How to build an effective shoulder stability workout

Any shoulder stability workout and overall program should consider the anatomical makeup of the shoulder - as well as the person whom the shoulder belongs too!

Cadaver inspection and histological analysis reveal high proportions of sensory nerve fibres in the labrum (Alashkham et al., 2018) and incomplete attachment of the labral surface to the glenoid from 11 to 2 o’clock (Alashkham et al., 2018). These have important implications.

Loss of labral surface from a lesion during the injury likely leads to impaired/or loss of sensory nerve function of those afferent nerve endings in that region and a reduction in purchase area of the glenoid with the humeral head. The most common sensory nerve afferent in the labrum appears to be golgi tendon organ-like structures (Witherspoon et al., 2014) suggesting the need for rehabilitation to incorporate heavy loading through all planes and high rate of force development. This is because we know Golgi tendon organ’s primary function is responding to the rate and magnitude of tension in the muscle-tendon unit (Tuthill & Azim, 2018).

Other sensory fibres that have been consistently observed in the labrum include Ruffini, Pacinian and Krause corpuscles (Withersoon et al., 2014). The function of these sensory fibres is related to both fast and slow-responding functions to end range joint positions and as such are often coined “limit detectors” (Tuthill & Azim, 2018). This suggests that in addition to magnitude and rate of force, we may want to consider placing the shoulder in the end range position for long, slow holds or dropping into that end range position quickly.

The muscles crossing the glenohumeral joint have different fibre type compositions ranging from 22% to 40% for Type IIx, 17% to 51% for Type IIa and from 23% to 56% for Type I (Srinivasan et al., 2007) with high heterogeneity in fibre type composition across different muscles and between people (Srinivasan et al., 2007; Lovering et al., 2008).

The impact of sensory and cognitive reweighting on task performance via the dual-task effect (DTE)

We seldom perform any task in an isolated environment that requires us only to move our limbs in the absence of performing some sort of cognitive task. Much of the time we find ourselves in dual-task situations defined here loosely as “the concurrent performance of two tasks that can be performed independently, measured separately and have distinct goals” (Bayot et al., 2018, pp 3) and often there is a cost or change in performance to one or both of the tasks as a result of this dual-tasking phenomenon.

In the upper limb examples may include a spin pass in rugby or a tackling an opponent coming off a ruck, or in netball, a defender may attempt to block to striker or predict where they are going to shoot and have their shoulder in a position of maximal flexion overhead. All the examples above typically require simultaneous cognitive and motor performance to successfully complete the task. The are three main domains of cognitive functioning that are typically looked at in the dual-task research are executive functioning, working memory and attention.

Adapted from Plummer et al., 2013. Table 1, pp 2567; Plummer et al., 2015. The red dots represent the theoretical pre-training scores of athletes scoring on an upper limb DT exercise and the green represents the theoretical post-training scores.

As highlighted above taken from the work by Plummer et al., there are variable consequences to the dual-task phenomenon, and it is quite clear that they are not all bad! The sparse number of upper limb studies examining the DTE appear to show mutual deterioration in performance across most participants when compared to a single task exercise (Bank et al., 2018; Srinivasan et al., 2015) and this magnitude of interference increases substantially with increasing task difficulty (Bank et al., 2018).

Following rehabilitation of the unstable shoulder, we want to fall either in the centre or top right quadrant - either of the others yields a net negative effect on performance. This may be somewhat context specific depending on the parameters that define task success; a greater emphasis motor performance for successful completion might mean cognitive performance can suffer more without total performance suffering and vice versa. However, for the game athlete or team sports, motor tasks are often performed concurrently or after executive function processes thus …

The goal of training is the rightward shift in the pre-post testing curve that intersects the middle as shown above!

Why vision, shoulder performance and the control-to-chaos continuum are critical components of shoulder rehab!

Vision plays a fundamental role in performance and goal-directed tasks, and it should be considered and trained as part of a return to sport phase of training.

The main visual skills involved in sports are eye–limb coordination, static and dynamic acuity, peripheral vision, spatial focus and judgement of speed and distance of subjects and objects in the environment (Buscemi et al., 2024).

Fitt’s law relates to the time taken to reach an object. It depends on the distance to the object to be reached and the size of that object in the environment. The athlete who is better able to judge distance and lock onto the target quicker can act sooner from a performance perspective.

Stereopsis, the perception of depth, is essential for the three-dimensional analysis of the environment and allows the athlete to judge the distance of objects relative to themselves and other targets. It will mainly involve dynamic stereopsis which is to perceive the depth of an object or subject while following it (e.g. tennis ball in tennis) or static stereopsis, where the target is fixed like putting in Golf. Depth perception discriminates athletes from non-athletes as they have a greater ability to integrate visual abilities and visual motor skills to control the action (Buscemi et al., 2024).

Concurrent to stereopsis, visual search capacity forms the second key element to Fitts Law and task performance. It is all well and good the combat athlete being able to judge the opponent’s depth of their head for throwing a hook, but if they are unable to lock onto the target area whilst tracking the opponents’ limbs it is unlikely they will be successful in their task execution. Tracking the target object in the environment requires visual search and is best explained by the Guided Search Theory (see Chan et al., 2013) that use both top-down and bottom-up mechanisms as shown in the figure below.

The sports vision training pyramid as proposed by as cited in Laby et al., (2018) shows five levels of visual capacities that can be trained, with the foundation of visual acuity and contrast sensitivity being paramount. Different aspects of visual function appear to correlate well with, and in some cases predict, on field performance in sports that rely heavily on upper limb performance including baseball, shooting, softball and basketball (Laby et al., 2021) and SVT leads to small and moderate effect sizes on reduced decision making response time and improved sport-specific performance respectively in athletes (Guo et al., 2025).

Taken together, sports vision training (SVT) and the dual task effect are critical elements to consider in the rehabilitation of the unstable shoulder and is best applied via the control-to-chaos model that has been developed for late-stage rehabilitation (Taberner et al., 2019, 2025). It is worth noting that the first two layers of the sports vision pyramid can and should be trained from day 0 in the clinic and can be easily incorporated into any training program as you will see below.

Putting it all together: The framework for rehabilitating the unstable shoulder from injury to performance

The clinic environment is busy. We often must make decisions on our feet and have little time to delve into concepts in depth, this where I feel loosely held frameworks are useful for giving the basics to build a program around the person in front of us. The framework below is something that I have found useful in helping guide progression through a shoulder stability workout over the phases of rehabilitation from injury and pain.

The benefits of long hold isometrics in building our movement maps and navigational ability

Long hold isometrics allow sustained exposure to body positions that the athlete may encounter in their sport and as discussed in a previous article here, we can attempt to deepen the well of the attractor states we want to try and push the athlete toward.

Force expression is a skill as it demands integrated, seamless effort of the contractile tissue and nervous system in unison. The increased time under tension in hold isometrics combined with the reciprocal perception-action coupling of the nervous system from holding an external load in a defined position without support, results in higher proprioceptive inputs and larger involvement of somatosensory areas in the CNS. This is highlighted by the greater variability in burst rate activations and interspike intervals of motor units, higher cortical demand and cyclical recruitment and de-recruitment of motor unit suggesting more nuanced control strategies compared to pushing isometrics (Oranchuk et al., 2024). 

So, for the wobbly shoulder, long hold isometrics in variations of the push up, frog pose, quadruped hold and hangs can be incredibly potent in the early stages for reintroducing load. Because long holds are easy to execute, you can add in dual-task exercises and vision drills easy in this phase and serves to start building an element of chaos into movement early in the athlete’s rehabilitation!

I liken this to map making. In the early stage, the athlete (and their body) has lost their movement map and their ability to navigate it. We want to rebuild the map (of movement) and our navigational ability; long hold isometrics are useful in this stage but need to be taken to near failure for maximal benefit!

Holding isometrics for extended periods of more than 2 minutes also builds the fatigue-resistant qualities of the body: improvements in aerobic capacity and local muscle endurance can be achieved with a block of long hold isometrics performed 2-3 times weekly.

The Triphasic approach to strength training and its benefits

The triphasic approach to programming comes from the work of Cal Dietz in his book ‘Triphasic Training’ and is something I would highly recommend reading!

Most sports involve acceleration, deceleration, change of direction and sometimes top end speed – the athlete must be able to break, change and express force incredibly quickly! This chain of movements on a neuromuscular level is eccentric, isometric (our amortization phase) and then concentric in nature. The athlete is better able to handle higher levels of eccentric breaking force, smaller amortization phase and a rapid shortening cycle is going to be faster out the blocks, performing cuts or decelerating ultimately giving themselves more time to execute their task! Hopefully the figure below showing two athletes, Joe (Red) and Bob (Blue), illustrates this nicely, both are tennis players, and if we were able to accurately measure their shoulder internal rotators during  a tennis serve then the force-time curve below depicts the differences in their parameters.  

Whilst we don’t change direction with the upper limb or decelerate in the same way as we do with our legs, the movements in sport still go through a triphasic nature. In a tennis serve, the shoulder internal rotators undergo rapid lengthening during the windup phase before then rapid shortening during the release phase. When making a tackle, the players arm will be forced backwards (eccentric action of the pectoral muscles, subscapularis, supraspinatus’s etc) before then reacting and shortening in response to the anterior-posterior movement of the humeral head on the glenoid.

Both supraspinal and spinal pathways have distinct patterns of control and modulation during eccentric contractions, shortening and isometric muscle actions (Duchateau & Enoka, 2008; 2016).

The difference in neuromuscular control provides an opportunity for skill acquisition of the three phases of muscle actions and training them in order of use, eccentric, isometric then concentric allows the layering of each skill on top of the other. Normally, I would run this block for a 6-week period, and the athlete would be far down the line in their rehab as this approach is intense! I’d say you want to run at least 6-12 weeks of traditional strength training of higher volume and a good aerobic base before progressing onto the triphasic stuff… for the upper limb I would use the main push and pull exercise for triphasic and leave the rest of it. See the table below for how I would program intensity and volume – I have found a clustered set approach when using triphasic helps with intra-set fatigue and keeps the bar speed crisp!

Upper limb plyometric exercises using the ABC model of plyometrics

I like things made simple! It is easier for my brain and often means I can pull concepts quickly and apply them to the person in front of me. Plyometrics are a bit of a mine field when it comes to categorising them and I won’t go into detail here but the way I split them is by determining what target output I am wanting to go after. Again, revisiting our force-time curve above but now we have split the phase up with the black line and intentionally separating the eccentric from the amortization and concentric phase.

My reasoning? Feeding nicely from the triphasic section above, the phases of muscle actions have distinct neural control AND purposes.

The negative phase looks at the capacity of the athlete and their neuromuscular system to handle high forces to slow them to a stop during any task of upper or lower limb. Visiting the paper again by Duchateau & Enoka (2016), absorb plyometrics look at reducing inhibitory drive from Ia and Ib afferent reflex pathways and increasing supraspinal facilitation.  

To do this we need regular exposure to very high negative forces ie. eccentric muscle actions so the body can get more efficient at handling higher eccentric breaking forces. Although the body does not absorb the force, I think its adjective description highlights the characteristics of movement that we are looking for. We know the force-velocity curve, the higher the force the slower the velocity thus super high negative forces coincide with slower positive expression, I am only focused on how much negative force the athlete can handle before their neuromuscular system says no! I am not concerned how quickly the athlete can cycle through the eccentric, amortization and concentric phase.  

For bounce plyometrics, as the name suggests, the focus is on how quickly how our athlete can cycle through the eccentric, amortization and concentric phase of a movement. is the key determinant of impulse and thus by extension performance. This is a quality is likened to bouncing a bouncy ball. You can throw the ball hard at the floor and it deforms quickly before returning to the original dropped position.

This method also addresses issue of upper limb plyometrics where ground contact time is always higher so the concept of splitting them by GCT <250ms does not apply. The adjective description of each better informs the athlete and coach of what characteristic output we are seeking to develop. That’s not to say that within both absorb and bounce that I don’t split into intensive and extensive, I do, and I’ll write more about that in future articles.

Why throwing is important!

Humans have thrown for as long as we have walked on two legs, and this ability has afforded us with the opportunity to hunt and fight. There is no movement for the upper limb that comes close to full kinetic chain integration than throwing, something that often lacks in end stage return to performance. We can build a strong shoulder but building one that can throw shit at high speeds, for long distances means we have a shoulder that can handle the extremes of end range joint positions with incredibly high forces! That is a shoulder that is bulletproof.

As you can see from the flow diagram presented earlier, I split throwing into movements that are fast with light weight or heavy and slow. Much the same as absorb plyometrics, the heavy throws aim to desensitise the inhibitory reflex pathways associated with high force, high intent movements but naturally the heavier weight means the throwing velocity is substantially slower and we do not (routinely) push into the end range position. Heavier throws often build confidence in the shoulder for impact-based athletes likely rugby.

The velocity-focused throws emphasise high impulse generation and use a light load throwing implement to ramp up release velocity; often we reach both joint end range extremes during velocity-focused throws and high speeds meaning we desensitise the limit detectors (Pacinian corpuscles and Ruffini endings). During velocity-focused throwing there is some focus on ensuring the athlete is able to unlock their hip-shoulder complex to allow to rotation of the shoulder relative to the lumbopelvic region, this is because the degree of rotation of the shoulder relative to the hip (coined hip-shoulder separation) predicts trunk rotation velocity, which in turn correlates with pitch velocity (Trasolini et al., 2022).

I would begin to incorporate velocity-focused throws early on in rehabilitation from the perspective of developing throwing technique and the hip-shoulder separation and lumbopelvic control. The heavy-slow throws would come later in rehabilitation once a level of tissue capacity has been developed, again, you could put this throw category into a triphasic block of training.

Hopefully the information above provides a loose framework around how you can approach training the wobbly shoulder and the different things to consider in the training!

References

Alashkham, A., Alraddadi, A., Felts, P., & Soames, R. (2018). Histology, vascularity and innervation of the glenoid labrum. Journal of Orthopaedic Surgery, 26(2), 2309499018770900.

Witherspoon, J. W., Smirnova, I. V., & McIff, T. E. (2014). Neuroanatomical distribution of mechanoreceptors in the human cadaveric shoulder capsule and labrum. Journal of anatomy, 225(3), 337-345.

Srinivasan, R. C., Lungren, M. P., Langenderfer, J. E., & Hughes, R. E. (2007). Fiber type composition and maximum shortening velocity of muscles crossing the human shoulder. Clinical Anatomy: The Official Journal of the American Association of Clinical Anatomists and the British Association of Clinical Anatomists, 20(2), 144-149.

Lovering, R. M., & Russ, D. W. (2008). Fiber type composition of cadaveric human rotator cuff muscles. journal of orthopaedic & sports physical therapy, 38(11), 674-680.

Tuthill, J. C., & Azim, E. (2018). Proprioception. Current Biology, 28(5), R194-R203.

Bayot, M., Dujardin, K., Tard, C., Defebvre, L., Bonnet, C. T., Allart, E., & Delval, A. (2018). The interaction between cognition and motor control: A theoretical framework for dual-task interference effects on posture, gait initiation, gait and turning. Neurophysiologie Clinique, 48(6), 361-375.

Plummer, P., Eskes, G., Wallace, S., Giuffrida, C., Fraas, M., Campbell, G., ... & Skidmore, E. R. (2013). Cognitive-motor interference during functional mobility after stroke: state of the science and implications for future research. Archives of physical medicine and rehabilitation, 94(12), 2565-2574.

Bank, P. J., Marinus, J., van Tol, R. M., Groeneveld, I. F., Goossens, P. H., de Groot, J. H., ... & Meskers, C. G. (2018). Cognitive‐motor interference during goal‐directed upper‐limb movements. European Journal of Neuroscience, 48(10), 3146-3158.

Plummer, P., & Eskes, G. (2015). Measuring treatment effects on dual-task performance: a framework for research and clinical practice. Frontiers in human neuroscience, 9, 225.

Buscemi, A., Mondelli, F., Biagini, I., Gueli, S., D’Agostino, A., & Coco, M. (2024). Role of sport vision in performance: systematic review. Journal of Functional Morphology and Kinesiology, 9(2), 92.

Chan, L. K., & Hayward, W. G. (2013). Visual search. Wiley Interdisciplinary Reviews: Cognitive Science, 4(4), 415-429.

Laby, D. M., & Appelbaum, L. G. (2021). Vision and on-field performance: a critical review of visual assessment and training studies with athletes. Optometry and Vision Science, 98(7), 723-731.

Laby, D., & Kirschen, D. (2018). A new model for sports and performance vision. Vision Development & Rehabilitation, 4(2), 91-7.

Taberner, M., Allen, T., & Cohen, D. D. (2019). Progressing rehabilitation after injury: consider the ‘control-chaos continuum’. British journal of sports medicine, 53(18), 1132-1136.

Taberner, M., Allen, T., O’keefe, J., Chaput, M., Grooms, D., & Cohen, D. D. (2025). Evolving the control-chaos continuum: part 2—shifting “attention” to progress on-pitch rehabilitation. Journal of Orthopaedic & Sports Physical Therapy, 55(3), 162-172.

Guo, Y., Chen, C., Peng, J., Deng, L., & Yuan, T. (2025). Does Visual Training Enhance Athletes' Decision‐Making Skills and Sport‐Specific Performance? A Systematic Review and Meta‐Analysis. Scandinavian Journal of Medicine & Science in Sports, 35(10), e70140.

Oranchuk, D. J., Nelson, A. R., Lum, D., Natera, A. O., Bittmann, F. N., & Schaefer, L. V. (2024). ‘Pushing’versus ‘holding’isometric muscle actions; what we know and where to go: A scoping and systematic review with meta-analyses. MedRxiv, 2024-11.

Duchateau, J., & Enoka, R. M. (2016). Neural control of lengthening contractions. Journal of Experimental Biology, 219(2), 197-204.

Duchateau, J., & Enoka, R. M. (2008). Neural control of shortening and lengthening contractions: influence of task constraints. The Journal of physiology, 586(24), 5853-5864.

Trasolini, N. A., Nicholson, K. F., Mylott, J., Bullock, G. S., Hulburt, T. C., & Waterman, B. R. (2022). Biomechanical analysis of the throwing athlete and its impact on return to sport. Arthroscopy, sports medicine, and rehabilitation, 4(1), e83-e91.