Metabolic systems.

During the laboratory tests, a stationary metabolic system (SMS), Oxycon Pro® (Carefusion GmbH, Hoechberg, Germany) was used in the mixing chamber mode for all metabolic measurements. The software used was JLAB 4.53. In the field environments, a mobile metabolic system (MMS), Oxycon Mobile, version 5.10, (CareFusion GmbH, Hoechberg, Germany) was used while measuring gas exchange variables and ventilation breath by breath. Both metabolic systems were switched on at a minimum of 30 minutes before data collection and calibrated immediately before and after each test using the automated procedures in accordance with the manufacturer’s recommendations. A high precision gas of 15.00% O₂ and 6.00% CO₂ (accuracy: O₂ ± 0.04% and CO₂ ± 0.1%; Air Liquid AB, Kungsängen, Sweden) was used for calibrating the gas analyzers in both the SMS and MMS. The calibrations were always performed in the same environmental settings as the two different measurement conditions.

A facemask with non-rebreathing air inlet valves (Combitox, Dräger Safety, Lübeck, Germany) was used in both the laboratory and field conditions. All measured metabolic variables were saved in averages of 15 seconds. Both the SMS and the MMS were carefully controlled as well as validated prior to and during the present study. This included checks with a metabolic calibrator to ensure that the two systems were interchangeable. For all methodological studies, as well as more detailed descriptions of these systems, see Rosdahl and colleagues [15], Salier Eriksson and colleagues [16], and Schantz and colleagues [17].

Heart rate monitor.

In both the laboratory and field conditions, HR was recorded using a Polar Wearlink 31 transmitter (Polar Electro Oy, Kempele, Finland). The HR values were displayed and saved in averages of 15 seconds and stored in the SMS or the MMS. For safety reasons, HR values were also stored in averages of 15 seconds in the Polar Electro S610i HR monitor watch (Polar Electro Oy, Kempele, Finland) during the field tests. However, the HR values were always used from the MMS, except for three periods (1–6.25 minutes), one in each of three participants, when values were missing due to technical problems. In these cases, the HR values from the watch substituted the missing values. To confirm that this was appropriate, all individuals’ watch HR values were compared with the corresponding HR values from the MMS. No significant differences were detected in this comparison.

Cycle ergometer.

A mechanically braked pendulum cycle ergometer (828E Monark Exercise AB, Vansbro, Sweden) was used for performing the cycle exercise in the laboratory. Before each test, the scale was zeroed while each participant sat on the saddle with his or her feet resting on the frame between the pedals. The saddle height was adjusted so that one knee of the participant was slightly bent when the foot was on the pedal in its lowest position. A digital metronome (DM70 Seiko S-Yard Co. Ltd, Tokyo, Japan) was used to instruct the participants to maintain correct cycling cadence. The workload was controlled every minute by checking the cadence and the braking force as indicated on the pendulum position.

Laboratory tests.

All the repeated tests in the laboratory kept the same standard procedures of measurements. Firstly, a few measurements at rest were carried out before the cycle ergometer exercise. This included measurements of body weight and height as well as resting HR. For determining the resting HR, the participants rested quietly in supine position on a treatment table for ten minutes. The values from the last five minutes were averaged and determined as the resting HR.

The submaximal cycle ergometer exercise consisted of three different workloads; 50, 100 and 150 watt (W) for the females, and 100, 150 and 200 W for the males. The cycling cadence was kept to 50 revolutions per minute (rpm) in accordance with Åstrand [18, p. 19]. For each workload, the participants cycled until steady state HR was attained for two consecutive minutes (total work time approximately six minutes), after which the resistance was increased. Steady state HR was defined as the average HR value of one minute, based on the 30 seconds before and the 30 seconds after each full minute, were 2–3 beats · min^-1 at two consecutive minutes. The third workload was increased to only 125 W or 175 W for the females and males, respectively, if, after the second workload, the participant’s HR was higher than 150 beats · min^-1 and their rated RPE exceeded 15 [1].

Between the second and third workloads, the participants sustained pedalling for one minute at a self-chosen low cadence, and with a low resistance of 5 Newton (N). They were then instructed to resume the cadence of 50 rpm while the investigator gradually increased the resistance until, after one minute, the third workload was achieved (resistance was increased to 50 W during the first 15 seconds, to 100 W the next 15 seconds and successively to the final workload during the last 30 seconds). During all submaximal workloads, the participants sat in an upright position with their hands laying on the handlebars. The RPE assessments for breathing and legs, respectively, took place during the latter part of the final minute of each workload. After this phase of submaximal exercise, and before the maximal exercise, a period of two minutes followed by continued cycling with a self-chosen low cadence at a resistance of 5 N.

The maximal exercise was carried out with a cycling cadence of 80 rpm, as this cadence has been shown to be associated with the longest time to exhaustion at maximal efforts [19]. During the first three minutes, the workloads were set to 60, 100, and 120 or 140 W for each minute. The latter alternatives depended on which third workload the participants had performed during the submaximal exercise. Thus, 120 W was selected if the third level had been 125 W or 175 W for the females and males, respectively, whereas 140 W was used if it had been 150 W or 200 W. After these initial three minutes, the resistance increased by 20 W every minute until voluntary exhaustion occurred and ended the test. The RPE assessments for breathing and legs, respectively, were made immediately after this completion. To ensure that the participants achieved their O₂max, at least two of the following criteria were met by each participant: 1) a plateau in O₂ despite increasing exercise intensity (defined as a O₂ increment of < 150 mL · min^-1), 2) a respiratory exchange ratio of ≥ 1.1, and 3) a RPE of ≥ 17 [1, 20, 21]. During all cycle ergometer exercise, the participants looked into a shield with a blue-grey colour without any pattern. Thus, the visual stimulation was at a low level.

Field tests.

The participants pedaled their own bicycles either to or from their workplace choosing themselves which direction and time suited them best. They were instructed to cycle at their ordinary commuting intensities for the entire route. Characteristics of the commuters’ bicycles, such as weight and number of gears, can be found in Schantz and colleagues [22]. The tests were carried out between June and November. 18 of the tests were performed during the morning rush hours, while the remaining two were conducted during the evening rush hours. The cycle commutes took place in the inner urban and suburban areas of Greater Stockholm, Sweden. This mixture of areas meant that the commuters experienced a variety of route environments, in terms of e.g. buildings, greenery, topography and traffic. The study setting for the field tests, in overall terms, is illuminated in S3 Appendix. For a more detailed description of these areas, see Wahlgren and Schantz [23].

The participants were met at their designated start address by one of the investigators who transported the measurement equipment. Prior to the commutes, the MMS was placed in a custom-made backpack on the participants. A GPS was also placed in the backpack to track the road. The starting time of the cycle commutes was synchronized with a second investigator waiting at the final address. When the commuters arrived at their destinations, the total trip time was noted (means were; 29 minutes for men and 23 for women). The participants were then immediately asked to rate RPE for breathing and legs, respectively, of the overall commute. They also stated the amount of stops they made at traffic lights as well as other stops [22]. Levels of ambient conditions in terms of temperature, relative humidity and wind speed for each individual cycle commute were obtained from the website of the Stockholm-Uppsala Air Quality Management Association [24], see Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Ambient conditions during the cycle commuting in field (mean ± SD).

https://doi.org/10.1371/journal.pone.0300776.t002

Assessments of the cycle route environments.

The Active Commuting Route Environment Scale (ACRES) was used for assessments of the participants’ perceptions and appraisals of their route environments. For most of the participants, this was done a year before the field tests. ACRES has been characterized by considerable criterion-related validity and reasonable test-retest reproducibility [23, 25]. In this study, two items from ACRES were used for overall assessments of the routes cycled to work. The first was whether the route environments inhibit or stimulate the cycling. This item was formulated as follows: “Do you think that, on the whole, the environment you cycle in stimulates/inhibits your commuting?”. The second item was whether the route environment was experienced as unsafe or safe for reasons of traffic. This was formulated as follows: “How unsafe/safe do you feel in traffic as a cyclist along your route?”. The items’ response scales ranged between 1 and 15, with adjectival opposites labelled as “inhibits a lot” and “stimulates a lot”, respectively, “very unsafe” and “very safe”. Along the scale, the position number 8 represented a neutral position labelled as neither nor e.g. inhibiting or stimulating. The questionnaire instructions included a drawn map that separated the areas into; inner urban and suburban–rural areas [25]. Participants were asked to differentiate their commuting experiences between these two areas, see the ratings in Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Route environment ratings in the inner urban and suburban–rural areas of Greater Stockholm, Sweden (mean ± SD).

https://doi.org/10.1371/journal.pone.0300776.t003

Cycling cadence.

With regard to the cycling cadence, even if this variable was not monitored in the field, it can be assumed that varying cadences have been used in the different measurement conditions. Most likely, the participants preferred to use an average higher cadence in the field compared to the fixed level of 50 rpm in the laboratory [27]. Hansen and colleagues [27] studied a range of cadences around 60–90 rpm at two different cycle ergometer workloads covering our average commuting intensity (65% of O₂max; Table 7). Based on their results, the relationship between RPE and O₂ does not appear to be affected by using different cadences. This is supported by Kounalakis and colleagues [28], who found no significant differences in RPE or HR at the same absolute O₂ values when comparing cadences of 40 and 80 rpm during 90 minutes of cycle ergometer exercise. On the contrary, an early study [29] noted that a cycling cadence of 40 rpm compared to 60 and 80 rpm generated higher RPE levels despite no differences in O₂. These findings were most evident at the higher workloads (corresponding to approximately 45 and 64% of O₂max). Given this discrepancy in findings, it cannot be ruled out that a use of varying cadences has contributed to the present deviations in RPE between the laboratory and field conditions.

Interestingly, numerically higher RPE mean differences between the two measurement conditions were noted in the rating for legs (2.5–2.7 RPE units) compared to breathing (1.4–1.6 RPE units). Possibly, these different results may be due to that varying cadences have been used. According to the two-factor categorization of perceived exertion [14], it is likely that the participants perceived that the local muscular strain in their legs affected them more differently, than their central breathing exertion did, when using varying cadences. Nevertheless, the fact that RPE for breathing differs significantly between the two conditions indicates that there may be more moderating factors affecting the present RPE results than the cycling cadence.

Exercise duration.

The commuting durations were about 29 and 23 minutes for males and females, respectively (Table 6). These are slightly longer time periods than the total exercise duration for the submaximal cycle ergometer exercise in the laboratory (about 18 minutes). Due to these time variations, it is possible that the present RPE differences between the laboratory and the field environment have been underestimated, as RPE has been shown to rise with increasing exercise duration of ergometer cycling and treadmill walking at work rates corresponding to 60% of O₂max [30, 31].

This presumed underestimation of RPE difference between the two conditions, is further supported by data from 1661 commuter cyclists of both sexes, and with varying ages. Multiple regression analyses showed that Borg’s 6–20 RPE scale increased with 0.58 RPE units per 10 minutes at a given cycling speed when taking speed, duration, age and sex into account (Peter Schantz, personal communication 2022-03-22). Furthermore, given that longer exercise durations are related to higher cycling speeds in the above mentioned large sample of cyclists [32], it can be anticipated that they who cycle longer durations have higher O₂max. Thereby, they can sustain a given cycling speed at a lower % of O₂max, and at a lower RPE as well. Thus, it is possible that the value from the multiple regression analyses stated above represents an additional form of underestimation of the duration effect on the present RPE results.

Constant versus variable exercise.

Given the naturalistic setting of the field tests in suburban-urban commuting environments, including varying topography and different traffic situations such as queuing, turning and red lights, etc., it is clear that the field cycling involved distinctly more variable and intermittent exercise compared to the constant submaximal cycle ergometer workloads in the laboratory. The question is whether this affects the relation between RPE and the overall workload responses in terms of %HRR and %O₂max. Based on previous studies, it appears that the overall workload during a given exercise duration, regardless of whether the intensity levels are constant or varying intermittently, is the important determinant of RPE in this respect [33, 34]. In such a case, the RPE responses in the laboratory and field conditions are comparable at equal mean exercise intensities. However, more studies of these issues are warranted.

Instructions and protocols.

The fact that the participants self-regulated their commuting cycling in field (regarding e.g. levels of cycling gears, cadence, and exercise intensity) versus that they followed a protocol in the laboratory, is also a methodological difference possibly affecting RPE. This is because a high degree of autonomy during high-intensity interval training as well as during submaximal exercise sessions has been associated with significantly lower RPE levels compared to conventional and prescribed exercise protocols [35, 36]. Thus, it is possible that also the present results can be due to differences in exercise autonomy.

Moreover, the participants rated RPE during the last minute of each of the submaximal workloads in the laboratory, whereas in the field setting, they made their RPE assessments for the total commute immediately thereafter. Since RPE assessments for longer exercise durations have been shown to reflect the exertion rate close to the end of the exercise rather than the total exercise period [37], the present relations between RPE and exercise intensities in the field may be questioned. Given this, we have compared mean values of %HRR and %O₂max, respectively, between the entire cycle commutes and the last 5-min periods for all participants in this study (see S5 Appendix). These analyses were based on a data set where the transition periods at both the start and the end of the cycle commutes have been excluded, i.e. the transitions from resting level to exercise intensity as well as from exercise intensity back to resting level. The comparisons showed that there were no differences in terms of % of O₂max between the entire cycle commutes and the last 5-min periods. On the other hand, % of HRR was slightly lower for the entire commutes compared to the last 5-min periods (averages: 68.9% versus 71.5%; P-value = 0.003; S5 Appendix). Thus, in line with Kilpatrick and colleagues [37], the present RPE ratings in the field may correspond to a somewhat higher mean level of %HRR than the one stated in this study. However, this would have led to an increased difference between the laboratory based predicted RPE and the rated RPE in the field, and therefore, it cannot explain the present results.

Environmental aspects.

The individual cycle commuting routes were in the inner urban and/or suburban areas of Greater Stockholm. Thereby, the participants experienced a variety of visual and auditory stimuli from the environment in terms of e.g. buildings, greenery, topography, traffic [cf. 23], and weather conditions. According to the ratings with the ACRES [25], the participants found that these environments ranged on average between 9.1–11.6 units in terms of inhibits or stimulates cycling (1 inhibits a lot; 8 being neutral; 15 stimulates a lot) (Table 3). These somewhat stimulating route environments should be seen in relation to the simple and standardized exercise laboratory environment, without the possibility of stimuli from outside. Thus, it can be concluded that the cycling commutes, compared to the laboratory exercise, had; (1) an overall higher level of external stimuli e.g. through traffic, (2) more natural components, including e.g. greenery and possibly water, and (3) more varying ambient conditions. So, how could these environmental differences have affected the present RPE results?

Already Pennebaker and Lightner [7] noted that symptoms of fatigue during exercise can be coupled to the degree of external stimuli. In their first experiment, they found that subjects who heard distracting street sounds via headphones during treadmill exercise reported less fatigue symptoms than subjects hearing an amplification of their own breathing. These findings were confirmed by their second experiment, in which subjects jogged a similar distance in two different environments, a cross-country path, and a monotonous lap course on a field. Although the cross-country jogging required more focus on the external environment than the lap course, higher speeds were demonstrated at comparable levels of fatigue symptoms. The findings of the present study can be related to Pennebaker and Lightner [7] in various ways. First of all, their effect of auditory stimuli can be related to the traffic situations that exposed the cycle commuters to varying ambient sounds. Traffic noise has been shown to adversely affect both cyclists and pedestrians [38–40], and Miedema [41] has shown that such forms of auditory impact can create high levels of annoyance, which might influence the perception of exertion.

Furthermore, the traffic situations in the present study demanded certain levels of visual and integrative attention. These distractions may have diminished the responsiveness to the internal sensations from the physical exercise, and thus led to a lower perceived exertion in the field compared to the laboratory environment [cf. 7].

Similarly, results by Ceci and Hassmén [8] and Mieras and colleagues [9] indicate that an increased level of external stimuli can reduce the perceived exertion, compared to less stimuli, at equal exercise intensities (cf. S4 Appendix). Both studies compared physiological and psychological responses to exercise in two different environments, a laboratory setting (with low external stimuli) and a field setting (with high external stimuli). The field trials in Ceci and Hassmén [8] took place in a green and blue recreational area, and the cycling in Mieras and colleagues [9] (cf. S2 Appendix) took place on a trail along a creek that in itself was a green and blue setting, while the overall framing setting for the trail was a mixture of primarily built up and green settings. Therefore, it is possible that the nature per se had an impact on the reduction of perceived exertion. The effect of soft fascination from nature and greenery on human wellbeing and restoration points in this direction [cf. 10, 11, 42]. In the present RPE assessments in field, the effect of greenery cannot, however, be isolated from the other commuting route environments, making it difficult to evaluate any potential effect of greenery on RPE. Therefore, further studies are needed to investigate this matter.

It is no doubt that the study by Ceci and Hassmén [8] involve the most green and beautiful external environment of the three studies using RPE measurements (cf. S1–S4 Appendices). At the same time, it is the study showing the greatest difference in physical performance. Whereas Mieras and colleagues [9] had 30% higher power outputs in field, the present study showed 19–30% higher relative levels of %HRR and %O₂max at given RPE levels, the average running speed Ceci and Hassmén [8] was 66% higher outdoors compared to indoors at given RPE levels (cf. S4 Appendix). This difference stimulates to a hypothesis that greenery and aesthetics can possibly have a specific and greater role in this context, a matter that deserves future studies.

Another potential cause for the present higher RPE levels in the laboratory compared to the field setting could be due to that the participants experienced the cycling commuting somewhat unsafe, which demanded directed attention. The participants’ ACRES ratings for unsafety-safety traffic of the overall route environments ranged on average between 5.6–10.9 (1 very unsafe; 8 neutral; 15 very safe) (Table 3). Consequently, a traffic-related stress could have caused an elevated HR relative to the workload and O₂ demand. This in turn could have led to the rated RPE levels in the field being compared with the laboratory based predicted RPE levels at a too high %HRR intensity. Contradictory to this, however, the relationships between %O₂max and RPE also showed significant RPE differences between the two conditions. Moreover, this theory is further refuted by the fact that relationships between HR and O₂ established during ergometer cycling in the laboratory have been shown to be valid for estimating intensity spectrums of O₂ based on HR measurements during cycle commuting in field for the same group of participants as in the current study [43].

Finally, the environmental differences in ambient conditions are also worth mentioning. In this respect, the average temperature during the cycle commuting was 10 and 12°C for males and females (Table 2), respectively, while in the laboratory it was kept around 20°C. Regarding this rather narrow temperature range, there is, to our knowledge, no study that clearly points in the direction of the present RPE results. On the contrary, Maw and colleagues [44] showed that when cool (8°C) and natural (24°C) temperature conditions were compared during constant ergometer cycling at the same absolute HR level corresponding to the currently used cycle commuting average (around 136 beats · min^-1; Table 7), the mean RPE difference between the two conditions was less than 0.5 RPE units.