Electromyography and clinical reasoning
Why do exercise treatments sometimes fail and how can we overcome this?
– Written by Einar Einarsson, Qatar
Electromyography (EMG) is a familiar tool in a research setting, but it is far less common to see it being used clinically. This article will specifically talk about the use of EMG in a clinical setting and describe the author’s experience with it. If we compare any two individuals, we can see at some level, differences in 0ur motor patterns. Many of them are inconsequential, but some of them are caused by pathology or pain. Often, as sports physiotherapists, we intervene with exercise therapy and typically we think we are targeting certain muscles with certain exercises, but rarely are we able to account for these individual differences in our exercise prescription. Is this a big problem?
In the author’s experience, using EMG to identify an individual's actual muscle activation while performing certain exercises can lead to significant changes in clinical outcomes. This is certainly not something which should be advocated in all patients, but it could be argued that when a patient is not responding well to a treatment programme, the use of EMG and clinical reasoning can be an invaluable addition to your toolkit. What follows is the result of more than 20 years’ of clinical experience of using EMG as part of the management strategy for selected subgroups of patients, in the hope that some of the ideas presented here may be of clinical use.
This paper describes the use of EMG in three separate but inter-related sections: identification of activation errors, retraining activation patterns and carry-over from therapeutic exercise to functional activity. EMG measures the electrical activity of a muscle, (surface or indwelling) which can then be compared during different exercises or activities. In practice, use of real-time feedback of bar graphs showing activation levels of the muscles of interest while exploring different movements and exercises can identify what works best on a given day for a particular patient. The use of this real-time feedback allows clinicians to make sensible clinical reasoning decisions about what movements are increasing or decreasing activation levels of certain muscles. Simply seeing these activation levels (visual feedback) can significantly aid in retraining movement patterns of patients as they explore the exercises.
THE MIDDLE CAN BE A MUDDLE – WHY THE GROUP AVERAGE CAN BE MISLEADING
Large bodies of research have examined muscle activation patterns during therapeutic exercises and typically report on the average activation of a group of individuals while performing some exercises (usually healthy male university students). Consequently, clinicians will prescribe exercises based on these results; showing which exercises activate certain muscles the most. If we scratch the surface of these papers, we can see that there is a lot of natural variability between subjects in the activation of different muscles – on average, a certain exercise will be best at activating a certain muscle, but within the group you will always see some variability. You can think of this as ‘responders’ and ‘non-responders’ to this particular exercise. If you choose the exercise that activates a muscle best on average, then on average you have made a good choice. Some patients, however, are ‘non-responders’ to a particular exercise, so while ‘on average’ a specific exercise would be considered beneficial to these patients, specific individuals will not benefit from it. Indeed, it seems impossible, even in normal healthy subjects, to find a particular exercise which activates a certain muscle maximally in every single subject1. When we have a patient in front of us, what we really want to know is what is the best way to activate certain muscles for this patient. Knowing what works on average is a good start, but it is definitely not always correct. In addition, most of the available literature is on healthy subjects and therefore it should also be noted that there is very little information about how patients (who have pain or pathology) are different to healthy, uninjured subjects.
IDENTIFICATION OF ACTIVATION ERROR – ‘WAKING UP’ THE MUSCLE
Some patients, for whatever reason, simply do not activate a certain muscle group. Motor learning research has taught us to start our training in a ‘cognitive’ phase and work through ‘associative’ stages to ultimately reach ‘autonomic’ control. Perhaps growing from the profession’s roots in treating polio patients, where specific isolated muscles are genuinely neurologically ‘knocked out’ and require careful painstaking isolationist exercises to be rebuilt, it is common practice to begin rehabilitation with an isolationist-type approach. However in our clinical experience we have found that it’s not necessarily the simplest isolated exercises or movements that are best to ‘wake up’ these apparently inhibited muscles.
As stated above research shows great variety in activation and muscle recruitment between individuals and in our clinical experience we have found the best way to ‘wake up’ a muscle, at least in difficult cases, is to find a movement pattern where the weak or underactive muscle is an integrated part of movement and then use different stimulating or supporting techniques to trigger activation of the involved muscle. The initial aim is to find an exercise that gives good recruitment in a healthy movement pattern. Once activation is established in such a movement, there seems to be a window of opportunity where EMG biofeedback can be used to teach the patient a cognitive understanding of the desired recruitment. Gradually the EMG feedback is withdrawn and the patient has to cue their performance to their own ‘feeling’ for the exercise, with occasional EMG checks. Ultimately this new feeling for the exercise is then integrated into their normal training routine.
This tailored, exploratory and collaborative approach that we have found useful is outlined in the clinical examples below.
We recently treated two recreational tennis players experiencing problems activating the vastus medialis oblique (VMO) after knee surgery, leading to slow progress in their postoperative rehabilitation. It was observed that when trying to activate VMO in a typical manner: in a sitting position pushing down to the floor while slowly creating an extension moment at the knee, the VMO was simply not activating in either athlete. After exploring a range of different strategies with each of the patients, different approaches worked with each: patient 1 achieved highest activation while stepping back and coming up from deep knee flexion (Figure 1) whereas patient 2 (Figure 2) showed the highest activation levels while on a seated knee extension machine in midrange working at relatively high speed and a moderate load. In both cases these initial specific ‘starter’ exercises ‘woke up’ the muscle, such that after only a relatively short period of exercise, the activation of VMO was observed during the more typical rehabilitation approaches and usual rehabilitation could be resumed.
RETRAINING ACTIVATION PATTERNS – TO GET TRULY SPECIFIC WITH YOUR EXERCISE, YOU WILL NEED TO EXPERIMENT
Physiotherapists usually have some favourite exercises they fall back on for certain conditions. As alluded to above, the bulk of exercises used in training for pathological conditions are based on research done in healthy individuals. We cannot assume that the responses of those with pain or pathology are going to be the same. Indeed it could be argued that there is almost certainly going to be a different motor pattern in the presence of pain than in healthy individuals, secondary to the altered movement patterns patients have adopted, perhaps in an attempt to consciously avoid pain or as a reflexive spinal inhibitory response. An example of this approach is O’Sullivan’s classification-based approach to low back pain which suggests that there are adaptive and maladaptive motor behaviours which need to be specifically identified and targeted during rehabilitation2. Rehabilitation may then be aimed at increasing or decreasing the activation of certain muscles during certain movements – clearly an objective real-time feedback of the activation levels of these muscles is important in such cases. While this may be achieved by simple manual palpation in the case of the lumbar spine extensors, things can become more difficult where the anatomy is a little more complex.
Exercise programmes for given pathologies each have an array of typical exercises. By using EMG biofeedback we can simply select which ones work, which will not necessarily be the same for each individual. If we are, for instance, concerned with overactive tensor fascia lata (TFL) in a patient with patellofemoral pain syndrome, EMG sensors could be used in a clinical setting while designing a programme for gluteal strengthening, to ensure the TFL does not activate excessively and ‘take over’ the exercise in favour of gluteal muscles. Through trial and error, we have stumbled across some good reactive neuromuscular techniques to suppress activation of the TFL in many individuals, which when used in combination with EMG biofeedback can verify that the exercise is having the desired effect.
Case example: gluteus maximus vs TFL
Recently we had two recreational triathlon athletes coming in to our clinic with patellofemoral pain syndrome which was aggravated after prolonged cycling. The treating therapists thought that an important part of their rehabilitation was a reduction in the activity of their TFL during cycling, with a concomitant enhancement of activity of the gluteus maximus. Unfortunately, usual approaches were ineffective. In most cases the TFL can be suppressed and gluteus maximus activation augmented in single-leg bridging exercises with a superimposed external rotation and adduction moment from an elastic tube or a pulley applied to the weight-bearing leg (Figure 3). In athlete 1, this exercise did not work at all, and the TFL was still highly active, however when the stimulus was changed and the athlete was asked to push into a ball while doing a single-leg bridge, it worked well to suppress the TFL (Figure 4). In athlete 2 the above exercise with tubing worked well but typical gluteus maximus ‘activation’ exercises in prone with hip extension and knee flexion were not effective (Figure 5). In both athletes, a tailored, specific identification of an appropriate exercise which demonstrated the desired activation patterns allowed the athletes to ‘learn the feeling’ associated with these movements, which was ultimately transferred to cycling, to good clinical effect.
During the exploratory phase of designing an appropriate exercise programme, we have been regularly surprised by how changes in segments remote to the muscles of interest can have strong effects on activation patterns. For example, when trying to improve activation of the gluteus medius muscle during stance phase gait, we have noticed that providing an external stabilising force to the ankle on the same foot (usually manually) or stabilising the hip with elastic tubing from medial to lateral can enhance activation of the gluteus medius (Figure 6) . In these cases we suspect some mechanism is inhibiting the gluteals and compensatory factors are in play. Either way, these approaches have regularly been met with marked increases in activity. Making the task easier for gluteals can often help, rather than the more common approach of increasing resistance against these same muscles.
Maybe your shoulder shouldn’t always be ‘back and down’
Clinicians need to tread very carefully when contemplating changing the sporting technique of elite athletes. Well-meaning efforts to reduce the perceived possibility of an injury could result in performance reductions which can be catastrophic for all concerned. That said, we present a case where EMG can have a performance-enhancing effect in the absence of any pathology. We recently had an elite-level freestyle swimmer who was noticing a unilateral reduction in power in the catch and early pull phases. EMG in water is problematic at best and often logistically impossible to implement, so we had to attempt to mimic aspects of the freestyle stroke on dry land as well as examine simpler movements to postulate as to why power production was decreased at these outer ranges of shoulder flexion and abduction. During the assessment, we were able to use the contralateral shoulder as a point of reference to examine for differences in activation patterns and levels. It transpired that the athlete had a marked increase in lower and middle trapezius activation during these outer range movements. Speculatively we thought this may have been related to a long-resolved injury, during the rehabilitation of which the athlete had been encouraged to ‘stabilise’ his scapula into a retracted and depressed position during exercises and training. In this position, the athlete was less efficient at generating the required ‘pull’ force during catch and early pull-through phases. EMG proved useful here in re-learning to activate his upper trapezius during these extremes at the same time as quieting his middle and lower trapezius allowing for a restoration of his normal (large) range of scapular upward rotation during elevation.
BUT DOES THE CLINICAL METHOD TRANSFER TO OTHER ACTIVITIES?
The clinic is a very different environment to the training and playing field and it cannot be assumed that changes shown in the relatively calm and focused environment of your clinic will transfer to the organised mayhem of competition and practice, with all the distractions present in those situations. We believe that this transfer from clinic to pitch is a key, under-investigated area of rehabilitation. We certainly do not claim to have all the answers here, but will point to some encouraging clinical examples in the hope that they are useful.
Figure 7 shows the EMG pattern of a football player with a history of repeated hamstring injury during a hamstring fatigue test, followed by running. During the fatigue test we see increased activation of the medial (compared to lateral) hamstrings, then during running we see the same pattern repeated. Clinically it was suspected that his repeated injury to the lateral hamstrings was related to the reduced activation levels and that restoration of increased activation during therapeutic exercise and training was an important treatment goal. The athlete then performed low-intensity lateral hamstring biased exercises intermittently for approximately 5 minutes, after which the firing patterns during running were reassessed. A transfer of the hamstring activation (to an increase in lateral hamstring activation) was observed during running. In this case the learning process was fast and transferred well to the track. We have often seen a similar pattern with athletes: as soon as they understand the problem and have usable feedback allowing them to change their activation patterns, this carries over to other activities. These changes can happen very rapidly, even within the first session – as was the case with an elite football player with a long-term chronic groin injury, rehabilitated at Aspetar. The athlete had undergone surgical intervention 1 year previously with good results, but was still feeling pain and some weakness. EMG assessment revealed inhibited lower abdominal recruitment and in turn upper abdominal dominance. The player understood the problem with the help of EMG – and lower abdominal exercises were prescribed. At the end of the first session the player had no pain and strength measurements had increased by 50%. Testing after 2 weeks showed the same results. When the initial stages of retraining are performed with this concurrent EMG feedback, we have anecdotally noticed a much shorter learning process and better retention of these patterns between sessions, as well as better transfer to functional activity, particularly in elite athletes with good body awareness.
SO SHOULD I BE PUTTING EMG ON ALL MY PATIENTS, ALL THE TIME?
Here we have hopefully shown that EMG can be used within a clinical reasoning framework for certain individuals, with certain problems. We are not advocating its use in all athletes, but suggest that when:
an athlete is not responding to an intervention as expected,
their examination is showing an unusual response during a routine movement pattern,
your exercise intervention or movement retraining is having difficulty,
the use of clinical EMG might be considered to improve outcomes and help understand the problem better. We suggest manipulating variables such as joint angles, speed, movement plane, external support and focus of force placement, while you attempt to identify optimal strategies for the athlete in front of you. The result of the intervention can be evaluated immediately to identify change in activation of key muscles and its relation to other parameters like pain, strength or overall quality of movement.
Einar Einarsson B.Sc., M.Sc., M.T.C.
Aspetar – Orthopaedic and Sports Medicine Hospital
Boettcher CE, Ginn KA, Cathers I. Standard maximum isometric voluntary contraction tests for normalizing shoulder muscle EMG. J Orthop Res 2008; 26:1591-1597.
O’Sullivan P. Diagnosis and classification of chronic low back pain disorders: maladaptive movement and motor control impairments as underlying mechanism. Man Ther 2005; 10:242-255.
Image by Helgi Halldorsson