I am a musculoskeletal physiotherapist, who graduated in Brazil a while ago. Now I work as an academic at the University of Otago – New Zealand. When I was an undergraduate student in Brazil, MWM had not been incorporated into the undergraduate curriculum yet. So, as a new graduate, I enrolled into one of the Mulligan Concepts workshops. When I first learned about mobilization with movement (MWM), I was impressed with the immediate reduction in pain, and improvement in range of motion observed in the clinic. But I was also intrigued and asked myself some questions … how can this technique work? What does it do? Does it have any impact on muscle activity levels? Does it matter?

I wondered “what was the mechanism underlying MWM in patients with musculoskeletal disorders?” and how that led to “pain-free, immediate, long-lasting effect”. At that time, I did not have the training to design a research project to answer these questions. A number of years later, I found myself asking that question again. This time, I was working as an academic in New Zealand. This time, I had the training and “freedom” to decide and to plan a number of research projects to gain a better understanding on what are the underlying mechanisms of MWM in patients with shoulder pain.

With that question in mind, I designed a series of research projects. To understand the underlying mechanisms of MWM in patients with shoulder pain, I thought it was important to plan studies using a step-wise approach. I planned initially laboratory-based studies, with the goal of using data from these studies to inform future lab-based studies, but also clinical trials. Adopting that strategy ensured that each new study I conducted would be informed by findings and knowledge gained from previous studies.

As a first step, we decided to conduct some studies with asymptomatic individuals and explored what was the effect of sustained glides on scapular and shoulder muscle activity levels. We conducted two studies (Ribeiro et al. , 2016, Ribeiro et al. , 2017). The first study (Ribeiro, Castro, 2016) was exploratory. We wanted to know whether it would be worthy pursuing this research question. Therefore, we monitored only four muscles (i.e. supraspinatus, infraspinatus, middle and posterior deltoid muscles) during some shoulder movements. On that study (Ribeiro, Castro, 2016), we compared shoulder abduction or elevation performed with or without sustained glide. Our findings suggested significant reductions on shoulder muscle activity levels. Motivated by our findings, we designed the second study (Ribeiro, Sole, 2017).

The first study had some limitations. We just assessed shoulder movements with and without sustained glide. We did not analysed what happens to muscle activity levels before, during and after the sustained glides. In a clinical setting, clinicians would usually assess the movement first, perform the MWM, and then re-assess. Such process was not included in our design for study 1.

The second study compared clinician-administered sustained glide with participant-administered sustained glides (Ribeiro, Sole, 2017). This time, we designed a slightly more complex study: (1) used a cross-over study design; (2) assessed muscle activity levels before, during and immediately after shoulder movement performed with the glide; (3) assessed scapular (upper and lower trapezius, serratus anterior) and shoulder (supraspinatus, infraspinatus, middle and posterior deltoid) muscles.

A cross-over study design was a really nice challenge to get involved with (from a research perspective). In cross-over designs, participants are randomly allocated to two different sequences of interventions (or experimental conditions). In our study, half of participants received the clinician-administered glide followed by the participant-administered glide; while the other half received the opposite sequence (i.e. participant-administered glide followed by clinician-administered glide). The main advantage of a cross-over design study is that participants are their own controls. This is a neat approach, as it means that between-subject variability is reduced. That results in a smaller sample size (Stoney and Johnson, 2012).

Designing and analysing the second study was exciting but also challenging. The challenges with cross-over design included: (1) the risk of carry-over effects (i.e. the effect of the first intervention are long enough to influence the measurements of the second intervention) (Stoney and Johnson, 2012) and defining the wash-out period; (3) designing and conducting the statistical analysis (this was a challenge for me as an early career researcher – may not be for others).

The challenges of carry-over effect and wash-out period were linked. Not much research had been done, particular on the effects of sustained glides on scapular and shoulder muscle activity levels. We did not find any previous study using a cross-over design and assessing effects of manual therapy on muscle activity levels. Hence, we based our decision on data from study 1. The wash-out period was set at 5 min. We did not know whether this would be enough. But had good reasons to believe it would. Turned out, 5 min was long enough, and we did not find any carry-over effect in our EMG data.

The statistical analysis. Study 2 was my first study with a cross-over design, and understanding and conducting the stats analysis was a massive learning curve. Specially because I decided to run the analysis using the R Software (as I do for all my projects). I get a lot of satisfaction and bit of frustration when using R (but will leave the joys and challenges of it for another moment). To summarize, conducting Study 2 was a massive research learning. Challenging. Really exciting journey.

Findings from Study 2 showed that asymptomatic individuals had significant reductions in scapular and shoulder muscle activity levels when moving the arm while the sustained glide was applied to the shoulder. This was especially true during the concentric phase of shoulder abduction. Findings from the second study (Ribeiro, Sole, 2017) confirmed the findings from the first study (Ribeiro, Castro, 2016). That was re-assuring and encouraging, suggesting that what we are observing in the lab is a “true” effect of the sustained glide on muscle activity levels. This will be strengthen further if another research group can replicate the same findings.

Our findings also showed that there were no differences between clinician-administered and participant-administered glide. Both conditions led to reductions on muscle activity levels. Hence, no “statistically significant” differences were found between the two experimental conditions. We did noticed, though, that the reductions in muscle activity levels were larger in the clinician-administered glide condition. We believe that is due to a more precise, controlled glide that is generated by the clinician, when compared to a less accurate, less precise glide performed indirectly by the belt. This was confirmed by participants who reported they could feel a “firmer” glide at their shoulder when the clinician applied the glide. The directions of the glides are also different (Hing et al. , 2015), and that needs to be taken into account.

Together, these studies helped us to gain a better understanding of the effects of sustained glides on scapular and shoulder muscle activity levels when pain is not present.

It is reasonable to expect that manual therapy techniques have mechanical, and neurophysiological effects on the body (Bialosky et al. , 2009, Vicenzino et al. , 2011). Based on our findings, we speculate that sustained glides might improve pain and ROM due to the following factors:

  1. Mechanical effects: postero-lateral sustained glides at the shoulder led to posterior displacement of the humerus (Ho and Hsu, 2009). This posterior displacement may affect moment arm of shoulder muscles. That may have an impact on muscle recruitment. As I wrote, this is only a speculation, with no data to support it. To assess the mechanical effect of sustained glides on the glenohumeral joint, ideally, one needs to measure the position of the humeral head with regards to the glenoid cavity with and without the glide. I do not have the research expertise on the area, but this certainly brings few methodological challenges. This mechanical effect (i.e. posterior displacement) may also affect spinal and transcortical reflex circuits that regulate excitability of motor neurons (Struyf et al. , 2015). We cannot discard the fact that the clinician’s hand provide some mechanical support to the glenohumeral joint. With this in mind, it is reasonable to think that less muscle contraction is required to provide support to the joint during shoulder movement. Together, these factors could explain why we observed reductions in scapular and shoulder muscle activity levels.
  2. Neurophysiological effects: cutaneous and proprioceptive afferent discharge and changes in descending drive to motor neurons (Bialosky, Bishop, 2009). Cutaneous afferents might inhibit and stimulate motor neurons located in the spinal cord, and can impact on motor neuron threshold and the recruitment order of motor units. This mechanism could partially explain the reductions in muscle activity levels when the sustained mobilization was applied to the shoulder. I am planning some further research to assess the effect of sustained shoulder glides on the neuro-system.We conducted studies 1 and 2 so that we could understand what happens when pain is not present. That is the “normal” response to the glide. When designing and planning a research programme, it is wise (that is my take on it at least), to undertake a step-wise approach: (1) start small; (2) conduct exploratory studies; (3) improve the design of the study by analysing data from previous studies; (4) understand the foundational mechanisms of an intervention; (5) understand how asymptomatic individuals respond to an intervention (in this case a sustained glide); (6) then move into research with symptomatic individuals.

By adopting this approach, we strengthened the research design. I improved my statistical analysis skills. We gained a better understanding of what is happening in the body when we apply a sustained postero-lateral glide on the shoulder. These knowledge can be used in future research.

Now, you may be thinking, “fair enough Dan, but patients have pain. So, what is the effect of sustained glides on patients with shoulder pain? After all, patients may respond differently to this technique”. That might be true. They might respond differently. I also ask myself that same question!

To answer your question about the effect of MWM on patients with shoulder pain, we designed our third study. In that study, we compared the effect of sustained glides with a “placebo” glide. We used a cross-over study design (again), and measured scapular and shoulder muscle activity levels. Findings are very interesting, and we hope to submit that study for publication soon.


Bialosky JE, Bishop MD, Price DD, Robinson ME, George SZ. The mechanisms of manual therapy in the treatment of musculoskeletal pain: a comprehensive model. Manual therapy. 2009;14:531-8.

Hing W, Hall T, Rivett DA, Vicenzino B, Mulligan B. The Mulligan Concept of Manual Therapy, Textbook of Techniques. 1 ed: Churchill Livingstone; 2015.

Ho KY, Hsu AT. Displacement of the head of humerus while performing “mobilization with movements” in glenohumeral joint: a cadaver study. Manual therapy. 2009;14:160-6.

Ribeiro DC, Castro MP, Sole G, Vicenzino B. The initial effects of a sustained glenohumeral postero-lateral glide during elevation on shoulder muscle activity: A repeated measures study on asymptomatic shoulders. Manual therapy. 2016;22:101-8.

Ribeiro DC, Sole G, Venkat R, Shemmell J. Differences between clinician- and self-administered shoulder sustained mobilization on scapular and shoulder muscle activity during shoulder abduction: A repeated-measures study on asymptomatic individuals. Musculoskelet Sci Pract. 2017;30:25-33.

Stoney CM, Johnson LL. Design of Clinical Studies and Trials. In: Gallin JI, Ognibene FB, editors. Principles and Practice of Clinical Research San Diego: Academic Press; 2012.

Struyf F, Lluch E, Falla D, Meeus M, Noten S, Nijs J. Influence of shoulder pain on muscle function: implications for the assessment and therapy of shoulder disorders. Eur J Appl Physiol. 2015;115:225-34.

Vicenzino B, Hing W, Rivett D, Hall T. Mobilisation with Movement: The Art and the Science. 1 ed: Churchill Livingstone; 2011.


Dan is a Senior Lecturer at the School of Physiotherapy – University of Otago (New Zealand). He conducts clinical research on musculoskeletal disorders, with a special interest in shoulder rehabilitation. He holds a “Sir Charles Hercus Health Research Fellowship” funded by the Health Research Council NZ. During his Fellowship, Dan will undertake a number of studies on process evaluation of clinical trials.  Please find a link to one of Dan’s articles https://www.sciencedirect.com/science/article/pii/S246878121730084X