INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ALTHOUGH LIGHT EXPOSURE is considered the most important regulator of the mammalian circadian system, studies of rodents and humans have indicated that physical activity and exercise also have a profound influence on circadian timing. This effect is particularly well established in rodents.

In rodents, physical activity can entrain the circadian system (8, 15, 36), enhance the rate of reentrainment to a shifted light-dark cycle (32, 34), alter circadian period (16, 36), and elicit dramatic phase shifts in constant conditions (18). Moreover, physical activity apparently moderates phase-shifting effects of other zeitgebers (24, 25). Physical activity and light-induced phase shifts in nocturnal rodents are mediated by different neuroanatomical pathways and are characterized by near-reciprocal phase-response curves. Whereas light elicits phase delays when administered in the early evening and phase advances in the late subjective night (23), physical activity elicits phase advances during the mid-day to early night and phase delays during the mid-night to early day (36).

Preliminary evidence in humans indicates that exercise can shift circadian rhythms in constant conditions (7, 39) and accelerate entrainment to a shifted sleep-wake schedule (14) or to a shortened sleep-wake period (31). Indeed, Van Reeth and colleagues (39) found that 3 h of moderate (50%), intermittent exercise (5 × 30-min intervals, with 6-min rest intervals) elicited approximately the same phase shift as a 3-h 5,000-lx light pulse. Thus exercise may provide an alternative or adjuvant phase-shifting stimulus with similar potency as bright light.

Particularly intriguing are findings that physical activity and light have phase-shifting interactions in rodents. Whereas these two stimuli can substantially antagonize each other's phase-shifting effects (sometimes completely) when administered within a few hours of each other (4, 26, 33, 35), they can have synergistic effects when separated by several hours (33, 34, 37). The interactions of physical activity and light are apparently complex and cannot be predicted from a simple summation of the phase-response curves for each stimulus (4, 26, 33). For example, profound antagonism can occur even at times in which the antagonizing stimulus elicits no independent phase shift (4, 26, 33).

Because humans and animals show homologous phase-shifting responses to exercise and light, it is plausible that these stimuli will elicit similar phase-shifting interactions in humans. Such interactions might have considerable practical utility for shiftworkers, transmeridian air travelers, and other individuals with circadian malsynchronization. Dramatic resetting of the circadian pacemaker might be possible after a few days of appropriately timed light and exercise. Conversely, alterations in physical activity levels might attenuate or defeat attempted phase-shifting by bright light exposure.

The interaction is also of interest, because people often experience exercise and bright light simultaneously when exercising outdoors or in well-lit indoor facilities. In humans, bright light generally elicits phase delays and phase advances during the hours before and after the circadian temperature nadir, respectively (10, 19, 25). Although a full PRC for exercise in humans has not been published, Van Reeth et al. (39) and Buxton et al. (7) reported that exercise elicits phase delays in humans when administered at times varying from ~4 h before to 4 h after the temperature nadir.

Only one previous study has examined the circadian interaction of exercise and bright light in humans. In an 8-day simulated night-shift study, Baehr et al. (3) used a two-by-two design comparing exercise vs. no exercise and bright light vs. dim light. Treatments were administered during the first 6 h of the first 3 night shifts. Light exposure involved 40-min 5,000-lx pulses each hour for 6 h. Exercise involved cycling during the last 15 min of each treatment hour at 50-60% heart rate reserve. Whereas light alone elicited a significant phase delay, exercise alone elicited no significant shift, and no significant interaction was observed between exercise and light treatments. However, the exercise may have been of insufficient intensity or continuity to interact with light. Both human and animal studies have suggested that larger phase shifts can be elicited by more vigorous and/or continuous exercise (5, 7). Animal work also shows an intensity-dependent antagonism of light-induced shifts (27, 28, 35).

Given the complex nature of the phase-shifting interaction between physical activity and light in rodents, an understanding of the interaction in humans will likely require a series of investigations. The focus of the present investigation was to observe whether vigorous, prolonged exercise could modulate phase-shifting effects of late night/early morning bright light. On the basis of evidence that the delay portions of the PRCs for exercise and bright light coincide at this time, this was considered an appropriate starting point for investigation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Volunteers

Design

Procedures

Peak O₂ consumption testing. One to two weeks before the experiment, participants completed a graded, maximal cycle ergometer test to volitional exhaustion. The attainment of O_2 peak was defined by a plateau of oxygen consumption with increasing work rate or by heart rate within 10 beats/min of age-predicted maximum (220  age), plus respiratory exchange ratio >1.10. Metabolic measurements were made using a MedGraphics Metabolic Cart. The purpose of this test was to verify the high fitness levels of the participants and to establish the peak heart rate, from which target work rates during the experimental cycling treatment were determined.

Home recording. During the week before each laboratory treatment, volunteers were requested to maintain a stable sleep-wake schedule consistent with their usual habits (i.e., bed and wake times not varying by >2 h). Adherence to this schedule was verified by a daily diary and Actillume wrist movement and illumination recording.

Laboratory recording. The laboratory protocols were scheduled from ~1700 on a Friday evening to 0800-1200 on the following Monday morning (3 nights and 2 days). Illumination was maintained at a constant level of 50 lx during wakefulness (except during the experimental treatments) and 0 lx during the sleep periods. Ambient temperature was maintained at 20-21°C.

Bright Light Alone. Bright light was administered for 3 h by 1,600-W cool-white fluorescent lighting built into the ceiling, providing 3,000-lx illumination evenly throughout the room. Volunteers were invited to read or watch TV in a seated position. Continuous video monitoring verified that the participants remained awake. As a control for necessary social interactions during the Bright Light + Exercise treatment, experimental staff interacted with the volunteers approximately every 10 min during the Bright Light Alone treatment.

Bright Light + Exercise. During the Bright Light + Exercise treatment, volunteers cycled on an ergometer for 3 h at 65-75% of heart rate reserve. For example, assuming a maximum heart rate of 200 beats/min and a resting heart rate of 50 beats/min, 65% of heart rate reserve would be (200  50) × 0.65 = 97.5 + resting heart rate = 147.5 beats/min, and 75% heart rate reserve would be (150 × 0.75) + 50 = 162.5 beats/min. Maintenance of this exercise intensity was verified by a Polar heart rate monitor, which stored heart rate every minute and sounded an alarm whenever the participant's heart rate was outside of this zone. Minor adjustments in work rate were made to maintain the appropriate exercise intensity. No participant spent >10 min outside of the target heart rate zone, including 1-3 min "warm-up" and "warm-down" periods at the beginning and end of the 3-h session. An electric fan was focused on the volunteers. The exercise treatments were monitored by laboratory personnel who entered the room approximately every 10 min, providing water, food, and encouragement. The room lighting was maintained at 3,000 lx during the 3 h of exercise. Many of the participants watched TV during the treatment. It was verified with video monitoring and staff visits that the participants kept their eyes open during the exercise.

Assays

Data Analysis

The circadian rhythm of 6-SMT excretion was derived from 24-h collection and was presumably less masked compared with the temperature acrophase. Therefore, timing of the experimental stimuli was defined relative to each individual's baseline 6-SMT acrophase. The circadian acrophase of urinary 6-SMT coincides closely with the circadian temperature nadir, both occurring ~60-90 min after the plasma melatonin acrophase (2). The circadian timing of treatment (treatment phase angle) was determined retrospectively by subtracting the time of midpoint of the 3-h stimulus from the time of the baseline 6-SMT acrophase. Thus treatments centered before or after the 6-SMT acrophase were considered negative or positive treatment phase angles, respectively.

Phase shifts in the circadian acrophases of 6-SMT and rectal temperature were determined by subtracting the acrophase for the last 24-h period in the laboratory (final phase) from the acrophase for the 24 h immediately preceding the treatment (baseline). Temperature nadir shifts between nights 1 and 3 were similarly calculated. According to convention, a negative phase shift indicated a delayed phase after treatment, whereas a positive phase shift indicated an advanced phase.

The correspondence of 6-SMT and rectal temperature as estimates of circadian phase was assessed with Spearman rank-order correlations, combining baseline data from both treatments.

Because of high expectation that bright light would elicit a phase delay, the significance of phase shifts in the acrophases of 6-SMT and temperature after Bright Light Alone were assessed with 1-tailed t-tests. The significance of 6-SMT and temperature rhythm shifts after Bright Light + Exercise were assessed with 2-tailed t-tests.

Phase shifts in the rhythms of 6-SMT and temperature were compared between Bright Light Alone and Bright Light + Exercise with paired t-tests. Furthermore, the slopes of the linear regressions between the timing of the stimuli and the phase shifts were compared with t-tests. The associations between shifts in the two phase markers elicited by each treatment were assessed with Spearman rank-order correlations. Power calculations were conducted with Sample Power, a program developed by Borenstein et al. (6).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Volunteer Characteristics and Data Selection

1

1

Correlations Between Baseline Phase Markers and Phase Shifts

Phase Angles Between Treatment Midpoint and Baseline 6-SMT Acrophase

6-SMT Phase Shifts

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Phase shifts in 6-sulphatoxymelatonin (6-SMT) acrophase after Bright Light Alone and Bright Light + Exercise. A: individual data points. The abscissa displays the timing of the midpoint of the 3-h stimuli relative to the baseline 6-SMT acrophases. The ordinate displays the phase shift from baseline to final assessment after the treatments. Positive and negative values represent phase advances and delays, respectively. B: the mean (±SE) phase shifts after the treatments.

View larger version (43K): [in this window] [in a new window]   Fig. 2.   6-SMT excretion patterns for a representative subject during Bright Light Alone (A-C) and Bright Light + Exercise (D-F). The shaded bars under the abscissa indicate the sleep intervals. The arrows in A and D indicate baseline estimates of 6-SMT acrophase during the first 24 h in the laboratory. Bright Light (B) and Bright Exercise (D) were administered from 2300 to 0200 on night 2. Final 6-SMT acrophase was assessed over the last 24 h in the laboratory; acrophase estimates are indicated by the arrows in C and F.

Temperature Nadir Shifts

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Phase-shifts in temperature nadir after Bright Light Alone and Bright Light + Exercise. A: individual data points. The abscissa displays the timing of the midpoint of the 3-h stimuli relative to the baseline 6-SMT acrophases. The ordinate displays the phase shift from baseline to final assessment after the treatments. Positive and negative values represent phase advances and delays, respectively. B: the mean (±SE) temperature nadir phase shifts after the treatments.

View larger version (46K): [in this window] [in a new window]   Fig. 4.   Rectal temperature patterns for a representative subject during Bright Light Alone (A-C) and Bright Light + Exercise (D-F). The shaded bars under the abscissa indicate the sleep intervals. The arrows in A and D indicate visual estimates of baseline temperature minimum during night 1 in the laboratory. Bright Light (B) and Exercise (D) were administered from 2330 to 0230 on night 2. Final baseline minima estimates during the last night are indicated by the arrows in C and F.

Post Hoc Assessment of Discrepancy Between 6-SMT and Temperature Data

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The shifts in 6-SMT acrophase and temperature nadir after Bright Light Alone were more variable than in previous reports (10, 25). The size and direction of phase shifts are dependent on the experimental conditions, as well as the intensity and duration of light exposure. In the present study, the phase delays to light were likely limited by entrainment to the subjects' usual light-dark and sleep-wake schedules. Studies under entrained conditions have found equivalent or smaller average daily phase shifts to light at similarly sensitive portions of the phase-response curve (9, 11, 13), although a study in our laboratory (12) led us to expect larger shifts. Greater effects of light have been observed in constant routine (10) and free-running protocols (19), especially when there were several consecutive treatments. Published PRCs describe phase delays and advances to light administered before and after the circadian temperature nadir, respectively (10, 25). Nonetheless, there are numerous examples (19, 38) of apparent irregularities in the time of transition from delays to advances and the amplitude of phase shifts that define a human PRC.

Bright Light + Exercise elicited no significant difference in phase shifts compared with Bright Light Alone. These data are consistent with those of Baehr et al. (3), which showed no significant interaction between bright light and relatively moderate exercise (~50-60% O_2 peak) when administered intermittently in six 15-min bouts 0-6 h after usual bedtime. The present study tested continuous exercise of twice the duration and ~20-40% greater relative intensity (~70% O_2 peak). This extreme exercise also failed to significantly modify the effects of bright light.

In the present study, detection of exercise modulation of light shifts was limited by low statistical power, due partly to a high variability in phase shifts. Simulations showed that adequate power (0.80) to detect significant treatment effects would have required shift differences of ~2-2.5 h. Because phase shifts in 6-SMT and temperature acrophases after Bright Light Alone were only 20 and 70 min, respectively, antagonism of light shifts with exercise could not have been demonstrated. Nonetheless, there was little indication that such an antagonism occurred. Compared with Bright Light Alone, the delay in temperature acrophase was 8 min less, but the delay in the 6-SMT acrophase was 48 min greater after Bright Light + Exercise.

On the other hand, on the basis of other studies in humans showing phase-shifting effects of 2-4 h after much less vigorous exercise at a similar circadian time (39) and rodent studies indicating synergism between light and physical activity, a 2-2.5 h potentiation of light shifts with exercise could reasonably have been anticipated. Some indication of such a synergism was found in the 6-SMT results. Indeed, the effect size for the shift in 6-SMT acrophase was three times greater for Bright Light + Exercise compared with Bright Light Alone. However, this was not a reliable finding; much of the difference could be attributed to the four largest shifts after Bright Light + Exercise. Unfortunately, confirmation of these large shifts was not possible with the temperature data, which were lost for these subjects.

Numerous rodent studies have indicated that exercise antagonizes phase-shifting effects of bright light when performed within a few hours of the light exposure (26-28, 35). However, these data do not necessarily contradict the human data, because antagonism has been clearly established only for phase-advancing effects; much weaker antagonism has been found for phase-delaying effects of light (26-29).

Apparent differences in stimulus interactions between human and animal studies may be partly explained by species differences in exercise PRCs. Whereas the physical activity PRC in nocturnal rodents is ~180° out of phase with the light PRC, both stimuli have phase-delaying effects in humans during the 4 h immediately before the circadian temperature nadir (7, 10, 39). Thus a simple synergistic summation of the PRCs might be expected at times at which both stimuli elicit phase delays. However, this effect was not reliably seen. One apparent similarity between human and rodent data is that the phase-shifting interaction between light and activity may not be easily predicted from their respective PRCs.

The lack of synergism between phase-delaying effects of light and exercise might also be interpreted as an antagonism of bright light by exercise (or vice versa). Such an interpretation would not be consistent with animal studies, in which these stimuli elicit a phase-dependent cancellation of each other's phase-shifting effects. However, these possibilities cannot be ruled out.

The sympathetic-activating effects of exercise could have elicited an increased pupillary diameter in the Bright Light + Exercise treatment. Whereas some research has found no effect of exercise on pupillary diameter (40), other research has revealed a significant pupillary dilation that was more pronounced when exercise was of high intensity, long duration, and occurring in a bright environment (21). Because greater pupillary diameter should result in a greater phase-shifting effect of light (17), the failure of exercise to reliably enhance light effects in the present study suggests that pupillary dilation was not a significant factor.

Without placebo treatments, the extent to which the observed shifts are attributable to a free-running environment, stress, timing of daytime illumination (50 lx), or other factors associated with the protocol remains unclear. However, the literature provides a strong rationale for expecting phase-delaying effects of late night/early morning bright light. Moreover, whereas free run and illumination explanations would predict systematically small phase delays in the current protocol, a mixture of responses was found, including large phase advances and large phase delays as well as smaller shifts.

The present study did not test how exercise and bright light compare as phase-shifting stimuli in humans or whether exercise alone shifted the circadian system. However, other evidence suggests that exercise can influence the human circadian system. Indeed, much more modest exercise than reported herein has elicited significant phase shifts (39), and dose-response phase-shifting patterns of exercise have been documented in humans (7) as well as rodents (5, 22). Moreover, rodent data indicate that physical activity can substantially modulate the phase-shifting effect of light, even at times at which activity elicits no independent phase-shifting effect (4, 26, 33).

This initial study reveals little about dose-response patterns of interaction between exercise and bright light in humans. Thus it is plausible that, although high-intensity exercise may not reliably influence phase responses to 3 h of 3,000-lx illumination, moderate exercise might interact substantially with light of lower intensity or duration. A complete understanding of the interaction between exercise and light would require a more systematic evaluation of different intensities and durations of exercise and light at different circadian times.

A caveat in generalizing the present results is the extremely high fitness levels of the volunteers. It is plausible that greater effects of exercise might occur in less fit individuals, who have been examined in other studies (7, 39). Thus exercise attainable by the general population may prove useful. A possible hamster analogy is that previously sedentary hamsters have shown the most dramatic phase-shifting responses to increased physical activity associated with their first opportunity to use a running wheel (18). The results in the present study are also limited by an inability to recruit more women volunteers, despite special efforts to recruit equal numbers of women and men.

In summary, exercise had no consistent moderating effect on phase-shifting effects of late night/early morning bright light. These data are consistent with both human research and with rodent research showing no significant effect of physical activity on phase-delaying effects of light. It remains plausible, however, that exercise and bright light might have synergistic or antagonistic phase-shifting effects in humans at other times, analogous to the numerous such reports in rodents.