Physiology&Behavior,Vol. 51, pp. 549-555, 1992
0031-9384/92$5.00 + .00 Copyright© 1992PergamonPress Ltd.
Printed in the USA.
Somatosensory Control of the Onset and Retention of Maternal Responsiveness in Primiparous Sprague-Dawley Rats H Y W E L D. M O R G A N , * A L I S O N S. F L E M I N G *l A N D J U D I T H M. STERN1-
*Erindale College, University of Toronto, Mississauga, Ontario, Canada L5L 1C6 and "~Department of Psychology, Rutgers-the State University of New Jersey, New Brunswick, NJ 08903 Received 24 M a y 1991
MORGAN, H. D., A. S. FLEMING AND J. M. STERN. Somatosensorycontrolofthe onsetand retentionofmaternalresponsiveness in primiparousSprague-Dawleyrats. PHYSIOL BEHAV 51(3) 549-555, 1992.--The role of perioral and ventral-trunk somatosensory stimulation from pups mediating the initial expression of maternal behavior and its long-term retention 8 days later, was investigated. Six groups of female rats were permitted to physically interact with four l-5-day-old foster pups for 1 h, 36 h after Cesarean delivery on gestation day 21. Prior to this maternal experience, dams were subjected to: reduced cutaneous rostral snout sensitivity (anaptia) by injection of lidocaine into the mystacial pads; reduced ventral-trunk sensations by occlusion of the entire ventrum with a full spandex jacket; both manipulations; or control manipulations. Additional groups of females not receiving a maternal experience (inexperienced) also received the somatosensory deprivation or control manipulations. During retention testing, rats in the singly manipulated experienced groups exhibited reduced latencies to become maternal in comparison to their inexperienced counterparts (approximately 3 days vs. 8 days). However, rats previously rendered both anaptic and ventrallyoccluded responded like inexperienced rats in showing a long latency to become maternal (8 days). Thus, reduction of either perioral or ventral somatosensory contact from pups did not block the maternal experience effect, but reduction of both of these inputs did. Maternal behavior
Somatosensation
Maternalexperience
MATERNAL behavior in Norway rats normally is initiated spontaneously during parturition in response to endogenous hormonal changes (5,24) and to sensory cues from pups (10,28). Maternal behavior also can be induced by sensory stimulation alone by housing a nulliparous adult with pups continuously for 5-10 days (sensitization) (20,23,38), the likelihood and speed of becoming maternal varying with strain and sex (28). The heightened maternal responsiveness of the parturient rat may be elicited shortly before parturition (19,25) or following late-term Cesarean delivery (20). If the litter is removed at the time of parturition, maternal responsiveness remains very high for at least 1 day, but wanes thereafter (22) in parallel with a precipitous decline in the levels of hormones responsible for the rapid onset of maternal behavior at term (20,21). The long-term retention of maternal behavior following separation from pups after parturition depends on physical interactions with the pups. Orpen and Fleming (20) found, for instance, that as little as 30 min of interaction with pups within 36 h of pregnancy termination (by Cesarean delivery) is adequate to sustain elevated maternal responsiveness for the lO-day period. Increasing mother-pup contact to 24 h resulted in heightened
Rats
responsiveness in mothers up to 30 days later (3,4,6,11). These results indicate that with very little postpartum experience, mothers retain a high level of maternal responsiveness for a considerable period. What is being acquired during that maternal experience, however, is unelcar. Although few studies have addressed this issue directly, there is quite a large literature concerned with pup cues and the sensory modalities mediating the initial expression and continued postpartum maintenance of the behavior after parturition (10,27,28). Some of these cues may also be important for the maternal-experience effect. In Norway rats, the disruption of visual, auditory, olfactory, or vomeronasal modalities, singly, does not substantially delay or alter the initial expression or maintenance of maternal behavior if dams have the opportunity to crouch over the pups (2,9,10,12; Fleming, Garvarth, and Cheung, submitted); however, if dams are prevented from physically interacting with pups, the combined visual, olfactory, and auditory input from pups is not adequate to sustain the elevated responsiveness that characterizes the newly-parturient female (13,20,26). The importance of proximal tactile stimuli from pups for both the initiation (33) and postpartum expression (14,15,29,32,34,35)
i Requests for reprints should be addressed to Alison S. Fleming, Department of Psychology,Erindale College, University of Toronto, Mississauga, Ontario, Canada L5L IC6.
549
550
MORGAN, FLEMING AND SI-I~RN
of maternal behavior is now well established. Maternal oral activities, such as retrieval and licking of pups, as well as hovering over pups, from which position young pups are able to suckle, require adequate trigeminal stimulation from pups for their expression, as shown in studies in which trigeminal inputs were acutely or chronically blocked (14,15,32,33,34,35). Nursing behavior, characterized by sustained quiescence and a high dorsal arch, requires adequate ventral-trunk somatosensory inputs from effective pups, including suckling (30,31,35). The actual display of one or more components of maternal behavior, resulting in a physical interaction between dam and pup, also sustains a long-term maternal responsiveness (3,4,6,11,13,20,26), even when the dam's nipples are removed previously (13,26). Thus, somatosensory stimulation from pups, but not necessarily suckling, probably contributes to the maternal-experience effect. The present study was conducted to determine whether perioral and ventral-trunk somatosensory stimulation from pups during a brief maternal experience are both required for the long-term retention of maternal responsiveness following a period of separation from pups. Dams were subjected to Cesarean delivery on gestation day 21 to preclude pup contact prior to the 1-h period of interaction with pups 36 h later. During pup contact, perioral somatosensation was reduced (anaptia) by injecting lidocaine into the mystacial pads, thereby desensitizing the infraorbital branch of the trigeminal nerve (8), and ventraltrunk somatosensation was reduced by fitting dams with jackets. Eight days after the maternal experience (and 10 days after Cesarean delivery), rats were tested for their latencies to become maternal. Comparisons were made with maternally experienced dams that received sham manipulations and with dams that had no prior experience with pups.
ml injection of 0.9% saline was given SC in a rear leg. In the saline condition the locations of the injections were reversed: under light ether anaesthesia 0.1 ml of 0.9% saline was injected into each mystacial pad, and 0.1 ml of 2% lidocaine was injected SC in a rear leg. This procedure was undertaken to control for systemic effects of the lidocaine. The location and dosage of the lidocaine injections were adopted because of their known effect of perioral anaptia (snout desensitivity) (14,34). Injections were given 10 min before introduction of pups. Rats in the lidocaine conditions were tested for perioral anaptia at the beginning and at the end of the exposure phase by applying pressure to the snout with a pencil approaching from the back of the head (14,15). The animal was considered to be anaptic if it did not orient away from or towards the pencil on any of three trials. Postparturient females in the ventral-trunk occlusion groups were lightly anaesthetized with ether and fitted with jackets for the experience phase. The jackets were constructed of double layered 100% nylon spandex material, and fitted so as to wrap entirely around the female covering all of her teats (and surrounding skin), while allowing maximum flexibility to move the head, arms or legs. The jackets were secured in place by velcro strips (on the dorsal trunk of the animal). Animals in control jacket conditions were fitted with jackets that had a large hole in the ventral-trunk region to allow pups access to the entire nipple region and contact with the ventrum. Because jackets were necessarily constricting and probably somewhat stressful, animals were habituated to the jackets for I h each day 3 days prior to mating, and for 3 days at the end of pregnancy, before the exposure period. Jacket sizes were adjusted to different body sizes.
METHOD
Ten groups of animals were tested, six experienced and four inexperienced with pups. The females were assigned to groups in the order in which they became pregnant. In the experienced groups, Cesarean-delivered females were presented with four foster pups, 36 h after surgery, group 1, experienced/lidocalne (E/L, n = 10) animals, were injected with lidocaine immediately before the pup exposure phase; group 2, experienced/saline (E/ S, n = 10) animals, were injected with saline immediately before the exposure; group 3, experienced/full jacket (E/FJ, n = 10) animals were fitted with ventral-trunk jackets immediately prior to exposure; and group 4, experienced/control jacket (E/C J, n = 10) animals, were fitted with sham jackets immediately before the exposure phase; group 5, experienced/lidocaine/full jacket (E/L/FJ, n = 10) animals received both the lidocaine andjacket manipulations; group 6, experienced/saline/control jacket (E/ S/CJ, n = 5) animals received the saline and sham-jacket manipulations. The inexperienced groups received the same single treatments, except pups were not presented for an exposure phase: group 7, inexperienced lidocaine (I/L, n = 9); group 8, inexperienced saline (I/S, n = 5); group 9, inexperienced full jacket (I/FJ, n = 5); and group 10, inexperienced control jacket (I/CJ, n = 5). Onset and retention of postpartum maternal responsiveness were tested in two phases. The pup exposure, or experience phase, occurred on day 2 after Cesarean delivery, when females were permitted a 1 h contact with pups. Inexperienced animals received no pup stimulation at this time. In the induction phase (day 10 after Cesarean delivery) all females were tested for their latencies to respond maternally to foster pups.
Rats and Housing Animals were Sprague-Dawley female rats, 60-90 days of age, randomly selected from a stock bred at Erindale College. Prior to tests for maternal behavior, females were housed in plastic cages (22 cm wide × 44 cm length × 15 cm height), with Purina rat chow and water available ad lib and maintained under a 12:12 illumination cycle (lights on at 0800 h), at a room temperature of 20-22°C. During the experience phase, females were housed in small Plexiglas cages (22 cm wide × 44 cm length × 15 cm height); during the induction phase rats were transferred into large Plexiglas cages (37 cm wide × 47 cm length × 21 cm height). Shredded paper was provided as nesting material.
Mating and Cesarean Deliveries Females were mated with proven stud males on the afternoon of late proestrus. Pregnancy was confirmed the following morning by presence of sperm in the vaginal smear (gestation day 1). All pregnant females had foeti removed by Cesarean delivery on day 21 of gestation, the day before the expected day of delivery. Cesarean delivery was performed under ether anaesthesia, using a single, 2-3 cm midventral-trunk incision. This procedure was adopted to eliminate all contact between dam and pups prior to the controlled l-h exposure or experience phase.
Manipulation of Dams Subjects in the perioral groups were given lidocaine or saline injections. In the lidocaine condition, rats were anaesthetized lightly with ether and given a 0.1 ml SC injection of 2% lidoeaine (Sigma Chemicals, MA, dissolved in saline and HC1 buffered solution, pH 6) in each mystacial pad (14,34). In addition, a 0.1
Experimental Design
Maternal Experience Phase For females in the experienced groups, four 1--4-day-old foster pups (taken from mothers of the same colony) were placed di-
SOMATOSENSORY CONTROL rectly into the dam's nest site to increase the likelihood that the females obtained interactive experience with pups. Maternal behavior was recorded over an 8-min period using a 0/1 time sampiing procedure in which the occurrence (or not) of each behavior was recorded for each successive 5 s interval. There were two observation periods, 1-8 min and 30-38 min after the introduction of pups. In addition, six spot checks were carried out at 10 min intervals between 10-30 min and 40-60 min. The mother's behaviors that were recorded during observations included ORAL behaviors: moving pups from one location to another, mouthing (but not moving) pups, and licking (general or genital). The primary ventral-trunk behavior was hovering over pups which was recorded when the animal stood over pups, usually in the nest; while hovering, the female often engaged in pup-licking or self-grooming. Unfortunately, since in this study we did not record immobility or consistently note pup attachment to the teats, we are unable to include Stern and Johnson's (32) crouch categories that characterize the dam when she is nursing pups. In addition nosing, sit-hover with pups-at-head, and lie-in-contact were recorded. Nosing pups involved probing pups with the tip of the snout; sit-hover with pups-at-head was seen exclusively in the ventral-trunk-jacketed groups; this response was recorded when the female was in a sit or hover posture and the pups were gathered around her head region; and lie-incontact was recorded when the female was seen lying adjacent to the pups, in contact with some of them. After 60 min the foster pups were removed and subjects were transferred back to a small opaque cage. Observations during the experience phase were not made for inexperienced animals; however, the same cage changes were made.
Induction of Maternal Behavior On day 10 after Cesarean delivery, maternal induction latencies were initiated for all females. Females were tested each morning, according to the procedure described below. Daily testing was continued until a female was designated as maternal, or for 10 days. Prior to each test for maternal behavior, any foster pups from the previous day were removed and returned to their lactating donor mothers. Then, four recently-fed foster pups, 1-5 days of age, were taken from another donor mother and placed in the opposite corner to the test female's nest or site where she slept. The behavior of the test female was recorded for the following 8 min, using the same 0/1 observation sheets as in the experience phase. The same behaviors were recorded. The female's position and behavior were also recorded in two spot checks during the remainder of the day and in a third spot check before removal of the foster pups and the beginning of maternal testing the following morning. The criteria for maternal behavior involved retrieval of four pups during the observation period on two consecutive days and a hover posture on at least one of these days. Maternal latency was the number of days elapsing between the first maternal induction test and the first of these two maternal days.
Data Analyses Each manipulation condition in the experience phase and in the induction phase was compared to its appropriate controls using nonparametric, Mann-Whitney U or chi-square tests. For analyses of behavior during the experience phase frequencies of behaviors across the two 8-rain observation periods were combined, since within group comparisons across the two time points indicated very few meaningful differences. In addition, the occurrence of the different behaviors during the six spot-checks
551 was noted and included in analyses comparing groups in frequency of spot-checks during which behaviors occurred as well as percent of animals exhibiting the different behaviors during the exposure phase. Analyses during the induction phase were done on latencies (in days) to become maternal and to exhibit individual maternal behaviors. In addition, some comparisons also involved the analyses of frequencies of the different behaviors during the first induction test. Results from spot-checks are reported in instances where they differ from the results of the continuous observations. RESULTS
Maternal Experience Phase The first set of analyses considered the effects of perioral anaesthesia or reduced ventral-trunk stimulation on the onset and initial expression of maternal behavior. Animals in all six experienced groups exhibited some components of maternal behavior; however, the pattern and intensity of components shown varied in different conditions. Perioral condition. All animals were found to be periorally anaptic at the beginning and the end of the exposure phase. As shown in Table l, there were no significant differences between groups in the percent of animals showing either oral or ventral behaviors. However, consistent with the observation that in comparison to the E/S group (n = 10), the E/L group (n = 10) spent a lower percent of the total observation time actively engaged in oral interactions with the pups (11.5% vs. 20.3%), E/L animals showed lower frequencies of oral behaviors (the sum of moving, mouthing and licking; U = 24, p < 0.05), due primarily to differences in pup licking (U = 26, p < 0.06). Comparisons of ventral behaviors indicate no group differences in either the frequency or percent of time in ventral contact with pups (ranging from 11% to 16%). With respect to the compensatory behavior, pup-nosing, a higher percent of E/L than E/S animals nosed pups (70% vs. 0%, p < 0.003), showing also a higher frequency of this behavior (U = 15, p < 0.01). Ventral-trunk condition. As can be seen in Table l, comparisons of the percent of animals showing the different behaviors in the two groups showed only one difference; a higher percent of Group E/FJ than group E/CJ moved pups to the nest site (p < 0.02). Group comparisons of total oral behaviors showed no differences, with both groups spending approximately 28% of the time engaged in these behaviors. However, as shown in Table l, in comparison to group E/CJ (n = 10), E/FJ animals (n = 10) did show a higher frequency of pup moves (U = 19.5, p < 0.0 l). In contrast, the two groups spent approximately 7-9% of the observation period in ventral-trunk contact with pups and did not differ from one another. With respect to the compensatory behaviors, group E/FJ spent more time than control animals sit-hovering with pups-at-head (U = 15, p < 0.001). Combined condition. Only one of the E/L/J animals did not appear to be completely anaptic at the end of the l-h exposure. All animals were judged to be anaptic at the beginning of the phase. The results of the combined procedure, were consistent with, while not identical to, the effects of each of the individual manipulations alone. As can be seen in Table 1 there were no differences in the percent of animals showing either oral or ventral-trunk behaviors during the exposure phase. However, group E/L/FJ displayed marginally less oral activity (U = 15, p < O. 1, 1 T), with groups E/L/FJ and E/S/CJ spending approximately 7% and 20% of the observation time, respectively, interacting orally with pups. With respect to ventral-trunk behaviors, group E/L/FJ engaged in less hovering (U = 8, p < 0.02) than did controls, although the percent of the total observation time en-
552
MORGAN, FLEMING AND STERN
TABLE 1 BEHAVIORSSHOWNDURINGTHE EXPOSUREPERIOD Perioral E/L Oralbehaviors Move Mouth Lick Ventral-trunk(hover) Compensatory behaviors Nose Sit/head Lie-in-contact
Perioral and Ventral-Trunk
Ventral-Trunk E/S
E/FJ
E/CJ
5b(90) I (20) 2 (80) 4 (80) 6 (50)
39_+ 7b(90) 2 -+ 1 (20) 8_+ 2 (90) 26-+ 4 (90) 22-+ 10 (60)
53_+7 (100) 5 _+2a (70c) 17-+3 (100) 29-+5 (90) 13-+6 (50)
55--+ 10 (100) 2 -+ Id (10c) 16-+ 3 (100) 36-+ 9 (90) 18_+ 6 (70)
3-+ lb(70a) 0 (0) 41 -+ 16 (70)
0b(0a) 0 (0) 36 -+ 12 (80)
0(0) 8 -+ 2d (70c) 4 -+ 4 (10)
0 (0) 0a (0~) 5 _+ 4 (30)
22_+ 2 -+ 5-+ 14-+ 16_+
E/L/FJ 14-+ 1 _+ 5_+ 8_+ 5_+
E/S/CJ
7 (80) 1 (10) 4 (50) 4 (40) 3f(30)
4(I-+22 (100) 0 (0) 22-+15 (60) 18-+12 (80) 17-+ 5r(80)
2_+ If(80e) 42 _+ 18f (90¢) 0 (0)
1_+ If(20e) 0f(0e) 0 (0)
Perioral groups (E/L and E/S), ventral groups (E/FJ and E/CJ), and combined perioral and ventral groups (E/L/FJ and E/S/CJ), mean + SEM intervals during which behaviors are shown, and (%) showing the behaviors. ~b Difference between E/L and E/S significant(p < 0.05): ~percentof animals or ~mean number of intervals. c.dDifferencebetween E/FJ and E/CJ significant(p < 0.05): c percent of animals or dmean number of intervals. e.fDifference between E/L/FJ and E/S/CJ significant (p < 0.05): *percent of animals or rmean number of intervals.
gaged in ventral-trunk contact with pups was low in both groups (3% vs. 9%, respectively). With respect to compensatory behaviors, a higher percent of Group E/L/FJ rats than Group E/S/CJ rats displayed pup-nosing (p < 0.05, 1T) and sit-with-pups-athead (p < 0.01); similar group differences were found for the frequency of these behaviors (U = 8, p < 0.002 and U = 2.5, p < 0.004, respectively).
Induction of Maternal Behavior Because there were no significant differences in any behaviors among the four inexperienced control groups and because their individual group sizes were small, groups I/S, I/L, I/C J, and I/ FJ were combined into a single Inexperienced Control (I/C) group (n = 24). Perioral condition. As shown in Fig. 1, there were no significant differences between the two experienced E/L and E/S groups in the latency to become maternal; there were also no differences in latencies of individual maternal behaviors or in the frequency of behaviors during the first induction day. However, consistent with previous results, significant differences in various measures of maternal behavior were found as a function of the experience status of the animals, between I/C and E/S animals, as well as between I/C and E/L rats. As can be seen in Fig. 1, in comparison to I/C animals (n -- 24), Group E/S (n = 10) had shorter latencies to maternal behavior (U = 55, p < 0.01). Consistent with these results Group E/S also had shorter latencies to retrieval (U = 60, p < 0.02), and licking (U = 54, p < 0.008). In comparison to group I/C, E/L animals also had shorter latencies to maternal behavior (U = 72, p = 0.06), as well as to hover (U = 70, p < 0.04), and lick pups (U = 62, p < 0.02). Comparisons between each of the experienced groups and group I/C in behavioral frequencies on the first induction day reflect the more rapid maternal behavior shown by the experienced groups; these animals exhibited a higher frequency of approaches, pup moves and retrievals, general and genital licking and mouthing (U = 64--91, p < 0.05-0.006). Ventral-trunk condition. There were no significant differences between E/FJ and E/CJ groups in lateneies to become maternal or to exhibit individual maternal behaviors (see Fig. 1). There were also no differences between groups in behavioral frequencies on the first induction day.
Comparisons between ventrally-occluded-experienced and -inexperienced animals showed that I/C animals (n = 24) had much longer latencies than the E/FJ animals (n = 10) in maternal behavior (U = 34, p < 0.001). They also had longer latencies to retrieval (U = 40, p < 0.002), hover (U = 77, p = 0.07) and lick (U = 31, p < 0.001). Similarly, the I/C (n = 24) animals had longer latencies than E/CJ animals (n = 10) for maternal behavior (U = 59, p < 0.02) as well as for retrieval (U = 71, p = 0.06), hover (U = 55, p < 0.008) and lick (U = 61, p < 0.02). Comparisons between groups E/FJ or E/CJ and IC in behavioral frequencies on the first induction day indicated that both experienced groups exhibited elevated levels of approach and pup licking and group E/FJ showed, in addition, heightened mouthing and retrievals (U = 18-75, p = 0.09-
0031-9384/92$5.00 + .00 Copyright© 1992PergamonPress Ltd.
Printed in the USA.
Somatosensory Control of the Onset and Retention of Maternal Responsiveness in Primiparous Sprague-Dawley Rats H Y W E L D. M O R G A N , * A L I S O N S. F L E M I N G *l A N D J U D I T H M. STERN1-
*Erindale College, University of Toronto, Mississauga, Ontario, Canada L5L 1C6 and "~Department of Psychology, Rutgers-the State University of New Jersey, New Brunswick, NJ 08903 Received 24 M a y 1991
MORGAN, H. D., A. S. FLEMING AND J. M. STERN. Somatosensorycontrolofthe onsetand retentionofmaternalresponsiveness in primiparousSprague-Dawleyrats. PHYSIOL BEHAV 51(3) 549-555, 1992.--The role of perioral and ventral-trunk somatosensory stimulation from pups mediating the initial expression of maternal behavior and its long-term retention 8 days later, was investigated. Six groups of female rats were permitted to physically interact with four l-5-day-old foster pups for 1 h, 36 h after Cesarean delivery on gestation day 21. Prior to this maternal experience, dams were subjected to: reduced cutaneous rostral snout sensitivity (anaptia) by injection of lidocaine into the mystacial pads; reduced ventral-trunk sensations by occlusion of the entire ventrum with a full spandex jacket; both manipulations; or control manipulations. Additional groups of females not receiving a maternal experience (inexperienced) also received the somatosensory deprivation or control manipulations. During retention testing, rats in the singly manipulated experienced groups exhibited reduced latencies to become maternal in comparison to their inexperienced counterparts (approximately 3 days vs. 8 days). However, rats previously rendered both anaptic and ventrallyoccluded responded like inexperienced rats in showing a long latency to become maternal (8 days). Thus, reduction of either perioral or ventral somatosensory contact from pups did not block the maternal experience effect, but reduction of both of these inputs did. Maternal behavior
Somatosensation
Maternalexperience
MATERNAL behavior in Norway rats normally is initiated spontaneously during parturition in response to endogenous hormonal changes (5,24) and to sensory cues from pups (10,28). Maternal behavior also can be induced by sensory stimulation alone by housing a nulliparous adult with pups continuously for 5-10 days (sensitization) (20,23,38), the likelihood and speed of becoming maternal varying with strain and sex (28). The heightened maternal responsiveness of the parturient rat may be elicited shortly before parturition (19,25) or following late-term Cesarean delivery (20). If the litter is removed at the time of parturition, maternal responsiveness remains very high for at least 1 day, but wanes thereafter (22) in parallel with a precipitous decline in the levels of hormones responsible for the rapid onset of maternal behavior at term (20,21). The long-term retention of maternal behavior following separation from pups after parturition depends on physical interactions with the pups. Orpen and Fleming (20) found, for instance, that as little as 30 min of interaction with pups within 36 h of pregnancy termination (by Cesarean delivery) is adequate to sustain elevated maternal responsiveness for the lO-day period. Increasing mother-pup contact to 24 h resulted in heightened
Rats
responsiveness in mothers up to 30 days later (3,4,6,11). These results indicate that with very little postpartum experience, mothers retain a high level of maternal responsiveness for a considerable period. What is being acquired during that maternal experience, however, is unelcar. Although few studies have addressed this issue directly, there is quite a large literature concerned with pup cues and the sensory modalities mediating the initial expression and continued postpartum maintenance of the behavior after parturition (10,27,28). Some of these cues may also be important for the maternal-experience effect. In Norway rats, the disruption of visual, auditory, olfactory, or vomeronasal modalities, singly, does not substantially delay or alter the initial expression or maintenance of maternal behavior if dams have the opportunity to crouch over the pups (2,9,10,12; Fleming, Garvarth, and Cheung, submitted); however, if dams are prevented from physically interacting with pups, the combined visual, olfactory, and auditory input from pups is not adequate to sustain the elevated responsiveness that characterizes the newly-parturient female (13,20,26). The importance of proximal tactile stimuli from pups for both the initiation (33) and postpartum expression (14,15,29,32,34,35)
i Requests for reprints should be addressed to Alison S. Fleming, Department of Psychology,Erindale College, University of Toronto, Mississauga, Ontario, Canada L5L IC6.
549
550
MORGAN, FLEMING AND SI-I~RN
of maternal behavior is now well established. Maternal oral activities, such as retrieval and licking of pups, as well as hovering over pups, from which position young pups are able to suckle, require adequate trigeminal stimulation from pups for their expression, as shown in studies in which trigeminal inputs were acutely or chronically blocked (14,15,32,33,34,35). Nursing behavior, characterized by sustained quiescence and a high dorsal arch, requires adequate ventral-trunk somatosensory inputs from effective pups, including suckling (30,31,35). The actual display of one or more components of maternal behavior, resulting in a physical interaction between dam and pup, also sustains a long-term maternal responsiveness (3,4,6,11,13,20,26), even when the dam's nipples are removed previously (13,26). Thus, somatosensory stimulation from pups, but not necessarily suckling, probably contributes to the maternal-experience effect. The present study was conducted to determine whether perioral and ventral-trunk somatosensory stimulation from pups during a brief maternal experience are both required for the long-term retention of maternal responsiveness following a period of separation from pups. Dams were subjected to Cesarean delivery on gestation day 21 to preclude pup contact prior to the 1-h period of interaction with pups 36 h later. During pup contact, perioral somatosensation was reduced (anaptia) by injecting lidocaine into the mystacial pads, thereby desensitizing the infraorbital branch of the trigeminal nerve (8), and ventraltrunk somatosensation was reduced by fitting dams with jackets. Eight days after the maternal experience (and 10 days after Cesarean delivery), rats were tested for their latencies to become maternal. Comparisons were made with maternally experienced dams that received sham manipulations and with dams that had no prior experience with pups.
ml injection of 0.9% saline was given SC in a rear leg. In the saline condition the locations of the injections were reversed: under light ether anaesthesia 0.1 ml of 0.9% saline was injected into each mystacial pad, and 0.1 ml of 2% lidocaine was injected SC in a rear leg. This procedure was undertaken to control for systemic effects of the lidocaine. The location and dosage of the lidocaine injections were adopted because of their known effect of perioral anaptia (snout desensitivity) (14,34). Injections were given 10 min before introduction of pups. Rats in the lidocaine conditions were tested for perioral anaptia at the beginning and at the end of the exposure phase by applying pressure to the snout with a pencil approaching from the back of the head (14,15). The animal was considered to be anaptic if it did not orient away from or towards the pencil on any of three trials. Postparturient females in the ventral-trunk occlusion groups were lightly anaesthetized with ether and fitted with jackets for the experience phase. The jackets were constructed of double layered 100% nylon spandex material, and fitted so as to wrap entirely around the female covering all of her teats (and surrounding skin), while allowing maximum flexibility to move the head, arms or legs. The jackets were secured in place by velcro strips (on the dorsal trunk of the animal). Animals in control jacket conditions were fitted with jackets that had a large hole in the ventral-trunk region to allow pups access to the entire nipple region and contact with the ventrum. Because jackets were necessarily constricting and probably somewhat stressful, animals were habituated to the jackets for I h each day 3 days prior to mating, and for 3 days at the end of pregnancy, before the exposure period. Jacket sizes were adjusted to different body sizes.
METHOD
Ten groups of animals were tested, six experienced and four inexperienced with pups. The females were assigned to groups in the order in which they became pregnant. In the experienced groups, Cesarean-delivered females were presented with four foster pups, 36 h after surgery, group 1, experienced/lidocalne (E/L, n = 10) animals, were injected with lidocaine immediately before the pup exposure phase; group 2, experienced/saline (E/ S, n = 10) animals, were injected with saline immediately before the exposure; group 3, experienced/full jacket (E/FJ, n = 10) animals were fitted with ventral-trunk jackets immediately prior to exposure; and group 4, experienced/control jacket (E/C J, n = 10) animals, were fitted with sham jackets immediately before the exposure phase; group 5, experienced/lidocaine/full jacket (E/L/FJ, n = 10) animals received both the lidocaine andjacket manipulations; group 6, experienced/saline/control jacket (E/ S/CJ, n = 5) animals received the saline and sham-jacket manipulations. The inexperienced groups received the same single treatments, except pups were not presented for an exposure phase: group 7, inexperienced lidocaine (I/L, n = 9); group 8, inexperienced saline (I/S, n = 5); group 9, inexperienced full jacket (I/FJ, n = 5); and group 10, inexperienced control jacket (I/CJ, n = 5). Onset and retention of postpartum maternal responsiveness were tested in two phases. The pup exposure, or experience phase, occurred on day 2 after Cesarean delivery, when females were permitted a 1 h contact with pups. Inexperienced animals received no pup stimulation at this time. In the induction phase (day 10 after Cesarean delivery) all females were tested for their latencies to respond maternally to foster pups.
Rats and Housing Animals were Sprague-Dawley female rats, 60-90 days of age, randomly selected from a stock bred at Erindale College. Prior to tests for maternal behavior, females were housed in plastic cages (22 cm wide × 44 cm length × 15 cm height), with Purina rat chow and water available ad lib and maintained under a 12:12 illumination cycle (lights on at 0800 h), at a room temperature of 20-22°C. During the experience phase, females were housed in small Plexiglas cages (22 cm wide × 44 cm length × 15 cm height); during the induction phase rats were transferred into large Plexiglas cages (37 cm wide × 47 cm length × 21 cm height). Shredded paper was provided as nesting material.
Mating and Cesarean Deliveries Females were mated with proven stud males on the afternoon of late proestrus. Pregnancy was confirmed the following morning by presence of sperm in the vaginal smear (gestation day 1). All pregnant females had foeti removed by Cesarean delivery on day 21 of gestation, the day before the expected day of delivery. Cesarean delivery was performed under ether anaesthesia, using a single, 2-3 cm midventral-trunk incision. This procedure was adopted to eliminate all contact between dam and pups prior to the controlled l-h exposure or experience phase.
Manipulation of Dams Subjects in the perioral groups were given lidocaine or saline injections. In the lidocaine condition, rats were anaesthetized lightly with ether and given a 0.1 ml SC injection of 2% lidoeaine (Sigma Chemicals, MA, dissolved in saline and HC1 buffered solution, pH 6) in each mystacial pad (14,34). In addition, a 0.1
Experimental Design
Maternal Experience Phase For females in the experienced groups, four 1--4-day-old foster pups (taken from mothers of the same colony) were placed di-
SOMATOSENSORY CONTROL rectly into the dam's nest site to increase the likelihood that the females obtained interactive experience with pups. Maternal behavior was recorded over an 8-min period using a 0/1 time sampiing procedure in which the occurrence (or not) of each behavior was recorded for each successive 5 s interval. There were two observation periods, 1-8 min and 30-38 min after the introduction of pups. In addition, six spot checks were carried out at 10 min intervals between 10-30 min and 40-60 min. The mother's behaviors that were recorded during observations included ORAL behaviors: moving pups from one location to another, mouthing (but not moving) pups, and licking (general or genital). The primary ventral-trunk behavior was hovering over pups which was recorded when the animal stood over pups, usually in the nest; while hovering, the female often engaged in pup-licking or self-grooming. Unfortunately, since in this study we did not record immobility or consistently note pup attachment to the teats, we are unable to include Stern and Johnson's (32) crouch categories that characterize the dam when she is nursing pups. In addition nosing, sit-hover with pups-at-head, and lie-in-contact were recorded. Nosing pups involved probing pups with the tip of the snout; sit-hover with pups-at-head was seen exclusively in the ventral-trunk-jacketed groups; this response was recorded when the female was in a sit or hover posture and the pups were gathered around her head region; and lie-incontact was recorded when the female was seen lying adjacent to the pups, in contact with some of them. After 60 min the foster pups were removed and subjects were transferred back to a small opaque cage. Observations during the experience phase were not made for inexperienced animals; however, the same cage changes were made.
Induction of Maternal Behavior On day 10 after Cesarean delivery, maternal induction latencies were initiated for all females. Females were tested each morning, according to the procedure described below. Daily testing was continued until a female was designated as maternal, or for 10 days. Prior to each test for maternal behavior, any foster pups from the previous day were removed and returned to their lactating donor mothers. Then, four recently-fed foster pups, 1-5 days of age, were taken from another donor mother and placed in the opposite corner to the test female's nest or site where she slept. The behavior of the test female was recorded for the following 8 min, using the same 0/1 observation sheets as in the experience phase. The same behaviors were recorded. The female's position and behavior were also recorded in two spot checks during the remainder of the day and in a third spot check before removal of the foster pups and the beginning of maternal testing the following morning. The criteria for maternal behavior involved retrieval of four pups during the observation period on two consecutive days and a hover posture on at least one of these days. Maternal latency was the number of days elapsing between the first maternal induction test and the first of these two maternal days.
Data Analyses Each manipulation condition in the experience phase and in the induction phase was compared to its appropriate controls using nonparametric, Mann-Whitney U or chi-square tests. For analyses of behavior during the experience phase frequencies of behaviors across the two 8-rain observation periods were combined, since within group comparisons across the two time points indicated very few meaningful differences. In addition, the occurrence of the different behaviors during the six spot-checks
551 was noted and included in analyses comparing groups in frequency of spot-checks during which behaviors occurred as well as percent of animals exhibiting the different behaviors during the exposure phase. Analyses during the induction phase were done on latencies (in days) to become maternal and to exhibit individual maternal behaviors. In addition, some comparisons also involved the analyses of frequencies of the different behaviors during the first induction test. Results from spot-checks are reported in instances where they differ from the results of the continuous observations. RESULTS
Maternal Experience Phase The first set of analyses considered the effects of perioral anaesthesia or reduced ventral-trunk stimulation on the onset and initial expression of maternal behavior. Animals in all six experienced groups exhibited some components of maternal behavior; however, the pattern and intensity of components shown varied in different conditions. Perioral condition. All animals were found to be periorally anaptic at the beginning and the end of the exposure phase. As shown in Table l, there were no significant differences between groups in the percent of animals showing either oral or ventral behaviors. However, consistent with the observation that in comparison to the E/S group (n = 10), the E/L group (n = 10) spent a lower percent of the total observation time actively engaged in oral interactions with the pups (11.5% vs. 20.3%), E/L animals showed lower frequencies of oral behaviors (the sum of moving, mouthing and licking; U = 24, p < 0.05), due primarily to differences in pup licking (U = 26, p < 0.06). Comparisons of ventral behaviors indicate no group differences in either the frequency or percent of time in ventral contact with pups (ranging from 11% to 16%). With respect to the compensatory behavior, pup-nosing, a higher percent of E/L than E/S animals nosed pups (70% vs. 0%, p < 0.003), showing also a higher frequency of this behavior (U = 15, p < 0.01). Ventral-trunk condition. As can be seen in Table l, comparisons of the percent of animals showing the different behaviors in the two groups showed only one difference; a higher percent of Group E/FJ than group E/CJ moved pups to the nest site (p < 0.02). Group comparisons of total oral behaviors showed no differences, with both groups spending approximately 28% of the time engaged in these behaviors. However, as shown in Table l, in comparison to group E/CJ (n = 10), E/FJ animals (n = 10) did show a higher frequency of pup moves (U = 19.5, p < 0.0 l). In contrast, the two groups spent approximately 7-9% of the observation period in ventral-trunk contact with pups and did not differ from one another. With respect to the compensatory behaviors, group E/FJ spent more time than control animals sit-hovering with pups-at-head (U = 15, p < 0.001). Combined condition. Only one of the E/L/J animals did not appear to be completely anaptic at the end of the l-h exposure. All animals were judged to be anaptic at the beginning of the phase. The results of the combined procedure, were consistent with, while not identical to, the effects of each of the individual manipulations alone. As can be seen in Table 1 there were no differences in the percent of animals showing either oral or ventral-trunk behaviors during the exposure phase. However, group E/L/FJ displayed marginally less oral activity (U = 15, p < O. 1, 1 T), with groups E/L/FJ and E/S/CJ spending approximately 7% and 20% of the observation time, respectively, interacting orally with pups. With respect to ventral-trunk behaviors, group E/L/FJ engaged in less hovering (U = 8, p < 0.02) than did controls, although the percent of the total observation time en-
552
MORGAN, FLEMING AND STERN
TABLE 1 BEHAVIORSSHOWNDURINGTHE EXPOSUREPERIOD Perioral E/L Oralbehaviors Move Mouth Lick Ventral-trunk(hover) Compensatory behaviors Nose Sit/head Lie-in-contact
Perioral and Ventral-Trunk
Ventral-Trunk E/S
E/FJ
E/CJ
5b(90) I (20) 2 (80) 4 (80) 6 (50)
39_+ 7b(90) 2 -+ 1 (20) 8_+ 2 (90) 26-+ 4 (90) 22-+ 10 (60)
53_+7 (100) 5 _+2a (70c) 17-+3 (100) 29-+5 (90) 13-+6 (50)
55--+ 10 (100) 2 -+ Id (10c) 16-+ 3 (100) 36-+ 9 (90) 18_+ 6 (70)
3-+ lb(70a) 0 (0) 41 -+ 16 (70)
0b(0a) 0 (0) 36 -+ 12 (80)
0(0) 8 -+ 2d (70c) 4 -+ 4 (10)
0 (0) 0a (0~) 5 _+ 4 (30)
22_+ 2 -+ 5-+ 14-+ 16_+
E/L/FJ 14-+ 1 _+ 5_+ 8_+ 5_+
E/S/CJ
7 (80) 1 (10) 4 (50) 4 (40) 3f(30)
4(I-+22 (100) 0 (0) 22-+15 (60) 18-+12 (80) 17-+ 5r(80)
2_+ If(80e) 42 _+ 18f (90¢) 0 (0)
1_+ If(20e) 0f(0e) 0 (0)
Perioral groups (E/L and E/S), ventral groups (E/FJ and E/CJ), and combined perioral and ventral groups (E/L/FJ and E/S/CJ), mean + SEM intervals during which behaviors are shown, and (%) showing the behaviors. ~b Difference between E/L and E/S significant(p < 0.05): ~percentof animals or ~mean number of intervals. c.dDifferencebetween E/FJ and E/CJ significant(p < 0.05): c percent of animals or dmean number of intervals. e.fDifference between E/L/FJ and E/S/CJ significant (p < 0.05): *percent of animals or rmean number of intervals.
gaged in ventral-trunk contact with pups was low in both groups (3% vs. 9%, respectively). With respect to compensatory behaviors, a higher percent of Group E/L/FJ rats than Group E/S/CJ rats displayed pup-nosing (p < 0.05, 1T) and sit-with-pups-athead (p < 0.01); similar group differences were found for the frequency of these behaviors (U = 8, p < 0.002 and U = 2.5, p < 0.004, respectively).
Induction of Maternal Behavior Because there were no significant differences in any behaviors among the four inexperienced control groups and because their individual group sizes were small, groups I/S, I/L, I/C J, and I/ FJ were combined into a single Inexperienced Control (I/C) group (n = 24). Perioral condition. As shown in Fig. 1, there were no significant differences between the two experienced E/L and E/S groups in the latency to become maternal; there were also no differences in latencies of individual maternal behaviors or in the frequency of behaviors during the first induction day. However, consistent with previous results, significant differences in various measures of maternal behavior were found as a function of the experience status of the animals, between I/C and E/S animals, as well as between I/C and E/L rats. As can be seen in Fig. 1, in comparison to I/C animals (n -- 24), Group E/S (n = 10) had shorter latencies to maternal behavior (U = 55, p < 0.01). Consistent with these results Group E/S also had shorter latencies to retrieval (U = 60, p < 0.02), and licking (U = 54, p < 0.008). In comparison to group I/C, E/L animals also had shorter latencies to maternal behavior (U = 72, p = 0.06), as well as to hover (U = 70, p < 0.04), and lick pups (U = 62, p < 0.02). Comparisons between each of the experienced groups and group I/C in behavioral frequencies on the first induction day reflect the more rapid maternal behavior shown by the experienced groups; these animals exhibited a higher frequency of approaches, pup moves and retrievals, general and genital licking and mouthing (U = 64--91, p < 0.05-0.006). Ventral-trunk condition. There were no significant differences between E/FJ and E/CJ groups in lateneies to become maternal or to exhibit individual maternal behaviors (see Fig. 1). There were also no differences between groups in behavioral frequencies on the first induction day.
Comparisons between ventrally-occluded-experienced and -inexperienced animals showed that I/C animals (n = 24) had much longer latencies than the E/FJ animals (n = 10) in maternal behavior (U = 34, p < 0.001). They also had longer latencies to retrieval (U = 40, p < 0.002), hover (U = 77, p = 0.07) and lick (U = 31, p < 0.001). Similarly, the I/C (n = 24) animals had longer latencies than E/CJ animals (n = 10) for maternal behavior (U = 59, p < 0.02) as well as for retrieval (U = 71, p = 0.06), hover (U = 55, p < 0.008) and lick (U = 61, p < 0.02). Comparisons between groups E/FJ or E/CJ and IC in behavioral frequencies on the first induction day indicated that both experienced groups exhibited elevated levels of approach and pup licking and group E/FJ showed, in addition, heightened mouthing and retrievals (U = 18-75, p = 0.09-
