Indirect EPOC Prediction Method Based on Heart Rate Measurement

White Paper by Firstbeat Technologies Ltd.

This white paper has been produced to review the method and empirical results of an indirect EPOC measurement method developed by Firstbeat Technologies Ltd. Parts
of this paper may have been published elsewhere and are referred to in this document.

TABLE OF CONTENTS
INTRODUCTION............................................................................................... 1 CONSTRUCTION OF THE MODEL FOR HEART RATE
Training load assessment ............................................................................. 1 BASED EPOC PREDICTION
EPOC in exercise sciences ........................................................................... 1 The EPOC model was constructed based on meta-analysis data of peer-
CONSTRUCTION OF THE MODEL FOR HEART RATE BASED EPOC reviewed articles. Only valid studies were carefully selected for this purpose.
PREDICTION ..................................................................................................... 1 The data included 48 different exercise settings, including a total of 158 trained
and untrained male and female subjects. Exercise durations ranged from 2 to
MODEL VALIDATION..................................................................................... 2
180 minutes and exercise intensities from 18 to 108% of VO2max. The modeling
Methods ........................................................................................................ 2 data included both continuous and intermittent exercises and consisted of
Results .......................................................................................................... 2 running, cycling and upper-body ergometer exercise.
EPOC: ESSENTIAL INFORMATION ON EXERCISE ................................... 3
APPLICATIONS OF EPOC IN TRAINING ..................................................... 3 EPOC is predicted only on the basis of heart rate derived information (see
Figure 2). The variables used in the estimation are current intensity (%VO2max)
Controlling the training load during a single exercise session .................... 3 and duration of exercise (time between two sampling points, ∆t) and EPOC in
Programming and periodization of training ................................................. 4 the previous sampling point. The model is able to predict the amount of EPOC
REFERENCES AND FURTHER READING.................................................... 5 at any given moment. No post-exercise measurement is needed (see Figure 1).
The model can be mathematically described as follows:

EPOC (t) = f(EPOC(t−1), exercise_intensity(t), ∆t). (1) (Saalasti 2003)

INTRODUCTION
This document describes a method for EPOC (excess post-exercise oxygen At low exercise intensity (<30-40%VO2max), EPOC does not accumulate
consumption) prediction based on heart rate (HR) measurement. EPOC is significantly after the initial increase at the beginning of exercise (see Figure
defined as the excess oxygen consumed during recovery from exercise as 5). At higher exercise intensities (>50%VO2max), EPOC accumulates
compared to resting oxygen consumption. The EPOC prediction method has continuously. The slope of accumulation gets steeper with increasing intensity.
been developed to provide a physiology-based measure for training load
assessment.

Training load assessment

It is difficult to select an optimal exercise dose. Sufficiently strenuous exercise
causes a disturbance in body’s homeostasis which after recovery results in
improved fitness (E.g. Brooks & Fahey 1984; Åstrand & Rodahl 1986). Too
easy training does not improve fitness but too hard training may in long term
lead to overtraining. It is therefore important to measure the training load.

Methods that are used in assessing training load may be broadly characterized
as subjective and physiological measures. Subjective measures are easy to
access, but do not always reflect physiological responses and recovery demand.
Traditional physiological measures, such as oxygen consumption (VO2), heart
rate and blood lactate, reflect mainly momentary intensity of exercise and not
length of exercise or cumulative exercise load. There are also training load
measures such as training impulse (TRIMP), but which does not have
physiological basis or scale and therefore may be difficult to interpret.

EPOC is a physiological measure (amount of oxygen consumed in excess after Figure 1 A. EPOC measured in laboratory (EPOCmeas). B. Predicted EPOC based on
exercise as measured in liters or ml/kg) that that reflects the recovery demand heart rate derived information (EPOCpred). EPOC prediction does not need any post-
and the disturbance of body’s homeostasis brought by the exercise. exercise measurements. VO2REC = recovery VO2, tREC = recovery time, VO2BL =
Measurement of EPOC has been possible only by analyzing respiratory gases Baseline (resting) VO2.
with laboratory equipment, thus being expensive, time consuming and not
applicable to everyday purposes.

The lack of valid and easy-to-apply physiology based method for the
assessment of training load has led us to develop a method to estimate EPOC
indirectly from heart rate measurement.

EPOC in exercise sciences

The first observation of an elevated resting metabolic rate after exercise was
made in 1910 by Benedict and Carpenter and was later studied as “oxygen
debt” (Hill and Lupton in 1923). The present name EPOC has been used not
only to represent oxygen repayment during recovery but also to reflect the
general exercise-induced disturbance of body’s resting metabolism (Gaesser &
Brooks 1984; Gore & Withers 1990) and resting homeostasis (Brehm & Gutin
1986): “the cause of Excessive Post-Exercise Oxygen Consumption (EPOC) is
the general disturbance to homeostasis brought on by exercise” (Brooks &
Fahey 1984).

EPOC reflects the body’s recovery requirements after exercise. Active oxygen-
consuming recovery processes occurring in the body are due to replenishment
of body’s resources (O2-stores, ATP, CP) and increased metabolic rate
(increased HR and respiratory work, elevated body temperature) caused by
metabolic by-products and hormones produced during exercise. (Brooks &
Fahey 1984; Åstrand & Rodahl 1986; Børsheim & Bahr 2003)

• EPOC reflects a general disturbance in body’s homeostasis caused by

exercise.
• EPOC is calculated by subtracting the area under resting VO2 from the
area under the recovery VO2 curve (see Figures 1 and 4). Figure 2. The estimation model of EPOC. HR(max) = (maximal) heart rate, RespR =
• EPOC gets higher with higher intensity and/or longer duration of respiration rate, %VO2max = percentage of maximal oxygen uptake, t = time, ∆t =
exercise (e.g. Børsheim & Bahr 2003) (see Figures 3 and 4). time between two sampling points.

Indirect EPOC Prediction Method Based on Heart Rate Measurement 1 © 2012 Firstbeat Technologies
Published: May 2005. Last update: March 2012. All rights reserved
Figure 3. EPOCmeas (shaded area) and EPOCpred (dotted line) from four sample exercises. Figures A to C represent typical constant-intensity exercises, whereas Figure D
represents a typical exercise during which the intensity changes naturally, for example according to speed (e.g. running, cycling, skating or rowing), work rate (e.g. indoor rowing or
cycle ergometry) or terrain (uphill/downhill).

A B
EPOC left (%)
EPOC (t)

%VO2max
EPOC (t-1)
Time (min)

Figure 4. Calculation models of the EPOCpred upslope component (A) and downslope component (B), fitted using meta-analysis data. The combination of upslope and downslope
equations determines the gradient of EPOCpred accumulation.

Measurements included two 10-min submaximal steady state exercise sessions

at 40% and 70% VO2max, with a constant load, and a maximal incremental
bicycle ergometer (Ergoline, Bitz, Germany) test to voluntary exhaustion.
Heart period data was collected beat-by-beat with an RR-recorder (Polar
Electro Ltd., Kempele, Finland) and VO2 data breath-by-breath with a Vmax-
analyzer (Sensor Medics, California, Palo Alto, USA).

Baseline (B) Exercise (E) Recovery (R)

5 min 15min
10 min
40% VO2max

10 min
70% VO2max

Figure 5. The accumulation pattern of EPOC at various intensities as a function of

exercise duration.
n * 2 min
Maximal
The accumulation formula of HR-based EPOC is a combination of upslope exercise test

(Fig. 4 A) and downslope (Fig. 4 B) formulas. When the intensity of exercise is

high, EPOC accumulates, whereas during periods of rest or low-intensity Figure 6. Submaximal and maximal exercise protocols on a cycle ergometer.
activity, the combination of these formulas results in decreasing EPOC.
Results
There may be a time lag of about 15 s between the cessation of exercise and HR-based EPOC was found to correlate with measured EPOC and the
reaching the peak value of heart beat derived EPOC. This is due to the slow goodness of fit (r2) value was 0.79 (see Figure 7). Mean absolute error (MAE)
recovery pattern of VO2 after exercise, which lags behind the true intensity (the values for the HR-based EPOC, when compared to the measured EPOC values,
calculation model is not able to recognize the exact end point of exercise). were 9.4, 14.0 and 16.9 ml/kg for 40% and 70% constant load exercise and for
maximal incremental exercise, respectively. For the pooled data, MAE was
13.7 ml/kg. HR-based EPOC was also tightly connected with blood lactate
MODEL VALIDATION (Data published, Rusko et al. 2003) levels, with the r2-value being 0.79 (see Figure 8).
Methods
Subjects were 32 healthy adults (8 fit and 8 less fit males and females), age
38±9 years (mean±SD), weight 69.6±10.8 kg, height 171.6±8.5 cm and VO2max
44.0±8.8 ml/kg/min. The procedure is presented in Figure 6.

Indirect EPOC Prediction Method Based on Heart Rate Measurement 2 © 2012 Firstbeat Technologies
Published: May 2005. Last update: March 2012. All rights reserved
 High EPOC-values are typically attained in exercise where cardiorespiratory
load and oxygen consumption remain at high level without possibility to
recover. Exercise that recruits large muscle mass, such as cross-country skiing
and running, results in higher EPOC values than exercise that recruits small
muscle mass. High EPOC values are also gained in intermittent exercise, such
as interval training, soccer or squash, if recovery periods are short and intensity
remains moderate. When applied to same exercise type, EPOC can be used to
compare the demand of different exercises.

EPOC reflects mainly aerobic properties of the exercise and therefore, does not
reflect optimally exhaustion due to local muscular fatigue and/or acidity. Thus,
in strength exercise, EPOC may be low although the individual would be
exhausted.

The day-to-day variation in the physiological training state of an individual can

be tracked with EPOC. Short-term changes in performance, environmental
factors and possible illnesses affect EPOC accumulation. EPOC is a sensitive
Figure 7. The correlation of heart beat derived EPOC (EPOCpred) with measured
measure for both cardiac and respiratory responses. Even slightly unusual
EPOC (EPOCmeas) during cycle ergometry exercise. (Modified from Rusko et al.
2003) responses can be tracked (see Table 2).

APPLICATIONS OF EPOC IN TRAINING

Controlling the training load during a single exercise session
EPOC can be applied across sports, as can be seen from Table 4. An individual
willing to improve his/her fitness level can try different sports and check which
ones are the best for his/her purpose.

EPOC can be used to confirm whether exercise fulfilled the purpose set before
the exercise session. Table 3 represents the main types of aerobic exercise and
the expected EPOC response. If the purpose is to enhance cardiorespiuratory
fitness, EPOC should be high (see Figures 10 A and B). During low-intensity
basic endurance exercise and separate warm-up exercises, EPOC should be
kept at a low level (see figure 9 A and B). During cool-down, a decline in
EPOC should be seen, indicating active recovery after exercise.

Table 1. EPOC depends on exercise properties.

Higher EPOC Lower EPOC
Figure 8. The correlation of HR-based EPOC (EPOCpred) with blood lactate levels
(Bla) during cycle ergometry exercise. (Modified from Rusko et al. 2003) Decreased intensity
Increased intensity
Shorter duration
Longer duration
EPOC: ESSENTIAL INFORMATION ON EXERCISE Continuous exercise
Discontinuous exercise
Table 1 summarizes the properties of exercise that determine the magnitude of Longer recovery periods during intermittent
Shorter recovery periods during intermittent
EPOC. During exercise, EPOC increases or decreases depending on whether exercise
exercise
disturbance or recovery in homeostasis is expected. EPOC starts to decrease if Active recovery during intermittent exercise
Passive recovery during intermittent
the intensity decreases enough during exercise. This implies that the exercise
physiological training load is not increasing further, but is decreasing instead. Whole body exercise
Lower/upper body exercise

Table 3. Summary of different types of targeted exercises and the related EPOC.

Exercise type and purpose Exercise characteristics Expected EPOC-response

Recovery exercise: Constant workload low-intensity exercise (30-50%VO2max) of short duration (15 to 30 min). Blood EPOC remains at low level or
To speed up lactate removal after exercise and cycle lactate (Bla) levels remain at resting level or decrease towards the resting level. No disturbance or decreases.
metabolic by-products in muscles. recovery of body’s homeostasis.

Basic endurance/slow distance training: Constant workload low-intensity exercise (40-60%VO2max) of long duration (1 to several hours). Bla Slow accumulation of EPOC and
To enhance the oxidation of fat and build up the remains at resting level. No significant disturbance in body’s homeostasis. a low EPOC peak.
endurance base (aerobic threshold).

Pace endurance/fast distance training: Constant workload exercise (60-85%VO2max) of moderate to long duration (30 min to 1 hour). Bla Rapid accumulation of EPOC and
To enhance the oxidation of carbohydrates and lactate increases above resting level. Significant disturbance of body’s homeostasis. a high EPOC peak.
clearance (anaerobic threshold).

VO2max training: Constant load or interval exercise with high intensity (>85%VO2max), short to moderate duration (15 Very rapid accumulation of
To improve maximal cardiorespiratory performance to 30 min). Bla increases rapidly and fatigue emerges quickly. EPOC and a high to very high
(VO2max): oxidation of carbohydrates, lactate tolerance EPOC peak.
and fast force production (specific to race pace) of
skeletal muscles.

Figure 9. Two examples of recovery exercises from an athlete. A) The intensity of this recovery exercise is slightly too high and EPOC accumulates rapidly at the beginning of
session. B) A successful recovery exercise during which intensity and EPOC accumulation remain relatively low throughout the exercise. Note the small difference in intensity but a
significant difference in the accumulated training load (peak EPOC) between the two exercises.

Indirect EPOC Prediction Method Based on Heart Rate Measurement 3 © 2012 Firstbeat Technologies
Published: May 2005. Last update: March 2012. All rights reserved
EPOC is useful in monitoring day-to-day changes in the physiological response Table 2. Factors causing higher or lower HR-based EPOC values when compared to
to training. If there is an unexpected EPOC response, the training program can usual values.
be adjusted depending on the cause of the different response (See table 2). EPOC is higher than usual

EPOC reflects changes in the level of cardiorespiratory fitness. If peak EPOC Possible cause Recommended action
is lower during the same exercise with the same workload (control exercise), Continue training. If possible, reduce the
the fitness level has probably improved because the disturbance of homeostasis Decreased cardiorespiratory fitness. EPOC level back to normal or slightly
is lower. Similarly, if EPOC is higher, the fitness level has most likely below it.
decreased. See Table 2 for additional explanations for higher or lower EPOC Environmental conditions: increased alti- Decrease absolute work rate to match pre-
levels than usual. More fit individuals are able to exercise at the same relative tude, temperature or humidity. vious EPOC levels.
intensity for a longer period of time than less fit individuals, which leads to
higher EPOC. Decrease training load and maximize
Not fully recovered from previous exercise.
recovery.

When coaching a team, it is important to get information on the physiological Do not exercise if you suffer from an
Illness.
responses of each individual. Team training sessions and games have a illness.
different impact on each player due to individual differences in e.g. the level of EPOC is lower than usual
cardiorespiratory fitness, position, game style and motivation. With EPOC, the
training load of each individual player can be monitored and the training Possible cause Recommended action
program adjusted (e.g. some players may need more intense training while
others need more rest after the games). See Figures 9 C and 9 D for an example Assure recovery in your training program.
Recent monotonous, high-volume training
Schedule both high- and low-intensity exer-
in soccer. period.
cise into your training program.

Programming and periodization of training Increase intensity and/or duration of

The overall load that accumulates during training periods can also be evaluated Increased cardiorespiratory fitness. training to reach or slightly exceed former
EPOC levels.
with EPOC. A schematic example of training load over a training period of an
endurance athlete is presented in Figure 10. The integration of training
intensity and duration enables easier quantitative analysis of training. More
accurate analysis of previous training loads helps in determining recovery
requirements and designing subsequent training sessions optimally: the training
load can be increased if the previous load is considered to have been too low,
or decreased if the load had been higher than planned.

Figure 9. Examples of the accumulation of HR-based EPOC in different exercise sessions: (A) High-intensity interval training session (Nordic walking/running in a steep uphill). (B)
High-intensity constant velocity running exercise. (C) A soccer player (defense) from the Finnish national league during a pre-season practice match. (D) A player from the same
team (mid-fielder) in the same match. Note the difference in physiological load between the two players. The match was preceded by a 20-min warm-up and there was a half-time
of about 10 min between the two halves.

Figure 10. A schematic representation of an endurance athlete’s training load during eight successive weeks (the columns represent daily values of EPOC). This two-month period
prepares the athlete for the most important races of the season. The daily values are highest during weekends mainly due to races. Note also the less loading days before race days.

Indirect EPOC Prediction Method Based on Heart Rate Measurement 4 © 2012 Firstbeat Technologies
Published: May 2005. Last update: March 2012. All rights reserved
 Gore, C.J. & Withers, R.T. (1990). The effect of exercise intensity and duration on the
Table 4. Accumulation of HR-based EPOC values in different sports. oxygen deficit and excess post-exercise oxygen consumption. European Journal of Applied
Impact on Typical Physiology 60: 169 – 174.
Sport/activity Typical exercise cardiorespiratory EPOC
fitness (VO2max) (ml/kg) Hawley, J.A., Myburgh, K.H., Noakes, T.D. & Dennis, S.C. (1997). Training techniques
to improve fatigue resistance and enhance endurance performance. Journal of Sports Sci-
ences 15: 325 – 333.
10-km race (27-50 min,
High – Very high 120-260
70-90% VO2max) Hill, A.V. & Lupton, H. (1923). Muscular exercise, lactid acid, and the supply and the
utilization of oxygen. Q J Med 16: 135 – 171.
Marathon 42 km (2h
Running 10min -5h, 65- High – Very high 120-650 Häyrinen, M., Luhtanen, P., Juntunen, J., Hynynen, E., Vänttinen, T., Lipponen, K.
85%VO2max) and Heliskoski J. (2007). An Evaluation of Physical Loading, Recovery and Stress in Youth
Soccer. Science for Success II congress. Poster.
60 min low intensity http://www.firstbeat.fi/userData/firstbeat/download/hayrinen_et_al_science_for_success_200
Low – Moderate 40-90
(55-65%VO2max) 7_congress.pdf

Laursen, P.B. & Jenkins, D.G. 2002. The Scientific Basis for High-Intensity Interval
Brisk walk 1h Training. Optimising Training Programmes and Maximising Performance in Highly Trained
Low 10-25
(40-50 %VO2max) Endurance Athletes. Sports Medicine 32 (1): 53 – 73.
Walking
Trekking with a back- Luhtanen, P., Nummela, A. & Lipponen, K. (2007). Physical loading, stress and recovery
pack in hilly terrain for Low – Moderate 25-75 in a youth soccer tournament. VIth congress on science in football.
5h (50-60% VO2max) http://www.firstbeat.fi/userData/firstbeat/download/luhtanen_2007_congress.pdf

Spinning session 40 min Pullinen, K. (2010). Cardiorespiratory and local tissue oxygenation responses to
Moderate – Very high 50-200 incremental exercise in type 1 diabetic patients and healthy controls. Master’s Thesis,
(60-80%VO2max)
Cycling University of Jyväskylä, Finland.
http://www.firstbeat.fi/userData/firstbeat/download/pullinen_master_thesis_2010.pdf
Cycling to work 20 min
Low 5-15
(30-50%VO2max) Rusko, H. (Ed.) (2003). Handbook of Sports Medicine and Science - Cross Country Skiing.
Blackwell Science.
Aerobic workout with
ergometer 30 min Moderate – High 60-120 Rusko, H. (2004). Influence of Increased Duration or Intensity on Training Load as
(65-75%VO2max) evaluated by EPOC and TRIMPS. ACSM congress presentation.
Rowing http://www.firstbeat.fi/userData/firstbeat/download/rusko_acsm_2004_congress.pdf
Warm-up at gym with
ergometer 10 min Low 15-25 Rusko, H., Luhtanen, P., Rahkila, P., Viitasalo, J., Rehunen, S. & Härkönen, M.
(55-65%VO2max) (1986). Muscle metabolism, blood lactate and oxygen uptake in steady state exercise at
aerobic and anaerobic thresholds. European Journal of Applied Physiology and
15-km race (35-60min,
High – Very high 130-320 Occupational Physiology 55: 181 – 186.
Cross- 70-85%VO2max)
Country Rusko, H.K., Pulkkinen, A., Saalasti, S., Hynynen, E. & Kettunen, J. (2003). Pre-pre-
Skiing diction of EPOC: A tool for monitoring fatigue accumulation during exercise? ACSM Con-
2h low intensity (40-
Low – Moderate 10-70 gress, San Francisco, May 28-31, 2003. Abstract: Medicine and Science in Sports and Exer-
60% VO2max) cise 35(5): Suppl: S183.

Game 90 min, position: Saalasti, S. (2003). Neural networks for heart rate time series analysis. Academic
Low – High 30-150 Dissertation, University of Jyväskylä, Finland.
back
Soccer
Sedlock, D.A. (1991). Postexercise Energy Expenditure Following Upper Body Exercise.
Game 90 min, position: Research Quarterly for Exercise and Sport 62 (2): 213 – 216.
High – Very high 150-300
mid- fielder, offence
Seppänen, M.J. (2005) Effect of increased velocity and duration of running on training load
as evaluated by EPOC. Master’s thesis, University of Jyväskylä, Department of Biology of
45 min aerobics class
Aerobics Moderate – Very high 70-200 Physical Activity. (In Finnish: Nopeuden ja keston vaikutukset tasavauhtisten juoksuharjoi-
(65-85 %VO2max) tusten kuormittavuuteen.)

Badminton, 1h game (70-80 Vänttinen, T., Blomqvist, M., Lehto, H. & Häkkinen, K. (2007). Heart Rate and Match
High – Very high 130-280 Analysis of Finnish Junior Football Players. VIth congress on science in football. Poster.
Squash %VO2max)
http://www.firstbeat.fi/userData/firstbeat/download/vanttinen_et_al_world_congress_of_scie
nce_football_2007.pdf
1h game (50 -
Tennis Low – High 25-130
70% VO2max) Vänttinen, T., Nummela, A. & Rupf, R. (2007). Practical Experiences From Measuring
Exercise Intensity and Recovery State with Heart Rate Monitoring in Team Sport. International
Playing 18 holes (about Symposium Computer Science in Sport. Poster.
Golf 3 hours, 30- Very low – Low 5-10
http://www.firstbeat.fi/userData/firstbeat/download/vanttinen_et_al_international_symposium_c
40% VO2max)
omputer_science_in_sport_2007.pdf

See also complete list of latest publications:

REFERENCES AND FURTHER READING http://www.firstbeat.fi/physiology/research-and-publications

Åstrand, P.-O. & Rodahl, K. (1986). Textbook of work physiology. Physiological bases of
exercise. 3rd ed. McGraw-Hill, USA.
White Papers:
Banister, E.W. (1991). Modeling Elite Athletic Performance. In: MacDougall, J.D.,
Wenger, H.A. & Green, H.J. (Eds.) Physiological Testing of High-Performance Athlete. 2nd Download white papers at http://www.firstbeat.fi/physiology/white-papers
ed. Champaign, Illinois: Human Kinetics.
- An Energy Expenditure Estimation Method Based on Heart Rate Measurement
Brehm, B.A. & Gutin, B. (1986). Recovery energy expenditure for steady state exercise in - EPOC Based Training Effect Assessment
runners and nonexercisers. Medicine and Science in Sports and Exercise 18 (2):205-210. - Heart Beat Based Recovery Analysis for Athletic Training
- VO2 Estimation Method Based on Heart Rate Measurement
Benedict , F.G. & Carpenter, T.M. (1910). The metabolism and energy transformations of
healthy man during rest. Washinton, DC: The Carnegie institute.

Brooks, G.A. & Fahey, T.D. (1984). Exercise physiology. Human bioenergetics and its
applications. New York: Macmillan Publishing Company. For more information:

Børsheim E. & Bahr R. (2003). Effect of exercise intensity, duration and mode on postex- Firstbeat Technologies Oy
ercise oxygen consumption. Sports Medicine 33(14): 1037-1060. Yliopistonkatu 28 A, 2. Kerros
FI-40100 Jyväskylä
Foster C., Florhaug J.A., Franklin J., Gottschall L., Hrovatin L.A., Parker S., Doleshal,
P., Dodge C. (2001). A new approach to monitoring exercise training. Journal of Strength
Finland
and Conditioning Research 15(1): 109-115.
info@firstbeat.fi
Gaesser, G. & Brooks, G. (1984). Metabolic bases of excess post-exercise oxygen con- www.firstbeat.fi
sumption: a review. Medicine and Science in Sports and Exercise 16: 29 - 43.

Indirect EPOC Prediction Method Based on Heart Rate Measurement 5 © 2012 Firstbeat Technologies
Published: May 2005. Last update: March 2012. All rights reserved