Thermal behavior is an important thermoeffector during situations in which body temperature increases such as when exposed to heat stress at rest (1) and during exercise (2). Thermal behavior is initiated upon sensations of thermal discomfort (3), secondary to increases in body temperatures (4,5). We have shown that thermal behavior is engaged during exercise and remains engaged during recovery from exercise, likely to help promote reductions in core temperature, despite that autonomic thermoeffector responses (sweating and skin blood flow) had already returned to preexercise levels (6). However, it is unknown if the utilization of thermal behavior differs between males and females during exercise and recovery.
Exercise prescription using a fixed rate of metabolic heat production is important for eliminating differences in body morphology when studying heat loss thermoeffectors (7). In a compensable environment, autonomic thermoeffectors are driven primarily by the evaporative heat loss required to achieve heat balance (8). Autonomic heat loss responses are similar when the rate of metabolic heat production does not differ between males and females at lower intensities of exercise (9). That said, compared with males, females have a greater sensitivity to thermal stimuli (10) and perceive a given thermal stimulus to be more unpleasant (10,11). It is speculated that these differences may be accounted for by a greater density of thermoreceptors per unit body surface area (11,12). Nevertheless, it is possible that, for a given change in body temperature, females desire a greater cooling stimulus to maintain thermal comfort levels. This implies that females may use thermal behavior to a greater extent than males during exercise, however this remains unknown.
We have shown during recovery from exercise, that thermal behavior is largely activated to oppose sustained elevations in core temperature, with skin temperature playing a relatively minor role (6). Others have observed that core temperature is elevated for a longer period in females compared with males after exercise (13). The mechanisms underlying these observations are not fully understood, but are likely due to differences in skin blood flow occurring subsequent to changes in baroreceptor sensitivity postexercise (13). Given the likelihood of greater core temperatures postexercise in females, there is also reason to believe that their thermal behavioral responses during the postexercise period would be greater than males.
The purposes of this study were twofold. First, we tested the hypothesis that thermal behavioral responses during exercise at a fixed rate of metabolic heat production would be greater in females compared to males due to greater thermal perceptual response (i.e. greater thermal discomfort) for a given change in body temperature during exercise. Our second aim was to test the hypothesis that thermal behavioral responses during recovery would also be greater in females compared to males due to persistently elevated core temperatures.
Twenty young, healthy adults (10 females, 10 males) participated in this study. All subjects were physically active, normotensive, nonsmokers, not taking medications, cognitively normal, and self-reported to be free from any known cardiovascular, metabolic, neurological or psychological illness. Female subjects were not pregnant, which was confirmed via a urine pregnancy test, and reported to be normally menstruating. Each subject was fully informed of the experimental procedures and possible risks before giving informed written consent. The study was approved by the Institutional Review Board at the University at Buffalo, and performed in accordance with the standards set by the latest revision of the declaration of Helsinki. It should be noted that a portion of these data from 12 subjects (six males, six females) have been published previously to test a unique hypothesis (6). An a priori power analysis showed that, at an alpha level of 0.05 and an effect size (f) of 0.33, (based on pilot data from our laboratory investigating the effect of sex on behavioral thermoregulation) a sample size of 10 subjects per group would result in a power of 0.96 that would be sufficient to detect a significant between-within subject interaction.
Subjects visited the laboratory on two occasions. Visit one was a screening and familiarization visit and visit two was the experimental trial. During the experimental trial, subjects arrived at the laboratory euhydrated, confirmed via urine specific gravity <1.020 (actual urine specific gravity: males, 1.010 ± 0.006; females, 1.009 ± 0.006) and having refrained from strenuous exercise, alcohol, and caffeine for 12 h, and food for 2 h. To control for menstrual cycle hormones, females were tested only during the first 10 d after self-identified menstruation or during the placebo phase of their oral contraceptives (n = 5), a period in which estrogen and progesterone are at their lowest levels. All experimental testing was conducted during the summer months in Buffalo, NY (mean outside temperature during experimental trials: 21°C ± 4°C). Subjects wore a standard short sleeved T-shirt and shorts (Lululemon Athletica, Vancouver, BC, Canada), and their own socks and athletic shoes (~0.4 clo). Females also wore a sports bra of their choosing underneath the standardized shorts and shirt.
Instrumentation and measurements
Height and weight were measured with a stadiometer and scale (Sartorius Corp. Bohemia, NY), and body surface area was calculated accordingly (14). Skinfold thickness was measured in triplicate at the chest, axilla, triceps, subscapula, abdomen, suprailliac, and thigh (Harpenden, Baty International, UK), and percent body fat was estimated from body density (15), which was calculated from the sum of skinfolds for males (16) and females (17). Physical activity level was estimated using the validated International Physical Activity Questionnaire (18). Cognitive ability was measured using the Montreal Cognitive Assessment (19) to confirm that there were no cognitive deficits that may confound the results.
All measures of temperature were obtained using thermocouples (Omega Engineering, Inc. Stamford, CT), unless otherwise noted. Mean skin temperature was measured as the unweighted average of 10 thermocouples attached to the right side of the body on the shin, posterior calf, posterior thigh, anterior thigh, abdomen, chest, scapula, hand, triceps, and forehead (20). The 10 site mean skin temperature calculation was chosen based on the recent recommendation that 10 sites are most appropriate for studies examining thermal comfort (21). Approximately 60 min before experimental testing, subjects swallowed a telemetry pill (HQ Inc., Palmetto, FL) for measurement of core temperature. Our previous findings revealed that the relative contributions of core and mean skin temperatures to thermal behavior during and after exercise were approximately 70% and 30%, respectively (6). Thus, mean body temperature was calculated as [0.7 × core temperature + 0.3 × mean skin temperature]. This provided an index of thermal afferent feedback stimulating thermal behavior, as has been recently employed (6,22).
Heart rate was continually measured via a three-lead ECG (DA100C, Biopac Systems, Inc. Goleta, CA). Skin blood flow was measured continually via integrated laser Doppler flowmetry (Periflux System 5010; Perimed, Stockholm, Sweden) on the dorsal surface of the left forearm. Skin blood flow was not measured under the clothing to remove any influence of the laser Doppler probes and wires from interfering with the skin to clothing interface. The arm was stabilized and held in the same position throughout the protocol. Blood pressure was measured using electrosphygmomanometry (Suntech, Tango M2, Raleigh, NC) on the right arm every 10 min. Mean arterial pressure was calculated as double the diastolic blood pressure value added to the systolic blood pressure value and divided by three.
Local sweat rate was measured by securing a plastic capsule that covered 3.9 cm2 of skin 4 to 5 cm below the axilla, on the midaxillary line and on the upper arm on the right side of the body. These capsules were perfused with dry nitrogen at a flow rate of 0.5 L·min−1. The water vapor of the gas exiting the capsules was continuously measured by capacitance hygrometry (HMT130; Vaisala, Woburn, WA), and local sweat rate was calculated by multiplying the humidity output by flow rate and dividing that value by the surface area of the capsule (23). The local sweat rates are reported as an average of these two locations.
Metabolic data were taken every 30 min during the preexercise, exercise, and postexercise periods, and were obtained via a mouth piece and two-way non-rebreathing valve (Han Rudoph, Inc., Shawnee, KS). Minute ventilation was calculated from expired airflow that was measured via a heated pneumotachometer (Hans Rudolph, Inc. Shawnee, KS), which was continually integrated over 30 s and corrected to STPD. The fraction of expired oxygen and carbon dioxide (Vacumed, Ventura, CA) was continually measured from a 3-L mixing chamber. Oxygen uptake and carbon dioxide production were calculated using the Haldane Transformation. The rate of metabolic heat production was calculated from oxygen uptake and the RER using a standard equation (7). These data are presented as absolute (W), normalized to body surface area (W·m−2) and to body mass (W·kg−1). The external workload was calculated by multiplying the cadence by the resistance load applied to the ergometer. The rate of evaporation required for heat balance was also determined in absolute (W) and relative to body surface area (W·m−2) during exercise and recovery using a standard equation (24).
Whole-body and neck thermal sensations (1, cold; 4, neutral; 7, hot (24)) and thermal discomfort (1, comfortable; 4, very uncomfortable (24)) were measured to the nearest 0.5 units using standard visual analogue scales. Ratings of perceived exertion were measured using the Borg Scale (6, no exertion; 20, maximal exertion) (25). All perceptual measures were recorded every 10 min.
Quantification of thermal behavior
Thermal behavior was objectively and continuously measured using techniques modified from those of Cabanac et al. (3,26) whereby a device was placed on the dorsal aspect of the neck as has been used previously in our laboratory (1,6). Neck skin temperature was controlled using a dual tubing system that was in direct contact with the subject’s neck. The first series of tubing comprised of four 8-cm lengths positioned in parallel, alternating with the second series of five 8-cm lengths. The first series of tubing was perfused with 34°C water at a constant flow rate of 2.2 L·min−1. The second series of tubing was perfused with −20°C fluid (antifreeze). The flow rate of the second series was directly controlled by the subject via a two-way ball valve, permitting flow rates of 2.2 L·min−1 when turned on. By opening this valve, the cool liquid ran through the tubing and cooled the skin on the subject’s neck. By closing this valve, the water warmed their neck skin temperature. The effective contact area was 20 × 10 cm and was secured using a Velcro strap around the neck with enough pressure that the whole device was in constant contact with the neck but not uncomfortable. The perception of a change in neck skin temperature was sensed within 15 to 20 s. Subjects were instructed to control the temperature of the dorsal aspect of the neck using this valve at any time necessary so that it was thermally comfortable throughout the experiment. The neck was chosen because it is the only skin area known to be equally and highly sensitive to both cooling and heating (27). Neck skin temperature was measured using a single thermocouple attached to the dorsal aspect of the neck, directly under the device. The temperature of the effluent fluid immediately after contact with the neck was measured using a single thermocouple embedded in the tubing. Previous work in our laboratory indicated that the temperature of the effluent fluid flowing through the neck device provided the most sensitive measure of thermal behavior (6). As a result, neck device temperature has been interpreted to reflect both the initiation and magnitude of thermal behavior.
At least 24 h before the experimental protocol, subjects were familiarized with the neck device as suggested by Cabanac et al. (3) and Schlader et al. (1). Subjects were seated in the chair, and the device was attached to their neck using Velcro straps. The tubing leading into the device was suspended from above so as to avoid adding any extra weight on the neck. Subjects were then instructed to start cycling at the required workload for the experimental trial (see below). While cycling, participants were familiarized with all perceptual questionnaires and with using the valve for the neck device. Specifically, the subjects were instructed to turn the valve on to receive cooling and off to turn the cooling off, as often as desired. This cycling familiarization lasted approximately 10 min per participant. Subjects were reminded that during the experimental trial, the valve was to be turned on when the dorsal neck was considered to be thermally uncomfortable and to turn the valve off when the thermal comfort at the neck was reinstated.
The experimental trial took place in a moderate thermal environment (24.4°C ± 0.5°C; 45% ± 10% relative humidity). After instrumentation, subjects rested while seated on a recumbent cycle ergometer for 30 min (preexercise). This was followed by 60 min of cycling (exercise) at 65 rpm using a constant workload of 1 kp that elicited a metabolic heat production of 221 ± 39 W (117 ± 18 W·m−2) for males and 221 ± 45 W (129 ± 21 W·m−2) for females. A fixed heatload was used to eliminate potentially confounding influences of body surface area on heat exchange and body temperatures (28). Exercise was followed by a 60-min seated resting period (postexercise). Mechanical fanning was not permitted to eliminate clothing billowing. Subjects watched nonstimulating documentaries throughout the entire protocol. Subjects were instructed to maintain their neck temperature at a thermally comfortable level regardless of how the rest of their body felt throughout the entire protocol.
Data and statistical analyses
Continually recorded data were binned as 60-s averages every 10 min. These temporal data were analyzed for changes between sex and over time using two-way repeated measure ANOVA. For the temporal data, when a significant F test was identified, a priori post hoc comparisons were made between baseline (30 min preexercise) and end-exercise (60 min exercise) time points. Temporal data were further analyzed for the area under the curve (AUC) for exercise (inclusive of preexercise 30 min to end-exercise 60 min time points) and recovery (inclusive of end-exercise 60 min to end-recovery 60 min time points). An unpaired t-test was used to determine differences between AUC in males and females during exercise and recovery. To directly determine the independent effect of exercise on thermal behavior during recovery, data were extracted the minute that mean body temperature returned to preexercise levels within each subject during the postexercise period, as we have done previously (6). Data were only included if the individual value of mean body temperature was within ±0.1°C of the final baseline temperature (i.e. 30 min baseline) for ≥5 min (i.e. five samples) and continuing to decrease or remain stable thereafter. These data were analyzed using a two-way mixed-model ANOVA. Furthermore, the effect of sex on absolute changes in the dependent variables from preexercise to postexercise was isolated by using unpaired t-tests. All analyses were carried out using Prism software (version 7; GraphPad Software Inc., La Jolla, CA). In all instances, when a significant F test was identified, post hoc comparisons were made using Sidak adjusted comparisons. For all analyses, a priori statistical significance was set at P ≤ 0.05 and actual P-values are reported where possible. Standardized effect size (ES; Cohen’s d) analyses were further used to interpret the magnitude of differences in AUC and the change in pre–post data. An ES was designated as trivial (d < 0.20), small (d = 0.20–0.49), moderate (d = 0.50–0.79), or large (d > 0.80).
Males and females did not differ in age (P = 0.20), reported levels of physical activity (P = 0.40), or cognitive ability (P > 0.99). Females were shorter (P < 0.01), had a smaller body surface area (P < 0.01), and smaller body mass (P < 0.01), but greater percent body fat (P < 0.01) compared with males (Table 1).
The exercise stimulus elicited similar rates of absolute metabolic heat production, metabolic heat production relative to body mass, and metabolic heat production relative to the body surface area of the subjects (P ≥ 0.33, Table 2). The absolute workloads eliciting these rates of metabolic heat production did not differ between sex (males: 69 ± 4; females: 63 ± 12 W, P = 0.16), nor did the external workload relative to body mass (males: 0.9 ± 0.1; females: 1.0 ± 0.2 W·kg−1, P = 0.10). The absolute evaporative heat loss requirement also did not differ between males and females (P = 0.95, Table 2), nor did it differ between sex relative to body surface area (P = 0.70, Table 2). However, both average heart rate (males: 106 ± 11; females: 133 ± 16, P < 0.01), and the percent of maximal heart rate (P < 0.01, Table 2) elicited during cycling was greater for females compared to males. Mean arterial pressure did not differ between sex (P = 0.15), or within sex for males (P ≥ 0.07) or females (P ≥ 0.86) during exercise. Mean arterial pressure did not differ within males during recovery, (P ≥ 0.20); however, females had a reduction in mean arterial pressure at the 60-min recovery time point compared to the end of exercise (P < 0.01). Finally, the rating of perceived exertion did not differ between sexes (P ≥ 0.37), but did increase throughout all exercise time points (P < 0.01) compared to preexercise or recovery.
Body temperature responses during exercise and recovery
There were no temporal differences in mean skin temperature (P = 0.66), core temperature (P = 0.57) or mean body temperatures (P = 0.84) between males and females during exercise or recovery (Fig. 1). Mean body temperature AUC did not differ between males and females during exercise (P = 0.10) as the increase in mean skin (males: +0.5 ± 0.7; females: +0.8°C ± 0.7°C, P = 0.36) and core (males: +0.7 ± 0.3; females: +0.9°C ± 0.3°C, P = 0.14) temperatures during exercise were not different between sexes. Females had a greater AUC for mean body temperature during recovery (P = 0.01), due to a greater reduction in mean skin (males: −0.6 ± 0.7; females: −1.1°C ± 0.6°C, P = 0.02) and core temperatures (males: −0.6°C ± 0.3°C; females: −0.8°C ± 0.2°C, P = 0.04) compared to males (Fig. 1).
Thermoeffector responses during exercise and recovery
Temporal changes in thermoeffector responses were not different between males and females throughout the protocol except for an interaction for neck device temperature (P = 0.02) at 40 min where neck device temperature was warmer for males (P < 0.01). Females showed a greater AUC for thermal behavior during exercise (P = 0.04). However, there were no differences for AUC for neck device temperature during recovery (P = 0.11). Autonomic heat loss effectors were not different during exercise between males and females (local sweat rate: P = 0.08; forearm skin blood flow: P = 0.10). However, during recovery from exercise, females had a greater AUC for local sweat rate (P = 0.02), but not forearm skin blood flow (P = 0.11) (Fig. 2).
Perceptual responses during exercise and recovery
Temporal analyses of whole-body thermal sensation (P = 0.17), whole-body thermal comfort (P > 0.99), neck thermal sensation (P = 0.67), and neck thermal comfort (P = 0.47) did not differ between sex throughout exercise and recovery. However, during recovery, females had a greater AUC for their whole-body thermal sensation (P = 0.05), suggesting that they felt cooler on average compared to males. Average whole body thermal discomfort was not different between males and females during exercise with no differences in total AUC (P = 0.09). Likewise, AUC for whole body thermal comfort did not differ between sexes during recovery (P = 0.27). The AUC for the local thermal sensation of the neck was not different between sexes (P = 0.066), although only females perceived an increase in warmth at the neck during exercise. In contrast, females felt that the neck was cooler on average during recovery, supported by a greater AUC (P = 0.02). Finally, although neck thermal comfort was not different between males and females during recovery, females had a greater AUC (P < 0.01) (Fig. 3).
Thermal behavior postexercise at the same mean body temperature
Mean body temperatures and data for all other variables were extracted from minute data at the 30-min preexercise time point and postexercise at the time point that mean body temperature returned to this same value (±0.1°C) in each individual. In three males and three females mean body temperature did not return to preexercise values and they were therefore excluded from analysis. In all subjects who were included for analysis (n = 14, 7 females, seven males), mean body temperature returned to within ±0.1°C of the preexercise value. The time it took for mean body temperature to recover to preexercise levels during recovery in females (27 ± 14 min) and males (39 ± 21 min) did not differ (P = 0.25).
The dynamics of mean body temperature returning to preexercise levels during recovery differed in females compared to males. Specifically, at the time that mean body temperature had returned to preexercise values in all subjects, females had elevated core temperatures (+0.20°C ± 0.15°C, P < 0.01) but males did not (+0.04°C ± 0.09°C, P = 0.66). The changes in temperature were different between sex (P = 0.01). In contrast, mean skin temperature was reduced in females, though not significant (−0.51°C ± 0.45°C, P = 0.060), whereas males did not have a large reduction in mean skin temperature (−0.04°C ± 0.12°C, P = 0.88). These changes in mean skin temperature were also different between sex (P < 0.01) (Table 3). At this point in time, thermal behavior was engaged in females as neck device temperature (24.1°C ± 3.6°C) remained reduced compared to males (27.0°C ± 1.5°C, P = 0.04). This was also reflected in a greater change in neck device temperature between males and females (P = 0.05), and in neck temperature for females (P < 0.01). Local sweat rates did not differ from preexercise to postexercise within (P = 0.25) or between (P = 0.46) sexes, nor was the change from preexercise to postexercise between males and females different for average local sweat rates (P = 0.33). Finally, forearm skin blood flow did not differ within (P = 0.10) or between (P = 0.09) sexes, nor did the change in forearm skin blood flow from preexercise to postexercise between males and females (P = 0.06) (Table 3).
By using a low-intensity exercise protocol that elicited similar rates of metabolic heat production between males and females, the evaporative heat loss requirements, temporal changes in body temperatures, sweating, and skin blood flow responses during exercise and recovery did not differ between groups. Despite that the autonomic thermoeffector responses to exercise did not differ between sexes, females used thermal behavior to a greater extent than males during exercise, as evidenced by a greater AUC for neck device temperature. Furthermore, temporal dynamics of body temperatures and thermoeffectors did not differ during recovery from exercise. However, at the point that the mean body temperature had returned to preexercise levels (i.e., the point at which the thermal afferent stimulus for behavior was the same preexercise vs postexercise), core temperature remained elevated and thermal behavior remained engaged in females, but not in males. These findings suggest that during exercise at a fixed rate of metabolic heat production, females use a greater magnitude of cooling compared to males, despite similar overall changes in body temperatures. Additionally, there appear to be differential dynamics in the recovery of core and mean skin temperature for females, compared to males, postexercise. It is possible that these differences contribute to thermal behavior seen during recovery from exercise in females, but not males, at the point when mean body temperature has returned to preexercise levels.
Sex differences in thermal behavior during exercise
By design, we used a fixed workload to elicit the same rate of metabolic heat production from both males and females, which eliminated the influence of body surface area on heat exchange and body temperatures (28) (Table 2). As a result, both AUC and dynamic temporal changes in mean skin, core, and mean body temperatures were not different between males and females during exercise (Fig. 1). Additionally, there were no differences in skin blood flow or local sweat rate (Fig. 2). These findings are supported by previous research demonstrating that 30 min of low-intensity exercise (~200 W·m−2 of metabolic heat production) does not elicit differences in body temperatures or autonomic thermoeffector activation between males and females (9). Interestingly, we have uniquely identified that when body temperatures are influenced to the same degree during exercise, females used thermal behavior to a greater extent compared to males, as evidenced by a greater AUC for the neck device temperature. Although this study was the first to identify sex differences during exercise and recovery from low-intensity, steady-state exercise, it is not without limitations. It is important to note that, although thermal behavior was greater (i.e., neck device temperature was lower) in females compared with males during exercise, the actual neck skin temperature measured under the device was not different between sexes (Fig. 2). The explanation for this finding is not obvious. However, it is possible that a given magnitude of change in neck device temperature, due to a rapid change in perfusate fluid temperature, may be less efficient when transferred to the neck in females. Therefore, neck skin temperature may not be influenced to the same extent between sexes. This contention is supported by greater subcutaneous fat at the posterior neck in females compared to males (30). Despite this, we interpret these findings to indicate that females required a greater cooling stimulus to maintain neck thermal comfort, thereby using thermal behavior to a greater extent during exercise when compared to males.
It has been shown that females (31) and males (32) have reduced sensitivity to cool stimuli during exercise, independently. However, whether there are differences in cool sensitivity between sexes during exercise has not been directly tested. That said, females have a greater overall perceptual sensitivity (i.e., thermal sensation and discomfort) to a given thermal stimuli compared with males at rest (10), and to warm stimuli compared with males during exercise (12). Therefore, it is possible that a given magnitude of increase in body temperatures may result in greater thermal discomfort for females. Such differences could help explain the greater thermal behavioral response by females in our study. Our data show that body temperatures were elevated to a similar, if not slightly greater degree in females. Moreover, females perceived increases in whole body and neck thermal discomfort toward the end of exercise, whereas in males, these perceptions did not differ from preexercise throughout exercise. Nevertheless, there were no differences between sexes. However, the effect sizes related to these variables can help to elucidate these differences, whereby the P value and effect sizes of core temperature (P = 0.14, d = 0.49) and skin temperature (P = 0.36; d = 0.15) do not suggest there were differences, whereas those for thermal discomfort (P = 0.09, d = 0.62) do approach significance, with a medium sized effect. Hence, it is possible that these changes in thermal discomfort over time in females may have stimulated greater behavioral thermoregulation, in response to similar increases in body temperature to males.
Another explanation could be that, at the same rate of metabolic heat production, females exercised at a greater relative intensity, determined from the percentage of their estimated maximal heart rate elicited by the exercise protocol. This is important because the hypoanalgesic effect of exercise, which is believed to cause attenuated thermal sensitivity during exercise, is greater with increases in relative exercise intensity (33). Notably, in the present study females cycled at approximately 70% of their estimated maximal heart rate, whereas males were only at approximately 58%. Additionally, females had a tendency for slightly greater whole body thermal discomfort and neck thermal discomfort throughout exercise. Collectively, these findings indicate that factors other than the direct thermal afferent stimulus (e.g., mean body temperature), may contribute to thermal behavior during exercise. Indeed, we previously reported that mean body temperature accounts for only approximately 50% of the variation related to thermal behavior (6). Thus, further investigation is warranted toward identifying if factors such as fitness, relative exercise intensity, or other afferent stimuli (e.g. skin wetness) (34) also contribute to thermal behavior during exercise.
Sex differences in thermal behavior during recovery
Temporal data did not reveal any differences between males and females in body temperatures or thermoeffectors during recovery, except that mean skin temperature, although not significant between sexes, dropped within the first 10 min of recovery for females, whereas mean skin temperature in males did not decrease until the final 10 min of recovery. A greater AUC for mean body temperature in females during recovery highlights these temporal differences, whereby both mean skin and core temperatures were reduced in females to a greater extent than in males. Interestingly, these findings are in contrast to previous research which has shown that core temperature in females remains elevated compared to males during recovery from exercise (13). This can be due to differences in biophysical characteristics, or in response to greater postexercise hypotension in females, which can elicit deleterious effects on skin blood flow and sweat rate responses, thus mitigating the release of heat from the body core. Our data do not reveal differences in postexercise hypotension (Table 2) nor reductions in skin blood flow or sudomotor activity during recovery, probably due to the relatively low exercise intensity. Therefore, it is appropriate to conclude that the temporal core temperature dynamics did not differ to males in the present study. Nevertheless, changes in mean body temperature are regarded to be the primary afferent stimulus to thermal comfort and thermal behavior. Therefore, to further investigate behavioral and autonomic thermoeffector activation during recovery, we analyzed our data the minute that mean body temperature returned to preexercise levels, as we have done previously (6). At this time, the mean thermal afferent stimulus for thermal behavior was effectively identical between preexercise and postexercise periods in all subjects included in the analysis, allowing for direct comparisons between males and females on thermoeffector responses during exercise recovery. This analysis revealed differential dynamics by which mean body temperature returned to preexercise levels between males and females. Specifically, core temperature remained elevated, but was countered by reductions in mean skin temperature, compared to preexercise in females. In contrast, core temperature and mean skin temperature in males were not different to preexercise levels at the same mean body temperature. Consequently, thermal behavior (i.e., neck device temperature) was engaged in females, but not males. Interestingly, every female whose core temperature remained elevated at this time point by at least +0.2°C (n = 4), preferred a neck device temperature that ranged from −1.4°C to −8.6°C. In contrast, no males had core temperatures greater than +0.2°C at this time point, nor a neck device temperature lower than −1.4°C (range, −1.4°C to 0.09°C). These data provide further insights into differences in thermal behavior during recovery elicited by prior exercise. Namely, it appears that when the afferent stimulus for thermal behavior returns to preexercise levels, an independent stimulus (e.g., core temperature) can continue to drive thermal behavior. In an earlier study, we included equal numbers of males and females, and reported that thermal behavior remains engaged during recovery from exercise, largely due to persistently elevated core temperatures (6). The present data indicate that this is driven largely by females.
There are several methodological considerations that warrant discussion. First, the ideal protocol to identify sex differences in thermoregulation employs a fixed rate of metabolic heat production relative to body mass to eliminate differences in body size and evaporative heat loss requirements (35). We did not intentionally match groups for body surface area, percent body fat or fitness. However, our data revealed no differences in the metabolic heat production and evaporative heat loss requirements relative to body mass or body surface area. Therefore, we believe our data reflect appropriate sex differences for thermoregulatory variables during exercise. Nevertheless, differences in percent body fat, and error associated with estimating body fat from skinfold thickness, are important factors considering the potential insulating properties of subcutaneous fat tissue, which should be further investigated. However, although a protocol matching for body size can help control important sex differences, it should not be discounted that there are population wide differences between males and females as it relates to body size, and some ecological validity may be lost in the practice of matching for body size in laboratory studies. In addition to this, it should be noted that exercising at a fixed rate of metabolic heat production is not always an ecologically valid protocol, as under typical conditions, individuals would be able to self-regulate their exercise intensity and pace based on relative exercise intensity. Nevertheless, the approach employed herein provides important insights underlying the potential for differential thermal behavioral responses between males and females during exercise. A second consideration is that we had females self-report their menstrual cycle. A consequence of this is that we do not definitively know their hormone levels at the time of testing. Furthermore, we only tested females within the follicular phase (first 10 d) of their menstrual cycle. Previous research indicates that females have elevated core temperatures at baseline and during exercise when in the luteal phase of their cycle (36) and therefore, it could be presumed that at other times during the menstrual cycle, thermal behavior during exercise may be relied on to an even greater extent. A third limitation resides within our core temperature measurements which, when indexed by a telemetry pill within the gastrointestinal tract, changes at a slower rate than esophageal temperature (37). Thus, there is a possibility that the dynamics of core temperature were underestimated. In this study, we evoked relatively minimal thermoeffector and hemodynamic responses using low-intensity exercise. The main limitation of this is that we did not see differences in postexercise hypotension between males and females that are hypothesized to be a large reason for core temperature remaining elevated in females (13), secondary to attenuated skin blood flow and sudomotor responses postexercise. It may be appropriate, therefore, to identify the effect of exercise intensity on thermal behavior, given that postexercise hypotension is greater with increases in relative exercise intensity (38). Finally, recent evidence indicates that differences in cutaneous thermal sensation between sexes may be a function of the relative surface area of skin stimulated (39). It cannot be excluded that the findings presented herein are not confounded by the greater relative surface area covered by the neck device in females. However, if this were the case, we would have expected that the females would have used thermal behavior to a lesser extent because a given amount of neck cooling to the same absolute neck surface area should feel cooler in females compared to males.
The findings from this study highlight sex-related differences during and after exercise, whereby females used thermal behavior to a greater extent that males to maintain thermal comfort. That said, it is important to further investigate how physical characteristics (e.g., body surface area, body composition, etc.) may contribute to the greater thermal behavioral response during exercise in females. Specifically, controlling for subcutaneous fat thickness at the dorsal neck would be particularly important when employing the model used in the present study. Additionally, it is important to elucidate the effects of thermal behavior when used during prolonged sport settings with intermittent recovery periods that occur in summer months and/or in hot and humid environments, to restore thermal comfort and improve performance (40). These findings also stimulate further questions in regard to whether or not populations with altered thermoregulatory processes use thermal behavior appropriately, and if it can help to alleviate symptomology, thereby increasing comfort and exercise compliance of an individual. Finally, it is possible that in different conditions (e.g., greater ambient temperatures and higher exercise intensities), an extended recovery and/or different modalities that promote behavioral thermoregulation to improve heat loss and reduce core temperature may be beneficial. These findings might be applicable to specific occupational settings, such as firefighting (29), and could lead to behavior-mediated therapies that promote greater thermal comfort during recovery from exercise.
Findings from this study reveal that females rely on thermal behavior to a greater extent when exercising at a fixed rate of metabolic heat production that increased body temperatures to similar levels as in males. Furthermore, during recovery, the dynamics of body temperatures returning to preexercise levels differ between males and females. This resulted in elevated core temperatures in females which appear to elicit a greater thermal behavioral response postexercise compared with males.
This study was funded by Lululemon Athletica Inc. The authors would like to thank both Lululemon Athletica Inc. and the subjects for their participation. Rob Gathercole is the Research Director for Whitespace at Lululemon Athletica Inc. Zachary Schlader, Rob Gathercole and Blair Johnson contributed to the conceptualization of this study. Chris Chapman, James Sackett, and Zachary Schlader were fully responsible for data collection. Nicole Vargas was responsible for data analysis and drafting of the article. All authors approved the final manuscript. Zachary Schlader has received travel reimbursement from Lululemon Athletica Inc. The results of this study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. The results do not constitute endorsement by ACSM.
1. Schlader ZJ, Sackett JR, Sarker S, Johnson BD. Orderly recruitment of thermoeffectors in resting humans. Am J Physiol Regul Integr Comp Physiol
2. Flouris AD, Schlader ZJ. Human behavioral thermoregulation
in the heat. Scand J Med Sci Sports
3. Cabanac M, Cunningham D, Stolwijk J. Thermoregulatory set point during exercise
: a behavioral approach. J Comp Physiol Psych
4. Bulcao CF, Frank SM, Raja SN, Tran KM. Relative contribution of core and skin temperatures to thermal comfort
in humans. J Therm Biol.
5. Schlader ZJ, Simmons SE, Stannard SR, Mündel T. The independent roles of temperature and thermal perception in the control of human thermoregulatory behavior
. Physiol Behav
6. Vargas NT, Chapman CL, Sackett JR, et al. Thermal behavior remains engaged following exercise
despite autonomic thermoeffector withdrawal. Physiol Behav
7. Cramer MN, Jay O. Biophysical aspects of human thermoregulation
during heat stress. Auton Neurosci Elsevier BV
8. Gagnon D, Kenny GP. Does sex have an independent effect on thermoeffector responses during exercise
in the heat? J Physiol
9. Gagnon D, Kenny GP. Sex differences
in thermoeffector responses during exercise
at fixed requirements for heat loss. J Appl Physiol
10. Golja P, Tipton MJ, Mekjavic IB. Cutaneous thermal thresholds—the reproducibility of their measurements and the effect of gender. J Therm Biol.
11. Lautenbacher S, Strian F. Sex differences
in pain and thermal sensitivity: the role of body size. Percept Psychophys
. 1991; 50(2):179–83.
12. Gerrett N, Ouzzahra Y, Coleby S, Hobbs S, Redortier B, Voelcker T, et al. Thermal sensitivity to warmth during rest and exercise
: a sex comparison. Eur J Appl Physiol
. 7 ed. 2014;114(7):1451–62.
13. Kenny GP, Jay O. Sex differences
in postexercise esophageal and muscle tissue temperature response. Am J Physiol Regul Integr Comp Physiol
14. Bois Du D, Bois Du EF. A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med
. 1916; 73:863–71.
15. Siri WE. Body composition from fluid spaces and density: analysis of methods. In: Brozek J, Henschel A, editors. Techniques for Measuring Body Composition
. Washington, DC: Nutrition (Burbank, Los Angeles County, Calif.); 1961. pp. 223–44.
16. Jackson AS, Pollock ML. Generalized equations for predicting body density of men. Br J Nutr
17. Jackson AS, Pollock ML. Generalized equations for predicting body density of women. Med Sci Sports Exerc
. 1980;12(3): 175–81.
18. Craig CL, Marshall AL, Sjostrom M, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc
19. Nassreddine ZS, Phillips NA, Bedirian V, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc
20. Stolwijk JA, Hardy JD. Partitional calorimetric studies of responses of man to thermal transients. J Appl Physiol
. 1966; 21(3):967–77.
21. Liu W, Lian Z, Deng Q, Liu Y. Evaluation of calculation methods of mean skin temperature for use in thermal comfort
study. Build Environ
22. Guéritée J, Redortier B, Tipton MJ, House JR. Thermal comfort
following immersion. Physiol Behav
23. Graichen H, Rascati R, Gonzalez RR. Automatic dew-point temperature sensor. J Appl Physiol
24. Gagge AP, Stolwijk JA, Hardy JD. Comfort and thermal sensations and associated physiological responses at various ambient temperatures. Environ Res
25. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc
26. Cabanac M, Massonnet B, Belaiche R. Preferred skin temperature as a function of internal and mean skin temperature. J Appl Physiol
27. Nakamura M, Yoda T, Crawshaw LI, et al. Relative importance of different surface regions for thermal comfort
in humans. Eur J Appl Physiol
28. Gagnon D, Jay O, Kenny GP. The evaporative requirement for heat balance determines whole-body sweat rate during exercise
under conditions permitting full evaporation. J Physiol
. 2013;591(11): 2925–35.
29. Colburn D, Suyama J, Reis SE, et al. A comparison of cooling techniques in Firefighters after a live burn evolution. Prehosp Emerg Care
30. Fournet D, Ross L, Voelcker T, Redortier B, Havenith G. Body mapping of thermoregulatory and perceptual responses of males and females running in the cold. J Therm Biol.
2013 Aug 1;38(6):339–44.
31. Gerrett N, Ouzzahra Y, Redortier B, Voelcker T, Havenith G. Female thermal sensitivity to hot and cold during rest and exercise
. Physiol & Behav
32. Ouzzahra Y, Havenith G, Redortier B. Regional distribution of thermal sensitivity to cold at rest and during mild exercise
in males. J Therm Biol
33. Koltyn KF. Exercise
-induced hypoalgesia and intensity of exercise
. Sports Med
34. Fukazawa T, Havenith G. Differences in comfort perception in relation to local and whole body skin wettedness. Eur J Appl Physiol
35. Cramer MN, Jay O. Selecting the correct exercise
intensity for unbiased comparisons of thermoregulatory responses between groups of different mass and surface area. J Appl Physiol (1985)
36. Pivarnik JM, Marichal CJ, Spillman T, Morrow JR. Menstrual cycle phase affects temperature regulation during endurance exercise
. J Appl Physiol
37. Mündel T, Carter JM, Wilkinson DM, Jones DA. A comparison of rectal, oesophageal and gastro-intestinal tract temperatures during moderate-intensity cycling in temperate and hot conditions. Clin Physiol Funct Imaging
38. Jones H, George K, Edwards B, Atkinson G. Is the magnitude of acute post-exercise
hypotension mediated by exercise
intensity or total work done? Eur J Appl Physiol
39. Filingeri D, Zhang H, Arens EA. Thermosensory micromapping of warm and cold sensitivity across glabrous and hairy skin of male and female hands and feet. J Appl Physiol (1985)
40. Chalmers S, Jay O. Australian community sport extreme heat policies: limitations and opportunities for improvement. J Sci Med Sport