In this blog, I explore the nuances of assessing load and damage at common running injury locations. We’ll examine the limitations of using external loads like ground reaction forces (GRFs) as proxies for internal tissue loading, delve into the role of musculoskeletal modelling in assessing tissue load, and discuss the pitfalls of using cumulative loading measures as a measure of injury risk. Finally, we’ll explore how wearable technology, such as instrumented insoles, can revolutionise the way by which we estimate tissue loading and damage, hereby offering new insights into injury prevention, conservative rehabilitation protocols, and performance optimisation.
External Loads vs. Internal Tissue Loads: A Complex Relationship
When runners strike the ground, the ground reacts with a force equal in magnitude and direction (i.e., Newton’s 3rd law of motion). This reaction force can be measured using force platforms. These so-called ground reaction forces are often considered indicative of the stress placed on the body during running. However, the relationship between these external forces and the internal mechanical loads experienced by tissues is far from straightforward. Indeed, GRFs provide valuable information about the overall external mechanical load during running, alongside being a critical determinant of performance across human locomotion, inclusive of running, sprinting, decelerating and changing direction. With that said, GRFs do not directly correspond to the loads experienced by specific tissues within the body, i.e., internal loads [1, 2]. This discrepancy arises because the internal forces depend on numerous factors, including muscle activity (e.g., co-activation), joint kinematics, and individual muscle moment arms. A similar discrepancy arises between net joint moments and internal tissue loading. This can be understood using a simple example: consider two individuals (with the same anthropometrics and lever arms) that are performing an isometric squat, but one individual voluntarily co-activates their quadriceps and hamstrings and thus experiences a higher patellofemoral contact force. However, the joint moment and external GRF is still the same and these do thus not accurately reflect the increased internal loading.
As a result, relying solely on external biomechanics like GRFs, joint moments or tibial accelerations [3] to assess injury risk can be misleading (Figure 1)! High magnitudes in external biomechanics do not necessarily translate to high internal tissue loading, and vice versa. This disconnect underscores the requirement for more sophisticated approaches to evaluating tissue-specific loads.
Musculoskeletal Modelling: A Window into Internal Tissue Loads
Musculoskeletal modelling offers a powerful tool for bridging the gap between external forces and internal tissue loads. By simulating the body’s musculoskeletal system, these models can estimate the forces experienced by individual muscles, tendons, and bones during running. This approach allows researchers to assess internal tissue loading with far greater precision than is possible through external biomechanics alone.
Musculoskeletal models are computational representations of the human body, incorporating data on muscle geometry, joint kinematics, and even neuromuscular control [4, 5] (Figure 2). These models use inputs like 3D motion capture data, GRFs, and electromyography (EMG) signals to estimate internal tissue forces within the musculoskeletal system.
One of the primary benefits of musculoskeletal modelling is its ability to provide detailed insights into the internal loading of specific tissues, without need for invasive measurements. For example, these models can estimate the forces experienced by the Achilles tendon during different phases of the running cycle, helping to identify periods of high loading that may contribute to injury risk. Musculoskeletal models also allow for the exploration of “what-if” scenarios, such as how changes in running technique or footwear might affect tissue loading. This capability is invaluable for designing personalised injury prevention strategies and optimising performance. Indeed, I have used musculoskeletal modelling to investigate how the force and damage (see later) change with increases in running speed, surface gradient and step frequency [6]. The results of this study can be seen in Figure 3. Briefly, increases in speed generally increased the damage on all three structures (patellofemoral joint, tibia, and Achilles tendon) and changes in surface gradient had variable effects, with increases in one structure but decreases in other structures. Finally, increases in step frequency decreased the damage on the three common running injury locations investigated. I have also used the same approach to investigate the muscle forces of the hamstrings during popular hamstring strength training exercises [7].
Load versus damage
When estimating the load on internal tissues, researchers and clinicians often use the so-called impulse. It is defined as the product of force and the time over which it acts and thus reflects the area under a waveform curve (Figure 4). The reason for its popularity is that is accounts for both the magnitude and the duration of a force.
However, using impulse as a proxy for injury risk can be problematic due to the non-linear relationship between load and tissue damage [6, 8]. Indeed, a high peak force applied over a short duration (as in a sudden, heavy impact) causes more damage than a lower force applied over a longer duration, even if the overall impulse is the same. This is easy to understand using an analogy: if you hit a bone very hard once with a hammer, it will break. But if you hit it 1000 times with a small impulse, but such that that cumulative (total) impulse over these 1000 hits is the same as the one big hit, the cumulative (total) damage will be much smaller. This non-linear relationship means that two steps with identical impulses could have vastly different injury risk implications, depending on how the force is applied. As a result, using impulse alone as a measure of injury risk may lead to misleading conclusions regarding internal tissue loading, and subsequently, implications of relative injury risk [6]. This is important to remember when reading papers that use impulse! In this context, it is important to recognise that the impulse can be useful to better understand other aspects of movement such as between-leg imbalances in force application.
There are ways to better account for the non-linear relationship between load and damage, for example using a weighted impulse. Although it is beyond the scope of this blog to explain this in detail, here is a short explanation: With the weighted impulse, each force value is first exponentiated using an empirically derived value, so that higher values get much more relevance in the weighted impulse, see Figure 4. Indeed, with this approach high peak forces have more relevance than lower values. For more details regarding the exponentially weighted impulse, I would advise the reader to this work [6, 8].
A final thing we must account for is the number of steps an athlete performs to complete a given distance: This is called the cumulative impulse, or cumulative weighted impulse (damage proxy). With this approach we have shown that increases in running speed increase the cumulative damage on the patellofemoral joint and Achilles tendon, with a smaller increase for the tibia (Figure 3) [6]. Therefore, even though an athlete is taking less steps to complete a given distance at higher speeds, the damage increases because each step has considerably more damage and this increased damage per step outweighs the decrease in the number of steps. This not only applies to constant speed running, but also to accelerations and decelerations. The cumulative impulse without this exponential weighting, however, would suggest that load decreases and thus misinform on injury risk (Figure 3)! Overall, these findings suggest that care should be taken when progressing movement velocity (and rate of change in velocity) to minimise injury risk. Indeed, for some structures, increases in distance may be better than increases in speed after rehabilitation of an injury due to the non-linear relationship between loading and damage.
Wearable Technology: Estimating Tissue Loading and Damage in Real-Time
While musculoskeletal modelling is a nice tool to estimate tissue load and damage, it is very time consuming and bound to a lab. Advancements in wearable technology have however opened new avenues for assessing estimated tissue loading and potential injury risk in real-time and in-field. These devices such as instrumented insoles and wearable sensors, provide continuous data on running biomechanics, enabling more accurate estimates of tissue loading and potential damage.
Instrumented insoles are one of the most promising wearable technologies for assessing tissue load in runners. These insoles are equipped with sensors that measure various parameters such as the pressure distribution, magnitude, temporal characteristics of foot contact and sometimes even accelerations and orientations (from an IMU). However, as explained before, these external biomechanics do not reflect internal tissue loading very well. We have overcome this limitation by combining musculoskeletal modelling with machine learning and wearable insoles as follows [9]: We estimated the tissue load and damage at three common injury locations using musculoskeletal modelling during various running conditions. At the same time, runners wore instrumented insoles and this insole data was used as input to an artificial neural network to predict the internal tissue loads estimated with the musculoskeletal modelling approach. We showed this approach to yield much better accuracy in estimating tissue loading than approaches that for example estimate tissue loading based on the distance ran or steps taken (Figure 4). Importantly, at the time of the study, the insoles were also commercially available and they will be re-introduced again soon so everybody can access this information.
While more work is required to further fine-tune these insoles (and other wearables), the ability to assess tissue loading in real-time has significant implications for both injury prevention, rehabilitation, and performance optimisation in runners. By providing immediate feedback on biomechanical parameters, including tissue loading, wearable devices can help runners and athletes in other sports identify and correct potentially harmful patterns before they lead to injury. Indeed, we have shown a significant reduction in injury risk when runners received real-time wearable-based feedback on estimated tissue loading [10].
Wearable technology also offers the potential for highly personalized training programmes. By continuously monitoring loading/damage, these devices can help runners periodise and optimise their training loads to match their individual capacity and goals. For example, an athlete recovering from an injury might use wearable technology to ensure that their training loads remain within safe and conservative limits, reducing the risk of re-injury. However, before such an approach is useful, wearable devices need to be able to estimate tissue capacity. We are currently working on ways to integrate this within wearable technology.
Conclusion
There are a few take home points from this blog worth repeating:
- External biomechanics do not always reflect internal biomechanics well;
- The impulse does not account for the non-linear relationship between load and damage and may misinform on potential running-related injury risk;
- Musculoskeletal modelling can be used to estimate internal tissue loading and damage, but is time consuming and generally restricted to laboratory environments;
- Wearables can be used to estimate tissue loading and damage in-field in real-time, but require further refinement and – for some applications – a combination with knowledge on tissue capacity.
References
1. Matijevich, E.S., et al., Ground reaction force metrics are not strongly correlated with tibial bone load when running across speeds and slopes: Implications for science, sport and wearable tech. PloS one, 2019. 14(1): p. e0210000.
2. Scott, S.H. and D.A. Winter, Internal forces of chronic running injury sites. Med Sci Sports Exerc, 1990. 22(3): p. 357-69.
3. Zandbergen, M.A., et al., Peak tibial acceleration should not be used as indicator of tibial bone loading during running. Sports Biomechanics, 2023: p. 1-18.
4. Lloyd, D.G. and T.F. Besier, An EMG-driven musculoskeletal model to estimate muscle forces and knee joint moments in vivo. Journal of biomechanics, 2003. 36(6): p. 765-776.
5. Delp, S.L., et al., OpenSim: open-source software to create and analyze dynamic simulations of movement. IEEE Trans Biomed Eng, 2007. 54(11): p. 1940-50.
6. Van Hooren, B., L. Van Rengs, and K. Meijer, Per-step and cumulative load at three common running injury locations: the effect of speed, surface gradient and cadence. Scandinavian Journal of Medicine and Science in Sports, 2024. 34(2): p. e14570.
7. Van Hooren, B., et al., Muscle forces and fascicle behavior during three hamstring exercises. Scand J Med Sci Sports, 2022. 32(6): p. 997-1012.
8. Edwards, W.B., Modeling overuse injuries in sport as a mechanical fatigue phenomenon. Exercise and sport sciences reviews, 2018. 46(4): p. 224-231.
9. Van Hooren, B., L. van Rengs, and K. Meijer, Predicting Musculoskeletal Loading at Common Running Injury Locations using Machine Learning and Instrumented Insoles. Med Sci Sports Exerc, 2024.
10. Van Hooren, B., G. Plasqui, and K. Meijer, The Effect of Wearable-Based Real-Time Feedback on Running Injuries and Running Performance: A Randomized Controlled Trial. Am J Sports Med, 2024. 52(3): p. 750-765.
