Seabome DE, Sullivan SJ. Changes in H-reflex amplitude during massage of triceps surae in healthy subjects. Journal of Orthopaedic and Sports Physical Therapy. 1990;14:5559. 6 Morelli M, Seaborne DE, Sullivan SJ. H-reflex modulation during manual muscle massage of human triceps surae. Arch Phys Med Rehabil. In press. 7 Schieppati M. The H o h a n n reflex: a means of assessing spinal reflex excitability and its descending control in man. Pmg Neumbiol. 1987;28:345-376.

This research was conducted while Dr Williams was the recipient of a Canadian Visiting Commonwealth Fellowship at Concordia University. The support of this agency is acknowledged, as is the support of the Dean of the Faculty of Arts and Science, Concordia University. We also express our appreciation to Francois Kemp, P eng, and Michel Goyette, P eng, of the Centre de Recherche, Institut de

8 Kukulka CG, Beckman SM, Holte JB, Hoppenwonh PK. Effects of intermittent tendon pressure on alpha motoneuron excitability. Phys Ther. 1986;66:1091-1094. 9 Leone JA, Kukulka CG. Effects of tendon pressure on alpha motoneuron excitability in patients with stroke. Phys Ther. 1988;68:475 480. 10 Bklanger AY, Morin S, Pkpin P, et al. Manual muscle tapping decreases soleus H-reflex amplitude in control subjects. Physiotherapy Canada. 1989;41:192-196. 11 Eke-Okoro ST. The H-reflex studied in the presence of alcohol, caffeine, force and fatigue. Electmmyogr Clin Neumphysiol. 1982;22:579589. 12 Morelli M, Sullivan SJ, Seaborne DE. Comparison of human triceps surae H-reflexes obtained from mid and distal recording sites. Electmmyogr Clin Neumphysiol. 1990;30:181186. 13 Mathieu PA, Kemp F, Goyette M, et al. Utilisation d'un micro-ordinateur pour la mesure du reflexe-H. In: Proceedings of the Canadian Medical and Biological Engineering Congress. 1988:75-76. 14 Wolf SL, Minkwitz JA Topical anaesthetics: effects on the Achilles tendon and H-reflexes, I: able-bodied subjects. Arch Phys Med Rehabil. 1989;70:531-536. 15 Hayes KC, Sullivan SJ. Tonic neck reflex influence on tendon and Hoffmann reflexes in man. Electromyogr Clin Neumphysiol. 1976;16:251-261.

r
16 / 560

Abstracts for research, theory, and special interest platform and poster presentations.

1 Proposals for 90-minute program sessions.

Deadline: All materials must be received at the APTA office by Monday, November 4,1991. For complete directions and required forms, consult the September and October issues of Physical Therapy.

Physical Therapyll70lume 71, Number 8/August 1991