Social Contact Synchronizes Free-Running Activity Rhythms of Diurnal Palm Squirrels

Physiology & Behavior, Vol. 66, No. 1, pp. 21–26, 1999 © 1999 Elsevier Science Inc. Printed in the USA. All rights reserved 0031-9384/99/$–see front matter

PII S0031-9384(98)00271-6

Social Contact Synchronizes Free-Running Activity Rhythms of Diurnal Palm Squirrels SHANTHA M. W. RAJARATNAM1 AND JENNIFER R. REDMAN Department of Psychology, Monash University, Clayton, Victoria 3168, Australia Received 22 April 1998; Accepted 31 August 1998 RAJARATNAM, S. M. W. AND J. R. REDMAN. Social contact synchronizes free-running activity rhythms of diurnal palm squirrels. PHYSIOL BEHAV 66(1) 21–26, 1999.—Social contact with conspecifics entrains rhythms of a number of species, although convincing demonstrations of the phenomenon in diurnal mammals are limited. The present study examined the question of whether social contact mutually synchronizes free-running locomotor activity rhythms of the diurnal Indian palm squirrel, Funambulus pennanti. Twelve male squirrels were housed individually, without visual contact, in two separate laboratories (six in each laboratory). The squirrels were initially held under opposing light–dark (LD) schedules (with an 11 h phase difference), and were then placed under constant bright light (LL). Squirrels from separate laboratories were paired together, and each pair was placed into a fresh cage on the day of the pairing. After 48 days of social contact, the squirrel pairs were separated, and returned to their original positions in the two laboratories in fresh cages. Free-running phase and period were assessed prior to and after the social contact for each squirrel. The phase difference in the free-running rhythms of pairs of squirrels was significantly decreased following social contact. Actogram records revealed strong evidence of social synchronization of free-running rhythms in four of the six pairs. For the remaining two pairs, the data were ambiguous. This study confirmed the findings in other species, that social cues are a potent zeitgeber for F. pennanti. © 1999 Elsevier Science Inc. Funambulus pennanti Indian palm squirrel Entrainment Diurnal

Circadian rhythm

Nonphotic

A number of features of social entrainment appear to be similar to the effects of other nonphotic zeitgebers, such as induced activity or cage changing (13,18). Usually only a few animals tested will entrain or phase shift, because phase response curves (PRCs) to the stimuli are generally very low amplitude. In addition, phases of entrainment and shape of the PRCs are similar. These similarities imply a common underlying mechanism for nonphotic entrainment (13). In view of the life history features of the Indian palm squirrel (Funambulus pennanti), this species is an excellent model for studying the ability of social stimuli to synchronize freerunning circadian activity rhythms. F. pennanti is strongly diurnal, and remains active throughout the year. Palm squirrels usually associate in a large group and apparently recognize group members (23), but pairing and maintenance of territory usually occurs during the reproductive season (C. Haldar, personal communication). Furthermore, the circadian activity rhythms of F. pennanti respond to other nonphotic entraining agents such as exogenous melatonin (15) and ambient temperature cycles (16). Therefore, we tested the hypothesis that

SOCIAL cues are known to influence the circadian system in a variety of species (3,5,9,11,12,13,21,22). The term social cue is applied broadly in these situations, referring to any periodic signal emanating from another animal that has the capacity to entrain rhythms, via a visual, acoustic, olfactory, or tactile stimulus, or by some combination of stimuli. For example, some pairs of house sparrows (Passer domesticus) and deer mice (Peromyscus maniculatus) mutually synchronize activity rhythms when individuals are housed in neighboring cages (9,12), or in the same cage (1). Recently, social cues were shown to facilitate reentrainment to LD phase shifts in diurnal Octodon degus (5). In addition, timed, daily presentation of social stimuli partially entrained activity rhythms of O. degus (6). Marked species differences in the sensitivity of the circadian system to social stimuli may occur. For example, Syrian hamsters (Mesocricetus auratus) (13) and microchiropteran bats (Hipposideros speoris) (11) are entrained by social contact with conspecifics, but rats (Rattus norvegicus) (7) and Australian sugar gliders (Petaurus breviceps) (10) are not affected. 1To

Social cues

whom requests for reprints should be addressed. E-mail: [email protected]

21

22

RAJARATNAM AND REDMAN

social cues from conspecifics can serve as entraining agents for activity rhythms of F. pennanti. MATERIALS AND METHODS

Animals and Housing Twelve male Indian palm squirrels (F. pennanti), wild trapped from Perth Zoo, Western Australia, were housed individually in wire cages (60 3 40 3 40 cm) under 12:12 LD cycles, in temperature-controlled (25 6 0.58C), sound-attenuated laboratories within the Department of Ecology and Evolutionary Biology Animal House, Monash University. Food, consisting of a mixture of powdered rat pellets (GR21, Ridley Agri Products, Pakenham VIC, Australia) and parrot breeders mix (Vi-Trex Sales Pty. Ltd., Footscray VIC, Australia), and water were available ad lib. Fresh fruit or vegetables were provided three times each week. During routine laboratory maintenance every 7 days, saw dust trays beneath each cage were replaced, and food and water were replenished. Gross locomotor activity was monitored continuously using a passive infrared motion detection system (15,16). Actograms were generated using the Wheel Running program (Department of Psychology, Monash University, Clayton VIC, Australia). The experiments were conducted in two identical laboratories (A and B). Illumination was provided by two 100-watt incandescent globes mounted above light-lock cabinets located at the entrance to each laboratory. In both laboratories A and B, squirrels were housed on both sides, with a maximum of four cages on one side (two rows of two squirrels) and eight cages on the other side (two rows of four squirrels). The distance between horizontally adjacent cages was approximately 40 cm, and approximately 35 cm between vertically adjacent cages. White perspex screens were placed between adjacent cages to prevent visual contact between squirrels, and a black curtain was suspended in the middle of the laboratory. Experimental Protocol All squirrels were initially held in laboratory A under 12:12 LD cycles (lights on 0800 h, lights off 2000 h; 100:0 lx) for 20 days. The squirrels were then randomly assigned to two groups (social I and social II, n 5 6 per group). No significant difference in phase angle difference (PAD) under LD entrainment between the two groups was found, t(10) 5 0.76, p 5 0.467. Social II squirrels were moved to laboratory B, while social I squirrels remained in laboratory A. Different 12:12 LD cycles were imposed in each laboratory in order to induce a phase difference of approximately 11 h between the two groups; laboratory A: lights on 1100 h, laboratory B: lights on 2400 h. After stable entrainment to the new LD regimes had been established for all squirrels (Day 12), LL was imposed to induce freerunning. The mean light intensity in laboratory A was 103.0 lx (range 5 61 to 148 lx), and in laboratory B was 107.5 lx (range 5 60 to 160 lx). No significant difference in the light intensity levels was seen in the two laboratories, t(10) 5 0.22, p 5 0.828. As a result of housing the squirrels under differing LD schedules, the phase of activity onsets in the two groups of squirrels were sufficiently different to allow the experimenter to determine whether synchronization could occur. However, no significant difference in t between the two groups was found prior to the social contact t(10) 5 0.53, p 5 0.608. Squirrels were then put into pairs, comprising one squirrel from each of the laboratories, such that light intensity levels were

matched and the largest possible phase difference was gained for each pair. Previous experience with palm squirrels had demonstrated that males can be housed in groups of two or more after long periods of isolation without the occurrence of fighting. The procedure for placing squirrels into pairs was as follows. All social I squirrels were transferred to clean cages on Day 26. Their positions in the laboratory, and hence, the light intensity levels they were exposed to, remained the same. Within 45 min of this procedure, social II squirrels were moved into Laboratory A, and were placed into the same cages as their respective pairs. The social contact stage was imposed for 48 days. During this time, the quantity of fresh fruit and vegetables provided to each cage was doubled. On Day 74, social II squirrels were removed from their cages and were placed into clean cages in their original positions in laboratory B. Social I squirrels, which remained in the same positions in laboratory A, were also placed into clean cages. All squirrels were then monitored under these conditions for a further 29 days to assess the effects of social contact. Individuals within each pair of squirrels were identified at the end of the social contact stage on the basis of either the difference in weight between the animals (in the case of pairs 1 and 3, where the individuals differed by more than 20 g), or by some distinguishing physical feature that the animal possessed (e.g., natural markings on coat). Data Analysis Because pairs of squirrels cohabited the same cage during the social contact stage, activity records that were obtained for each cage during this time consisted of two squirrels’ data. The resources were not available to monitor individual behavior during the social contact. The group data, however, could be interpreted with some degree of accuracy, because the t and free-running phase were known for each individual prior to and after the social contact. A comparison of the t and phase of activity rhythms for the two squirrels in each pair prior to and after the social contact provided a measure of social effects. Furthermore, there was a large phase difference between members of pairs prior to social contact (Table 1); thus, activity rhythms of individuals in each pair were usually discernible during the social contact.

TABLE 1 DIFFERENCES IN THE PHASE OF ACTIVITY ONSET (MINUTES) BETWEEN THE TWO INDIVIDUALS IN EACH PAIR, CALCULATED IMMEDIATELY PRIOR TO AND FOLLOWING THE PERIOD OF SOCIAL CONTACT

Pair

1 2 3 4 5 6 Mean SE

A Phase Difference Before Social Contact (min)

584.5 683.9 588.0 634.0 645.8 708.9 640.85 20.43

B Phase Difference After Social Contact (min)

257.7 120.0 647.7 157.1 435.9 30.0 274.73 93.63

B2A ∆ Phase Difference (min)

2326.8 2563.9 159.7 2476.9 2209.9 2678.9 2366.12 109.00

The change in phase difference (minutes) as a result of the social contact is also reported.

SOCIAL SYNCHRONIZATION OF RHYTHMS Two independent, trained judges estimated t for 10 days prior to and after the social contact, and the mean value calculated. Activity onset was used as the reference phase for all except two squirrels, where activity offset was used because onsets were not clearly defined. The mean was also taken of the judges’ estimates of the free-running phase on Days 25 and 74, the days immediately prior to and following the period of social contact. Activity onset was used as the reference phase for phase determination in all cases. Previous studies have used the difference between the postsocial contact free-running phase and the phase that is predicted from the rhythm prior to social contact as a method of assessing social effects on the activity rhythm (7). In F. pennanti, spontaneous changes may occur in t and rhythm phase under constant conditions (Rajaratnam and Redman, unpublished observations), which would be incorrectly attributed to social effects according to this method of assessment. To overcome this problem, a different method of assessing social effects was used in the present study. To assess whether social entrainment between pairs of squirrels had occurred, the phase difference (in minutes) prior to and after the social contact was calculated for each pair. No significant difference in t was seen between social I and social II squirrels prior to the social contact; therefore the assumption was made that a significant change in the phase difference between activity onsets of squirrels in each pair would reflect the effects of the treatment. The treatment consisted of a cage change, followed by the period of social contact, fol-

23 lowed again by a cage change. A repeated-measures t-test was used to compare phase differences prior to and after the social contact. To determine whether any aftereffects of social entrainment had occurred, a one-tailed repeated-measures t-test was used to compare t prior to and after the social contact. RESULTS

Four of the six pairs showed strong evidence of social synchronization (Fig. 1). In the remaining two pairs, the data were ambiguous (Fig. 2). In one representative pair of squirrels that did show social entrainment, individual activity rhythms were not clearly discernible during the first part of the social contact stage (Fig. 1). A “combined rhythm” appeared after approximately Day 60. When the social contact ceased, the rhythms of the two squirrels free-ran from approximately the same phase. This phase was predictable from the phase of the combined activity rhythm during the social contact stage, strongly suggesting that social entrainment had occurred. After the period of social contact, one squirrel (#3/4) showed either a phase delay or a temporary change in t (Days 75 to 81). In contrast, two pairs did not conclusively entrain. In one such pair, when the squirrels were separated after the social contact stage (Day 74), a 436-min phase difference was seen between their activity onsets (Fig. 2). Although the activity rhythm of one of the pair (#3/2) free-ran from a phase that was predictable from the phase of the combined rhythm during the social contact stage, the rhythm of the other squirrel

FIG. 1. Activity records for Pair 6 (a—# 3/4; b—#4/7), in standard double plot actogram format, showing social synchronization of free-running activity rhythms. Experimental days are marked on the left-hand side of the actograms, and experimental stages are marked on the right-hand side: 1—individually housed under 12:12 LD cycles, 2—individually housed under LL, 3—social contact stage under LL, 4—individually housed under LL. Clock time (hours) is represented on the abscissa. During the social contact period, data obtained from the pair of squirrels is presented in each squirrel’s individual actogram.

24

RAJARATNAM AND REDMAN

(#5/0) did not. It is possible that the activity rhythm of the squirrel that did not free-run from a phase predicted by the combined rhythm was negatively masked during the latter part of the social contact stage. According to this interpretation, there would be no evidence of social entrainment. Alternatively, the two squirrels may have shown synchronous freerunning rhythms during the latter part of the social contact stage, and an abrupt phase advance in response to the cage change and/or the separation may have occurred in one of the pair. It is also noted that the rhythm of one of the pair (#5/0) showed a spontaneous change in several days after the squirrels were separated. A Pearson product moment correlation revealed that the interobserver reliability for phase estimates was very high (r 5 0.98, p , 0.001). Phase differences for individuals in each pair, prior to and after the social contact, are shown in Table 1. Marked differences occurred between the free-running phases of the two animals in each pair prior to the social contact stage. After the period of social contact, the phase difference decreased in all pairs of squirrels except for Pair 3, which showed an increase (159.7 min) that was relatively small in comparison to phase changes observed in other pairs (Table 1). The decrease in phase difference between members of each pair from prior to the social contact to after the social contact was found to be significant, t(5) 5 3.36, p , 0.05. This strongly suggests that social synchronization of the rhythms had occurred. A Pearson product moment correlation revealed that as for phase estimates, interobserver reliability for t estimates was also very high (r 5 0.92, p , 0.001). The mean t prior to social contact was 24.24 h (SE 5 0.03), and after social contact

was 24.48 h (SE 5 0.14). This increase in t was found to be significant, t(11) 5 2.09, p , 0.05. DISCUSSION

This study demonstrated that social cues involving direct physical contact may serve as a zeitgeber for diurnal F. pennanti. Clear evidence of social entrainment was observed in four of six pairs of squirrels. In the remaining two pairs the data were not conclusive because the free-running phases adopted after the social contact stage were not entirely predictable from the phase of the “combined rhythm” during the social contact. The significant decrease in phase difference between members of each pair after the social contact stage provides further evidence that social synchronization occurred. An explanation for the lack of convincing evidence of entrainment in two pairs of squirrels may be individual differences in the sensitivity of the circadian system to social cues. A common finding in nonphotic entrainment studies is that entrainment is not observed in all individuals (13,18). For example, only three of seven palm squirrels showed stable entrainment to cycles of high and low ambient temperature (16). The present findings with F. pennanti support previous research in other species. The current evidence for social entrainment may be organized according to the various protocols used to demonstrate the phenomenon: 1) mutual social entrainment under free-running conditions, with constant exposure to conspecifics (1,3,9,11,12,24); 2) social entrainment in enucleated or congenitally blind animals (20,21); 3) entrainment to cycles of presence/absence of conspecifics (13,22); and 4) socially induced accelerated reentrainment to LD cycle phase shifts (5).

FIG. 2. Activity records for Pair 5 (a—# 3/2; b—# 5/0). Conventions as for Fig. 1. Although there appears to be synchronization during the social contact stage, the squirrels show different free-running phases when they were separated.

SOCIAL SYNCHRONIZATION OF RHYTHMS

25

An additional finding in the present study is that t was significantly longer in squirrels after the social contact stage. t-lengthening may reflect aftereffects of entrainment (14), as observed in rhesus monkeys (Macaca mulatta), which were socially entrained under self-selected LD cycles (24). However, it is also possible that the observed changes in t in the present study were a function of the duration of exposure to LL (14). A number of squirrels from both groups (i.e., social I and social II) showed phase changes after the social contact stage. Again, these changes may reflect aftereffects of entrainment, or may be due to spontaneous changes in the coupling of circadian oscillators. Direct physical contact was imposed in this study to maximize the likelihood of occurrence of social effects. When animals have physical contact, information concerning the behavioral status of the other member of a pair is available via all sensory modalities (i.e., visual, acoustic, olfactory, and tactile). In addition, animals are able to exert control over resources available in the cage, which may also contribute to social entrainment. In a number of previous studies that have reported a lack of social entrainment, physical contact was not used (2,6,19). In the marmoset Callithrix j. jacchus, acoustic cues were necessary for social entrainment, and in one pair direct physical contact was required (3). It is possible that physical contact increases the potency of social cues, and hence, increases the likelihood of social entrainment. The relevant cues for social entrainment in palm squirrels cannot be ascertained from this study. When direct physical contact is allowed, it is not possible to systematically manipulate social cues. Previous studies have sought to examine this question in other species. Acoustic stimuli entrain circadian rhythms of several avian species (8,12). Recently, Goel and Lee postulated that olfactory cues, rather than acoustic, tactile, or visual cues, mediate the effects of social contact on rate of reentrainment in diurnal O. degus (4). Therefore, there appear to be interspecies differences in the specific aspects of social contact that cause entrainment. A further problem with the use of direct physical contact as the social stimulus is that it was not possible to accurately assess the phase of each squirrel’s rhythm during the social contact stage, because group data were collected during this time. Individual data during the social contact stage would have been useful to examine how social entrainment occurred. In other words, was there a gradual change in t, did rhythms phase shift in order to entrain, or was there a combination of both of these mechanisms? Future studies may use telemetry devices that permit concurrent monitoring of two animals’ activity whilst they cohabit the same cage. The present study used bright LL (ø100 lx) to minimize the disruption to rhythms that occurs when palm squirrels are

maintained under lower light intensities (Rajaratnam and Redman, unpublished observations). It may be argued that the apparent social synchronization of rhythms actually reflects a form of photic entrainment, because acoustic cues from the other member of the pair cause squirrels to wake up and be exposed to light during the time they would normally be inactive. Reebs (17) examined this hypothesis using house sparrows (P. domesticus) that were exposed to daily cycles of acoustic stimuli under DD (0 lx). Activity rhythms of P. domesticus were entrained to both playbacks of conspecific vocalizations, as well as to mechanical noise under these conditions; therefore refuting the hypothesis that exposure to light is necessary for social (acoustic) entrainment. Studies reporting circadian effects of social contact in blind animals (20,21), and in animals maintained under DD (7,11) also refute the hypothesis that social effects occur via exposure to light, and indicate that social entrainment constitutes a nonphotic event. Nevertheless, the question of whether LL is a requirement for social entrainment in F. pennanti remains an empirical one. The protocol adopted for this study was selected to avoid the potentially confounding effects associated with periodic exposure paradigms, in particular, the disturbance caused by the experimenter entering the laboratory each day at the same time. However, one methodological issue relating to use of the constant exposure paradigm is that the effects of social contact could not be differentiated from the phase shifting effects of cage changing (13). Although cage changing alone is unlikely to account for the synchronous rhythms observed in F. pennanti after the period of social contact, we are currently examining this issue. In conclusion, this study provided clear evidence that social cues are an effective zeitgeber for diurnal F. pennanti. F. pennanti may, therefore, be a valuable model for examining the potential use of social and other behavioral cues in circumstances where light input from the environment is inadequate in humans. ACKNOWLEDGEMENTS

The authors are grateful to Perth Zoo, Western Australia, for supplying them with palm squirrels, Mr. Graeme Farrington and the Department of Biological Sciences (formerly Ecology and Evolutionary Biology), Monash University, for providing the research laboratories, Mrs. Dawn Johnson and Ms. Cheryl Roberts for care and maintenance of animals, and Ms. Rosemary Williams for assistance with figures. This study complied with the National Health and Medical Research Council (Australia) guidelines for the care and use of animals in research. S.M.W.R. was the recipient of a Monash Graduate Scholarship and a Monash Postgraduate Publications Award.

REFERENCES 1. Crowley, M.; Bovet, J.: Social synchronization of circadian rhythms in deer mice (Peromyscus maniculatus). Behav. Ecol. Sociobiol. 7:99–105; 1980. 2. Ekert, H. G.; Nagel, B.; Stephani, I.: Light and social effects on the free-running circadian activity rhythm in common marmosets (Callithrix jacchus; primates): Social masking, pseudo-splitting, and relative coordination. Behav. Ecol. Sociobiol. 18:443–452; 1986. 3. Erkert, H. G.; Schardt, U.: Social entrainment of circadian activity rhythms in common marmosets, Callithrix j. jacchus (Primates). Ethology 87:189–202; 1991. 4. Goel, N.; Lee, T. M.: Olfactory bulbectomy impedes social but not photic reentrainment of circadian rhythms in female Octodon degus. J. Biol. Rhythms 12:362–370; 1997.

5. Goel, N.; Lee, T. M.: Social cues accelerate reentrainment of circadian rhythms in diurnal female Octodon degus (Rodentia-Octodontidae). Chronobiol. Int. 12:311–323; 1995. 6. Goel, N.; Lee, T. M.: Social cues modulate free-running circadian activity rhythms in the diurnal rodent, Octodon degus. Am. J. Physiol. 263:R797–R804; 1997. 7. Greenwood, K. M.: Social entrainment of circadian oscillators in rodents. PhD dissertation, La Trobe University, Melbourne; 1985. 8. Gwinner, E.: Periodicity of a circadian rhythm in birds by speciesspecific song cycles (Aves, Fringillidae: Carduelis spinus, serinus serinus). Experientia 22:765–766; 1966. 9. Kavanau, J. L.: The study of social interaction between small animals. Anim. Behav. 11:263–273; 1963.

26 10. Kleinknecht, S.: Lack of social entrainment of free-running circadian activity rhythms in the Australian sugar glider (Petaurus breviceps: Marsupialia). Behav. Ecol. Sociobiol. 16:189–193; 1985. 11. Marimuthu, G.; Rajan, S.; Chandrashekaran, M. K.: Social entrainment of the circadian rhythm in the flight activity of the Microchiropteran bat Hipposideros speoris. Behav. Ecol. Sociobiol. 8:147–150; 1981. 12. Menaker, M.; Eskin, A.: Entrainment of circadian rhythms by sound in Passer domesticus. Science 154:1579–1581; 1966. 13. Mrosovsky, N.: Phase response curves for social entrainment. J. Comp. Physiol. 162:35–46; 1988. 14. Pittendrigh, C. S.; Daan, S.: A functional analysis of circadian pacemakers in nocturnal rodents. I. The stability and lability of spontaneous frequency. J. Comp. Physiol. 106:223–252; 1976. 15. Rajaratnam, S. M. W.; Redman, J. R.: Effects of daily melatonin administration on circadian activity rhythms in the diurnal Indian palm squirrel (Funambulus pennanti). J. Biol Rhythms 12:339– 347; 1997. 16. Rajaratnam, S. M. W.; Redman, J. R.: Entrainment of activity rhythms to temperature cycles in diurnal palm squirrels. Physiol. Behav. 63:271–277; 1998. 17. Reebs, S. G.: Acoustical entrainment of circadian activity rhythms in house sparrows: Constant light is not necessary. Ethology 80:172–181; 1989.

RAJARATNAM AND REDMAN 18. Reebs, S. G.; Mrosovsky, N.: Effects of induced wheel running on the circadian activity rhythms of Syrian hamsters: Entrainment and phase response curve. J. Biol. Rhythms 4:39–48; 1989. 19. Sulzman, F. M.; Fuller, C. A.; Moore–Ede, M. C.: Environmental synchronizers of squirrel monkey circadian rhythms. J. Appl. Physiol. 43:795–800; 1977. 20. Takahashi, K.; Inoue, K.; Kobayashi, K.: Hayafuji, C.; Nakamura, Y.; Takahashi, Y.: Mutual influence of rats having different circadian rhythm of adrenocortical activity. Am. J. Physiol. 3:E515– E520; 1978. 21. Takahashi, K.; Murakami, N.; Hayafuji, C.; Sasaki, Y.: Further evidence that circadian rhythm of blinded rat pups is entrained by the nursing dam. Am. J. Physiol. 246:R359–R363; 1984. 22. Viswanathan, N.; Chandrashekaran, M. K.: Cycles of presence and absence of mother mouse entrain the circadian clock of pups. Nature 317:530–531; 1985. 23. Wright, J. M.: The Biology of Funambulus pennanti Wroughton. Feral in Western Australia. Honours dissertation, University of Western Australia, Perth; 1972. 24. Yellin, A. M.; Hauty, G. T.: Activity cycles of the rhesus monkey (Macaca mulatta) under several experimental conditions, both in isolation and in a group situation. J. Interdiscipl. Cycle Res. 2:475–490; 1971.