Abstract
nical power may act as a key indicator for physiological and mechanical changes during running. In this scoping review, we examine the current evidences about the use of power output (PW) during endurance running and the di erent commercially available wearable sensors to assess PW. The Boolean phrases endurance OR submaximal NOT sprint AND running OR runner AND power OR power meter, were searched in PubMed, MEDLINE, and SCOPUS. Nineteen studies were nally selected for analysis. The current evidence about critical power and both power-time and power-duration relationships in running allow to provide coaches and practitioners a new promising setting for PW quanti cation with the use of wearable sensors. Some studies have assessed the validity and reliability of di erent available wearables for both kinematics parameters and PW when running but running power meters need further research before a de nitive conclusion regarding its validity and reliability. Keywords:biomechanics; endurance runners; long-distance athletes; wearable device 1. Introduction Endurance running events are on the apex of a performance revolution, with the sub-2-h marathon barrier just broken (i.e., Vienna in 2019). In the same way the power meter changed training and racing in cycling [1] by providing a fair tool to assess performance with accurate replication, it might also change the way runners compete and train. Power, a term originated in classical physics, is de ned as the product of force and velocity [2]. Despite training
broken (i.e., Vienna in 2019). In the same way the power meter changed training and racing in cycling [1] by providing a fair tool to assess performance with accurate replication, it might also change the way runners compete and train. Power, a term originated in classical physics, is de ned as the product of force and velocity [2]. Despite training delivers stress on the body, the way runners measure this level of stress has been very limited. The faster a runner goes, the higher the stress for a certain level of tness. Training intensity is the true marker to tness (i.e., capacity to deal with a particular amount of stress) [3]. The application of mechanical load (i.e., external training load factors) and psychological and physiological e orts (i.e., internal training load factors) are a ected by training stress [4]. In running, some external load factors including volume and pace are widely used, while physiological internal load factors consider perceived exertion scales, heart rate, or blood lactate level [4]. On multiple training days, running distance alone could overshadow the accumulated training stress and, eventually, misinterpret the overall training stress [4]. Pace might be as clear as volume but, indeed, it is not easy to assess as Sensors2020,20, 6482; doi:10.3390 /s20226482 /journal/sensors
Sensors2020,20, 6482 2 of 20 the running settings (i.e., surface; slope gradient) as well as weather conditions (i.e., wind velocity) or individual internal factors (i.e., stress, sleep, illness) may a ect pace considerably and, therefore, challenge pace intensity quanti cation. None of these variables provides a fair and repeatable method to measure training intensity and, when training stress is measured imprecisely, injury risk may be increased and performance negatively altered. Given that new wearable devices allow to measure external load metrics apart from both volume and pace, there should be a growing focus on a combination of both biomechanical external (i.e., power output (PW)) and internal load metrics in the future of athletes monitoring [4]. Running, as cycling, is cyclical in nature. When running, three dimensional movements are needed. Normally, the body describes a forward movement, vertical oscillation, and a bilateral rotation over the running cycle. For such movements, mechanical work is required accounting vertical and forward movements for most of it. Throughout such movements, a runner acquires both kinetic energy and potential energy changes. The applied work runners develop over the loading phase and the subsequent take-o push to lift their body at every stride to work against environmental factors (i.e., ground reaction force, gravity force, and surface) refers to the external mechanical work. Then, the foot absorbs energy when colliding with the ground and produces power when pushing o . During running, expensive equipment such as speci c instrumented treadmills [5] have been utilised to acquire force data. Despite their proved accuracy, most coaches and practitioners are forced to avoid their use due to economic issues. Over the last years, inertial measurement units (IMUs) emerged, allowing the quanti cation of performance, providing coaches and athletes an easy-to-use tool to monitor PW during running (e.g., Runscribe (Scribe Lab. Inc., Half Moon Bay, CA, USA), Stryd (Stryd Inc. Boulder, CO, USA) or Myotest (Myotest SA, Sion, Switzerland)). Previous works have demonstrated the direct relationship between anthropometric measures (e.g., body mass) and spatiotemporal parameters and kinetics and kinematics [68]. Samozino and colleagues [9] attempted to supply an a ordable method to
to monitor PW during running (e.g., Runscribe (Scribe Lab. Inc., Half Moon Bay, CA, USA), Stryd (Stryd Inc. Boulder, CO, USA) or Myotest (Myotest SA, Sion, Switzerland)). Previous works have demonstrated the direct relationship between anthropometric measures (e.g., body mass) and spatiotemporal parameters and kinetics and kinematics [68]. Samozino and colleagues [9] attempted to supply an a ordable method to assess force-velocity and power-velocity pro les, using anthropometric and spatiotemporal data along over-ground sprint acceleration. However, Samozino's approach is inapplicable to submaximal velocities. Currently, an increasing number of systems allow the assessment of running power (new heart rate monitors by Polar (Polar Electro Ltd., Kempele, Finland) and Garmin (Garmin Ltd., Olathe, KS, USA)). Nevertheless, there is a lack of scienti c evidence testing either its validity or reliability, as well as limited insights on the use and interpretation of power in endurance runners, being this reduced to a few books [3,10], and further information provided by the devices' manufacturers (e.g., Stryd, https://blog.stryd.com/tag/validation-white-papers/; Myotest, //www.myotest.com/technology; RunScribe, //runscribe.com/blog/; Stryd, //blog.stryd.com; Polar: //www.polar.com/es/ smart-coaching/running-power). Although the validity and reliability of a wide array of wearable sensors have been shown for running spatiotemporal parameters measurement and they seem to be related with PW estimation [1115] , a deeper knowledge on PW in endurance running and a proper understanding on the use of power meters to quantify workload would be an outstanding step forward towards a new boundary within running training and performance. There is a need to measure training intensity with precision and wearable sensors might help monitor the training-induced stress and, although previous review articles have been focused on power data while running [16,17], none of those concentrated on validity and reliability of such wearables for running PW analysis. Advances in the knowledge of endurance running PW would allow the assessment and monitor of power not only in laboratory settings, but in the eld as well. Therefore, the aim of this scoping review was to critically examine the available running power meters and the current evidences about their use and application to endurance running performance.
of endurance running PW would allow the assessment and monitor of power not only in laboratory settings, but in the eld as well. Therefore, the aim of this scoping review was to critically examine the available running power meters and the current evidences about their use and application to endurance running performance.
Sensors2020,20, 6482 3 of 20 2. Materials and Methods A review of the literature was conducted following the guidelines of the Cochrane Collaboration and taking into consideration the guidance provided by previous studies focused on scoping reviews [18,19] . This design (i.e., scoping review) was selected in order to have a broader approach with the aim of mapping literature characterized by a variety of study designs. Additionally, ndings were reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) for scoping reviews [20]. 2.1. Eligibility Criteria Despite the limited evidence on this topic, some a priori inclusion criteria were considered for this scoping review: (i) only peer-reviewed articles were included; (ii) studies that were not published in English were not explored; (iii) no restrictions for age or sex of participants were applied. Additionally, no limitations regarding the study design were established. All manuscripts related to running with power or power meters were considered, regardless the study design, except literature reviews (e.g., systematic reviews or metanalysis). 2.2. Information sources A systematic search was conducted in the electronic databases PubMed, MEDLINE and SCOPUS for relevant studies until 1 June 2020. Keywords were collected through experts' opinion, a systematic literature review, and controlled vocabulary (e.g., Medical Subject Headings: MeSH). Boolean search syntax using the operators AND and OR was applied. The words endurance, running, runner, power, and power meter were used. Following is an example of a PubMed search: ((((((endurance) OR submaximal) NOT sprint) AND running) OR runner) AND power) OR power meter; Filters: Publication date from 1 January 2000; Humans; English. After an initial search, accounts were created in the respective databases. Through these accounts, the lead investigator received automatically generated emails for updates regarding the search terms used. These updates were received on a daily basis (if available), and studies were eligible for inclusion until the initiation of manuscript preparation on 5 June 2020. Following the formal systematic searches, additional hand-searches were conducted. Grey literature sources (e.g., conference proceedings) were also considered if a full-text version was available. In addition, the reference lists of included studies and
used. These updates were received on a daily basis (if available), and studies were eligible for inclusion until the initiation of manuscript preparation on 5 June 2020. Following the formal systematic searches, additional hand-searches were conducted. Grey literature sources (e.g., conference proceedings) were also considered if a full-text version was available. In addition, the reference lists of included studies and previous reviews and meta-analyses were examined to detect studies potentially eligible for inclusion. 2.3. Study Selection In selecting studies for inclusion, the three-step method was followed [21]. The rst step, according to this procedure, was an initial restricted search of the appropriate database collection, followed by an analysis of the text words included in the title and abstract, and the index terms used to characterize the document. A second search using all known keywords and index terms was performed through all included databases. Finally, the reference list of all the selected studies and reports has been checked for additional studies. The authors included the aforementioned lters (i.e., the language and the publication date limitations). 2.4. Methodological Quality in Individual Studies To analyse the methodological quality in studies, the recommendations by Cochrane Review Groups were taken into consideration [22]. Since all the studies examined show a cross-sectional design, quality was assessed using the modi ed version of the Quality Index developed by Downs and Black [23]. The original scale was reported to have good testretest (r=0.88) and inter-rater (r=0.75) reliability and high internal consistency (KuderRichardson Formula 20 (KR-20)=0.89). The modi ed version of the Downs and Black Quality Index is scored from 1 to 14, with higher scores indicating
Sensors2020,20, 6482 4 of 20 higher-quality studies. Two independent reviewers (DJC-FGP) performed this process and, in the event of a disagreement about the methodological quality, a third reviewer (LERS) checked the data and took the nal decision on it. Agreement between reviewers was assessed using a Kappa correlation for methodological quality. The agreement rate between reviewers wask=0.93 which can be interpreted as almost perfect [24]. It is worth noting that the study by Snyder and colleagues [25] was excluded as it is a letter to the editor in response to Aubry and colleagues' [26] work. 3. Results 3.1. Study Selection Figure were initially identi ed: 640 from PubMed, 378 from SCOPUS, and 263 from MEDLINE. Additionally, 6 studies were identi ed through other resources. From these 1287 studies, 674 after duplicates removed. The 613 studies excluded after titles and abstracts revisions were essentially based on a lack of relationship with the research interests of this review. After full-text revision, only 19 studies which included either validity or reliability of running wearable sensors suppling running PW and/or the speci c discussion of such wearable sensors were considered for the current work.Sensors 2020, 20, x FOR PEER REVIEW 4 of 21 3. Results 3.1. Study Selection Figure 1 provides a graphical schematization of the study selection process. A total of 1281 studies were initially identified: 640 from PubMed, 378 from SCOPUS, and 263 from MEDLINE. Additionally, 6 studies were identified through other resources. From these 1287 studies, 674 after duplicates removed. The 613 studies excluded after titles and abstracts revisions were essentially based on a lack of relationship with the research interests of this review. After full-text revision, only 19 studies which included either validity or reliability of running wearable sensors suppling running PW and/or the specific discussion of such wearable sensors were considered for the current work. Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram. 3.2. Study Characteristics The main characteristics of the studies included in this review (n = 19) are presented in the Tables 1 and 2. Table 1 shows a summary of 12
running PW and/or the specific discussion of such wearable sensors were considered for the current work. Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram. 3.2. Study Characteristics The main characteristics of the studies included in this review (n = 19) are presented in the Tables 1 and 2. Table 1 shows a summary of 12 studies using wearable sensors with the capacity of measuring power during different running exercises. Whereas three of those studies [11,27,28] examine the PW kinetics during different running protocols, the other four studies [15,25,26,29] investigate the relationship between PW and physiological parameters such as oxygen consumption (VO 2) at different intensities. Additionally, two further works [30,31] analyse the application of mathematical models, based on power laws, to predict running performance, whereas a recent study [32] assesses the agreement level between two mathematical models and five power meter devices through different running conditions. Other studies examined some parameters provided by the RunScribe power meter to describe the effects of the fatigue induced over a marathon [33,34] and the influence of different types of ankle treatments on running biomechanics [35]. Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) ow diagram.
Sensors2020,20, 6482 5 of 20 3.2. Study Characteristics The main characteristics of the studies included in this review (n=19) are presented in the Tables. Table measuring power during di erent running exercises. Whereas three of those studies [11,27,28] examine the PW kinetics during di erent running protocols, the other four studies [15,25,26,29] investigate the relationship between PW and physiological parameters such as oxygen consumption (VO2) at di erent intensities. Additionally, two further works [30,31] analyse the application of mathematical models, based on power laws, to predict running performance, whereas a recent study [32] assesses the agreement level between two mathematical models and ve power meter devices through di erent running conditions. Other studies examined some parameters provided by the RunScribe power meter to describe the e ects of the fatigue induced over a marathon [33,34] and the in uence of di erent types of ankle treatments on running biomechanics [35].
Sensors2020,20, 6482 6 of 20 Table 1.Studies (n=12) involving the use of wearable sensors with the capacity of measuring power during running protocols. References Subject Description Aim System Used Protocol Outcome Measures Results Dobrijevic et al. (2017) [15] 30 physical education students (15 men and 15 women) To explore the properties of the F-V relationship of leg muscles exerting the maximum pulling F at a wide range of V on a standard motorized treadmill Motorized treadmill using externally xed strain gauge dynamometer (CZL301, ALL4GYM, Serbia) connected to the subject wearing a wide and hard weightlifting belt Walking and running on a treadmill at di erent velocities (1.4 3.3 m.s 1 ), and maximum pulling F exerted horizontally were recorded Leg muscle capacities for producing maximum F, V, and power The F-V relationship of leg muscles tested through a wide range of treadmill V could be strong, linear, and reliable. Moreover, the two-velocity method could provide reliable and ecologically valid indices of F, V, and P producing capacities of leg muscles. Garc½a-Pinillos et al. (2019) [17] 49 endurance runners To examine how the PW changes while running at a continuous comfortable velocity on a motorized treadmill by comparing running power averaged during di erent time intervals Stryd system (foot pod) Runners performed a 3 min running protocol at comfortable velocity and P was examined over six recording intervals within the 3-min recording period: 0 10 s, 0 20 s, 0 30 s, 0 60 s, 0 120 s and 0 180 s Running PW P during running is a stable metric with negligible di erences, in practical terms, between shorter (i.e., 10, 20, 30, 60 or 120 s) and longer recording intervals (i.e., 180 s) Aubry et al. (2018) [14] 24 male runners (13 recreational, 11 elite) To investigate the applicability of running power (and its individually calculated run mechanics) to be a useful surrogate of metabolic demand (Vo 2), across di erent running surfaces, within di erent caliber runners. - Stryd system (chest strap) - Gas exchange measures (Cosmed Quark CPET and Cosmed K5 systems) 2 di erent test at
male runners (13 recreational, 11 elite) To investigate the applicability of running power (and its individually calculated run mechanics) to be a useful surrogate of metabolic demand (Vo 2), across di erent running surfaces, within di erent caliber runners. - Stryd system (chest strap) - Gas exchange measures (Cosmed Quark CPET and Cosmed K5 systems) 2 di erent test at 3 di erent paces, while wearing a Stryd on both an indoor and an outdoor test: -Treadmill vO 2test: running at 3 speeds for 2 min each -Outdoor vO 2test (on track): identical speeds for 4 min (1 min rest) - Spatiotemporal parameters - Running PW - vO 2 Running power (with Stryd) is not a great re ection of the metabolic demand of running in a mixed ability population of runners Snyder et al. (2017) [13] Manuscript clari cation: Request for clari cation to Aubry et al. (2018) Some major methodological aws in the mentioned paper are detected. The authors concluded that data analysis and, thereby, data interpretation are misleading Austin et al. (2018) [18] 17 well-trained distance runners To measure the correlations between running economy and P and form power at LT pace. - Stryd system (foot pod) - Gas exchange measures (Parvo Medics TrueOne 2400) Participants ran two 4 min trials: one with a self-selected cadence, and one with a target cadence lowered by 10% - Gas exchange measures - RPE - Power - Form power - SF RE is positively correlated with Stryd's power and form power measures yet the footpod may not be su ciently accurate to estimate di erences in the running economy of runners Garc½a-Pinillos et al. (2019) [36] 18 recreationally-trained male endurance runners To determine if the P-V relationship in endurance runners ts a linear model when running at submaximal velocities, as well as to examine the feasibility of the two-point method for estimating P at di erent velocities Stryd system (foot pod) Incremental running protocol on a treadmill. Initial speed was set at 8 km.h 1 , and speed increased by 1 km.h 1 every 3 min until exhaustion PW (W)
Description
A review of wearable sensors for measuring power output in endurance running.