Health efficacy of electrically operated automated massage on muscle properties, peripheral circulation, and physio-psychological variables: a narrative review

Manual massage, commonly used by healthy individuals for well-being, is an ancient practice requiring the intervention of a trained and experienced physiotherapist. On the other hand, automated massage is carried out by machines or modalities without or with minimal control of a human operator. In the present review, we provide a literature analysis to gather the effects of automated massage on muscle properties, peripheral circulation and psychophysiological variables as reported through psychometric and neurophysiological evaluations of each modality ranging from massage beds and whole-body vibrations to robotic massage. A computerized search was performed in Google Scholar, PubMed, and ResearchGate using selected key search terms, and the relevant data were extracted. The findings of this review indicate that for vibration massage, whole-body vibration exposure with relatively lower frequency and magnitude can be safely and effectively used to induce improvements in peripheral circulation. As for massage chair and mechanical bed massage, while most studies report on positive changes, the lack of strong clinical evidence renders these findings largely inconclusive. As for robotic massage, we discuss whether technological advances and collaborative robots might reconcile active and passive modes of action control during a massage and offer new massage perspectives through a stochastic sensorimotor user experience. This transition faculty, from one mode of control to the other, might definitely represent an innovative conceptual approach in terms of human–machine interactions.


Effects of muscle properties
Two studies investigated on the effect of mechanical bed massage on lower back muscle fatigue, strength, and recovery. In one study by Zhong et al. [52], the researchers induced muscle fatigue through reverse sit ups in prone position followed by mechanical bed massage and evaluated back muscle endurance immediately, and 24 h after the intervention. The sample for this study comprised entirely of athletes. In another study by Do-Kim et al. [21], trunk extension (TE) was used as an index for back muscle fatigue. In addition to TE, the group also recorded EMG signals to analyze the strength of erector spinae muscles. However, Do-Kim et al. [21] had incorporated heating to the mechanical bed massage intervention to examine the combined effect. Zhong et al. [52] reported no statistically significant difference in back muscle extension (BME) between massage and control group. The study noted the slower rate of decline in BME in massage group, which they associated with increased blood flow and elimination of metabolites, but the evidence was not conclusive. On the other hand, Do-Kim et al. reported a significant improvement in TE and EMG signal after combined application of massage with heating as an indicator of improved flexibility and strength in back muscles.
The potential for vibration-based massage therapy to aid recovery from exerciseinduced muscle stiffness has also been studied. Imtiyaz et al. [19] and Pournot et al. [44], in two separate studies, investigated the effect of vibro-massage in treating stiffness of the biceps brachii muscle generated from weighted loading. Pournot et al. massaged the subject's biceps 5 min after exercise and 5 min after recovery period. They used the one arm as control/passive arm and evaluated the shear elastic modulus (SEM) of the other arm after massage. Their study lacked to observe any positive impact of vibration on SEM, and therefore, on muscle stiffness. Furthermore, Imityaz et al. compared the effect of vibration to manual massage and reported increased rate of recovery in groups receiving vibration massage as indicated by lowered level of blood serum LDH (lactate dehydronase). However, the study also reported absence of any significant difference between massage group and vibration group on ROM (range of motion) and level of CK (creatine kinase). In another study by Anna et al. [65], the effect of lower limb vibration on cyclist is documented. In this study, the authors applied vibration for 60 min after exercise, and reported a gradual decrease in blood lactate level as well as significant difference in other biochemical markers (IL-6, Mb, MMP-2) in experimental groups after intervention. Both Imtiyaz et al. and Anna et al. had varied time points in collection of blood samples ranging from 1 to 72 h after intervention.
Hiraiwa et al. [18] explored the possibility of robotic massage to treat masseter muscle stiffness in a controlled clinical study. The study population constituted patients suffering from either unilateral or bilateral myofascial pain stemming from masseter muscle stiffness. In this controlled study, the treatment duration ranged from 4 to 12 weeks and authors reported significant reduction in muscle stiffness in the treatment group. Only one study by Durkin et al. [9] involved conducting massage through mechanically designed chair on a car seat. In this study, the authors separately evaluated different massage system integrated to the car seat and found no statistically significant difference.
The Waseda Asahi Oral-Rehabilitation Robot 1 (WASO-1), developed by the Waseda University and Asahi Roentgen Company, aimed at recovery of temporomandibular disorders. The authors approach to unilateral and bilateral temporomandibular disorder was targeted to a specific population, highly technical and very expensive. Other studies on investigating the robotic massage effects on masseter and temporal muscles had varying degrees of pressure exerted, which ranged from 100 g to 1.5 kg per square area [2,3,20,23,46,60]. The most effective outcome was derived at approx. 800 g on small facial muscles. The authors further reported that the robotic massage decreased muscle pain and promoted functional motor recovery in patients. Another study by Luo and Chang [30] showed that patients had improved muscle activity after undergoing multi-finger robotic massage. However, the validity of these experimental results remains questionable due to the small sample sizes and complex study designs. Despite the autonomy of these robotic devices and lack of intervention by the physiotherapist, the robotic massages still do not allow for individualization of massage. The robots were designed to follow respective preprogrammed massage commands and did not incorporate a pain threshold for applied pressure considering patient morphology. Instead, like manual massage techniques, the patients had to assume a passive position when receiving massage [55] (Table 1).

Effects on peripheral circulation
A summary characterizing the interventional protocol of different massage modalities on blood circulation is given in Table 2. The larger segment of these studies involve vibrational massage focused on whole-body vibration (WBV). In addition, the electrically operated chair and air-cuff-based massage has also been reported (Table 3). Variability in terms of exposure session and duration to WBV is stark in the reported studies. Several studies had multiple bouts within the same session. Intra-session variation in exposure from WBV ranged up to 15 min [47]. While some studies were relatively short term, the duration for long-term intervention ranged from 2 weeks [34] to 12 weeks [32]. In these longitudinal studies, there were up to three intervention session each week of the study period. Differences with regard to vibrational intensity for the WBV interventions ranged from 26 to 44 Hz with a median exposure of ≤ 30 Hz. All but one study was randomized controlled and, in that study, the exact location of vibrational massage has not been reported. In addition to measuring blood flow, Menendez et al. [33] and Sonza et al. [47] reported on the skin temperature measurements at both baseline and post-intervention. In both cases, an infrared thermography-based temperature measurement was taken on the lower extremity. While Menendez et al. [33] reported on positive changes to side-altering vibration massage, the effect of vertical synchronized WBV according to the other studies, have no significant changes to circulation. However, Mitchel et al. [34] did not report on vertically synchronized vibration. Sonza et al. [47] and Menendez et al. [33] reported on significant alterations in peripheral circulation at vibrational stimulus between 25 and 30 Hz compared to lower intensities. However, Manimmanakorn et. al. and Mitchel et al. [32,34] did not observe any significant improvement in terms of long-term intervention with vibrational massage. There are several methodical limitations observed in these studies. Vibrational apparatus for all studies either imparted side-altering or vertical vibration. Given that intensity of the vibrational stimulus in side-altering vibration can change the impact of WBV due to the variability in foot positioning or distance between feet on the vibrational platform, the lack of reporting on these details in almost all studies create some ambiguity         in interpreting the findings. Furthermore, in some studies, the use of footwear by subjects has not been reported. Another limitation indicated by Mahbub et al. [58] is on the knee flexion whilst the subject is placed on the vibration platform. As suggested by Brooke-Wavell and Mansfield [59], assuming an extended knee on the platform may cause deleterious effect on the spine and cervical joints because the vibration transverses toward the vertebral column. The International Society of Musculoskeletal and Neuronal Interactions (ISMNI) recommended listing parameters such as frequency/amplitude and acceleration level while describing the vibration involving exposure of subjects to WBV (Rauch et al. [57]). However, most studies have reported vaguely on the measures of vibrational amplitude. Additionally, the placement of a transducer, which is used for measurement of vibrational intensity, is not reported accurately in majority of the studies. This unfortunately creates discrepancy in the reported vibrational parameters and raises implication of replicating the studies in the future. Furthermore, the available research works involving human exposure to WBV did not consider or discuss the recommendations of the international standard and in most cases, the adoption of safety limits when exposed to WBV is not clearly reported. Another limitation is the lack of reporting in environmental condition by the reviewed studies. As mentioned by Mahbub et al. [58], vascular response to vibration massage can change in accordance with differences in room temperature and therefore should be informed.

Effects on psycho-physiological variables
Majority of studies involving automated massage modalities have focused on the finding impact of these modalities in treating pain, discomfort, stress, and relaxation. The nature of outcomes measured in these studies is largely qualitative.
Kim et al. [22] studied the efficacy of mechanical massage chair therapy (MCT) in relieving chronic lower back pain. As compared to physiotherapy, the authors reported that MCT showed similar improvement in pain management where the result were based on a set of questionnaires. In a study by Lim et al. [27], the authors sought to find the effect of MCT with binaural beats by analyzing the electroencephalogram (EEG) and cognitive tests. Their study revealed a significant reduction in mental fatigue in patient as indicated by theta/beta as well as theta + alpha/beta indices in the treatment group for MCT combined with binaural beats. Moreover, there were significant improvement in cognition and memory for individuals who received MCT or MCT combined with binaural beats. Muller et al. [37,38] and Willeke et al. [50] in three similar randomized controlled studies investigated if MCT and mental training could reduce stress and anxiety among office employees. Muller et al. in their initial study found a significant reduction in anxiety and stress susceptibility in groups which received MCT or MCT along with mental training. The results acquired from this study were largely based on questionnaires. Likewise, in their following study, Muller et al. measured the employee's heart rate, blood pressure and fingertip temperature. While they did not find any significant change in blood pressure and temperature, they reported that MCT combined with mental training induced reduction in heart rate over the course of the study. Surprisingly, in a very similar study by Willeke et al. [50] the authors could not find any significant change in heart rate variability (HRV). In contrast, they found a significant decrease in the blood cortisol levels in the MCT group. Furthermore, their study suggested a Paul et al. EURASIP J. Adv. Signal Process. (2021) 2021 possible positive correlation of MCT with systolic blood pressure (SBP) and cortisol levels, and a negative correlation with HRV through a post hoc analysis. Another study by Lee et al. [25] on a relatively large number of subjects showed a decrease in the level of anxiety after MCT. However, their result being only from a single-subject measurement reduces its validity. Franz et al. [68] explored the effects of a lightweight massage system (LWMS) integrated to a car seat on driver comfort and relaxation. Their experiment recounts reduced activity in trapezius muscle group, but no changes in rhomboideus muscle as measured with EMG. The authors relate such reduction in muscular activity as suggestive of decreased amount of stress due to mechanical massage.
Luo et al. [31] reported reduced delta and alpha power of electroencephalogram (EEG) frequency band after robotic hand massage. However, the evidence is weak as their study lacked a trial design and involved a very small sample size.
Two studies investigated the impact of mechanical bed massage on stress. While Do-Kim et al. [21] looked into physiological variables as HRV and EEG frequency band power, Lee et al. [26] also tracked changes in the biochemical markers of stress such as cortisol and norepinephrine. The study by Do-Kim et al. was a pre-test-and post-testbased trial design, contrary to Lee et al. who devised a treatment plan for 2 weeks and measured outcomes 2 weeks after the end of trial. Do-Kim et al. reported that automatic bed massage decreased HR and alpha-beta frequency power on EEG as well as improved scores on visual analogue scale (VAS) and state anxiety index (STAI). The study by Lee et al. had also showed a decrease in HRV, cortisol and norepinephrine levels, although the changes did not recur 2 weeks after the end of study for all parameters except for the norepinephrine level. Yang et al. [67] Wang et al. [66] and Ramalingam et al. [65] investigated the use of local vibration massage on pain management. The results were derived from questionnaires exhibited gradual improvement of pain perception throughout the study period.
In a recent review by Kerautret et al. [55] the authors discussed on the possibility of entirely autonomous massage modalities with ability to conciliate for both proactive and retroactive mode of motor control for user's massage experience. Earlier studies have largely involved highly technical and expensive robotic massage which may not be commercially feasible. However, in recent times there has been development of multiple automated massage modalities with added advantages of precision, availability, privacy and user choice. For instance, the iYU ® Pro from Capsix Robotics, which started development in 2016, has complete autonomy and with aid of remote interface can switch from completely automated to user directed control for massage. However, further empirical evidence is required to conform to user safety and incurred benefits from such collaborative massage modalities. Safety is one of the prime issues, therefore, given the fact that with the freedom of use, the level of autonomy granted to massage robots can account for unwanted risk to users if pressure level or trajectories of massage and similar parameters are not adjusted prior to intervention, and subsequently monitored by attending professional. Lima et al. [63] have emphasized on the importance on the understanding of histological structure and physiological process to be accounted for during massage intervention. The issue is further complicated by the varying ergonomics of individuals in terms of muscle mass, personal preference in massage routine and possible presence of undiagnosed medical dysfunctionality. Therapeutic massage, hence, may never be complacent with only automated massage modalities and without involvement of the physiotherapist. Automated massage modalities, from a preventive viewpoint may still be relevant when considering the importance of workload for physiotherapist and the limited medical resources available. Golovin et al. [61] suggest that such robotic massage modalities will provide an alternative means of assistance in cases when medical professionals are burdened with increased workload. In such cases these autonomous devices can administer simple massage techniques at a limited scale for preventive measures. Complex maneuvers such as stretching, strengthening and joint mobilization, which cannot be performed by robotic devices can be administer by a professional therapist, and thus providing compounding benefits. Hence, robotic massage can reduce risk of musculoskeletal injuries which occur from lack of joint mobilization and subsequently, improve well-being and health by allowing for regulated self-massage. It remains under scrutiny whether manual massage can be entirely substituted by its robotic counterpart. Even with the continuing advances in AI, it is questionable if physiotherapy can be robotized. Traditionally, manual massage therapy practice depends on feedback-based adjustment of manipulation of the patient's body by the therapist. The readjustment to ongoing feedback on the scar tissue ensures for treatment efficacy. Further studies on automated massage modalities are required and until then, mechanical apparatus such as massage chair and robots cannot be entirely integrated and therefore, remain another tool at the disposal of medical practitioners.

Conclusion
In this review of automated massage therapy, a range of massage modalities ranging from automated massage beds to robotic massage have been discussed pertaining to muscle properties, peripheral circulation and physio-psychological measures in different groups and health conditions. In many of these studies, a treatment or experimental group have been compared to a control group on basis of different types of psychological, physical, physiological and biochemical measures to evaluate massage effects. The studies have similarities and differences in terms of massage techniques involved such as kneading, rolling, and vibration. In almost all of these studies, massage has been shown to impart beneficial effects in both clinical and non-clinical experimental trials. It has been suggested that the beneficial effects incurred are induced from massage stimulation of pressure receptors on the body leading to enhanced vagal activity and reduced cortisol levels. However, in many cases, these effects are observed over a short period of time and/or involving small sample sizes. It is recommended that future studies look into the long-term effects of massage and design protocols conforming to international standards.