INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CIRCADIAN RHYTHMS IN MAMMALS are generated and maintained by an internal clock located in the suprachiasmatic nuclei (SCN) of the anterior hypothalamus (21, 31). The phase of this clock is regulated directly by photic information relayed from the retina to the SCN via the retinohypothalamic tract (15, 27) and indirectly via a projection from the intergeniculate leaflet (IGL), the geniculohypothalamic tract (5, 14). The IGL projection to the SCN is also essential for conveying nonphotic (behavioral) information to the circadian clock (2, 11, 12). It is now widely held that nonphotic as well as photic inputs are important for circadian clock regulation. Both stimuli have pronounced phase-resetting actions on the circadian clock, albeit with different phase-response characteristics. For example, nonphotic phase-response curves (PRCs), such as those produced from novelty-induced locomotion or triazolam, have phase-delay portions toward the end of the subjective night and phase-advance portions in the mid-to-late subjective day (23-25, 30, 33). Photic PRCs, on the other hand, are characterized by phase delays in the early subjective night and phase advances in the late subjective night (13). The distinct behavioral and physiological differences between photic and nonphotic shifting mechanisms could be a general feature of mammalian circadian pacemakers. From an ecological standpoint, these differences could underlie the capacity to respond adaptively to a variety of circadian time-keeping cues (zeitgebers).

Despite abundant literature on circadian entrainment, there is surprisingly little information on this subject in diurnal mammals. Contributing to a lack of understanding in this area is the problem of separating photic from nonphotic effects in diurnal species. If the nonphotic PRC is tied to the active phase of the sleep-wake cycle, as some have assumed (8, 32), then the advance portion of the diurnal nonphotic PRC would be 180° out of phase with that of nocturnal animals. In hamsters and mice, phase advances can be induced by behavioral arousal during the mid-to-late subjective day; therefore, it would be expected that such phase advances in diurnal mammals would occur during the mid-to-late subjective night. However, this corresponds to the phase-advancing region of their photic PRC. As this is a phase when light resets the clock, it must be ensured that manipulations intended as nonphotic do not result in a change in the amount of light received by the subject. Conversely, manipulations intended as photic must be made so as not to induce behavioral responses, because light may have a positive masking effect that increases arousal and activity.

The above concerns are exemplified by previous experiments undertaken in a diurnal monkey, the common marmoset (Callithrix jacchus). In two such studies (26, 34), pulses of induced activity produced light-like nonphotic PRCs in marmosets free running under constant light (LL). It is possible that the animals received more light during the activity pulse than if they had remained asleep. Conversely, in an earlier study, in which a photiclike PRC was generated by light pulses in marmosets under constant dim light (35), increased activity could have occurred in response to the light pulse. In view of the potential for confounds between different zeitgebers when generating PRCs in diurnal mammals, the present study was undertaken to examine phase resetting associated with manipulations designed to produce events that were primarily photic (a brief bright light pulse to minimize behavioral influence) or nonphotic (an activity pulse induced under total darkness to minimize light influence). The second aim was to exploit these procedures to determine if the nonphotic PRC of this diurnal species is phased relative to that of nocturnal hamsters or mice. We report here that the phase responses to entraining photic and nonphotic stimuli in the marmoset appear to be qualitatively similar to those in nocturnal mammals.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals

Activity Rhythm Measurements and Calculations

Circadian Entrainment Procedures

Light-pulse protocol. In the first experiment, marmosets (n = 4) were maintained under constant dim incandescent light (0.5 lx) for a minimum of 2 wk before light-pulse treatment. The free-running circadian activity rhythm was measured throughout the experiment. Photic stimulation consisted of a 20-s pulse delivered from overhead fluorescent tubes (~450 lx at cage level) together with high-intensity stroboscopic flashes (5 Hz; 100-µs flash duration; estimated ~100 lx flash intensity at cage level) from a xenon source pointed at the animals from a distance of ~4 m. The stroboscopic flashes were included to increase the impact of the short photic stimulus. Viewing of videos confirmed that the animals were awake during this manipulation, with their eyes open and facing toward the nest box entrance. Observations with infrared goggles as well as analyses of infrared motion detector output also confirmed that the majority of animals resumed resting after the pulse without exhibiting locomotor activity beyond their normal circadian activity period. For studying photic entrainment, the light-pulse procedure was initiated on day 1 near the end of the animals' subjective day (1500 external time) and was continued at the same external time over 12 consecutive days. The animals were then left undisturbed for a subsequent 2-wk period.

Behavioral pulse protocol. Marmosets (n = 6) were initially maintained under DimDim for a 3-wk period before treatment. The free-running circadian activity rhythm was monitored throughout the experiment. The behavioral pulse consisted of the continuous arousal of the animals by an investigator present in the experimental chamber over a 1-h period from 1400 to 1500 external time. Continuous activity was stimulated as needed by intermittent sprinkling with water and agitation of the cages. The constant dim lighting was turned off from 1400 to 1530 external time to ensure that the induced activity, as well as any residual activity after the disturbance, occurred in total darkness. Infrared night vision goggles were used to facilitate the experimental manipulations. This treatment was undertaken continuously over a 76-day period, after which time the animals were left undisturbed for 2 wk under DimDim. A schedule of control pulses of total darkness from 1400 to 1530, in the absence of induced activity, was undertaken over a subsequent 42-day period also under DimDim.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

General Characteristics of Light-Entrained and Free-Running Circadian Activity Rhythms

Photic Entrainment

View larger version (67K): [in this window] [in a new window]   Fig. 1.   Double-plotted actograms of the 4 animals used in the light-pulse entrainment trials. A 20-s light pulse was delivered over 12 consecutive days (days 10 to 21, designated by the vertical lines) in animals with a free-running activity rhythm under constant dim light conditions (DimDim). Days are indicated vertically from top to bottom, and time of day is indicated horizontally. Each horizontal span represents 2 consecutive days. Identification numbers of the animals are at the top left corner of the actograms.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Phase responses to a single 20-s light pulse delivered at various times throughout the circadian day in animals with a free-running activity rhythm under DimDim. The point at circadian time (CT) 10.9 represents the calculated entraining phase response (means ± SE) to the 12-day light-pulse regimen in Fig. 1. For comparison,  represents the calculated entraining phase response (means ± SE) to the induced activity pulse regimen in Fig. 3.

Nonphotic Entrainment

View larger version (88K): [in this window] [in a new window]   Fig. 3.   Double-plotted actograms of all 6 animals [5 males/1 female (animal 13)] used in the induced activity pulse experiment. Data are plotted as described in Fig. 1. Animals with a free-running activity rhythm under DimDim were subjected to a daily 1-h pulse of induced activity administered with an accompanying 1.5-h period of complete darkness denoted by the top vertical bars. This was carried out over 76 consecutive days (days 24 to 100). After an intervening 2-wk period under DimDim when no treatment was applied (days 101 to 114), the animals received daily control dark pulses delivered without induced activity as denoted by the lower vertical bars (days 115 to 160). Identification numbers of the animals are at the top left corner of the actograms.

View larger version (33K): [in this window] [in a new window]   Fig. 4.   Double-plotted actograms of the period between days 66 and 90 shown in Fig. 3 for animals 1, 13, and 16. During this period, there was a lengthening of the free-running period () as the phase of activity onset (subjective lights on; CT0) approached the scheduled activity pulse. The phase-delaying effect of the pulse is initially apparent near CT19-20.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present experiments show that nonphotic events falling in the late subjective night or early day are capable of producing phase delays in the common marmoset. This finding contrasts with two other studies, one from this laboratory, with the same species. In these studies, phase advances were produced by arousing stimuli falling in the late subjective night. The arousing stimuli were taped calls of conspecifics together with presentation of a mirror (34) and exposure to a novel cage (26). In both cases, the animals were kept in constant bright light. It is possible that the 2-h periods of behavioral arousal in the late subjective night would increase the amount of light received by animals normally asleep at this time. If that were the case, then the phase advance may have been primarily photic rather than nonphotic. This contention is consistent with a number of points. First, the phase-advance portion of the photic PRC peaks in the CT20-22 region (35; present results). Second, with marmosets kept in LL, phase advances of a magnitude similar to those induced by activity pulses at CT22 have been produced at the same time by continuous gentle waking of the animals, making them open their eyes but with little associated locomotor activity (J. D. Glass, unpublished data). Third, in the present work with the activity induced in darkness, there were no phase advances when this manipulation occurred toward the end of the subjective night. Collectively, these points attest to the need to consider the possible interference of light when assessing nonphotic events. In the present experiments, it would appear that this problem has been solved by the use of complete darkness during the arousing events and by the use of photic stimuli too brief to induce locomotor behavior. Although it is impossible to completely rule out an arousing (nonphotic) phase-shifting effect of the 20-s light pulse in the absence of induced activity, that different types of zeitgebers were successfully applied is evident from the fact that the phase angles of entrainment for the arousing stimuli and the photic stimuli were so different.

Although the input necessary for photic phase shifting is relatively well understood, the aspects of nonphotic arousing stimuli critical for phase shifting remain uncertain. Possible candidates involve being awakened, locomotion, novelty, and the arousing or motivating states accompanying behavioral manipulations (22-24). The present results are of some interest in this respect. In hamsters, it has been found that locomotion induced by cold is not as reliable in producing shifts as spontaneous locomotion in a novel wheel (12, 23). This has led to the suggestion that the extent to which the activity is subjectively rewarding may be important. However, in the present case, during the arousal, the marmosets made repeated attempts to retreat to their nest boxes and to avoid moving around. This indicates that in these experiments, positive motivation was not the cause of the activity-related nonphotic phase-shifting response.

It is possible, of course, that there are species differences in the type of nonphotic input effective in phase shifting. Scheduled locomotion on a treadmill seems less effective in rats (20) than mice (18). In ground squirrels, entrainment occurred in some individuals that hardly used wheels at all during a 3-h opportunity (10), whereas in hamsters the number of wheel revolutions is, in general, a good predictor of phase shifts (23). However, assessing species differences is difficult without knowing whether optimal zeitgebers have been used. For example, in the present experiments, it should not be assumed that the procedure used for arousing the marmosets is the most effective nonphotic phase-shifting manipulation. One feature of the present results is that the daily 60-min bouts of induced activity did not produce long-lasting entrainment. As is evident in Fig. 3, entrainment to the pulses did not persist, but lasted ~3 wk. Perhaps the marmosets became habituated to the procedures, although a second period of entrainment to the behavioral zeitgeber several weeks after the initial loss of synchronization would argue against this. Another possibility is that the duration and/or intensity of the stimuli were insufficient to maintain entrainment. With hamsters, the duration of novelty-induced locomotion is an important predictor of phase shifting (3). Nevertheless, even though the entrainment was not maintained, the data demonstrate unequivocally that when entrainment did occur, it was through delays produced by the pulses in the late subjective night.

This is relevant to attempts to study nonphotic phase shifting and entrainment in people. In such studies, attention has been focused on the early subjective night as a time for exercise (1, 32; reviewed in Ref. 28). This would appear to have been partly based on the assumption that nonphotic PRCs in people would be 180° out of phase with those in nocturnal hamsters. On the basis of the present data with diurnal primates, exploration of possible nonphotic effects at other phases of the cycle, including the subjective day, deserves more attention.

The need to produce purely photic or nonphotic stimuli in studies with people also should be addressed. For example, in the work of van Reeth et al. (32), exercise took place in an illumination of <300 lx. Their experiments were made before Boivin et al. (4) demonstrated that phase shifting of human rhythms could be obtained with illumination of only 180 lx. Although the eyes should have been kept open during a constant routine, it is not clear whether the amount of light reaching the retina was the same during the exercise. Notwithstanding such cautions, the report of schedule-induced entrainment in a single blind subject is enough to suggest that purely nonphotic entrainment is possible in people (17).

Although the present discussion has focused on possible species differences, differences between experiments, and potential confounds between variables, it is also beginning to become apparent that there is some commonality in nonphotic phase shifting, just as there is in photic shifting. In experiments with diurnal ground squirrels with a free-running period longer than the frequency of presentation of a running wheel (every 23.5 h), entrainment occurred by phase advances. Such advances fell in the late subjective day or early subjective night and were preceeded by activity. On the other hand, the present results indicate that in the marmosets with a free-running period less than the frequency of induced activity (every 24.0 h), entrainment occurred by phase delays. These delays fell in the late subjective night or early subjective day and were followed by activity. Although a multiple-point nonphotic PRC was not generated in the present study, the different phase angles of entrainment between ground squirrels and marmosets are nevertheless consistent with their nonphotic PRCs being similar and also similar to those of nocturnal rodents [i.e., in hamsters (24, 30), mice (6, 7, 18, 19), and rats (20), nonphotic advances have been produced in the mid-to-late subjective day and nonphotic delays in the late subjective night]. It has also been reported that in degus exhibiting a diurnal activity pattern, scheduled access to a wheel can phase delay the circadian clock at similar times of day as in nocturnal rodents (16).

In conclusion, these results in a diurnal primate are broadly similar to those discussed above for nocturnal rodents studied so far. Limited data for sparrows are anomalous (29), but for mammals the data are consistent with the view that nonphotic phase resetting is keyed to the cycles of day and night, or subjective day and night, regardless of whether the species is nocturnal or diurnal.