Keep cool: Memory is retained during hibernation in Alpine marmots

Keep cool: Memory is retained during hibernation in Alpine marmots

Physiology & Behavior 98 (2009) 78–84 Contents lists available at ScienceDirect Physiology & Behavior j o u r n a l h o m e p a g e : w w w. e l s e...

779KB Sizes 1 Downloads 71 Views

Physiology & Behavior 98 (2009) 78–84

Contents lists available at ScienceDirect

Physiology & Behavior j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / p h b

Keep cool: Memory is retained during hibernation in Alpine marmots L.E. Clemens ⁎, G. Heldmaier, C. Exner Department of Animal Physiology, Philipps-Universität Marburg, Karl-von-Frisch Straße 8, 35043 Marburg, Germany

a r t i c l e

i n f o

Article history: Received 11 December 2008 Received in revised form 17 March 2009 Accepted 16 April 2009 Keywords: Marmota Behavior Learning, Operant conditioning Positive reinforcement Open field test Familiarity detection Locomotor activity

a b s t r a c t Hibernators display severe changes in brain structure during deep torpor, including alterations in synaptic constitution. To address a possible effect on long-term memory, we examined learning behavior and memory of the hibernator Marmota marmota. In two operant conditioning tasks, the marmots learned to jump on two boxes or to walk through a tube. The animals were trained during their active season. Performance improved during the training phase and remained stable in a last test, four weeks before entrance into hibernation. When retested after six months of hibernation, skills were found to be unimpaired (box: before hibernation: 258.2 ± 17.7 s, after hibernation: 275.0 ± 19.8 s; tube: before hibernation: 158.4 ± 9.0 s, after hibernation: 137.7 ± 6.3 s). Contrary to these findings, memory seemed to be less fixed during the active season, since changes in test procedure resulted in impaired test performance. Besides the operant conditioning, we investigated the animals' habituation to a novel environment by repeated open field exposure. In the first run, animals showed exploratory behavior and thus a high locomotor activity was observed (63.6 ± 10.7 crossed squares). Upon a second exposure, all animals immediately moved into one corner and locomotion ceased (7.2 ± 1.9 crossed squares). This habituation was not altered even after hibernation (6.1 ± 1.1 crossed squares). We thus conclude that long-term memory is unaffected by hibernation in Alpine marmots. © 2009 Elsevier Inc. All rights reserved.

1. Introduction Alpine marmots are highly seasonal rodents that spend half of the year in hibernation, a physiological state characterized by alternating phases of normo- and hypometabolism. In deep torpor, metabolic, respiratory and heart rate are actively reduced, followed by a drop down in body temperature to near ambient values. Thereby energy demand in winter is decreased by about 80% as compared to summer, enabling the marmots to live entirely on endogenous fat reserves during the hibernation season [1,2]. Metabolic depression during deep torpor is accompanied by profound changes within the brain. Brain blood flow [3] and metabolism [4,5] are massively reduced and no peripheral EEG can be detected [6,7]. Beyond this, length and branching of dendritic arbors as well as the number of dendritic spines are decreased in the hippocampus [8–10], thalamus and cortex [9,11] and synaptic proteins in these areas seem to dislocate from the postsynaptic density zone [12–14]. These structural changes are found to be reversed during each interbout arousal and the terminal arousal in spring, in a temperature dependent manner. Since synaptic plasticity is essential for memory formation [15,16], a possible impact of degeneration and regeneration on memory must be considered. It has been shown that low body temperature, hypometabolism and a short photoperiod can influence learning and memory. In nonhibernators, there is evidence for impaired memory consolidation after ⁎ Corresponding author. Tel.: +49 6421 28 23496; fax: +49 6421 28 22052. E-mail address: [email protected] (L.E. Clemens). 0031-9384/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.physbeh.2009.04.013

cooling [17–20], spontaneous daily torpor [21] and at short day length [22], but already stored information seems to be unaffected. The few studies addressing the influence of hibernation on memory led to conflicting results: A study on ground squirrels, which either displayed torpor or did not show torpor following 11 days of cold exposure, demonstrated improved memory retention of the torpor group [23]; another study points to impaired memory in squirrels which had experienced a complete hibernation season, compared to a group which wintered at warm ambient temperatures [24]; furthermore, the discrimination between familiar and unfamiliar conspecifics was found to be unaffected by hibernation in one study [24], while others reported on a deficit in social recognition of unrelated individuals after hibernation [25]. To resolve the actual impact of hibernation on long-term memory, we investigated learning behavior and memory of the hibernator Marmota marmota. For this purpose, two common behavioral test approaches, operant conditioning and the open field test, were employed. The conditioning experiments included climbing two boxes in the correct sequence or walking through a tube, in order to obtain a food reward. The open field test was used to ascertain the habituation to a novel environment. The training was carried out during the active season of the animals and the individual performance was tested four weeks before entrance and four weeks after terminal arousal from six months of hibernation for the purpose of comparison. 2. Animals and methods The study was performed in accordance with the guidelines of the German animal welfare act (Deutsches Tierschutzgesetz).

L.E. Clemens et al. / Physiology & Behavior 98 (2009) 78–84

79

Fig. 1. Experimental time course. A: Conditioning experiments; animals were trained in the box task ( ) or the tube task ( ) during their active season and the performance was tested before hibernation in fall and after hibernation in spring. B: Open field test; 3 groups of marmots were tested in three consecutive years; the first ( ) and second run ( ) were carried out during the animals' active season and a third run ( ) was conducted after hibernation.

within the home pens, feeding the animals and observing their behavior.

2.1. Animals The study was performed with 10 Alpine marmots bred at the Philipps-Universität Marburg, Germany. Six adult males and 4 adult females out of 3 family groups were used. Ages ranged from 2–6 years at the beginning of the study. The marmots were housed in family groups. During the active season (April–October), the animals were kept in 12.5–35 m2 sized pens under natural climatic and photoperiodic conditions (Marburg, Germany: 50°49 N, 09 E). Food and water was provided ad libitum. The diet consisted of corn, shredded soy and varying fruit and vegetables. During hibernation (November–March), the family groups were kept in standard rabbit cages (56 × 102 × 60 cm3) in a climate chamber with an ambient temperature of 6 ± 1 °C in constant darkness and without food and water. General locomotor activity was monitored with passive infrared motion detectors, providing continuous information about the hibernation status (i.e. deep torpor, interbout arousal or terminal arousal). 2.2. Experimental time course Prior to the main experiments, a two week long familiarization period was carried out, during which the operator spent 1 h daily

2.2.1. Conditioning experiments The conditioning experiments were performed in two consecutive years with 8 animals in total. The first group of marmots (n = 3) was trained in the first task at the end of the animals active season, between August and September (in the following referred to as “fall group”). A second group of marmots (n = 5), which served as a naïve, age- and family group-matched control, was trained in the first task in the following summer, between June and July (in the following referred to as “summer group”). Between August and September, both the fall and summer group were trained to learn a second task. The ultimate test session of both tasks was conducted four weeks before entrance into hibernation and memory retention was examined four weeks after terminal arousal from hibernation (Fig. 1A). 2.2.2. Open field test Data of the open field test were collected in three consecutive years with 10 animals in total. The first run (novelty training) of the initial group of marmots (n = 3) was conducted in fall. Novelty training of a second group of marmots, serving as naïve control animals (n = 5), was carried out in the following spring and a third group of marmots

Fig. 2. Experimental set-up. A: Box task. B: Tube task. C: Open field arena; the bold line represents the border of the center fields.

80

L.E. Clemens et al. / Physiology & Behavior 98 (2009) 78–84

(n = 2) was trained in spring one year later. In all cases, the second exposure to the open field arena (second run) was carried out in fall, not later than four weeks before entrance into hibernation. Memory retention was tested four weeks after terminal arousal from hibernation in spring (Fig. 1B). 2.3. Experimental procedure 2.3.1. Conditioning experiments Animals were trained to accomplish two operant learning tasks. One was comprised of jumping on two boxes in the correct sequence and the other one of walking through a tube, in order to obtain a food reward (Fig. 2A and B). The conditioning experiments were held within the home pens. Animals were trained individually while the rest of the family group was kept in their “burrow” (a wooden box adjacent to the pen). The time measurement started, when a marmot approached the experimental set-up within a range of about 1 m and obviously paid attention to the test situation. The measurement stopped after the particular task had been performed 10 times in a series. Each accurate single performance, either jumping on the boxes or walking through the tube, was rewarded, which resulted in 10 rewards in total by completion of the measurement. If an animal did not engage in the test within 15 min, the test was terminated. In order to facilitate learning, two secondary reinforcers were adopted during the learning phase. A bamboo stick, to which a small piece of banana was attached as a food reward, was presented above the boxes or in front of the tube and served as a visual reinforcer. Simultaneously, a clicker served as auditive reinforcement. The click tone was given each time the animal showed the desired behavior. After introduction of the reinforcers, a small plastic box was introduced, and the food reward was presented above the box (Fig. 2A). In the initial training phase, the stick was presented each of the 10 sequences of jumping on the box or walking through the tube. When the animals showed stable training times, the stick was only presented for the first 5 cycles, and finally it was no longer presented at all. After the animal had learned to jump on the box without visual reinforcement, a bigger plastic box was added, which differed from the first box not only in size, but also in shape and color.

For successful accomplishment of the task, animals had to first jump on the small box and from there jump to the bigger box to be rewarded. In the second task, animals had to walk through a tube to obtain the food reward (Fig. 2B). Again, the time to complete 10 cycles of walking through the tube and getting rewarded was recorded. It was not defined, from which side the marmots should enter the tube in order to be able to test the effect of hibernation on individual performance strategies. Here, the visual reinforcer was only applied on the first training day. Training and test sessions were recorded with a webcam attached to a portable PC to further analyze marmot behavior. 2.3.2. Open field test To investigate memory on a familiar environment, animals were repeatedly exposed to an open field arena. The arena consisted of an empty room, 4 m2 in size, which was visually subdivided into 25 squares by markings on the floor (Fig. 2C). Locomotor activity was calculated by counting the number of crossed squares. A decrease in locomotor activity was considered as an indicator for habituation. In each run, the test animal was taken from the home pen and placed into a carrier box for transport to the open field arena. The measurement started, when the animal was placed within the central square of the arena and lasted for 5 min. The route of the animal was recorded continuously and its position was noted every 30 s to check for intrasession changes in locomotor activity. After each run, the floor and walls of the arena were cleaned from possible odors with a multi purpose cleaner. 2.4. Statistics All analyses were executed with Sigma Stat 3.5. The data sets were tested for normality and equality of variance. For comparison between groups, the t-test was used (parametric data). The analysis of temporal effects, either the difference in performance between training days or the test performance before and after hibernation, was investigated with a one-way repeated measures analysis of variance (RM ANOVA). Results were expressed as mean ± sem (normally distributed), or in one case as median ± st dev (not

Fig. 3. Arousal pattern during hibernation of a group of two marmots. A: Arousal pattern during hibernation; data on locomotor activity were collected with passive infrared motion detectors in the hibernacula of the separately housed family groups; the graph shows the activity of a group consisting of two marmots for each day during the hibernation season. B: Arousal pattern during January; the data observed with the motion detectors during January (day 50: Jan 01, day 79: Jan 30) are further illustrated in a double plot; one line corresponds to two successive days, with the second day (on the right side) being repeated in the line below (on the left side).

L.E. Clemens et al. / Physiology & Behavior 98 (2009) 78–84

81

Fig. 4. Individual training curves, box task. The colored dots refer to an individual's performance (time to complete the task in s) during the training phase. A: Fall group (n = 3); the stepwise omission of the stick was performed on day 9 and 13. B: Summer group (n = 3); the stepwise omission of the stick was realized on day 5 and 9.

normally distributed). Data reached the criterion for significance, when p-values were smaller than 0.05. The graphs were created with Sigma Plot 10.0 or Adobe Illustrator CS3.

3. Results Activity patterns during hibernation were monitored, using infrared motion detectors. Data revealed inactive phases, regularly interrupted by bursts of activity. The overall activity level differed among family groups, but showed the same time course. Activity was high in November, in the beginning of the hibernating season and decreased until December. Activity increased again at the end of February, reaching highest values in March prior to the end of hibernation (Fig. 3A and B).

3.1. Conditioning experiments Before the onset of the experiments, the marmots were familiarized with the presence of the operator within the home pens. Depending on their social status, the animals lost timidity sooner (dominant) or later (subdominant). After this familiarization period, the clicker and stick were introduced. The marmots showed individual differences concerning the time to approach the operator (training day 2: 26–365 s) and to a lesser degree to complete the task (training day 2: 104–283 s). Duration for picking up the banana from the stick 10 times in a series decreased significantly during the five training days in all individuals (training day 1: 279.6 ± 26.1 s, training day 5: 125.6 ± 11.3 s, p = b0.001, RM ANOVA). The main experiments started with the introduction of the first box. All marmots were found to improve continually in the course of training (Fig. 4A and B). In the fall group, mean duration decreased by more than 75% within the course of six training days (fall group training day 1: 890.3 ± 91.8 s, fall group training day 6: 198.7 ± 28.1 s, p = 0.028, RM ANOVA; Fig. 4A). The summer group needed significantly less time for completion of the test on the first training day (summer group training day 1: 259.3 ± 29.6 s, fall group training day 1: 890.3 ± 91.8 s, p = 0.003, t-test; Fig. 4B) and thus did not show a similar steep decline in training time in the course of training. The summer group solved the task on their first training day already as fast as the fall group on day 5 (summer groupday 1: 259.3 ± 29.6 s, fall groupday 5: 276.7 ±17.1 s).

All animals responded to the stepwise omission of the stick by a slight impairment of their individual performance, but improved after further training (Fig. 4A and B). The addition of the second box also resulted in longer training times (before: 171.8 ± 20.6 s, after: 255.2 ± 29.6 s, p = 0.031, RM ANOVA), but in this case, no improvement was found after further training (training day 1: 255.2 ± 29.6 s, training day 5: 224.7 ± 25.6 s). Two of the five summer group marmots gained aversions against the test situation because of repeated external disturbances by noise during the training phase. They hid immediately after being placed into the pen and did not approach the experimental set-up during the 15 min exposure. In the test session four weeks before hibernation, all trained animals retained their skills (training day 5: 224.7 ± 25.6 s, test before hibernation: 258.2 ± 17.7 s), while the two aversive marmots still did not attend the experiments. After hibernation, there was no difference in the performance of any trained individual compared to the test before hibernation (before: 258.2 ± 17.7 s, after: 275.0 ± 19.8 s; Fig. 5). The aversive marmots showed the same behavior as before hibernation, which differed from the behavior of untrained control animals. Naïve controls did not show the desired behavior (jump up and down the boxes) during a 15 min exposure to the experimental set-up, but moved around within the pen and explored the experimental set-up by sniffing and scratching.

Fig. 5. Performance in the box task before and after hibernation in comparison (n = 6).

82

L.E. Clemens et al. / Physiology & Behavior 98 (2009) 78–84

Fig. 6. Performance in the tube task (n = 6). A: Individual training curves; the colored dots refer to an individual's performance (time to complete the task in s) during the training phase. B: Comparison of the mean performance of the marmots before and after hibernation.

In the second task, both groups were trained to walk through a tube. Time to accomplish the task decreased within the training phase (training day 1: 330.8 ± 52.9 s, training day 5: 144.8 ± 11.6 s, p = 0.036, RM ANOVA; Fig. 6A) and remained at that level until the second test before hibernation (training day 5: 144.8 ± 11.6 s, test before hibernation: 158.4 ± 9.0 s). No change in training time was found after hibernation (before: 158.4 ± 9.0 s, after: 137.7 ± 6.3 s; Fig. 6B). The marmots showed three types of performance: 1) enter the tube only from one side, 2) walk from one side to the other and 3) no distinct strategy. Irrespective of the individual strategy, the custom was retained after hibernation. In contrast to the unchanged skills and habits after hibernation, performance was found to be more flexible during the training period. In addition to the slight impairment of performance after omission of the food presentation, animals required longer for the box task after introduction of the tube (before: 299.8 ± 21.5 s, after: 561.2 ± 127.4 s; Fig. 7A). Five training sessions were conducted for the marmots to relearn the task (Fig. 7B).

3.2. Open field test When the marmots were exposed to the novel environment, a high locomotor activity was observed (Fig. 8). Animals mainly moved along the walls of the open field arena (total: 63.6±10.7 squares, centre: 7.6±2.1 squares, Fig. 2C). Both, total number of crossed squares and the number of crossed squares of the centre varied notably among individuals (total: 23–127 squares; centre: 1–25 squares). Further analysis of the data revealed no age or sex effect on locomotor activity in novelty training. A comparison of the fall (n=3) and spring (n=7) trained animals indicates a seasonal effect, with the fall group tending to be less active than the spring group (fall group: 33.3 ±6.1 squares, spring group: 76.6±12.0 squares). During the 5 min exposure, locomotor activity decreased (minute 1: 18.3 ±2.2 squares, minute 5: 7.5±2.7 squares; p =0.03, paired t-test): Most animals (8 of 10) sat down in one corner of the arena and ceased locomotion before the 5 min were over. In the second run before hibernation, all animals immediately moved into one corner and locomotion ceased (run 1: 63.6 ± 10.7

Fig. 7. Effect of the introduction of the tube on the performance in the box task (n = 6). A: Mean performance of the marmots in the box task before and after introduction of the tube; the difference failed to reach significance (p = 0.096, RM ANOVA). B: Individual performance in the box task during five training days after introduction of the tube; the colored dots represent individual animals.

L.E. Clemens et al. / Physiology & Behavior 98 (2009) 78–84

Fig. 8. Locomotor activity in the open field test in the first run, the second run before hibernation (habituated) and the third run after hibernation (n = 10). Mean locomotor activity in the first run differs significantly from both the second run before hibernation (habituated) and the third run after hibernation (⁎⁎⁎p = b 0.001, RM ANOVA).

squares; run 2: 7.2 ± 1.9 squares; p = b0.001, RM ANOVA). When tested after hibernation, the animals again ran to one corner and no further locomotion occurred (6.1 ± 1.1 squares; Fig. 8). Although rearing behavior was not counted, frequent rearing was observed in the first run, which was absent in the second run before hibernation, as well as after hibernation. Experienced animals differed significantly from naïve controls in the test after hibernation (experienced: 5.5 ± 3.6 squares; inexperienced: 71.0 ± 31.8 squares, p = b0.001, Mann–Whitney Rank Sum Test, Fig. 9). 4. Discussion The aim of this study was to elucidate a possible effect of hibernation on long-term memory, in the hibernator M. marmota. The marmots were kept in family groups in outdoor enclosures and were trained during their active season. The training phase was extended over several months to ensure the formation of long-term memory. With the beginning of the hibernation season in November, the animals were transferred to a cold chamber at 6 °C. During hibernation, motion detectors in the hibernacula revealed arousal patterns that reflected the pattern of body temperature and metabolic rate of Alpine marmots as observed either in a cold chamber [26] or in the field [27]. After hibernation in April the next year, the marmots were returned to their outdoor pens. Behavioral tests were carried out four weeks after terminal arousal. In all behavioral tests after hibernation, the marmots showed the same performance as before hibernation. Time to climb two boxes and to walk through a tube to obtain a food reward remained unchanged. Even individual characteristics were retained, such as the strategy to walk through the tube. Further, test situation associated aversion of two marmots was still present after hibernation. Naïve control animals showed a completely different behavior than either the trained or the training averse marmots. The open field test also revealed intact memory retention: Habituation to the open field manifested in lowered locomotor activity during the test and was still present after six months of hibernation, indicating intact familiarity detection. The finding that inexperienced control animals showed a high locomotor activity in the open field in spring ruled out the possibility that locomotion was per se low after hibernation. The few previous studies on the influence of hibernation on memory were all performed with ground squirrels and led to conflicting results [23,24,28]. In two early studies, it was concluded that memory of ground squirrels, which had hibernated was equal

83

[28] or even better [23] than that of individuals which were either prevented from hibernation [28] or had not hibernated following few days of cold exposure [23]. However, both studies only investigated the effect of short phases of torpor on memory retention. Another study, concerning the influence of hibernation on social recognition in Belding's ground squirrels (Spermophilus beldingi), is indicating impaired social recognition of familiar, unrelated individuals, but intact memory regarding relatives, based on odor discrimination [25]. The most informative and also well documented examination of the influence of hibernation on memory was a study by Millesi et al. in 2001 [24]. They tested the performance of the European ground squirrel (Spermophilus citellus) in a Skinner box and a maze test, as well as its social recognition by nest-share preference. Squirrels allowed to hibernate following training were compared to subjects, which were prevented from hibernation by wintering at warm ambient temperatures. Four weeks after arousal from hibernation, the performance of the cold-housed individuals in the operant and spatial task was impaired in comparison to the non-hibernating group, while the social recognition of familiar individuals seemed to be unaffected. Taken together, including the outcomes of our study, these findings suggest that 1) distinct memory contents are differentially affected by hibernation, 2) that there are species-specific differences in memory storage during hibernation or 3) that the results are biased by the different methods that were employed. Concerning a difference in species, one possible explanation could be the different minimum body temperature reached during hibernation. At an ambient temperature of 6 °C, the body temperature of S. citellus is about 8 °C [29], while M. marmota retains a body temperature of about 11 °C [26]. For non-hibernators, it has been shown that the memory deficit increases at lower body temperatures [17,20,21], though it was concluded that this concerned only memory consolidation and not long-term memory. Another suggestion for the different outcomes is related to the use of different behavioral tests: In order to facilitate and accelerate learning, we used positive reinforcement to train the marmots. Furthermore, we carried out many training sessions with the objective of long-term memory formation. We observed that the marmots not only learned what to do to be rewarded, they also improved their motor skills, since climbing the boxes became smoother during the training phase. Thereby, the trained behavior might have become a matter of routine. Both the reinforcement and the intensive training could have favored retention of the acquired skills. However, it should also be noted that behavioral tests are generally biased by motivation. For instance, a lower motivation for

Fig. 9. Difference in locomotor activity between experienced (n = 10) and inexperienced (n = 7) animals (median ± st dev). Mean locomotor activity of experienced animals differs significantly from that of inexperienced control animals (⁎⁎⁎p = b 0.001, Mann–Whitney Rank Sum Test).

84

L.E. Clemens et al. / Physiology & Behavior 98 (2009) 78–84

food intake or a higher impulse to explore might drive animals in a maze to walk around, instead of solving the task as fast as possible, even if the information “where to find the food” is remembered. In the study on social recognition [25], a conclusion about memory retention was drawn from comparing the intensity of exploration of conspecific odors. Before hibernation, squirrels spent more time exploring odors from unfamiliar individuals than they did with odors of familiar subjects. After hibernation, odors from both familiar and unfamiliar animals were explored more intensively and the difference between these two groups was no longer significant. This result might not necessarily imply a memory deficit, but could also be indicative of a generally high exploratory activity after hibernation. In our study we noticed that after hibernation, the marmots took more time before they paid attention to the experimental set-up, compared to the test before hibernation. The performance in the test itself was unimpaired. Besides the finding of an unchanged memory after hibernation, we observed that the marmots required less time to learn the box task, when they were trained early in the active season. Although this is based on a comparison of only few animals, it indicates that the capacity of learning and memory might underlie seasonal changes. In accordance with this, neurophysiological studies on ground squirrels have demonstrated that neuronal parameters such as hippocampal dendritic length, arboring and the number of dendritic spines differ in the course of the year, reaching highest annual values immediately after hibernation [9,11,13,14]. This suggests a seasonal variation in the requirement of cognitive abilities, which might not be surprising in regard to the characteristics of a hibernator's annual cycle: In Alpine marmots, as in many other seasonal animal species, the daily routine is much more eventful in spring, where mating and breeding take place [30] and the above-ground activity, which is mainly comprised of feeding [31] and social interactions [32], is highest. However during fall, marmots reduce their above-ground activity [32,33]. Such a decrease was also found in our study: The marmots showed a lower locomotor activity in fall compared to spring in the first open field test. Seasonal characteristics of memory became further apparent in a difference between active and hibernation season: During the active season, the performance of M. marmota was not consistent, since alterations in the training procedure led to inferior test results. A causal explanation for this memory flexibility during the active season might be found in the physiological plasticity of the central nervous system, which provides the basis for adaptations to changing environmental conditions [34]. On the other hand, hibernation could exert a protective effect on memory because the animals are shielded from external stimuli that may interfere with previous memories [23,28]. Thus, during hibernation, memory seems to be “frozen”. It has previously been phrased that many global biochemical processes exploit the cold to minimize reaction rates and thereby energetic costs, while being able to reach full activity after rewarming [35]. Viewed in this light, long-term memory is part of the team: Memory is unaffected by six months of hibernation in the Alpine marmot and it can be harnessed right after arousal. Acknowledgment We are grateful to Nicole Steinberg for her help with the preparation of the manuscript, particularly with regard to the English wording. References [1] Ortmann S, Heldmaier G. Regulation of body temperature and energy requirements of hibernating alpine marmots (Marmota marmota). Am J Physiol, Regul Integr Comp Physiol 2000;278:698–704. [2] Heldmaier G, Ortmann S, Elvert R. Natural hypometabolism during hibernation and daily torpor in mammals. Respir Physiol Neurobiol 2004;141:317–29.

[3] Frerichs KU, Kennedy C, Sokoloff L, Hallenbeck JM. Local cerebral blood flow during hibernation, a model of natural tolerance to ‘cerebral ischemia’. J Cereb Blood Flow Metab 1994;14:193–205. [4] Reid MS, Kilduff TS, Romero LM, Florant GL, Dement WC, Heller HC. Monoaminic and metabolite levels in the cerebrospinal fluid of hibernating and euthermic marmots. J Sleep Res 1992;1:45–50. [5] Henry P, Russeth KP, Tkac I, Drewes LR, Andrews MT, Gruetter R. Brain energy metabolism and neurotransmission at near-freezing temperatures: in vivo(1)H MRS study of a hibernating mammal. J Neurochem 2007;101:1505–15. [6] STRUMWASSER F. Thermoregulatory, brain and behavioral mechanisms during entrance into hibernation in the squirrel, Citellus beecheyi. Am J Physiol 1959;196:15–22. [7] Walker JM, Glotzbach SF, Berger RJ, Heller HC. Sleep and hibernation in ground squirrels (Citellus spp): electrophysiological observations. Am J Physiol 1977;233: 213–21. [8] Popov VI, Bocharova LS, Bragin AG. Repeated changes of dendritic morphology in the hippocampus of ground squirrels in the course of hibernation. Neuroscience 1992;48:45–51. [9] Von der Ohe CG, Darian-Smith C, Garner CC, Heller HC. Ubiquitous and temperaturedependent neural plasticity in hibernators. J Neurosci 2006;26: 10590–8. [10] Popov VI, Medvedev NI, Patrushev IV, Ignat'ev DA, Morenkov ED, Stewart MG. Reversible reduction in dendritic spines in CA1 of rat and ground squirrel subjected to hypothermia–normothermia in vivo: a three-dimensional electron microscope study. Neuroscience 2007;149:549–60. [11] Ruediger J, Van der Zee EA, Strijkstra AM, Aschoff A, Daan S, Hut RA. Dynamics in the ultrastructure of asymmetric axospinous synapses in the frontal cortex of hibernating European ground squirrels (Spermophilus citellus). Synapse 2007;61: 343–52. [12] Popov VI, Bocharova LS. Hibernation-induced structural changes in synaptic contacts between mossy fibres and hippocampal pyramidal neurons. Neuroscience 1992;48:53–62. [13] Strijkstra AM, Hut RA, de Wilde MC, Stieler J, Van der Zee EA. Hippocampal synaptophysin immunoreactivity is reduced during natural hypothermia in ground squirrels. Neurosci Lett 2003;344:29–32. [14] Von der Ohe CG, Garner CC, Darian-Smith C, Heller HC. Synaptic protein dynamics in hibernation. J Neurosci 2007;27:84–92. [15] Martin SJ, Grimwood PD, Morris RG. Synaptic plasticity and memory: an evaluation of the hypothesis. Annu Rev Neurosci 2000;23:649–711. [16] Dudai Y. Molecular bases of long-term memories: a question of persistence. Curr Opin Neurobiol 2002;12:211–6. [17] Soumireu-Mourat B, Destrade C, Cardo B. Deep hypothermia in mice: effect of the interval between learning and treatment. Physiol Behav 1974;13:471–3. [18] Fulton D, Kemenes I, Andrew RJ, Benjamin PR. Time-window for sensitivity to cooling distinguishes the effects of hypothermia and protein synthesis inhibition on the consolidation of long-term memory. Neurobiol Learn Mem 2008;90:651–4. [19] Riccio DC, Hodges LA, Randall PK. Retrograde amnesia produced by hypothermia in rats. J Comp Physiol Psychol 1968;66:618–22. [20] Andjus RK, Knopfelmacher F, Russell RW, Smith AU. Effects of hypothermia on behavior. Nature 1955;176:1015–6. [21] Palchykova S, Crestani F, Meerlo P, Tobler I. Sleep deprivation and daily torpor impair object recognition in Djungarian hamsters. Physiol Behav 2006;87:144–53. [22] Pyter LM, Reader BF, Nelson RJ. Short photoperiods impair spatial learning and alter hippocampal dendritic morphology in adult male white-footed mice (Peromyscus leucopus). J Neurosci 2005;25:4521–6. [23] McNamara MC, Riedesel ML. Memory and hibernation in Citellus lateralis. Science 1973;179:92–4. [24] Millesi E, Prossinger H, Dittami JP, Fieder M. Hibernation effects on memory in European ground squirrels (Spermophilus citellus). J Biol Rhythms 2001;16: 264–71. [25] Mateo JM, Johnston RE. Retention of social recognition after hibernation in Belding's ground squirrels. Anim Behav 2000;59:491–9. [26] Heldmaier G, Ortmann S, Körtner G. Energy requirements of hibernating alpine marmots. In: Carey C, Florant GL, Wunder B, Horwitz B, editors. Life in the cold. Ecological, physiological and molecular mechanisms. Boulder: Westview Press; 1993. p. 175–83. [27] Arnold W. Social thermoregulation during hibernation in alpine marmots (Marmota marmota). J Comp Physiol, [B] 1988;158:151–6. [28] Mihailovic L, Petrovic-Minic B, Protic S, Divac I. Effects of hibernation on learning and retention. Nature 1968;218:191–2. [29] Strijkstra AM, Daan S. Ambient temperature during torpor affects NREM sleep EEG during arousal episodes in hibernating European ground squirrels. Neurosci Lett 1997;221:177–80. [30] Arnold W, Psenner H. Murmeltiere. In: Grzimek B, editor. Enzyklopädie der Tierwelt: Säugetiere. München: Kindler Verlag, vol. 3; 1988. p. 33–48. [31] Barash DP. Social behaviour and individual differences in free-living alpine marmots (Marmota marmota). Anim Behav 1976;24:27–35. [32] Johns DW, Armitage KB. Behavioral ecology of Alpine yellow-bellied marmots. Behav Ecol Sociobiol 1979;5:133–57. [33] Cazzola M, Panseri M. Behaviour of Alpine marmot (Marmota marmota) during the summer months: time budgets from a colony in valle Brembana (Bergamo, Italy). In: Ramousse R, Allaine D, Le Berre M, editors. Adaptive strategies and diversity in marmots. Stratégies adaptives et diversité chez les marmottes. Lyon: International Marmot Network Publisher; 2003. p. 39–40. [34] Gold PE. The many faces of amnesia. Learn Mem 2006;13:506–14. [35] Carey HV, Andrews MT, Martin SL. Mammalian hibernation: cellular and molecular responses to depressed metabolism and low temperature. Physiol Rev 2003;83: 1153–81.