Anchoring procedures are not as extensive for measurement of naturally occurring muscle pain during exercise compared to those required for perceived exertion metrics. The memory anchoring procedures can be quite similar for exercise-induced pain as described above for perceived exertion. Prior to exercise, while presenting a standardized instructional set, the individual is asked to think about the pain experienced in the active muscles during previous exercise or physical activity. Then, the individual is asked to remember times when levels of muscle pain equal to the low and high anchor points on the scale were experienced. During subsequent bouts of exercise, the individual is asked to rate muscle pain levels based on the memory of muscle pain at the low and high anchor points.

However, an exercise anchoring procedure cannot be used in conjunction with a pain scale in the same manner that it is used with a perceived exertion scale. The psychophysical concept underlying perceived exertion scale anchoring is based on the predictions of Borg’s Range Model. The basic tenets of this model assume that as individuals undertake exercise intensities across their entire performance range they are able to link physiological responses to corresponding and interdependent RPE values. This assumes that maximal RPE (e.g., ten on the OMNI Scale) is linked to attainment of maximal exercise intensity (e.g., POmax, VO₂max, 1RM). However, the achievement of maximal level of exercise-induced muscle pain as required for exercise anchoring procedures is not always possible. In studies by Cook and colleagues (1997, 1998), individuals did not detect muscle pain during load-incremented exercise until they attained 50–60 % of peak exercise capacity, with the pain threshold of some individuals not occurring until 90 %. Peak muscle pain values averaged ~5.5 in females and 8–8.5 in males using the 0–10 Pain Intensity Scale (Cook et al. 1997, 1998). In addition, although an individual may be able to remember a high level of muscle pain sensation experienced during previous exercise, for both clinical and physiological reasons it may not be possible to elicit such a response in certain individuals. Such limitations render the use of exercise anchoring procedures for category pain scales impractical.

Ratings of affective responses (AR) and PA enjoyment (PAE) during exercise are recognized as psychosocial correlates of perceived exertion. However, category scales to measure these constructs cannot be anchored at very low and very high exercise intensities as is the accepted procedures when anchoring an RPE scale. Individuals cannot be instructed to link AR or PAE values to any specific exercise intensity because these responses have been uniquely shaped over time in each individual. Previous PA experience, subjective behavioral norms, and values pertaining to PA adherence can vary greatly between individuals. This can result in interindividual differences between specific psychosocial domains that dominate the affective and enjoyment experience to exercise intensity.

Research results are conflicting regarding the intensities of exercise that result in the most positive AR during exercise. Kirkcaldy and Shephard (1990) proposed an inverted-U paradigm, predicting that moderate intensity exercise produces an optimal AR. Similar findings have been reported by Moses et al. (1989). Low exercise intensities may be insufficient to evoke positive changes in AR and high exercise intensities may produce significant negative shifts in AR. More recent evidence refutes this relation such that both high (Tate and Petruzzello 1995) and low intensity (Ekkekakis et al. 2000) exercise programs have led to positive changes in AR.

Ekkekakis’ (2003) “dual-mode” model explains the interindividual variability in AR that occurs across exercise intensities, specifically as it relates to the anaerobic threshold (AT). The AR during low to moderate intensity exercise (i.e., below the AT) is primarily shaped by cognitive processes that are unique to the individual. Above the AT, interoceptive cues driven by the increasing demand for energy supplied by anaerobic pathways dominate the AR. Therefore, AR at exercise intensities at or somewhat below the AT are rather heterogeneous, but AR at exercise intensities above the AT become increasingly less positive/more negative and are relatively homogeneous (Ekkekakis 2003; Ekkekakis et al. 2005; Hall et al. 2002).

Research has confirmed the marked interindividual differences in AR during exercise intensities below the AT, especially involving moderate intensity exercise. In a study by Van Lunduyt and colleagues (2000), participants estimated AR during moderate intensity cycle exercise (60 % VO₂peak). Results indicated that 44.4 % of subjects experienced an increase in AR, 41.3 % experienced a decrease in AR, and 14.3 % experienced no change in AR. Other studies have confirmed the shift from heterogeneity in AR at intensities below the AT to homogeneity in AR above the AT. In response to separate 15-min bouts of treadmill exercise, 47 % of subjects exhibited a decline in AR at intensities below the ventilatory threshold (VT) and 80 % of subjects exhibited a decline in AR at intensities above the VT (Ekkekakis et al. 2005). Similar results were found in response to 20 min of treadmill exercise. AR was more positive and stable below the AT with only 25 % of subjects exhibiting a decline in AR during performance at these intensities. Above the AT, 83 % of subjects exhibited a negative shift in AR (Parfitt et al. 2006).

When exercise-induced muscle pain or affect are part of a perceptual research paradigm, it may be beneficial to ask the individual to practice rating these variables along with perceived exertion during exercise anchoring procedures or a practice exercise test. This will allow the individual to practice rating all three of these independent constructs within a close time-frame during exercise. In addition, such orientation procedures present an opportunity to provide feedback to an individual prior to fitness testing or experimental exercise procedures regarding psychophysical appropriateness of his/her rating responses. It may be especially beneficial for children, enabling them to more accurately link the exercise intensity range to their own pain and affective experience (Robertson et al. 2009).

The neurophysiological mechanisms for naturally occurring, exercise-induced pain in healthy, uninjured individuals involve stimulation of mechanical and biochemical nociceptive systems in skeletal muscle. Pain threshold is defined as the onset of pain sensation and varies between individuals. Once pain threshold is reached, ratings of exercise-induced muscle pain should increase with physical measures of exercise intensity, such as PO and weight lifted. This measure of pain sensation occurs in conjunction with the accumulation of noxious by-products of metabolism such as blood lactate, hydrogen ions, and bradykinin, all of which increase as a function of increasing exercise intensity. Early exercise-induced muscle pain studies used the Borg (0–10) CR10 Scale to measure “aches and pain in the legs” during load-incremented and constant PO cycle exercise (Borg et al. 1985; Ljunggren et al. 1987). The investigations demonstrated evidence of concurrent validity of the CR10 Scale to measure pain sensations. Pain ratings were moderately correlated to blood lactate concentration at high PO’s during load-incremented exercise, with r = 0.45 at 200 W and r = 0.39 at 240 W (Borg et al. 1985), and at the end of constant PO exercise, with r = 0.54 (Ljunggren et al. 1987).

Later studies confirmed concurrent validity of the Pain Intensity Scale developed by Cook and colleagues (1997). The Pain Intensity Scale employs construct-specific verbal descriptors that are linked to the same numerical categories as appear on the original Borg CR10 Scale. In Cook’s investigation, pain ratings increased as a positively accelerating function of exercise intensity once pain threshold was achieved. It was noted that pain threshold ranged from 9 to 95 % of POpeak, indicating marked interindividual differences during load-incremented cycle ergometry (Cook et al. 1997, 1998). Mean pain threshold was ~50 % POpeak in males (Cook et al. 1997, 1998) and ~60 % POpeak in females (Cook et al. 1998). In males, pain ratings derived from the Pain Intensity Scale increased from a mean of ~2 at 60 % of POpeak to ~8–8.5 at 100 % of POpeak. In females, pain ratings increased from a mean of ~1 at 60 % of POpeak to ~5.5 at 100 % of POpeak (Cook et al. 1997, 1998). Robertson and colleagues (2009) developed the OMNI-Muscle Hurt Scale to measure exercise-induced muscle pain in children. This investigation found evidence for concurrent scale validity during isotonic resistance exercise performed by young children. High correlations were exhibited between weight lifted and pain ratings for biceps curl resistance exercise and knee extension resistance exercise, with r values across sets ranging from 0.67 to 0.87 (Robertson et al. 2009). In addition, construct validity was evidenced in Cook’s original study using the Pain Intensity Scale during load-incremented cycle exercise. High correlations ranging from r = 0.79–0.94 were found at intensities from 60 to 100 % POpeak (Cook et al. 1997).

Hardy and Rejeski (1989) demonstrated both construct and content validity of the Feeling Scale (FS) in college-aged males and females. The Multiple Affective Adjective Checklist (MAAC) employs a set of 132 adjectives. Subscales of the MAAC were used to compute criterion scores for both positive and negative affect. One group of subjects was instructed to choose adjectives describing a good feeling during exercise, while the other group chose adjectives describing a bad feeling during exercise. The results of the study found that subjects identified different affective states having good and bad feelings during exercise. The AR appropriately represented items at either end of the pleasure–displeasure continuum. The differentiated AR continuum was seen in 97 % of subjects who were asked to identify adjectives matching bad feelings and 94 % of subjects asked to identify adjectives matching good feelings (Hardy and Rejeski 1989). Kenney and colleagues (1987) conducted an investigation that also provided construct validity evidence for the FS in college-aged females. The study involved a cognitive-behavioral distress management training (DMT) program. The DMT program was administered to half of the participants between separate treadmill exercise bouts performed to exhaustion. The subjects who were administered the DMT program rated a more positive AR than subjects who did not receive the DMT when measures were obtained at the end of the treadmill run to exhaustion. However, RPE values were similar between subject groups (Kenney et al. 1987).

A few investigations have tested the validity of recently developed single-item PA enjoyment (PAE) scales (Haile et al. 2012; Stanley et al. 2009). These investigations correlated PAE ratings with AR measured using the FS. During both a load-incremented cycle ergometer protocol terminating at VO₂peak (Haile et al. 2012) and during a 20-min moderate intensity constant load cycle ergometer protocol (Stanley et al. 2009), significant positive relations were demonstrated between PAE and FS responses. In the investigation by Haile et al. (2012), PAE was measured using an 11-category scale that used the same format as the FS, with responses ranging from −5 to 5. The observed correlation coefficient between PAE and FS ratings was r = 0.92. In the investigation by Stanley et al. (2009), PAE was measured using a seven-point scale that employed a format different than the FS. The observed correlation coefficients between PAE and FS ratings ranged from r = 0.48–0.55.

The comparatively higher correlation coefficients reported by Haile et al. (2012) may be due to their use of a scale with similar format for the measurement of both AR and PAE. This argument has been employed to avoid the measurement of independent perceptual constructs using the same scale (Cook et al. 1997). For example, previous investigations measured both perceived exertion and pain during exercise using the CR10 Scale (Borg et al. 1985; Ljunggren et al. 1987). The resultant high correlation coefficients between RPE and pain intensity ratings may have been a “demand artifact” resulting from use of the same perceptual scale format to measure the two independent perceptual constructs (Cook et al. 1997). AR and PAE cannot be labeled as independent constructs similar to perceived exertion and pain. Rather, PAE is a specific domain of overall affect that may dominate the AR to exercise in many individuals. In addition, the PAE rating scale, although having a similar format to the FS, has verbal descriptors specific to enjoyment (Haile et al. 2012). Regardless, since acute exercise enjoyment is a novel construct, further research is necessary to study the measurement of AR and PAE simultaneously during exercise. In some populations in which enjoyment is a primary mediator of the overall affective experience during PA, it may be appropriate to measure PAE only.

Symptomatic pain has been used routinely to identify tolerable limits of exercise for clinical populations such as those with peripheral artery disease who experience intermittent claudication in active limbs. However, little research has focused on the use of exercise-induced pain as a target for exercise intensity prescription. O’Connor and Cook (2001) had young female adults perform 20 min of cycle ergometer exercise at a target muscle pain intensity rating of 3 on Cook’s (1997) 0–10 Pain Intensity Scale. A rating of 3 corresponds to the verbal descriptor “moderate pain.” On average, the target level of muscle pain was associated with a relative aerobic metabolic rate of 73.9 % VO₂peak at 6 min of continuous exercise, decreasing to 68.5 % VO₂peak at 20 min (O’Connor and Cook 2001).

The long-term adherence to exercise prescriptions that are based on muscle pain response are unknown. Pain experience of any kind during exercise may be a major factor contributing to sedentary behavior in many individuals. Therefore, prescribing exercise at intensities below the individual’s pain threshold may promote adherence to PA programs. However, previous research has found great interindividual variability in the pain threshold, as evoked during exercise. This makes it difficult to identify a group-normalized pain response that corresponds to a target physiological outcome and applies to a variety of activities for a wide range of individuals in a manner such as been shown for the RPE at the VT (Goss et al. 2003). Studies by Cook and colleagues (1997, 1998) determined that the pain threshold during load-incremented exercise occurred at 50–60 % of peak exercise capacity, but values ranged from 9 to 95 % of POpeak. Therefore, exercise intensity prescription based on pain ratings should take an individual approach, recognizing that the procedure may not be appropriate in those with a low pain threshold. Athletes performing high intensity exercise in which exercise-induced muscle pain is expected are a healthy population for which the prescription of exercise intensity using target muscle pain ratings has the most utility.

It has been shown that the amount of time spent during a given situation can depend on the affect experienced during the activity (Emmons and Diener 1986). Therefore, the acute AR to an initial exercise performance may influence future exercise participation. Exercise perceived as feeling pleasant may promote future participation. On the other hand, exercise perceived as feeling unpleasant could decrease future participation or lead to withdrawal from the activity altogether (Parfitt et al. 2006). The goal, then, is to maximize the positive AR that an individual experiences during exercise. This goal recognizes that a positive affective experience is an important link in the chain between exercise adoption and maintenance (Van Lunduyt et al. 2000).

A study by Da Silva and colleagues (2011) determined the AR corresponding to exercise intensities spanning the VT in sedentary normal weight, overweight and obese women. This application of the AR in exercise prescription was similar to methods used for calculation of RPE-VT. FS ratings were assessed throughout a graded treadmill exercise test to measure VO₂max. The FS ratings corresponding to 90 %, 100 % and 110 % of the VT were identified. Group average FS ratings for the entire sample were ~2.7, ~1.6, and ~0 corresponding to exercise intensities at 90 %, 100 % and 110 % of the VT, respectively. The AR were similar between normal weight and overweight groups at each intensity. The FS ratings were approximately 3, 2, and 1 at 90 %, 100 % and 110 % of the VT, respectively. The obese group had similar FS ratings to the normal weight and overweight groups at 90 % of the VT, but their ratings were significantly less positive at 100 % of the VT (mean FS rating = 0.5) and 110 % of the VT (mean FS rating = −1.95). These data indicate a positive affective experience at intensities spanning the VT in sedentary normal weight and overweight women, but obese women may require exercise intensities below the VT to experience positive AR (Da Silva et al. 2011). It must be noted that there was considerable variability in FS responses at each intensity, so even when the average FS rating was positive some subjects rated a negative affective experience. In addition, these data were collected during graded treadmill exercise, unlike normal continuous intensity or interval exercise bouts prescribed for health-fitness benefits.

Performing at exercise intensities that span the VT has resulted in significant changes in FS ratings during exercise. Ekkekakis and colleagues (2008) studied the AR of young adults during 15 min of continuous treadmill exercise at intensities corresponding to the VT, 20 % below the VT, and 10 % above the VT. In the condition where intensity was below the VT, 50 % of subjects experienced no change in AR throughout exercise while 43 % experienced a decrease in AR. At intensities equal to and above the VT, 77 % and 80 % of subjects experienced a decrease in AR throughout exercise, respectively. In every condition, however, a small number of subjects experienced an increase in AR during exercise (Ekkekakis et al. 2008).

The identification of a group-normalized AR at the VT may prove difficult for exercise intensity prescription due to the marked interindividual variability in AR across exercise intensities. This wide response variability is similar to that evidenced for exercise-induced pain ratings. The variability in AR is especially evident during moderate intensity exercise (Van Lunduyt et al. 2000). It is of note that moderate intensity exercise is recommended by professional organizations as the optimal level for PA programs designed to produce health-fitness benefits (ACSM 2013). According to Ekkekakis’ “dual-mode model,” exercise intensity above the VT results in the lowest interindividual variability in AR. This is due to the comparative dominance of noxious properties of physiological signals over cognitive processes in shaping the affective experience. Unfortunately, the comparatively more homogenous AR to exercise above the VT is typified by progressively more negative feelings (Ekkekakis et al. 2005). A negative AR during exercise indicating a displeasurable experience most likely contributes to poor program adherence.

Various investigations have shown that optimal AR may occur at low, moderate, or even high exercise intensities (Ekkekakis et al. 2000; Moses et al. 1989; Tate and Petruzzello 1995). As such, the development of an exercise prescription using AR measured separately for each individual may be a necessary approach to maximize PA adherence. Exercise prescriptions should identify the appropriate exercise intensity by choosing a target HR or RPE based on the optimal AR, or even by prescribing exercise intensity using a target FS rating. Rose and Parfitt (2008) asked sedentary women to perform separate 30-min treadmill exercise bouts at target FS ratings of 1 and 3. On average, the women chose an exercise intensity similar to the VT for both target FS ratings, indicating that the women felt the treadmill exercise was pleasurable (Rose and Parfitt 2008). The implications for program adherence using prescribed target FS ratings are unknown, but hold promise from a public health perspective. Monitoring and adjusting PA programs to continually optimize AR may be necessary to promote long-term habitual PA participation.