Physiology&Behavior,Vol. 52, pp. 189-193, 1992
0031-9384/92 $5.00 + .00 Copyright © 1992PergamonPress Ltd.
Printed in the USA.
BRIEF COMMUNICATION
Maternal Behaviour in the Wistar Rat Under Atypical Zeitgeber I. S C H E L S T R A E T E , *
E. K N A E P E N , t
P. D U T I L L E U L : ~ A N D M. H. W E Y E R S *l
*Facultd de Psychologic, Universit£ Catholique de Louvain, Louvain-La-Neuve, Belgium, "?Facultd de Psychologic, Universitd de LiOge, LiOge, Belgium, and ~.Facultd des Sciences Agronomiques, Universitd Catholique de Louvain, Louvain-la-Neuve, Belgium R e c e i v e d 11 J u n e 1991 SCHELSTRAETE, i., E. KNAEPEN, P. DUTILLEUL AND M. H. WEYERS. Maternal behaviour in the Wistar rat under atypicalZeitgeber. PHYSIOL BEHAV 52( 1) 189-193, 1992.--In previous experiments, neonatal rats subjected to atypical Zeitgeber (light and temperature) produce adult animals that present a different reactivity in some test situations. On the other hand, a great body of evidence indicates that physiological and psychological processes, including the maturation of the circadian system, are regulated by maternal behaviour. In order to investigate in which manner the atypical Zeitgeber, mentioned above, influences maternal behaviour, mothers with their young were observed systematically. More precisely, the question was: is the changed reactivity due to the fact that the young experience a situation of classical early privation (e.g., a decrease in social or other sensory stimulation or in nutrition) or is it due to a situation in which only the temporal pattern of the maternal behaviour is modified and thus disturbs the proper development of the circadian system of the young. Results tend to show that the only difference between the experimental and the control group is a modification in the circadian rhythm of several behavioural items. Only one activity (licking of the young) shows a significant difference in the overall mean value but in favour of the experimental group. We would suggest that the modified temporal pattern of the maternal care could be a poor timegiver for the young, in such a way that the development of their circadian rhythms is disturbed. The unusual reactivity in adulthood could depend on this abnormal development of the circadian system and not on a care privation. Maternal behaviour
Development
Early influence
Circadian rhythm
IN previous experiments, Knaepen and Weyers (13) have shown that the manipulation of the Zeitgeber ( 180 ° phase shifts of temperature and light cycles every 72 h and a permanent phase opposition between both cycles) during the first 18 days after birth causes significant modifications in body weight and in the ultradian components of locomotor and drinking circadian rhythm in adulthood (4 months). These modifications were obtained in the adult in a stressful situation. Early manipulation of this kind also affected the adult rat's learning capacity in a Y-maze (14). Thus, it would appear that the treatment with atypical Zeitgeber within a restricted period (before weaning) can produce a different reactivity in adulthood. Since the atypical environmental conditions are imposed during the nursing period, it was natural to look for reasons for the long-term effects in a modification of the maternal behaviour. Consequently, two different hypotheses are possible: either a modification in the quantity of maternal behaviour, that is to say a mild care privation, or a modification of the rhythmic pattern, that is to say
Finite Fourier transform
an incorrect timegiver during developmental stage of the circadian system, or eventually both. Since the early experiment with maternal deprivation, enough evidence has been gathered for the mother to be seen as a regulator of physiological and behavioural function in her developing offspring. A p h e n o m e n o n of protest followed by despair after maternal separation has been shown in m a n y species [review in (12)]. Several studies have shown that responses such as heart rate (1 l), development of pathogenetic factors (1), glucocorticoid level following novelty exposure (20), etc., are influenced by the same kind of separation. Although our experiment does not include a separation, it could be that the atypical temporal conditions have disturbed the mother in such a way that she could not maintain normal mothering. Thus, in that case, the young would suffer from mild care privation. On the other hand, the lactating female shows a rhythmicity in her maternal behaviour; this behaviour is more concentrated during daylight hours while activities concerning only the mother
1 Requests for reprints should be addressed to M. H. Weyers, Unit6 de Psychobiologie, Universit6 Catholique de Louvain, Place Croix du Sud 1,
B- 1348 Louvain-La-Neuve, Belgium.
189
190
SCHELSTRAETk ET AI..
herself maintain a nocturnal rhythm (3,9,10). When the nursing female is placed in LL (constant conditions), she continues to show a free-running rhythm for this activity with an average period of approximately 24 h (15). It has been shown that these maternal rhythmic stimuli affect the biological rhythms of their offspring and play an essential role in synchronization throughout lactation (2,17,21 ). It has also been well demonstrated that nursing mothers play a predominant role in fixing the phase of circadian rhythms in their young. For example, Viswanathan (22) has shown that the principal Zeitgeber during the first few days was the presence-absence cycle of the mother. This maternal Zeitgeber seems more important than the environmental ones. i.e., light and temperature cycles (22). The circadian rhythm of rodents progressively develops from the prenatal period to the 6th or 7th postnatal week, leading to a synchronization with the nycthemeral rhythm (6). This suggests that the development of the overt circadian rhythms of the young could be altered by a modification of the rhythm of the maternal activity, itself induced by atypical ambient conditions. The purpose of the present experiment is to evaluate in which manner--quantitative or temporal--the atypical Zeitgeber influences the maternal behaviour of the nursing rat. METHODS
Subjects and Procedure Nuiliparous females (243 + 20 g) of the Wistar strain, which is a derived strain, bred in our laboratory, were used. Each female was isolated in a macrolon cage (42 × 24 × 15 cm) with a male, and the male was withdrawn from the cage after a fortnight. On the day of parturition (day 1), the young were distributed equally among all the females having given birth within the same 12-h period, the number being cut to 7 young per female. The same day, the mothers and their young were transferred to a laboratory and put in a wire observation cage (45 × 26 × 13 cm). The cages were placed on a white PVC tray (53 × 37 × 7 cm) containing sawdust. At the back of each cage a mirror (15 × 30 cm) was suspended, enabling us to observe the entire area of the cage. Food and water were available ad lib. Rats in the experimental group (EXP) were exposed to light and temperature cycles (12: 12, 0800-2000 h, 400-0 lux; 19-25°C) in phase opposition. Both cycles were reversed by 180 ° every 3 days. For the control group (CTRL), the light and temperature cycles were in phase and were never shifted. There were six females and their litters in the experimental group, and four in the control group. Observations were started on the second postpartum day and were repeated every 3 days until weaning on the 19th day. Six observation sessions of 15 min each were equally distributed throughout the cycle, i.e., beginning 0930 h, 1330 h, 1730 h, 2130 h, 0130 h, and 0530 h. The duration of each activity was recorded with a stop-watch. When two behavioural patterns appeared simultaneously, their respective durations were recorded separately. The observations were made and recorded by the same observer throughout all the sessions.
Behavioural Items Activities of the mother oriented towards the young were: 1. 2. 3. 4.
Nursing: active and inactive nursing is included. Licking of young: the mother licks any part of the young. Sniffing of young. Nest-building: the female manipulates the sawdust from inside the nest. Activities oriented towards the environment or self-oriented were:
I. 2. 3. 4. 5. 6.
digging, grooming, walking, drinking, feeding, lying alone: the mother lies immobile and is not in contact with any of the young, 7. sniffing.
Statistical Analyses Differences in overall mean durations were assessed using a one-way ANOVA. The technical aspects oftbe statistical analyses based on the finite Fourier are given in the appendix. Periodogram analysis of time series data consists in searching for a sharp periodicity in a given set of periodicities and is based on the squared amplitude of the cosinusoid adjusted in a cosinor model (18). Combining amplitude and phase provides a complete reparameterization of the finite Fourier transform. In order to evaluate the effect of the atypical Zeitgeber on both the amplitude and the phase of the circadian rhythm of the nursing rat, the analysis of variance of the finite Fourier transform of the individual behaviour series is carried out at the circadian frequency component. Another application and recent developments of the present methodology may respectively be found in Deswysen et al. (7) and Dutilleul (8). RESULTS
Duration of the Activities: Overall Mean Values The effects of the atypical Zeitgeber on the overall mean duration of each activity over the three weeks are summarized in Table I. No significant difference between the two groups was observed except for the licking of the young. The experimental group's mothers spent more time grooming their young than those in the control group, F(1, 8) = 6.60, p < 0.05. In spite of the radically different conditions of the Zeitgeber, the two groups showed hardly any quantitative difference between them.
Periodogram Analysis Periodogram analysis showed lower circadian frequency amplitudes for all the behavioural items for the experimental group when compared with the control group. Only one behaviour--
TABLE 1 MEAN DURATIONSAND STANDARDERRORS*
Nursing Licking of young Sniffingof young Nest-building Digging Grooming Walking Drinking Feeding Lying alone Sniffing
Control Group
Experimental Group
464 (16) 56 (12) 5 ( 1) 23 (2) l0 (3) 118 (19) 77 (9) 39 (6) 105 (16) 62 (19) 65 (2)
480 (18) 93 (6) 8 (2) 7 (9) 5 (2) 175 (18) 51 (5) 32 (4) 86 (9) 66 (10) 63 (8)
* For each group, overall mean durations were computed for the differentkinds of behaviourover the 3 weeksand are expressed in s. Values in parentheses are the standard errors.
MATERNAL BEHAVIOUR RHYTHM
191
NURSING
NURSING 1000 A z 0 I,¢¢ ,~-
-
•
800000
EXP CTRL
800-
i
!
8
12 TIME
OF
i 24
i 30
I 36
CTRL
600000
,,,,
300
o
200000
t
200
0
0 3
6
9
12
15
18
COMPONENTS
FIG. 2. Periodogram computed for the 36 mean durations for the nursing activity as a function of time: © = control group (CTRL); • = experimental group (EXP). The circadian rhythm component correslxmds to the 6th component of six cycles over the behaviour series of 36 observations spaced by 4 h in time. The peak observed for the control group points out a sharp 24-h periodicity while no marked peak is observed for the experimental group.
500-
400
400000
RHYTHM
600 -
I,0 t::l
EXP
~
n-, u,,i Q.
200
0
w
•
,.= %
600-
400 Z ,< U.I :E
A
,'2 TIME
OF
1'8
;o
OBSERVATION
FIG. 1. Upper panel: mean duration (s/15 min) for the nursing activity as a function of time: (3 = control group (CTRL) (n = 4); • = experimental group (EXP) (n = 6). Lower panel: a 24-h cosine function was fitted to the raw data: (3 = control group; • = experimental group.
g r o o m i n g - - i n the EXP group showed a significant circadian frequency component. Figures l and 2 show this general tendency for the nursing. The raw data are shown in the upper panel of Fig. l and the result of the cosinor analysis in the lower panel. A 24-h cosine function was fitted to the raw data by conventional methods of least-squares regression. A nonsignificant fit was found for the E X P group, F(2, 33) = 2.42, NS, and a significant one for the C T R L group, F(2, 33) = 12.42,p < 0.0001. The decrease in amplitude and phase shift in the E X P group are evident in the cosine curves. Figure 2 shows the corresponding periodogram for the same behaviour. The C T R L group showed a high peak for the 6th frequency c o m p o n e n t that corresponded to the 24-h periodicity while no marked peak was observed for the E X P group. It appeared, therefore, that the lactating females in the experimental group did not keep a typical 24-h rhythm.
Analysis of Variance of the Finite Fourier Transform The circadian frequency c o m p o n e n t was analyzed in detail in model No. 2 given in the appendix. Results are given in Table 2. Significant differences appear both in amplitude and in the peak time (acrophase). This peak time was computed starting from the initial time of the observations. In the category of behaviour oriented towards the young, only the nursing showed a significant difference, F(2, 7) = 9.78, p < 0.01. Using 1900 h, the onset of darkness for the control group, mothers in the experimental group showed a decrease in amplitude and a 5-h phase delay. In the control group, the ac-
rophase for this behaviour appeared in the middle of the photofraction, which was also found to be the case by Grota and Ader (12) and Leon (17). In the category of behaviour not oriented towards the young, and again in the case of the mothers in the experimental group, lower amplitudes and phase advances with respect to the onset of darkness appeared for digging [A~ = 0915 h, F(2, 7) = 5.64, p < 0.05], grooming [A~ = 0815 h, F(2, 7) = 9.40, p < 0.01], walking [A~ = 0730 h, F(2, 7) = 27.11,p < 0.001], and drinking [A~ = 1045 h, F(2, 7) = 46.66, p < 0.0001 ]. Sniffing also showed a weaker amplitude but, unlike the other activities mentioned, showed a phase delay [A~ = 0800 h,/'(2, 7) = 20.14, p < 0.001 ]. Lying alone exhibited only a phase delay [A~ = 0945 h, F(2, 7) = 6.8 l, p < 0.05]. DISCUSSION Data analysis clearly shows that females submitted to a double modification of Zeitgeber (180 ° LD cycle shift every 3 days and
TABLE 2 AMPLITUDE
AND
ACROPHASE
Control G r o u p
Nursing Licking of young Sniffing of young Nest building Digging Grooming Walking Drinking Feeding Lying alone Sniffing
Experimental Group
AMP
~
AMP
196" 28.4 3.83 21 14.7 75.0* 83.5* 22.2* 55.8 64.7 49.0*
1245 h 0330 h 0415 h 0600 h 0045 h 2345 h 0130 h 0045 h 0400 h 1500 h 0130 h
114 31.3 1.93 3.97 3.17 51.5" 29.1 14.0 14.4 70.4 39.3
1745 h 1300 h 0730 h 0500 h 1000 h 0800 h 0900 h 1130 h 1230 h 0045 h 0930 h
Amplitude and acrophase of the circadian frequency component of the behaviour series of 36 mean values (* significant circadian frequency component, p < 0.05).
192
SCHELSTRAEIE E f AI.
a phase opposition between the light and temperature cycles) manifest drifts that only concern temporal aspects of their behaviour. Indeed, the imposed atypical rhythms do not seem to affect quantitatively maternal care, but rather modify the temporal distribution as the adjustment to the 24-h standard model is greatly decreased. During this period the young are very sensitive and their circadian system needs to develop according to a normal temporal pattern. Thus a disruption in that pattern could greatly affect a normal development. Consequently, it is this temporal drift, which is predictable under these conditions, and not the care privation that is probably the cause of the changed reactivity in adulthood. The origin of the long-term effects could therefore be an abnormal development of the circadian system as a result of a different time distribution of care-giving. As far as the influence of temperature on maternal behaviour is concerned, it is possible that the phase opposition between the light and temperature cycles is very important. In another unpublished experiment where litters and their mothers were only submitted to LD shifts with a constant ambient temperature, we did not obtain the long-term effects previously obtained by Knaepen and Weyers. The weight increase of the experimental group remained positive during the period of stress. The amplitude of the circadian and ultradian rhythms of motor activity and drinking was not modified by the atypical treatment. One explanation could be found in the results of Leon et al. (16). These authors demonstrated that the termination of a contact episode with the young was determined by thermal factors. Lactating females interrupted contact with their young for reasons connected with cerebral hyperthermia. This hyperthermia is influenced, by, among other factors, ambient temperature and the chronic elevation in the heat load. Mothers spent much less time nesting in a warm surrounding (26°C) than in a cool one (18°C), and less time also during the night than during the day. Since the nocturnal brain temperature cycle is negatively correlated with the onset and duration of nursing, mothers might be more vulnerable to acute brain hyperthermia during the night. Thus, in the present experiment, it could be that a warm ambience during the night and a cool ambience during the day produced a very disturbed presence-absence cycle. Moreover, it is well known that after phase shifts of the Zeitgeber, activity rhythm is more rapidly resynchronized than body temperature cycle (23). The mothers in the experimental group could have been in a state of internal desynchronization which would have accentuated the modification of their behavioural rhythms. In conclusion, the present experiment tends to focus on the importance of the temporal pattern of maternal care in the development of the young, suggesting that a privation of the normal temporal pattern at a critical period could affect the development of the circadian system to such an extent that the effects are still seen in psychobiological processes in the adult rat.
the value at hour 4t, its finite Fourier transtorm at frequency w (0 < w < p) is defined (21) by ~6
in which e/~ = cos(~, t) + i. sin(o~, t). Subsequently, we will adopt the following notation:
Z y~cos(~,t)1~-36 Y(o~p) =
The authors wish to thank H. S. Sweetman for her help in the revision of the manuscript. APPENDIX: STATISTICALANALYSES I. Finite Fourier Transform And Periodogram Analysis If {Yt; t = l . . . . . 36} represents a time series of 36 observations of one behavioural item at 4-h intervals where Ytdenotes
36 ytsin(~p, t ) / V 2 ~ . ~
t
Y2(~p)
I
Given the time series ~.l't: t - I . . . . . 36}, the value of the periodogram at frequency c~ (0 < w < p) is defined (21) by
For mathematical reasons, the evaluation of the pefiodogram is limited to frequencies ~p = 2rr.p/36
(p = 0, 1. . . . .
18)
corresponding either to the temporal mean (p = 0), or to an integer number ofp cycles over the observation period (p = 1, . . . . 18). This integer number ofp cycles over the observation period will identify the corresponding frequency
0031-9384/92 $5.00 + .00 Copyright © 1992PergamonPress Ltd.
Printed in the USA.
BRIEF COMMUNICATION
Maternal Behaviour in the Wistar Rat Under Atypical Zeitgeber I. S C H E L S T R A E T E , *
E. K N A E P E N , t
P. D U T I L L E U L : ~ A N D M. H. W E Y E R S *l
*Facultd de Psychologic, Universit£ Catholique de Louvain, Louvain-La-Neuve, Belgium, "?Facultd de Psychologic, Universitd de LiOge, LiOge, Belgium, and ~.Facultd des Sciences Agronomiques, Universitd Catholique de Louvain, Louvain-la-Neuve, Belgium R e c e i v e d 11 J u n e 1991 SCHELSTRAETE, i., E. KNAEPEN, P. DUTILLEUL AND M. H. WEYERS. Maternal behaviour in the Wistar rat under atypicalZeitgeber. PHYSIOL BEHAV 52( 1) 189-193, 1992.--In previous experiments, neonatal rats subjected to atypical Zeitgeber (light and temperature) produce adult animals that present a different reactivity in some test situations. On the other hand, a great body of evidence indicates that physiological and psychological processes, including the maturation of the circadian system, are regulated by maternal behaviour. In order to investigate in which manner the atypical Zeitgeber, mentioned above, influences maternal behaviour, mothers with their young were observed systematically. More precisely, the question was: is the changed reactivity due to the fact that the young experience a situation of classical early privation (e.g., a decrease in social or other sensory stimulation or in nutrition) or is it due to a situation in which only the temporal pattern of the maternal behaviour is modified and thus disturbs the proper development of the circadian system of the young. Results tend to show that the only difference between the experimental and the control group is a modification in the circadian rhythm of several behavioural items. Only one activity (licking of the young) shows a significant difference in the overall mean value but in favour of the experimental group. We would suggest that the modified temporal pattern of the maternal care could be a poor timegiver for the young, in such a way that the development of their circadian rhythms is disturbed. The unusual reactivity in adulthood could depend on this abnormal development of the circadian system and not on a care privation. Maternal behaviour
Development
Early influence
Circadian rhythm
IN previous experiments, Knaepen and Weyers (13) have shown that the manipulation of the Zeitgeber ( 180 ° phase shifts of temperature and light cycles every 72 h and a permanent phase opposition between both cycles) during the first 18 days after birth causes significant modifications in body weight and in the ultradian components of locomotor and drinking circadian rhythm in adulthood (4 months). These modifications were obtained in the adult in a stressful situation. Early manipulation of this kind also affected the adult rat's learning capacity in a Y-maze (14). Thus, it would appear that the treatment with atypical Zeitgeber within a restricted period (before weaning) can produce a different reactivity in adulthood. Since the atypical environmental conditions are imposed during the nursing period, it was natural to look for reasons for the long-term effects in a modification of the maternal behaviour. Consequently, two different hypotheses are possible: either a modification in the quantity of maternal behaviour, that is to say a mild care privation, or a modification of the rhythmic pattern, that is to say
Finite Fourier transform
an incorrect timegiver during developmental stage of the circadian system, or eventually both. Since the early experiment with maternal deprivation, enough evidence has been gathered for the mother to be seen as a regulator of physiological and behavioural function in her developing offspring. A p h e n o m e n o n of protest followed by despair after maternal separation has been shown in m a n y species [review in (12)]. Several studies have shown that responses such as heart rate (1 l), development of pathogenetic factors (1), glucocorticoid level following novelty exposure (20), etc., are influenced by the same kind of separation. Although our experiment does not include a separation, it could be that the atypical temporal conditions have disturbed the mother in such a way that she could not maintain normal mothering. Thus, in that case, the young would suffer from mild care privation. On the other hand, the lactating female shows a rhythmicity in her maternal behaviour; this behaviour is more concentrated during daylight hours while activities concerning only the mother
1 Requests for reprints should be addressed to M. H. Weyers, Unit6 de Psychobiologie, Universit6 Catholique de Louvain, Place Croix du Sud 1,
B- 1348 Louvain-La-Neuve, Belgium.
189
190
SCHELSTRAETk ET AI..
herself maintain a nocturnal rhythm (3,9,10). When the nursing female is placed in LL (constant conditions), she continues to show a free-running rhythm for this activity with an average period of approximately 24 h (15). It has been shown that these maternal rhythmic stimuli affect the biological rhythms of their offspring and play an essential role in synchronization throughout lactation (2,17,21 ). It has also been well demonstrated that nursing mothers play a predominant role in fixing the phase of circadian rhythms in their young. For example, Viswanathan (22) has shown that the principal Zeitgeber during the first few days was the presence-absence cycle of the mother. This maternal Zeitgeber seems more important than the environmental ones. i.e., light and temperature cycles (22). The circadian rhythm of rodents progressively develops from the prenatal period to the 6th or 7th postnatal week, leading to a synchronization with the nycthemeral rhythm (6). This suggests that the development of the overt circadian rhythms of the young could be altered by a modification of the rhythm of the maternal activity, itself induced by atypical ambient conditions. The purpose of the present experiment is to evaluate in which manner--quantitative or temporal--the atypical Zeitgeber influences the maternal behaviour of the nursing rat. METHODS
Subjects and Procedure Nuiliparous females (243 + 20 g) of the Wistar strain, which is a derived strain, bred in our laboratory, were used. Each female was isolated in a macrolon cage (42 × 24 × 15 cm) with a male, and the male was withdrawn from the cage after a fortnight. On the day of parturition (day 1), the young were distributed equally among all the females having given birth within the same 12-h period, the number being cut to 7 young per female. The same day, the mothers and their young were transferred to a laboratory and put in a wire observation cage (45 × 26 × 13 cm). The cages were placed on a white PVC tray (53 × 37 × 7 cm) containing sawdust. At the back of each cage a mirror (15 × 30 cm) was suspended, enabling us to observe the entire area of the cage. Food and water were available ad lib. Rats in the experimental group (EXP) were exposed to light and temperature cycles (12: 12, 0800-2000 h, 400-0 lux; 19-25°C) in phase opposition. Both cycles were reversed by 180 ° every 3 days. For the control group (CTRL), the light and temperature cycles were in phase and were never shifted. There were six females and their litters in the experimental group, and four in the control group. Observations were started on the second postpartum day and were repeated every 3 days until weaning on the 19th day. Six observation sessions of 15 min each were equally distributed throughout the cycle, i.e., beginning 0930 h, 1330 h, 1730 h, 2130 h, 0130 h, and 0530 h. The duration of each activity was recorded with a stop-watch. When two behavioural patterns appeared simultaneously, their respective durations were recorded separately. The observations were made and recorded by the same observer throughout all the sessions.
Behavioural Items Activities of the mother oriented towards the young were: 1. 2. 3. 4.
Nursing: active and inactive nursing is included. Licking of young: the mother licks any part of the young. Sniffing of young. Nest-building: the female manipulates the sawdust from inside the nest. Activities oriented towards the environment or self-oriented were:
I. 2. 3. 4. 5. 6.
digging, grooming, walking, drinking, feeding, lying alone: the mother lies immobile and is not in contact with any of the young, 7. sniffing.
Statistical Analyses Differences in overall mean durations were assessed using a one-way ANOVA. The technical aspects oftbe statistical analyses based on the finite Fourier are given in the appendix. Periodogram analysis of time series data consists in searching for a sharp periodicity in a given set of periodicities and is based on the squared amplitude of the cosinusoid adjusted in a cosinor model (18). Combining amplitude and phase provides a complete reparameterization of the finite Fourier transform. In order to evaluate the effect of the atypical Zeitgeber on both the amplitude and the phase of the circadian rhythm of the nursing rat, the analysis of variance of the finite Fourier transform of the individual behaviour series is carried out at the circadian frequency component. Another application and recent developments of the present methodology may respectively be found in Deswysen et al. (7) and Dutilleul (8). RESULTS
Duration of the Activities: Overall Mean Values The effects of the atypical Zeitgeber on the overall mean duration of each activity over the three weeks are summarized in Table I. No significant difference between the two groups was observed except for the licking of the young. The experimental group's mothers spent more time grooming their young than those in the control group, F(1, 8) = 6.60, p < 0.05. In spite of the radically different conditions of the Zeitgeber, the two groups showed hardly any quantitative difference between them.
Periodogram Analysis Periodogram analysis showed lower circadian frequency amplitudes for all the behavioural items for the experimental group when compared with the control group. Only one behaviour--
TABLE 1 MEAN DURATIONSAND STANDARDERRORS*
Nursing Licking of young Sniffingof young Nest-building Digging Grooming Walking Drinking Feeding Lying alone Sniffing
Control Group
Experimental Group
464 (16) 56 (12) 5 ( 1) 23 (2) l0 (3) 118 (19) 77 (9) 39 (6) 105 (16) 62 (19) 65 (2)
480 (18) 93 (6) 8 (2) 7 (9) 5 (2) 175 (18) 51 (5) 32 (4) 86 (9) 66 (10) 63 (8)
* For each group, overall mean durations were computed for the differentkinds of behaviourover the 3 weeksand are expressed in s. Values in parentheses are the standard errors.
MATERNAL BEHAVIOUR RHYTHM
191
NURSING
NURSING 1000 A z 0 I,¢¢ ,~-
-
•
800000
EXP CTRL
800-
i
!
8
12 TIME
OF
i 24
i 30
I 36
CTRL
600000
,,,,
300
o
200000
t
200
0
0 3
6
9
12
15
18
COMPONENTS
FIG. 2. Periodogram computed for the 36 mean durations for the nursing activity as a function of time: © = control group (CTRL); • = experimental group (EXP). The circadian rhythm component correslxmds to the 6th component of six cycles over the behaviour series of 36 observations spaced by 4 h in time. The peak observed for the control group points out a sharp 24-h periodicity while no marked peak is observed for the experimental group.
500-
400
400000
RHYTHM
600 -
I,0 t::l
EXP
~
n-, u,,i Q.
200
0
w
•
,.= %
600-
400 Z ,< U.I :E
A
,'2 TIME
OF
1'8
;o
OBSERVATION
FIG. 1. Upper panel: mean duration (s/15 min) for the nursing activity as a function of time: (3 = control group (CTRL) (n = 4); • = experimental group (EXP) (n = 6). Lower panel: a 24-h cosine function was fitted to the raw data: (3 = control group; • = experimental group.
g r o o m i n g - - i n the EXP group showed a significant circadian frequency component. Figures l and 2 show this general tendency for the nursing. The raw data are shown in the upper panel of Fig. l and the result of the cosinor analysis in the lower panel. A 24-h cosine function was fitted to the raw data by conventional methods of least-squares regression. A nonsignificant fit was found for the E X P group, F(2, 33) = 2.42, NS, and a significant one for the C T R L group, F(2, 33) = 12.42,p < 0.0001. The decrease in amplitude and phase shift in the E X P group are evident in the cosine curves. Figure 2 shows the corresponding periodogram for the same behaviour. The C T R L group showed a high peak for the 6th frequency c o m p o n e n t that corresponded to the 24-h periodicity while no marked peak was observed for the E X P group. It appeared, therefore, that the lactating females in the experimental group did not keep a typical 24-h rhythm.
Analysis of Variance of the Finite Fourier Transform The circadian frequency c o m p o n e n t was analyzed in detail in model No. 2 given in the appendix. Results are given in Table 2. Significant differences appear both in amplitude and in the peak time (acrophase). This peak time was computed starting from the initial time of the observations. In the category of behaviour oriented towards the young, only the nursing showed a significant difference, F(2, 7) = 9.78, p < 0.01. Using 1900 h, the onset of darkness for the control group, mothers in the experimental group showed a decrease in amplitude and a 5-h phase delay. In the control group, the ac-
rophase for this behaviour appeared in the middle of the photofraction, which was also found to be the case by Grota and Ader (12) and Leon (17). In the category of behaviour not oriented towards the young, and again in the case of the mothers in the experimental group, lower amplitudes and phase advances with respect to the onset of darkness appeared for digging [A~ = 0915 h, F(2, 7) = 5.64, p < 0.05], grooming [A~ = 0815 h, F(2, 7) = 9.40, p < 0.01], walking [A~ = 0730 h, F(2, 7) = 27.11,p < 0.001], and drinking [A~ = 1045 h, F(2, 7) = 46.66, p < 0.0001 ]. Sniffing also showed a weaker amplitude but, unlike the other activities mentioned, showed a phase delay [A~ = 0800 h,/'(2, 7) = 20.14, p < 0.001 ]. Lying alone exhibited only a phase delay [A~ = 0945 h, F(2, 7) = 6.8 l, p < 0.05]. DISCUSSION Data analysis clearly shows that females submitted to a double modification of Zeitgeber (180 ° LD cycle shift every 3 days and
TABLE 2 AMPLITUDE
AND
ACROPHASE
Control G r o u p
Nursing Licking of young Sniffing of young Nest building Digging Grooming Walking Drinking Feeding Lying alone Sniffing
Experimental Group
AMP
~
AMP
196" 28.4 3.83 21 14.7 75.0* 83.5* 22.2* 55.8 64.7 49.0*
1245 h 0330 h 0415 h 0600 h 0045 h 2345 h 0130 h 0045 h 0400 h 1500 h 0130 h
114 31.3 1.93 3.97 3.17 51.5" 29.1 14.0 14.4 70.4 39.3
1745 h 1300 h 0730 h 0500 h 1000 h 0800 h 0900 h 1130 h 1230 h 0045 h 0930 h
Amplitude and acrophase of the circadian frequency component of the behaviour series of 36 mean values (* significant circadian frequency component, p < 0.05).
192
SCHELSTRAEIE E f AI.
a phase opposition between the light and temperature cycles) manifest drifts that only concern temporal aspects of their behaviour. Indeed, the imposed atypical rhythms do not seem to affect quantitatively maternal care, but rather modify the temporal distribution as the adjustment to the 24-h standard model is greatly decreased. During this period the young are very sensitive and their circadian system needs to develop according to a normal temporal pattern. Thus a disruption in that pattern could greatly affect a normal development. Consequently, it is this temporal drift, which is predictable under these conditions, and not the care privation that is probably the cause of the changed reactivity in adulthood. The origin of the long-term effects could therefore be an abnormal development of the circadian system as a result of a different time distribution of care-giving. As far as the influence of temperature on maternal behaviour is concerned, it is possible that the phase opposition between the light and temperature cycles is very important. In another unpublished experiment where litters and their mothers were only submitted to LD shifts with a constant ambient temperature, we did not obtain the long-term effects previously obtained by Knaepen and Weyers. The weight increase of the experimental group remained positive during the period of stress. The amplitude of the circadian and ultradian rhythms of motor activity and drinking was not modified by the atypical treatment. One explanation could be found in the results of Leon et al. (16). These authors demonstrated that the termination of a contact episode with the young was determined by thermal factors. Lactating females interrupted contact with their young for reasons connected with cerebral hyperthermia. This hyperthermia is influenced, by, among other factors, ambient temperature and the chronic elevation in the heat load. Mothers spent much less time nesting in a warm surrounding (26°C) than in a cool one (18°C), and less time also during the night than during the day. Since the nocturnal brain temperature cycle is negatively correlated with the onset and duration of nursing, mothers might be more vulnerable to acute brain hyperthermia during the night. Thus, in the present experiment, it could be that a warm ambience during the night and a cool ambience during the day produced a very disturbed presence-absence cycle. Moreover, it is well known that after phase shifts of the Zeitgeber, activity rhythm is more rapidly resynchronized than body temperature cycle (23). The mothers in the experimental group could have been in a state of internal desynchronization which would have accentuated the modification of their behavioural rhythms. In conclusion, the present experiment tends to focus on the importance of the temporal pattern of maternal care in the development of the young, suggesting that a privation of the normal temporal pattern at a critical period could affect the development of the circadian system to such an extent that the effects are still seen in psychobiological processes in the adult rat.
the value at hour 4t, its finite Fourier transtorm at frequency w (0 < w < p) is defined (21) by ~6
in which e/~ = cos(~, t) + i. sin(o~, t). Subsequently, we will adopt the following notation:
Z y~cos(~,t)1~-36 Y(o~p) =
The authors wish to thank H. S. Sweetman for her help in the revision of the manuscript. APPENDIX: STATISTICALANALYSES I. Finite Fourier Transform And Periodogram Analysis If {Yt; t = l . . . . . 36} represents a time series of 36 observations of one behavioural item at 4-h intervals where Ytdenotes
36 ytsin(~p, t ) / V 2 ~ . ~
t
Y2(~p)
I
Given the time series ~.l't: t - I . . . . . 36}, the value of the periodogram at frequency c~ (0 < w < p) is defined (21) by
For mathematical reasons, the evaluation of the pefiodogram is limited to frequencies ~p = 2rr.p/36
(p = 0, 1. . . . .
18)
corresponding either to the temporal mean (p = 0), or to an integer number ofp cycles over the observation period (p = 1, . . . . 18). This integer number ofp cycles over the observation period will identify the corresponding frequency
