Comparison of different measures to monitor week-to-week changes in training load in high school runners

Abstract

Training load is commonly used to monitor training stress and is the product of external and internal physiological loads experienced by an athlete. With emerging wearable technology, it is possible to evolve existing external load measurement from duration or distance to runner-specific biomechanical data, which when combined with existing measures of internal load such as session rating of perceived exertion (sRPE), may improve the quantification of training stress. This study compared week-to-week changes in training duration with different training loads obtained from common and more individualized measures of external load. The training of nine male high school cross-country runners from the same team undertaking the same training program was monitored for two consecutive weeks. This two-week cycle included a “coach-prescribed” low and high training load week. Training loads were calculated with sRPE and external load measures including: duration (minutes), Step Count, “Bone Stimulus” (IMeasureU), and cumulative vertical force. Weekly distances (in miles) were also measured. Between-week percent change (%Δ) was compared among training loads and minutes using paired t-tests and Cohen’s d effect size. Different %Δ were found between sRPExMinutes (%Δ=65 ± 25%; p = 0.002, d = 1.83), sRPExStep Count (%Δ=66 ± 31%; p = 0.006, d = 2.06), sRPExForce (%Δ=66 ± 29%; p = 0.002, d = 1.91), and miles (%Δ=28 ± 13%; p = 0.019, d = 0.71) compared to minutes (%Δ=20 ± 8%). These findings highlight that only using weekly volume can greatly misrepresent changes in training stress in runners. We therefore encourage coaches and practitioners to consider training monitoring approaches beyond just weekly distance or duration. Simple measures of training load that include duration and sRPE might be sufficient.

Keywords

Cross-country coaching,endurance running,inertial measurement unit,rating of perceived exertion,youth sport

Introduction

Running training can be manipulated by changing training volume, intensity, frequency, and density (i.e., frequency of intense training) to improve fitness and ultimately, running performance.^1,2 By manipulating these training variables, coaches aim to produce a specific training stress in their runners.³ Historically, most runners and coaches have used external load factors such as volume (e.g., distance or duration) and intensity (e.g., pace) to quantity running training as these methods are engrained in the running culture.^4–7 However, there may be more appropriate and effective methods to monitor training stress in runners. A better understanding of and improvements in methods to quantify training stress should help coaches optimize running performance. For decades, sport scientists and coaches have used internal physiological load factors such as rating of perceived exertion or average heart rate as a measure of the physiological response to external loads.⁶ Training load, defined as the product of external and internal physiological loads experienced by an athlete, is a popular monitoring approach among coaches and sport scientists to monitor training stress.^7,8 Monitoring training load between training periods or cycles in runners might provide a better quantification of training stress to enable the guidance of necessary training adjustments to ensure optimal performance and minimize risk of over- or under-training.⁹ Further, small changes in training load in runners can lead to improvements in running pace in long races¹⁰ and thus, should be considered by running coaches. This is expected since it has been suggested that the internal load is the primary stimulus for adaptation in response to running training.¹¹ One of the current challenges of using training load is identifying which external and internal load factors provide the most useful and practical information for coaches and sport scientists.

Heart rate (HR) is commonly used by distance runners to assess internal load since it can be easily measured with chest- or wrist-worn instruments. However, it may not be related to changes in endurance performance, and may not be appropriate to monitor periods of intense training.¹² Session rating of perceived exertion (sRPE), the global RPE for an entire training session, encompasses both the psychological and physiological response to training⁶ and has been used to assess internal load since it is highly correlated with blood lactate levels even as training intensifies.^13,14 In fact, the combination of sRPE and training duration to measure training load is correlated with blood lactate measured 30 minutes following maximal training sessions while training impulse (TRIMP) measured from HR intensity distribution is not.¹⁴ Therefore, the combination of sRPE and duration to measure training load is not only convenient but might also be a more appropriate method to assess training stress compared to other methods requiring measurements of HR (e.g., TRIMP). Running session duration, distance, and average pace are the most common methods to monitor external load since coaches and runners can monitor these measures with a global positioning system (GPS) wristwatch. These measures of monitoring external training loads are simple, convenient, and cost effective making them valuable for coaches and runners to monitor training. Although these measures are valuable and practical when measuring training loads, they have a number of limitations⁶ as they do not necessarily reflect the specific and/or individual biomechanical demands of running. With the emergence of wearable technology, it is possible to measure runner-specific biomechanical data without the need for advanced and costly laboratory instruments. For example, wireless inertial measurement units (IMUs) secured to the distal tibia can measure step-by-step axial and resultant peak tibial accelerations,¹⁵ while wireless vertical force insoles can collect peak vertical force and vertical impulse.¹⁶ Peak tibial acceleration^17–19 and peak vertical force^20–22 are common measures of external load in the running biomechanics literature and, have been retrospectively and prospectively related to running injuries.^19,23–25 Thus, the combination of these biomechanically-specific external load measures and the commonly used internal load measure of sRPE might provide more individualized approaches to quantify training stress in runners. In fact, the combination of sRPE with external load measures provides a more individualized measure of average week-to-week changes in training stress in a broad population of runners following a self-directed program.²⁶ Training-intensity distribution (i.e., time spent within specific intensity zones) is another approach used to quantify running training but input intensity metrics (e.g., HR, speed, and RPE) greatly influence the resulting training-intensity distribution in middle-distance runners.⁴ Similarly, the use of biomechanically-specific external load measures could yield different week-to-week changes in training load in a homogenous group of runners who have been prescribed the same training program.

The primary purpose of this study was to compare week-to-week percent changes in coach-prescribed training duration with training load measures obtained from sRPE and different measures of external load in high school male runners. We hypothesized that with the inclusion of sRPE, the different training load measures would yield different week-to-week changes in training compared to training duration alone. A secondary observational purpose of this study was to assess individual runner responses in between-week changes among different training load measures. We expected a wider range of responses in week-to-week changes in training load compared to duration or distance alone. In testing these hypotheses, we aimed to provide evidence for coaches and practitioners to think beyond training duration or distance and to more closely consider individual responses when monitoring running training.

Methods

Participants

A convenience sample of twelve male high school (14-18 years) cross-country runners from the same team were recruited for the study. Runners were excluded if they sustained a lower extremity injury within the past six months. Injuries were defined as events that prevented training for at least seven days, or three consecutive scheduled training sessions.²⁷ All runners and their parents were informed verbally, and in writing, of all testing procedures, potential risks, and benefits of the study. Participants signed an assent form while their parents signed a consent form for procedures approved by the Institutional Review Board for Human Participants Research.

Procedures

Collection of training-related data over a two-week period occurred during the middle of the cross-country season and all running session data were collected on the high school campus or training locations during scheduled team sessions. This two-week data collection period occurred during a low training load week followed by a high training load week as planned and prescribed by the team coach (Table 1). To achieve a higher training load during the second week, the training prescription for week 2 included 1) an increase in weekly volume, and 2) a slight increase in overall intensity. All runners completed their specific coach-prescribed training duration and training intensity, the latter being prescribed using perceived effort and not pace or speed.

Table 1.

Training details regarding the two-week coach-prescribed low and high training weeks for one runner.

Week 1: Total volume: ∼175 minutes (5–6 runs and/or cross-training)
Monday	OFF
Tuesday	30–35 minutes
Wednesday	5-minute tempo with 3 minutes rest + 6–8 x 400 m at mile effort pace with 90 seconds rest + drills
Thursday	30–40 minutes
Friday	35–45 minutes
Saturday	Long Run 25% of your weekly volume (∼45min)
Sunday	Shakeout Day 15–20 minutes
Week 2: Total volume: ∼240 minutes (5–6 runs and/or cross-training)
Monday	OFF
Tuesday	35–40 minutes
Wednesday	Tempo (30 minutes total)
Thursday	30–40 minutes
Friday	6 x 100 m fast strides + easy volume + drills
Saturday	4–6 x 1 km at 3000 m effort pace with 90 seconds rest (45 minutes total)
Sunday	Long Run 25% of weekly volume (∼60 min)

Session internal and external load measures were obtained daily. Internal physiological load was monitored using sRPE on a 10-point numerical rating scale (Figure 1) within 10 minutes of the end of each running session.²⁸ Although the use of sRPE has many limitations,²⁹ it is important to note that these runners had been using the specific rating scale used in this study with their team coach, and were therefore accustomed to the scale to rate sRPE consistently. External loads measured included session 1) run duration, 2) run mileage, 3) Step Count, 4) “Bone Stimulus” (iMeasureU, VICON, USA), and 5) estimated cumulative vertical force. Running minutes or miles per training session were recorded using the runners’ own GPS wristwatches (different models and brands). It is important to note that training duration (minutes) was the coach-prescribed training metric. Step Count was measured as the product of session minutes and average session cadence (1000 Hz, Blue Thunder, IMeasureU/VICON, USA). “Bone Stimulus” was collected using the IMU (Blue Thunder) and calculated from the IMeasureU Step software³⁰ during each training session using bilateral IMUs secured to the distal third of the medial tibias before the start of each run (Figure 2). Vertical force insoles (200 Hz, Loadsol, Novel, USA) were used to obtained estimated cumulative force per session. Running session cumulative force was estimated due to the difficulty in collecting force insole data in multiple athletes over extended training durations. Estimated cumulative force was individualized for each runner based on treadmill testing completed prior to testing. In this testing session, following a self-selected warm up athletes completed one minute intervals at three coach selected paces which are aligned to training sessions used throughout the training analysis: 1) easy running pace, 2) tempo pace, and 3) 5 km race pace. Average peak vertical force from insoles were collected from the final 30 seconds of these running bouts. These testing sessions took place at the team’s high school training facility and all testing was completed in the runners’ own training footwear

Figure 1.

Session rating of perceived exertion (sRPE) scale used in the current study. Note that this sRPE scale was regularly used by the team coach for training monitoring.

Figure 2.

Position of inertial measurement unit on distal tibial.

Data analyses

Session RPE (i.e., internal load) was multiplied with four external load measures (minutes, Step Count, “Bone Stimulus”, and estimated cumulative vertical force) to obtain four different training load measures. In this study, the product of sRPE and training duration in minutes (sRPExMinutes) is considered a “simple” method to quantify training load as it only requires a visual scale and a stopwatch. To obtain more “complex” or more biomechanically-specific measures of training load, sRPE was multiplied with Step Count, “Bone Stimulus”, and the estimated cumulative force obtained from the advanced instruments. The use of bilateral “Bone Stimulus” and peak vertical force was expected to provide more biomechanically-specific and individualized measures of external load to calculate training load. From the manufacturer site, “Bone Stimulus” is dependent on “total number of steps taken” and “the size of the impact derived from each individual step”.³¹ However, “Bone Stimulus” is “more responsive to the size of step impact rather than the number of steps taken” which means that high impact steps contribute more to “Bone Stimulus” than low impact steps.³² Finally, the average peak vertical force data from the treadmill testing session at the beginning of the collection period were used to estimate session cumulative force. Since the relationship between peak vertical force and running speed can be defined well with a second order polynomial,³³ a trend line from a second order polynomial from the average peak vertical force versus running speed curve was constructed to estimate the corresponding average peak vertical force at session average run paces for each runner. Figure 3 provides an example of a average peak vertical force and speed curve fitted with a polynomial trend line. For each run, the average force per step was estimated based on the corresponding average speed on the average peak vertical force vs speed curve for each runner. Then estimated cumulative force was obtained by multiplying daily run step count and average force per step. Each of the six different measures to quantify running training (1. Minutes, 2. Miles, 3. sRPExMinutes, 4. sRPExStep Count 5. sRPEx“Bone Stimulus”, 6. sRPExForce) was used to determine the week-to-week percent change (%Δ) in running training for each measure during the two-week collection period. Training distance in miles was included as a training metric since it is a commonly used measure of monitoring running training.

Figure 3.

Example of a participant’s average peak vertical force and running speed (i.e. runner-specific easy, tempo, and 5 K race pace) curve fitted with a second order polynomial trend line to identify a vertical force value corresponding to the average session running speed.

Statistical analysis

Paired t-tests were used to compare between-week percent change (%Δ) in running training quantified using minutes run with other training load measures, as well as miles (p ≤ 0.05). Cohen’s d effect sizes were also computed to assess the magnitude of mean differences (small = 0.2; medium = 0.5; large = 0,8; very large = 1.3).³⁴

Results

Of the twelve male runners who began the study, data from three runners were not included in analyses due to missing data collection during the two-week period (high school Fall break).

Between-week %Δ differed from minutes (%Δ=20 ± 8%) for sRPExMinutes (%Δ=65 ± 25%; p = 0.002; d = 1.83), sRPExForce (%Δ=66 ± 29%; p = 0.004; d = 1.91), sRPExStep Count (%Δ=66 ± 31%; p = 0.006; d = 2.06), and miles (%Δ=28 ± 13%; p = 0.019; d = 0.71) while there was no difference in between-week %Δ for sRPEx“Bone Stimulus” (%Δ=41 ± 29%; p = 0.113; d = 0.74) compared to minutes (Figure 4).

Figure 4.

Box plots of week-to-week percent change (%Δ) among the different methods of quantifying running training. Force (estimated weekly cumulative peak vertical force from wireless force insole peak vertical force and session cadence), “Bone Stimulus” (proprietary metric from IMeasureU IMU-Step software), Step Count (minutes x average cadence), and Minutes were used as the external load measures in combination with sRPE to calculate four different training load metrics. Weekly Miles were also monitored and compared to Minutes. *: different than Minutes (i.e., the coach-prescribed training metric) (p ≤ 0.05). Box plots include the maximum and minimum values, first and third quartiles, and median.

Of the external load measures, between-week %Δ “Bone Stimulus” (p < 0.001; d = 2.85) and miles (p = 0.019; d = 0.71) were different than %Δ in minutes while between-week %Δ Step Count (p = 0.490; d = 0.12) and Force (p = 0.662; d = 0.17) were not different than between-week %Δ coach-prescribed minutes (Table 2). Average weekly running pace was faster for week 2 (8min34s per mile) compared to week 1 (8min1s per mile; p = 0.009; d = 0.89).

Table 2.

Average weekly external loads for weeks 1 and 2 (mean±SD), and week-to-week percent change for each external load measure.

External loads	Week 1	Week 2	% Change
Minutes	194.2 ± 31.2	234.2 ± 43.1	20%
Miles	22.8 ± 4.2	29.3 ± 5.4	28%^a
Step Count	31,081 ± 4368	36,971 ± 6674	19%
“Bone Stimulus” (unitless)	3,191 ± 479	3,262 ± 503	2%^a
Cumulative Force (10⁷ N)	4.26 ± 0.60	5.22 ± 1.11	22%

% Change: week-to-week percent change.

^aDifferent than Minutes (i.e., the coach-prescribed training metric) (p ≤ 0.05).

When sRPE was combined with the external load measures, there were larger ranges in individual responses to week-to-week %Δ in training load measures (76% to 97%) compared to the two common measures to quantify training in runners: minutes (27%) and miles (41%) (Figure 5). The standard deviation among training load measures was between 25–31% while minutes run (8%) and miles (13%) had smaller standard deviations, emphasizing the wider range among training loads.

Figure 5.

Individual runner week 1 to week 2 percent change (%Δ) responses for common external load measure (Minutes and Miles) and all four different training load measures. Week 1 was normalized to 0%. This figure illustrates the larger range of individual responses when using training load measures (lower block of figures) compared to external load measures (upper block of figures).

Discussion

The primary purpose of this study was to compare week-to-week changes in training duration with different training loads obtained from common and more biomechanically-specific measures of external load during coach-prescribed weeks of low and high training load in high school cross-country runners. Our primary hypothesis that the different training load measures would yield different week-to-week percent changes and direction of change (e.g., increase or decrease) compared to training duration alone, was supported. Week-to-week %Δ was larger for all training load measures (which includes the internal physiological load of sRPE), except for sRPEx“Bone Stimulus”, compared to the %Δ in training duration (Figure 4). Assessments of different training load measures to assess between-week or between-cycle changes in training stress in team sports such as rugby, soccer, rowing, and cycling have been studied in depth for decades^7,35,36 but different measures of training load to quantify between-week changes in training stress in distance running have not been studied as extensively. Our current results are in agreement with recent findings that training load measures, combining various external load metrics and sRPE, yield different week-to-week %Δ in training compared to using training duration alone.²⁶ These findings are explained by the fact that quantification of running training by duration alone does not account for the physiological response to training and only provides general information regarding the external training load of an athlete.^7,26,35,36 Thus, if a coach only quantifies external load metrics such as training duration or distance, they may be vastly misrepresenting the training stress experienced by runners. For example, a runner’s training volume could remain constant during two consecutive training days or weeks despite increases in intensity and resulting physiological response.⁵ This misrepresentation of training stress could lead to maladaptation or periods of non-functional overreaching in runners. On the other hand, training load measures likely provide a more individualized assessment of the physiological training response¹³ and overall training stress experienced by a runner.⁵ Thus, we urge coaches to consider thinking beyond just weekly volume and to incorporate measures of training loads in their monitoring approaches. This might be particularly important for inexperienced coaches, those who oversee the training of many runners, and those who oversee remote training (i.e., limited in-person and direct communication with athletes). Doing so may provide better quantification of training stress in runners and could help to avoid unplanned elevated training stress and optimize training responses. It is also important to mention that although certain running sessions require prescribed paces or interval splits, the use of prescribed session efforts instead of paces is likely useful to avoid unnecessary high internal physiological loads (e.g., sRPE) as a result of insisting on maintaining a specific pace on “easy” or recovery training sessions. In fact, RPE can be used to prescribe treadmill running intensities during stages of a graded exercise test in active men.³⁷ However, the use of sRPE to prescribe intensity during an entire training session in runners still remains to be studied.

Additionally, training load measured with training minutes and sRPE yielded the same average %Δ compared to training duration as the other training load measures including more complex biomechanical measures of external load (e.g., Step Count and Force) (Figure 4). These findings suggest that more complex biomechanical measures to monitor external loads may not be necessary. Thus, the use of a convenient external load measure such as training duration for calculating training load would be recommended for coaches instead of more complex biomechanical measures of external load.

We only studied %Δ in training load from coach-prescribed low to high training load weeks, but it is possible that a similar underestimation of training stress could be observed during planned periods of rest or reduced training load (e.g. “recovery weeks” or pre-competition “taper”) when only using training volume. In fact, recent evidence on a large population of recreational runners suggests that despite reductions in week-to-week %Δ in training volume (i.e., duration), increases in training stress—quantified with the training load of training duration and sRPE—can occur.²⁶ The pre-competition taper is heavily influenced by preceding training stress experienced by runners³⁸ as RPE might remain elevated during subsequent training sessions and thus, inappropriate quantification of training stress could greatly influence the effectiveness of a pre-competition taper. Future work on the use of different training load measures to quantify “tapering” or “recovery” periods would be highly valuable for coaches and sport scientists.

Although all other week-to-week %Δ in training load measures were different than training duration, week-to-week %Δ in sRPEx“Bone Stimulus” was not different than %Δ in prescribed training duration. This is in part the result of how “Bone Stimulus” was calculated. The proprietary “Bone Stimulus” metric uses the 3 D resultant acceleration of the IMU placed on the distal tibiae to calculate a surrogate measure of the “mechanical stimulus that would cause the bone to respond and remodel”. The calculated “Bone Stimulus” from the proprietary algorithm reaches a plateau or saturation after repeated cyclical loading over the course of a training run since it is based on the biology of bone cells’ adaptation to mechanical loading and remodeling.³⁰ Thus, the measured “Bone Stimulus” metric remains relatively constant with only an average of 2% week-to-week change between training cycles and might not be the most appropriate external load metric to include in training load quantification.

The secondary aim of this study was to assess individual runner responses in between-week changes among the different training load measures using an observational approach. Our secondary hypothesis was also supported. We observed that the combination of sRPE with the various external load measures led to larger ranges in individual responses to week-to-week %Δ in training load measures (76% to 97%) when compared to the two common measures to quantify training in runners: minutes (27%) and miles (41%) (Figure 5). The smaller range in week-to-week %Δ external loads (i.e., minutes and miles) conveys the message that the runners are experiencing similar training responses which can be misleading as observed in the current study. It is important to note that mileage did appear to have higher variance in week-to-week %Δ compared to minutes. This is due to 1) the fact that the team coach prescribed training in minutes and 2) the differing fitness levels and resulting average run paces for each runner. However, the standard deviations of the team averages for the week-to-week %Δ for the various training load measures ranged between 25-31%. Comparatively, the standard deviations in week-to-week %Δ for miles and minutes were 13% and 8%, respectively. The larger standard deviations for the week-to-week %Δ from the training load measures highlights the concept that inclusion of sRPE (i.e., an individualized measure of internal load) can quantify the different responses to coach-prescribed training.

Although training load measures seem to provide a more sensitive measure of week-to-week changes in training stress, %Δ among the different training load measures appears to be variable for individual runners. Indeed, Figure 6 demonstrates how week-to-week %Δ of the different training load measures varies greatly among individual runners. Specific training load measures might provide a slightly different quantification of training stress within specific runners which might highlight the need to individualize training quantification methods to ensure optimal training monitoring. Many factors can contribute to these varied responses for individual runners in %Δ among different training load measures including emotional/psychological state (e.g., family, school, relationship stress) and different changes in specific external load metrics due to higher training load (e.g., different cadence, different intensity of steps, different pace, etc.).⁵ Thus, future work should assess individual factors that might contribute to this individual response to various training load measures.

Figure 6.

Week-to-week percent change (%Δ) among training quantification methods (i.e., each black circle) for all runners included in the study. Black vertical rectangles illustrate the range in week-to-week changes in training among training monitoring methods for two specific runners (runners 3 and 4) to highlight individual responses in training stress among monitoring methods.

While using more biomechanically-specific and individualized external load measures to track training (e.g. Force, “Bone Stimulus”, and Step Count) might give a better representation of individual biomechanical loads, the measures used in the current study have limitations. The session average peak vertical force used as an external load measure was an estimation of the force experienced by the runners at an average running pace using a second-order polynomial and, therefore, is not a direct measure of average vertical force for each run. Unfortunately, daily wear of the insoles is currently not practical given the costs and expertise needed to operate the insoles. Finally, the use of Step Count may be limiting because using Step Count assumes that all steps produce the same mechanical load on the lower limb. In reality, some steps lead to greater magnitudes and different loading direction of the external force applied to the body. However, Step Count is a simple and convenient method to use as an external training load. The use of foot- or tibia-mounted IMUs to quantify the intensity/magnitude of the tibial acceleration will allow better quantification and weighing of running steps (e.g., Impact Load by IMeasureU) over the course of a running session.

The major limitation of this study is the small sample size (n = 9). However, all runners were part of the same team and training was prescribed by the same coach. This was highly important in this study to ensure a well-controlled training setting to ensure appropriate week-to-week comparisons in running training among methods. Finally, the two-week study period was short. However, we were not intending to track an entire training cycle but merely to highlight the fact that only monitoring volume (duration), between a low and a high training week that were prescribed by the team coach, can greatly underestimate training stress in runners.

Practical applications

Our findings suggest that the use of an internal physiological load (i.e., sRPE) combined with certain external loads, result in different outcomes in week-to-week changes in training in runners compared to using duration alone. We therefore suggest that coaches should consider factors beyond just weekly distance and could benefit from the use of training load assessments in their monitoring approaches. Further, the use of more complex measure of external loads used in the current study might be unnecessary and thus, we recommend that coaches can use simple measures to quantify external loads such as session duration in addition to a simple and convenient measure of internal load such as sRPE to more optimally and individually monitor training stress in runners. Finally, since different training monitoring methods yield different week-to-week changes in training load within some runners, individualized approaches to quantify training might be needed.

Footnotes

Acknowledgements

The authors would like to thank coach Nick Dwyer and the entire high school cross country team for taking part in the study, and Adriana Miltko, Tayler Vickery, and Richard Beltran for assisting with data collection. The authors would also like to thank IMeasureU/VICON for providing the inertial measurement units.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Christopher Napier

Max R Paquette

References

Kiely

Periodization theory: confronting an inconvenient truth. Sports Med 2018; 48: 753–764.

Viru

Nature of training effects. In: Garret

Kirkendall

(eds). Exercise and sport science. Philadelphia, PA: Lippincott Williams & Williams, 2000, pp. 67–95.

Coutts

Crowcroft

Kempton

Developing athlete monitoring systems: theoretical basis and practical applications. In: Kellmann

Beckmann

(eds). Sport, recovery and performance: interdisciplinary insights. London: Routledge, 2017, pp. 19–32.

Bellinger

Arnold

Minahan

Quantifying the training-intensity distribution in middle-distance runners: the influence of different methods of training-intensity quantification. Int J Sports Physiol Perform. Epub ahead of print 13 June 2019. DOI: 10.1123/ijspp.2019-0298

Paquette

Napier

Willy

, et al. Moving beyond weekly ‘distance’ – optimizing quantification of training load in runners. J Orthop Sports Phys Ther 2020; 50: 564–569.

Impellizzeri

Marcora

Coutts

AJ.

Internal and external training load: 15 years On. Int J Sports Physiol Perform 2019; 14: 270–273.

Foster

Monitoring training in athletes with reference to overtraining syndrome. Med Sci Sports Exerc 1998; 30: 1164–1168.

Gabbett

Infographic: the training-injury prevention paradox: should athletes be training smarter and harder?

Br J Sports Med 2018; 52: 203–2017.

Halson

SL.

Monitoring training load to understand fatigue in athletes. Sports Med 2014; 44 Suppl 2: S139–S147.

10.

Matos

Clemente

Brandao

, et al. Training load, aerobic capacity and their relationship with wellness status in recreational trail runners. Front Physiol 2019; 10: 1189.

11.

Mujika

Quantification of training and competition loads in endurance sports: methods and applications. Int J Sports Physiol Perform 2017; 12: S29–S217.

12.

Ten Haaf

Foster

Meeusen

, et al. Submaximal heart rate seems inadequate to prescribe and monitor intensified training. Eur J Sport Sci 2019; 19: 1082–1091.

13.

Abe

Yoshida

Ueoka

, et al. Relationship between perceived exertion and blood lactate concentrations during incremental running test in young females. BMC Sports Sci Med Rehabil 2015; 7: 5.

14.

Falk Neto

Tibana

de Sousa

NMF

, et al. Session rating of perceived exertion is a superior method to monitor internal training loads of functional fitness training sessions performed at different intensities when compared to training impulse. Front Physiol 2020; 11: 9–10.

15.

Brayne

Barnes

Heller

, et al. Using a wireless consumer accelerometer to measure tibial acceleration during running: agreement with a skin-mounted sensor. Sports Eng 2018; 21: 487–491.

16.

Burns

Deneweth Zendler

Zernicke

RF.

Validation of a wireless shoe insole for ground reaction force measurement. J Sports Sci 2019; 37: 1129–1138.

17.

Edwards

Ward

Meardon

, et al. The use of external transducers for estimating bone strain at the distal tibia during impact activity. J Biomech Eng 2009; 131: 051009.

18.

Aubol

Hawkins

Milner

CE.

Tibial acceleration reliability and minimal detectable difference during overground and treadmill running. J Appl Biomech. EPub ahead of print 10 August 2020. DOI:10.1123/jab.2019-0272.

19.

Milner

Hawkins

Aubol

KG.

Tibial acceleration during running is higher in field testing than indoor testing. Med Sci Sports Exerc 2020; 52: 1361–1366.

20.

Clark

Ryan

Weyand

PG.

Foot speed, foot-strike and footwear: linking gait mechanics and running ground reaction forces. J Exp Biol 2014; 217: 2037–2040.

21.

Logan

Hunter

Feland

, JT JTHet al. Ground reaction force differences between running shoes, racing flats, and distance spikes in runners. J Sports Sci Med 2010; 9: 147–153.

22.

Phan

Grisbrook

Wernli

, et al. Running quietly reduces ground reaction force and vertical loading rate and alters foot strike technique. J Sports Sci 2017; 35: 1636–1642.

23.

Davis

Bowser

Mullineaux

DR.

Greater vertical impact loading in female runners with medically diagnosed injuries: a prospective investigation. Br J Sports Med 2016; 50: 887–892.

24.

Pohl

Mullineaux

Milner

, et al. Biomechanical predictors of retrospective tibial stress fractures in runners. J Biomech 2008; 41: 1160–1165.

25.

Tenforde

Hayano

Jamison

, et al. Tibial acceleration measured from wearable sensors is associated with loading rates in injured runners. PM R 2020; 12: 679–684.

26.

Napier

Ryan

Menon

, et al. Session RPE in combination with training volume provides a better estimation of training responses in runners. J Athlet Train. Epub ahead of print October 2020. DOI: 10.4085/1062-6050-573-19.

27.

Yamato

Saragiotto

Lopes

AD.

A consensus definition of running-related injury in recreational runners: a modified Delphi approach. J Orthop Sports Phys Ther 2015; 45: 375–380.

28.

Fanchini

Ghielmetti

Coutts

, et al. Effect of training-session intensity distribution on session rating of perceived exertion in soccer players. Int J Sports Physiol Perform 2015; 10: 426–430.

29.

Halperin

Emanuel

Rating of perceived effort: methodological concerns and future directions. Sports Med 2020; 50: 679–687.

30.

Besier

IMU bone stimulus. IMeasureU, https://imeasureu.com/wp-content/uploads/2019/10/IMU_Bone_Stimulus.pdf (2019, accessed 16 October 2020).

31.

Wong

Finch

IMU-step dashboard summary. IMeasureU, https://imeasureu.com/wp-content/uploads/2018/02/IMUStepDashboardSummary.pdf (2018, accessed 16 October 2020).

32.

Swain

Kelleran

Graves

, et al. Impact forces of walking and running at the same intensity. J Strength Cond Res 2016; 30: 1042–1049.

33.

Tongen

Wunderlich

Biomechanics of running and walking. In: Gallian J (Ed.) Mathematics and sports: Mathematical Association of America. Mathematical Association of America, 2010, pp. 315–328. DOI: 10.5948/UPO9781614442004.026.

34.

Cohen

Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Erlbaum, 1988.

35.

Rodriguez-Marroyo

Villa

Garcia-Lopez

, et al. Comparison of heart rate and session rating of perceived exertion methods of defining exercise load in cyclists. J Strength Cond Res 2012; 26: 2249–2257.

36.

Tran

Rice

Main

, et al. Convergent validity of a novel method for quantifying rowing training loads. J Sports Sci 2015; 33: 268–276.

37.

Glass

Knowlton

Becque

MD.

Accuracy of RPE from graded exercise to establish exercise training intensity. Med Sci Sports Exerc 1992; 24: 1303–1307.

38.

Spilsbury

Fudge

Ingham

, et al. Tapering strategies in elite British endurance runners. Eur J Sport Sci 2015; 15: 367–373.