Validity of Rating of Perceived Exertion Scales in Relation to Movement Velocity and Exercise Intensity During Resistance-Exercise: A Systematic Review

Abstract

Context:

Movement velocity (MV) may be a valid tool to evaluate and control the load in resistance training (RT). The rating of perceived exertion (RPE) also enables practical load management. The relationship between RPE and MV may be used to monitor RT intensity.

Objective:

To evaluate the validity and practicality of RPE scales related to MV and training intensity in resistance exercise. We hypothesize a positive correlation among RPE, MV, and load intensity in RT. Therefore, RPE may serve as a supplementary indicator in monitoring RT load.

Data Sources:

Boolean algorithms were used to search several databases (SPORTDiscus, EBSCO, PubMed, Scopus, and Google Scholar).

Study Selection:

Studies published from 2009 to 2023 included clinical trials (randomized or not) in healthy female and male subjects that analyzed the relationship between different RPE scales and MV in basic RT exercises.

Study Design:

Systematic review.

Level of Evidence:

Level 3.

Results:

A total of 18 studies were selected using different RPE scales with reported MV training loads. Participants included RT and untrained male and female subjects (15-31 years old). Two RPE scales (OMNI-RES and repetitions in reserve) were used. The selected studies showed moderate positive correlations among these RPE scales, MV, and training load (eg, percentage of 1-repetition maximum [%1-RM]). In addition, equations have been developed to estimate %1-RM and MV loss based on the OMNI-RES scale.

Conclusion:

Studies show that RPE scales and MV constitute a valid, economic, and practical tool for assessing RT load progression and complementing other training monitoring variables. Exercise professionals should consider familiarizing participants with RPE scales and factors that might influence the perception of exertion (eg, level of training, motivation, and environmental conditions).

Resistance training (RT) is a complex and dynamic process that requires applying individualized workload parameters based on training status and goals and the proper progression of training loads over time to optimize training adaptations.^16,43 The intensity of RT is typically prescribed based on the percentage of 1-repetition maximum (1-RM) a person can lift in a given movement/exercise.^41,42,46 This allows the prescription of higher- or lower-intensity workloads based on the person’s capacity. However, several authors have proposed that the 1-RM may vary daily in the same person due to different factors (eg, fatigue accumulation) or change quickly within an RT program (eg, in novice athletes). Because of this, the 1-RM must be obtained frequently, which takes time and is often impractical.^15,19,23,35 More recently, movement velocity (MV) has been proposed as an objective method to monitor training load in RT.^{3,18,26,34,37} MV can be assessed using devices (some with more excellent reliability and validity than others), such as optical motion detection systems, linear position or velocity transducers, camera-based optoelectronic systems, and smartphone video-based systems.^11,36,38,50 In this way, analyzing and providing feedback to the user regarding lifting velocity and estimating the applied force and power (ie, from MV and displaced load) is possible.^38,49

Although these systems are commonplace in high-level athlete strength and conditioning facilities, most people who do fitness-related RT do not have access to them. Therefore, identifying an inexpensive way to estimate MV-related intensity may have practical applications. One possible practical solution is using rating of perceived exertion (RPE) scales to help gauge MV-related intensity.

For several years, different RPE scales have been proposed as tools for RT to help prescribe training load or predict 1-RM.^{2,12,14,17,24,25} In this regard, RPE has been used in children, adults, and older adults of both sexes to gain insight into perceptions of effort during exercise.^17,39,40 The OMNI-RES perception scale has “broadly generalizable properties” with a numerical response ranging between 0 and 10. ⁴⁰ However, it may also be helpful to examine the relationship among RPE, MV, and 1-RM within the various contexts of RT. Thus, this systematic review evaluated the relationship among the different RPE scales, MV, and the percentage of 1-RM (%1-RM) of basic strength exercises.

Methods

Protocol

This systematic review was developed and reported according to the parameters established in the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) in Exercise, Rehabilitation, Sport medicine and SporTs science (PERSiST) guidelines. ¹ This review was not eligible to be registered in PROSPERO, as it focuses on physical performance. The summary information of the protocol was uploaded to Figshare to make it publicly accessible and avoid unnecessary duplication (doi: 10.6084/m9.figshare.20432085).

Eligibility Criteria

The inclusion criteria for this systematic review were as follows: (1) studies published from 2009 to 2023; (2) clinical trials (randomized or not) in healthy female and male subjects; (3) published in specialized scientific journals hosted in databases; (4) published in peer-reviewed journals in English or Spanish; and (5) studies that analyzed the relationship between different RPE scales and MV in basic RT exercises. Articles that did not correspond to original research (eg, editorials, notes, reviews, etc) were excluded.

Information Sources

The search was performed in the following databases: SPORTDiscus, EBSCO, PubMed/MEDLINE, and Scopus. A manual search using Google Scholar was also performed to identify any potential studies using free terms.

Search Strategy

The search for the studies was performed using the following Boolean algorithms: (1) “resistance training" AND (“rate of perceived exertion” OR “rate of perceived exertion”) AND “movement velocity",” (2) “strength training” AND “rate of perceived exertion” AND “movement velocity,” (3) “RPE” AND “velocity,” and (4) “OMNI-RES” AND “velocity loss.” The search was enriched with the terms: NOT “sprint” NOT “gait” NOT “running” NOT “recovery.”

Study Selection

The eligibility process was carried out in 2 separate stages: (1) The authors independently selected the titles and abstracts of all nonduplicate articles and excluded those that did not meet the inclusion criteria. A definitive list was established, and discrepancies were resolved by consensus among the authors. When there was no consensus, a fourth author acted as a mediator. (2) Articles that passed the evaluation were downloaded (full-text), and 3 authors independently assessed eligibility. Duplicate articles, studies with no intervention, languages other than English or Spanish, and studies without analysis of the 2 main study variables were eliminated.

Data Extraction and Quality Assessment

The selected articles were subjected to a detailed reading and data extraction. Relevant information was extracted following some recommendations given in the literature. ¹³ All relevant data were prepared and presented in a synthesis table, including the research aim, sample characteristics, RT exercise, RPE scale used, intervention protocol, and principal conclusions. The risk of bias assessment of the studies was assessed using the Risk of Bias in Nonrandomized Studies - of Interventions (ROBINS-I) tool. ⁴⁵ Interpretation of domain-level and overall risk of bias judgments in ROBINS-I include: “Low risk of bias,” low risk of bias for all domains; “Moderate,” the study is judged to be at low or moderate risk of bias for all domains; “Serious risk of bias,” the study is judged to be at serious risk of bias in at least 1 domain, but not at critical risk of bias in any domain; “Critical risk of bias,” the study is judged to be at critical risk of bias in at least 1 domain; “No information,” there is no clear indication that the study is at serious or critical risk of bias and there is a lack of information in ≥1 key domains of bias. ⁴⁵

Results

In the first stage of the search strategy, 147 articles were identified. After removing duplicates, a sample of 34 potentially eligible articles was obtained. However, 12 articles did not fulfill the eligibility criteria and were excluded after a detailed abstract or full-text review. A total of 18 studies met the inclusion criteria and were included in this review. A flowchart of the article selection is shown in Figure 1.

Figure 1.

PRISMA flowchart. PRISMA, Preferred Reporting Items for Systematic reviews and Meta-Analyses.

The risk of bias assessment of the 18 included studies is presented in Figures 2 and 3. According to the ROBINS-I framework, the studies were at moderate (16 studies) and serious (2 studies) risk of bias. The most significant bias in the studies was due to sample size (eg, calculation for adequate statistical power), management of confounding variables, and reporting of missing cases.

Figure 2.

ROBINS-I risk of bias assessment graph. ROBINS-I, Risk of Bias in Nonrandomized Studies - of Interventions.

Figure 3.

ROBINS-I risk of bias assessment graph for each methodological quality item of each included study. ROBINS-I, Risk of Bias in Nonrandomized Studies - of Interventions.

Appendix Table A1 (available in the online version of this article) summarizes the subjects evaluated, the methodological aspects of the design, and the type of RPE scale used in the selected studies.

Discussion

This systematic review explored the relationship between different RPE scales and MV and the percentage of 1-RM of basic RT exercises (eg, bench press and squat). The most important findings indicate a high correlation between RPE, MV, and RPE and %1-RM in RT exercises. Based on this, it is proposed that self-perception scales of effort or MV constitute a practical and valid indicator of intensity that can be implemented in RT programs in both beginners and advanced resistance-trained athletes.

RPE OMNI-RES (0-10) Related to MV and %1-RM

A study by Naclerio et al ³⁰ used MV evaluation to estimate mechanical power and explore its possible relationship with RPE and load. No differences in the initial OMNI-RES (ie, between 1 and 3 initial repetitions of each set) were found for loads between 30% and 70% of 1-RM. However, there were differences between these intensities and higher loads (>70%) in the OMNI-RES. Likewise, agreement was reported between the variations of power and RPE values in the sets until muscle failure was reached. Unfortunately, this was not analyzed with the robust statistical models necessary to establish the relationship between these variables.

On the other hand, Bautista et al ⁵ developed a perceived velocity scale for the bench press that showed positive correlations with light, moderate, and high loads, and the correlations strengthened as training sessions progressed. The same scale was validated for the full-back squat exercise, ⁷ where linear and positive correlations with actual velocity over a range of intensities were found. However, the scale had poorer discrimination at velocities of low intensities, slight overestimation of MV at moderate intensities (eg, <40% 1-RM), and underestimation at higher intensities, which the authors attributed to insufficient familiarization with the scale. Bautista et al ⁶ also analyzed the relationship between mean MV and OMNI-RES scale values, finding that RPE was proportional to training load.

To further investigate the relationship between the relative load variables (%1-RM), MV, and OMNI-RES, linear regression models have been used to predict relative load from OMNI-RES and MV (Table 1).^31,32,47,48 Based on these findings, it can be suggested that RPE and MV can be used to predict variables for RT load progression.

Table 1.

Equations to estimate %1-RM and loss of MV from RPE scales

Reference	Estimated Variable	Population	Exercise	Equation	R 2	SEE
Babiloni-Lopez et al 2	Load (kg)	Physically active M&F (n = 18)	Smith machine squat	WP 35.21 + (6.07 × RPE) EBT 46.01 + (5.90 × RPE)	0.35 0.24	20.97 22.62
	Number of repetitions			WP 15.46 + (-1.38 × RPE) EBT 15.78 + (-1.48 × RPE)	0.45 0.48	3.80 3.35
	%1-RM			WP 36.31 + (5.14 × RPE) EBT 41.53 + (6.14 × RPE)	0.58 0.61	11.07 10.63
	MV (m/s) first repetition			WP 0.89 + (-0.05 × RPE) EBT 0.99 + (-0.04 × RPE)	0.57 0.46	0.11 0.10
	MV (m/s) last repetition			WP 0.90 + (-0.05 × RPE) EBT 0.90 + (-0.03 × RPE)	0.34 0.29	0.12 0.10
Naclerio et al 31	%1-RM	Strength-trained M&F (n = 308)	Bench press	29.03 + 7.26 × OMNI-RES	0.93	5.07
Naclerio and Larumbe-Zabala 32	%1-RM	Strength-trained athletes M&F (n = 290)	Back squat	5.07 + 9.63 × OMNI-RES	0.86	8.17
Naclerio and Larumbe-Zabala 29	%1-RM	Trained M (n = 154)	Power clean	31.10 + 7.26 × OMNI-RES	0.88	4.92
Varela-Olalla et al 47	Loss of velocity	Olympic wrestlers M&F (n = 5)	Bench press	2.294(OMNI-RES 2 ) - 25.68 (OMNI-RES) + 99.29	0.76	5.45
Varela-Olalla et al 48	%Rep	Physically active M (n = 7)	Bench press	1.13RPE 2 - 4.58RPE + 22.44	0.89	9.85

1-RM, 1-repetition maximum; %1-RM, percentage of 1-RM; EBT, elastic band training; M, male subjects; M&F, male and female subjects; %Rep, percentage of repetitions performed with respect to the maximum number of repetitions possible; MV, mean propulsive velocity in the given exercise; OMNI-RES, OMNI-Resistance; RPE, rating of perceived exertion; SEE, standard error of the estimate; WP, Smith machine training.

In turn, Chapman et al^8,9,10 analyzed the effectiveness of the OMNI-RES scale and electromyography signal to monitor changes in mean accelerative velocity during a set-to-muscle failure performed with different relative loads in the bench press and back squat exercise. Based on these studies, the OMNI-RES scale is a viable method for distinguishing different load zones. It reflects initial losses of acceleration/velocity during a continuous set to muscular failure in upper-body and lower-body resistance exercises. Athletes may be instructed to exercise with a maximal intended velocity, considering that OMNI-RES values of 6 or 7 indicate ≥10% velocity loss. ¹⁰

Finally, the type of load has been evaluated. For instance, when comparing the parallel squat exercise performed with weights on Smith machines and using elastic bands (EBs) with matched relative loads, differences in RPE and MV were observed between the 2 protocols. ² The RPE was lower in EB than using weights, while the speed was higher when using an EB. This difference could be due to the elongation coefficient, as EB provides less external load in the lower phases of the squat. In addition, it was possible to predict the values of load (ie, kg), number of repetitions, %1-RM, and mean propulsive MV from the RPE of the first repetition. ²

Table 1 summarizes the equations for estimating relative load and velocity loss based on the OMNI-RES scale from the studies analyzed. While these equations are helpful, it should be noted that there are some differences in statistical models used, and results are applicable only to the populations and conditions studied (eg, exercise, participants studied, measurement techniques employed, etc). Therefore, researchers and coaches should develop and validate their equations or cross-validate those available in the literature to the populations and training protocols they are using.

RPE Scale Based on Repetitions in Reserve

An athlete’s accuracy in assessing RPE improves with training experience, so novice athletes may not accurately assign RPE. ²¹ Therefore, implementation of RPE requires proper education, familiarization, and an adequate learning curve. ⁵² In search of more objective and practical strategies for RT control, the RPE has been related to repetitions in reserve (RIR).^21,52 In general terms, RIR (Table 2) is a perception scale based on the repetitions remaining in “reserve” concerning a maximum value of repetitions at each of the intensities of the RPE scale. ⁵²

Table 2.

RIR RT-specific RPE ^a

Rating	Description of perceived exertion
10	Maximum effort
9.5	No further repetitions but could increase load
9	1 repetition remaining
8.5	1-2 repetitions remaining
8	2 repetitions remaining
7.5	2-3 repetitions remaining
7	3 repetitions remaining
5-6	4-6 repetitions remaining
3-4	Light effort
1-2	Little to no effort

RIR, repetitions in reserve; RPE, rating of perceived exertion; RT, resistance training.

^aReproduced from Zourdos et al. ⁵²

Zourdos et al ⁵² examined this relationship in experienced and novice squat lifters. In the experienced subjects, the mean propulsive velocity at 1-RM load was slower (0.24 ± 0.04 m s⁻¹) compared with novices (0.34 ± 0.07 m s⁻¹). Likewise, when all repetitions and MV data from the trained group were pooled, the mean MV at all 1-RM percentages had a strong inverse correlation with RPE/RIR (r = -0.88; P < 0.01). Meanwhile, in novice subjects, a strong inverse correlation was observed between the mean propulsive velocity in all %1-RM and RPE values (r = -0.77; P < 0.01). This demonstrates the efficacy of an RIR-based RPE scale during strength exercises for use with “self-regulated” training loads.

Helms et al ²² compared concentric mean propulsive velocity and perceived exertion index based on RIRs. In this study, there was a strong correlation (r = 0.88-0.91) between the %1-RM and RPE values in each lift; moreover, MV was inversely correlated with RPE and the %1-RM in each lift (r = -0.79 to -0.87 and -0.90 to -0.92, respectively).

The authors suggested that using the RPE scale of RIR might be very useful in prescribing and modifying training loads instead of using a model based solely on %1-RM. In addition, it may allow the athlete to better adapt to the overload of his/her condition during training.

In this line of research, Balsalobre-Fernández et al ⁴ found moderate relationships among load and MV, RPE, and RIR in powerlifters. These variables were able to predict relative overload with comparable accuracy. Also, Ormsbee et al ³³ evaluated the efficacy of the RPE scale based on the RIR across different %1-RM loads of inexperienced and novice subjects on the bench press, and both the experienced and novice observed inverse solid correlations between mean velocity and RPE/RIR at all intensities. Thus, it seems that MV-overload, RIR-overload, or RPE-overload relationships can provide a more accurate estimate of relative overload than relationships obtained from generalized models.

One aspect that should be highlighted is that various factors such as personality traits, fatigue, exercise feedback, instructions, supplements and medications, and environmental conditions can influence a person’s perception of load during exercise.^20,21 These factors should be considered when using the OMNI-RES scale. ²⁰ In this sense, studies have shown that performing a velocity-based RT at moderate altitude (2320 m) increases both internal (RPE) and external load, increasing physiological stress imposed by hypoxia. ⁴¹ In addition, a randomized crossover trial found that hypoxia influences the perceived intensity of exercise and can substantially increase the training load experienced. ²⁷ Regarding fatigue, it has been shown that RPE may be a possible predictor of muscle fatigue in exercises such as the squat; however, MV loss may not accurately reflect muscle fatigue when participants are unable or not required to perform squats explosively with weight. ⁵²

Available evidence widely reports that the RPE scales are related to the intensity of RT as it increases with the load (eg, in terms of %1-RM), the number of repetitions and sets performed, and the duration of the exercise.^28,51 We have documented that RPE is also related to MV in basic strength exercises such as squats and bench press in healthy and physically active men and women. The RPE scale is a viable variable in training load control and, using other indicators such as MV or %1-RM may have limitations in certain situations, particularly with beginners, muscle hypertrophy optimization, group training, and rehabilitation/physiotherapy programs. ⁵² However, controlled trials are required to evaluate the efficacy of RPE-based RT, considering different training load variables, covariates, and robust statistical models that allow an optimal analysis of the results.

Limitations of the Study

In this systematic review, we must consider several fundamental limitations that influence the interpretation of the results. First, most of the included studies exhibited a moderate risk of bias and were not controlled trials, introducing potential confounders that could skew the correlation between RPE and MV for RT training load. In addition, the prevalent small sample sizes among the studies may limit the generalizability of the findings.

Practical applications

Although more research is needed, the available evidence indicates that, due to their relationship with MV and load in common strength exercises, the different RPE scales (eg, OMNI-RES and RIR) are valid and practical tools for assessing RT intensity. ⁴⁴ We therefore offer the following recommendations for strength and conditioning specialists:

• Clearly explain the purpose of RPE scales (eg, OMNI-RES and RIR) about MV and exercise intensity, familiarize participants with the use of the scales, and provide a learning period to enhance accuracy.

• The RPE scale can be a valuable additional indicator in various populations and conditions, including beginners, those participating in muscle hypertrophy programs, persons undergoing rehabilitation and/or physical readaptation, and those engaging in group training sessions.

• The precision in estimating 1-RM from the RPE scales is higher at medium and high intensities than at lower intensities.

• The RPE value should be obtained between the first and third repetition, depending on the magnitude of the overload, before a marked decrease in MV occurs. This is because, in a series of repetitions, a decrease in MV is observed due to a progressive decrease in the force applied in each repetition. However, the decrease in MV (eg, 10%, 20%, 30%) and the repetitions per set should be established according to each athlete’s objective.

• The use of RPE and OMNI-RES scales during training can be useful for monitoring fluctuations in the perception of training loads and performance (Figure 4). For example, RPE increases at the same load or the same perception at a higher load may indicate adaptation or maladaptation, respectively.

• Use population-specific prediction equations rather than generalized models to estimate relative loads where possible. Accordingly, key variables should be considered, such as age, gender, and performance level (eg, athletes with strength experience), sports specialty (eg, powerlifting), and the type of exercise performed (eg, squat, bench press).

• RPE scales are useful as an individual monitoring tool or can be used in conjunction with devices that assess MV (eg, linear velocity or position transducer, accelerometer, among others) (see Appendix Table A1), given that they increase the precision of the estimates by acting as predictor covariates.

• Finally, when using RPE and MV in strength training, it is essential to individualize and characterize the training groups, considering the data variability between persons. This implies analyzing the differences in physical fitness, experience level, and goals that each athlete may have. When analyzing and using data based on RPE, as with any other load indicator, this variability must be considered to accurately tailor the training sessions for each athlete.

Figure 4.

Perceived exertion and movement velocity scales for resistance training.

Conclusion

The RPE scales - OMNI-RES and RIR - correlated highly with MV and %1-RM in common RT exercises such as bench press and squats. In addition, they revealed a significant correlation with the loss of velocity caused by fatigue. Thus, these RPE scales constitute a useful, practical, and valid tool to monitor RT programs in beginners and advanced resistance-trained athletes, as long as there is adequate familiarization with the use of these scales and factors that can influence the perception of exertion, such as training experience, competition level, motivation, fatigue level, environmental conditions, etc, are considered. Controlled trials should evaluate the efficacy of RPE-based training, considering training load variables, covariates, and statistical models that allow robust evaluation of performance and fatigue over time.