Fluctuations in central and peripheral temperatures associated with feeding behavior in rats

Abstract

We examined the pattern of temperature fluctuations in the nucleus accumbens (NAcc), temporal muscle, and skin, along with locomotion in food-deprived and nondeprived rats following the presentation of an open or closed food container and during subsequent eating or food-seeking behavior without eating. Although rats in food-deprived, quiet resting conditions had more than twofold lower spontaneous locomotion and lower temperature values than in nondeprived conditions, after presentation of a container, they consistently displayed food-seeking behavior, showing much larger and longer temperature changes. When the container was open, rats rapidly retrieved food and consumed it. Food consumption was preceded and accompanied by gradual increases in brain and muscle temperatures (∼1.5°C) and a weaker, delayed increase in skin temperature (∼0.8°C). All temperatures began to rapidly fall immediately after eating was completed, but NAcc and muscle temperatures returned to baseline after ∼35 min. When the container was closed and rats were unable to obtain food, they continued food-seeking activity during the entire period of presentation. Similar to eating, this activity was preceded and accompanied by gradual temperature increases in the brain and muscle, which were somewhat smaller than those during eating (∼1.2°C), with no changes in skin temperature. In contrast to trials with eating, NAcc and muscle temperatures continued to increase for ∼10 min after the container was removed from the cage and the rat continued food-seeking behavior, with a return to baselines after ∼50 min. These temperature fluctuations are discussed with respect to alterations in metabolic brain activity associated with feeding behavior, depending upon deprivation state and food availability.

all living organisms require food, and they all spend a significant part of their life engaged in food-seeking activity, which culminates in food consumption. Feeding could also be viewed as a representative motivated behavior, which depends on the preceding deprivation state (hunger/satiation), current food availability, and its accessibility (ability to consume perceived food). Therefore, feeding includes both food-seeking behavior (motivated search) and food consumption (eating).

To learn more about the central mechanisms associated with different forms and components of feeding behavior, in the present study, we used high-resolution thermorecording from the brain and two peripheral locations. Although it is commonly believed to be a tightly regulated and highly stable homeostatic parameter, brain temperature shows relatively large fluctuations (∼3°C) within the normal range of behavioral activity in temperature-stable conditions (for a review, see Ref. 14). Early thermorecording studies revealed that feeding behavior is accompanied by an increase in brain temperature (1, 6), which is evident in different structures and generally correlates with core body temperature. Although brain temperature and metabolic rate correlatively increased during eating (1), metabolic neural activation associated with food-seeking behavior appears to be the primary factor determining brain hyperthermia. In fact, Moritz Schiff was first to show that simple presentation of meat to a hungry dog results in ∼1°C increase in brain temperature [1868; cited by James (12) 1892]; this increase disappeared when the animal was satiated. Since feeding behavior involves several prerequisites and components (deprivation state, food availability, food-seeking activity, eating), it remains unclear why brain temperature increases, how it correlates with peripheral temperatures and animal behavior, and what the functional significance of these temperature fluctuations is.

To answer these questions, we developed a simple paradigm of feeding behavior and used repeated recordings from the same animals, which were either food-deprived or satiated and exhibited different behavior depending on food accessibility. Male rats were exposed to the same food container, which could be either open, allowing the rats access to the food, or closed, not allowing direct contact with the food despite continuous food-seeking activity. We were interested in several questions: 1) How is food deprivation reflected in basal locomotor activity and basal temperatures and how does it affect an animal's behavioral response to food-related stimuli? 2) What changes in temperature and general activity occur during eating? 3) How do these changes differ from those occurring during food-seeking activity without any food consumption?

To represent brain temperatures, we chose the nucleus accumbens (NAcc), a ventrally located structure involved in somato-sensory integration and organization of behavioral processes (43). Possibly via this involvement in mechanisms underlying motivational processes, the NAcc is heavily implicated in various aspects of drug-seeking and drug-taking behavior (see 13, 37). Temperature recordings were also made from facial skin, providing a sensitive measure of peripheral vasoconstriction/vasodilatation, a primary factor determining heat loss to the external environment (2, 4, 40). Temperature recordings from the brain and skin were combined with those from the temporal muscle, which is not involved in head movement and is supplied, as is the brain, by arterial blood from the common carotid artery. This muscle, however, is involved in jaw-closing, suggesting its possible activation and heat production during food consumption. Therefore, analysis of brain-muscle and skin-muscle temperature differentials could show whether heat generation is local (due to metabolic activation) or global (blood supply), an approach previously used to study the source of brain hyperthermia and factors contributing to its generation (16). Finally, temperature recordings were supplemented by monitoring locomotor activity, thus allowing us to correlate behavioral and physiological parameters.

MATERIALS AND METHODS

Subjects.

Six Long-Evans male rats (Taconic, Germantown, NY), weighing 400–460 g and housed under a 12:12-h light-dark cycle (lights on at 0700), with ad libitum access to food and water, were used. Prior to surgery, rats were put through a pilot trial of the feeding procedure for both training and selective purposes (see Experimental protocol). Protocols were performed in compliance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, Publications 865-23) and were approved by the Animal Care and Use Committee, National Institute on Drug Abuse-Intramural Research Program.

Surgery.

All animals were implanted with three thermocouple electrodes, as previously described (18). Animals were anesthetized intraperitoneally with 3.3 ml/kg of Equithesin (pentobarbital sodium, 32.5 mg/kg and chloral hydrate, 145 mg/kg) and mounted in a stereotaxic apparatus. Four holes were drilled through the skull: three for securing screws and one for thermocouple insertion over the NAcc shell (1.2 mm anterior to bregma, 0.9 mm lateral to bregma) using the coordinates of Paxinos and Watson (32). The dura mater was retracted, and the thermocouple probe was slowly lowered to the desired target depth (7.4 mm, measured from the skull surface). A second thermocouple probe was implanted subcutaneously along the nasal ridge with the tip ∼15 mm anterior to bregma. Although the tail is the primary organ of heat dissipation in rats, facial skin has dense vascularization, and this location provides mechanical stability of the electrode, an essential condition for long-term, artifact-free temperature recordings. A third thermocouple probe was implanted in the deep temporal muscle (musculus temporalis). The probes were secured with dental cement to the three stainless-steel screws threaded into the skull. Rats were allowed 3 days of recovery and two more days of habituation (6-h sessions) to the testing environment before the start of testing.

Experimental protocol.

During pilot trials that preceded surgeries, rats were mildly food deprived for ∼5 h in the testing box and then given a small (2.0 cm × 1.5 cm × 8.0 cm), clear, perforated container with two open ends and containing a 1.5-g piece of food (Rodent NIH-07 Open Formula, which is regularly used for feeding in our animal colony; Zeigler Bros., Gardners, PA). If the rats were able to retrieve and consume the food within 1 h, they were selected for the experiment. All tests occurred inside a Plexiglas chamber (32 × 32 × 32 cm) equipped with four infrared motion detectors (Med Associates, Burlington, VT) placed inside of a sound attenuation chamber. Rats were brought to the testing chamber at ∼9:30 AM and attached via a flexible cord and electrical commutator to thermal recording hardware (Thermes 16; Physitemp, Clifton, NJ). Temperatures were recorded with a time resolution of 10 s, and movement was recorded as the number of infrared beam breaks per 1 min. Room temperature was maintained at 23–24°C and was controlled by another thermocouple located inside of the recording chamber. Each of the 6 rats underwent 4 to 6 recording sessions. To produce a desire to eat, the rats were deprived of food for 24 h before the session. During each session, after habituation to the testing chamber (∼2 h), a food container was placed in the center of the chamber. The container was either open or sealed from the sides. In all sessions, rats were given a ∼10-min window to somehow show interest in the container. If the container was closed, its lifting and chewing (obvious attempts at retrieving the food) were recorded as positive interest, while more subtle acts such as sniffing were not. In the case of an open container, the times when the rat retrieved the food and finished eating were recorded. Two of these food-presentation sessions were performed per day, with ∼2 h between them. In one session, the container was closed, while in the other, it was open. Despite differences in behavior, the durations of exposure to either open or closed containers were approximately equal in each pair of animals and each session.

Histology and data analysis.

When recording was completed, all rats were anesthetized, decapitated, and had their brains removed for sectioning and confirmation of probe placement. Brains were cut on a cryostat into 50-μm slices and placed on glass slides. All probes were located within the medial portion of the ventral striatum (NAcc shell), as described in Paxinos and Watson (32).

Temperature and movement data were analyzed with 20-s and 60-s time bins and were presented as both absolute and relative changes with respect to the moment of stimulus presentation or key behavioral events (start of eating, start of food-seeking activity, completion of eating, removal of a container, etc.). One-way ANOVA with repeated measures, followed by post hoc Fisher tests, was used for statistical evaluation of temperature and locomotion changes preceding and following events of interest. Student's t-test was used for comparisons of between-site and between-condition differences in temperature and locomotion. Between-state differences were evaluated based on statistical comparisons of basal temperatures and locomotion averaged for 6 min preceding the container's presentation. The use of the words “increase” or “decrease,” as well as “significant” refers to the presence of a statistically significant change in the parameter or differences between the compared groups or conditions (with at least P < 0.05) revealed by either ANOVA or Student's t-test.

RESULTS

Satiated and hungry animals: basal locomotion, basal temperatures, and responses to food-related stimuli.

As shown in Fig. 1A and Table 1, rats in food-deprived and nondeprived conditions during quiet rest had no statistically significant differences in basal temperatures in each recording location, although all values were lower in food-deprived conditions. Similarly, food-deprived rats showed more than a twofold lower spontaneous locomotion (1.47 ± 0.34 counts/min) compared with nondeprived rats (3.57 ± 1.46 counts/min); these differences were also statistically insignificant.

Fig. 1.

A: basal temperatures in the NAcc, temporal muscle and facial skin in food-deprived and nondeprived rats under quiet resting conditions. The relationships between NAcc-muscle (B), NAcc-skin (C), and muscle-skin (D) temperatures are analyzed separately for food-deprived and nondeprived conditions. Each graph shows regression lines, regression equations, and coefficients of correlation. Hatched line marks the line of equality. Regression lines are shown only for parameters with a significant correlation (**P < 0.01; ***P < 0.001).

Table 1.

Changes in temperature associated with different behavioral patterns in satiated and hungry animals

	Satiated	Hungry
	No Interest	Eating	No Eating
n	18	15	12
NAcc
basal	36.49±0.10	36.25±0.13	36.21±0.11
peak	37.06±0.11	37.71±0.08††	37.42±0.15†
change	+0.58	+1.45	+1.21
mean	0.68±0.08	1.47±0.10††	1.28±0.11††
Muscle
basal	35.55±0.15	35.26±0.14	35.16±0.15
peak	36.03±0.16	36.89±0.10††	36.23±0.17
change	+0.48	+1.62	+1.07
mean	0.58±0.08	1.72±0.09††	1.10±0.11††**
Skin
basal	34.55±0.13	34.18±0.14†	34.33±0.14
peak 1	34.24±0.11 (3)	33.95±0.14 (2)	33.99±0.14 (2)
change 1	−0.31°C	−0.23°C	−0.34°C
mean	−0.47±0.07	−0.26±0.04	−0.35±0.08
peak 2	34.75±0.18 (22)	35.13±0.14 (15)	34.66±0.16 (15)
change 2	+0.20°C	+0.93°C	+0.33°C

Mean change was calculated based on peak values of each parameter in each test. Significant differences between eating and no eating (*P < 0.05; **P < 0.01). Significant differences between satiated and both subgroups of hungry rats (†P < 0.05; ††P < 0.01).

In both conditions, there were stable and similar differences between temperatures in each recording location, with maximal values in the NAcc (36.49 ± 0.10 and 36.25 ± 0.08°C for nondeprived and deprived conditions), lower values in the muscle (35.55 ± 0.15 and 35.23 ± 0.10°C; P < 0.05 vs. NAcc), and lowest values in the skin (34.55 ± 0.13 and 34.29 ± 0.08°C; P < 0.05 vs. muscle and NAcc). NAcc and muscle temperatures, moreover, were tightly and linearly correlated in both conditions (Fig. 1B). Regression lines were virtually fully superimposed and parallel to the equality line (hatched), suggesting that temperatures in the NAcc remain ∼1°C higher than in the muscle within the entire range of temperature variation. The relationships between muscle-skin and skin-NAcc temperatures were more complex and state dependent (see Fig. 1, C and D). Although there was no correlation between these parameters in satiated conditions, NAcc and skin temperatures, as well as skin and muscle temperatures, were positively correlated (r = 0.52 and 0.60; P < 0.01) in food-deprived conditions.

Food deprivation significantly affected the animal's response to presentation of a food container. While satiated rats did not usually show any specific interest to the presented container (18/20 or 90%; in 2/20 cases or 10% they consumed food), hungry rats consistently showed food-seeking activity, manifesting in active interaction with the container. In 13 cases, rats removed food from an open container and consumed it, and in 12 cases, rats continued food-seeking behavior, attempting to get food from closed containers.

Temperature and locomotor responses to a food container in satiated conditions.

Although in all 18 cases, satiated rats did not actively interact with the presented container, its placement in the cage (first vertical line in Fig. 2) increased NAcc and muscle temperatures, decreased skin temperature, and induced locomotor activation. Skin and locomotion showed the shortest latency (20–40 s and 60 s, respectively), but the increase in brain and muscle temperatures became significant at later times (100–120 and 200–220 s, respectively). Both locomotion and temperatures increased again when the container was removed from the cage (second vertical hatched line in Fig. 2); these were clearly weaker than those associated with its presentation. Decreases in skin temperature were more transient than increases in NAcc and muscle temperatures, which peaked at ∼10 min after the container's removal (∼15 min) and returned to the baselines at 20–30 min. Temperature changes in the NAcc were not only more rapid than in the muscle but also larger in amplitude (∼0.6 and 0.5°C, respectively; see Table 1), resulting in a significant rise in the NAcc-muscle temperature differential (Fig. 2B). In contrast, the skin-muscle differential rapidly decreased and maintained at lower values for about 20 min after stimulation onset (or ∼9 min after the container's removal).

Fig. 2.

Changes in brain (○), muscle (□), and skin (▵) temperatures, as well as locomotion following presentation of a food container to satiated rats. Relative changes in temperatures with respect to prestimulus baseline (A; in°C), brain-muscle and skin-muscle temperature differentials (B; in°C), and locomotion (C). Because the duration of the container's presentation slightly varied in each trial, data were calculated separately with respect to the moment of presentation (first vertical dashed line) and its removal from the case (second vertical dashed line). Solid symbols show values significantly different from the prestimulus baseline. Each row of data was analyzed using a one-way ANOVA with repeated measures with subsequent use of Fisher test.

Two behavioral responses in food-deprived conditions: eating and sustained food-seeking activity.

Deprived animals always attempted to retrieve the food from the container introduced in their cage. When the food container was open, food-seeking activity always resulted in food retrieval and its consumption (n = 15), although latency to eating greatly varied in different trials (110 s to 620 s; mean = 325.3 ± 48.5 s). When the container was closed (n = 12), rats began to interact with it, trying to get the food, until this container was removed from the cage. In this case, latency to the first active contact to the container varied from 60 to 580 s (mean 218.3 ± 41.7 s). Duration of eating also greatly varied (280 to 800 s; mean = 487.33 ± 41.33 s). Because of this variability, our analysis of temperature and locomotion was performed for each consecutive behavioral event (see Fig. 3). Mean values and their between-group differences are shown in Table 1.

Fig. 3.

Changes in brain (○), muscle (□), and skin (▵) temperatures, as well as locomotion following presentation of a food container to food-deprived rats. Left: changes during presentation of an open container with subsequent eating. Right: changes following presentation of closed container, with no food consumption. Absolute changes in temperatures (A; in°C), relative changes in temperature with respect to prestimulus baseline (B), brain-muscle and skin-muscle temperature differentials (C; in°C), and locomotion (D). Because the duration of eating and the container's presentation varied in each trial, data were calculated separately with respect to the moment of presentation (first vertical dashed line), the beginning of eating (second dashed line on left panel), the beginning of active interaction with the container (second dashed line on right panel), and the end of eating (third dashed line on left panel), or the container's removal from the case (third dashed line on right panel). Solid symbols show values significantly different from the prestimulus baseline.

Presentation of an open container to hungry rats resulted in a rapid decrease in skin temperature, a gradual increase in brain and muscle temperatures, and a gradual increase in locomotion (Fig. 3, A, B, and D; left). Eating was associated with strong increases in the NAcc, muscle, and skin temperatures, along with a relative decrease in locomotion. Muscle temperature showed the most rapid increase (∼1.2°C over 5 min, or 0.23°C/min), clearly exceeding that in the NAcc (∼0.8°C over 5 min, or 0.15°C/min) and skin (0.7°C over 5 min). These differences resulted in a robust decrease in the NAcc-muscle and skin-muscle temperature differentials (Fig. 3C). After eating was completed, temperatures in the muscle (90 s) and skin (50 s) rapidly dropped. In the NAcc, the increase was momentarily stopped, but despite a continuous decrease, it became significant only on the 6th min. The peak of locomotor activity with intense grooming, washing, and rearing was typical of this period (Fig. 3D). Although temperatures in all locations were still higher at the end of the observation period (33 min), they always returned to baseline within 1 h. These changes were significantly greater than those seen in satiated rats, which showed no interest in food (see Table 1 and compare Figs. 2A and 3B). This behavioral pattern was also related to stronger and more prolonged changes in the NAcc-muscle and skin-muscle temperature differential (Fig. 3C).

Presentation of a closed container also resulted in a rapid decrease in skin temperature, gradual increase in brain and muscle temperatures, and a gradual increase in locomotion—the pattern seen in all other situations (Fig. 3, A, B, and D; right). When the rat began to interact with the container, locomotor activation accelerated, brain and muscle temperatures continued to increase at the same pace (∼1.2°C over ∼10 min), but skin temperature slowly returned to baseline or slightly higher values. Importantly, both locomotion and brain and muscle temperatures peaked after the container was removed from the cage. Although significant for ∼9 min, this increase was incomparably weaker than that occurring during interaction with the container. Brain and muscle temperatures were still significantly higher than baseline at the end of the analysis interval (33 min after presentation of a container), but always returned to baseline within 1 h. While qualitatively similar, the changes in temperature and locomotion were significantly greater than those seen in satiated rats, which showed no interest in food (see Table 1 and compare Figs. 2 and 3). This behavioral pattern was also related to stronger and more prolonged changes in the NAcc-muscle and skin-muscle temperature differential (Fig. 3C).

As shown in the Table 1, with quite similar baseline values, both NAcc and muscle temperature showed significantly greater increases in cases of eating than no eating. In contrast, both subgroups showed similar decreases in skin temperature, which occurred after the container was presented to the rats (peak at 2nd min in both groups). While skin temperature rapidly returned to baseline or somewhat higher levels in animals that were unable to retreive food, animals engaged in eating showed a significant and prolonged increase in skin temperature, with an almost 1°C increase over baseline. Animals that could not obtain food had clearly higher locomotor activity than rats that were able to eat (compare Fig. 3D, right and left).

Individual events of feeding behavior: food-arousing stimuli, eating, and sustained food-seeking activity with no eating.

When temperature changes were analyzed with high temporal resolution (20 s), we found that the initial temperature response to presentation of a food container to food-deprived and nondeprived rats is virtually identical in each recording location (Fig. 4A). In each case, skin temperature rapidly decreased (0–40 s), while NAcc temperature increased. While skin response was equally rapid and similar in magnitude in all three subgroups (see Table 1 and Fig. 4, A–C), NAcc temperature response was stronger in both subgroups of hungry animals, both in terms of latency (20–40 s vs. 100–120 s) and relative amplitude (0.17 and 0.19°C at 7th value vs. 0.13°C in satiated rats). Temperature in the muscle also increased in each group but slower than that in the brain; it became higher than baseline at 2–3 min, when the rat was engaged in either food-seeking or eating behavior. Temperature change in the skin was not only much stronger than that in muscle or NAcc, it was also very transient, being evident for ∼40 s following stimulation (B). This change also had the shortest onset latency, being significantly lower than baseline at the first post-stimulus data point (0–20 s).

Fig. 4.

The initial changes in temperatures following the food container's presentation in rats in food-deprived (A and B) and nondeprived conditions, with open container (C and D) and with closed container (E and F). Top graphs show relative changes in temperature in each recording location analyzed with 20-s time bins; bottom graphs show rate of temperature changes analyzed for each subsequent 20-s time bins. Solid symbols show value significantly different from baseline.

As shown in Fig. 5A, eating was preceded and accompanied by gradual and strong temperature increases in each recording location. The pre-eating increase was strongest in the NAcc, but during eating, the temporal muscle showed the highest rate of increase (most likely because of chewing). The completion of eating resulted in immediate termination of temperature increase in each location. While temperature began to gradually decrease in the muscle and skin, the decrease in the NAcc became significant only at the 5th min.

Fig. 5.

Temperature changes associated with eating (A and B) and food-seeking activity without eating (C and D) in food-deprived conditions. Top graphs show relative temperature changes and bottom graphs show rate of temperature changes for each subsequent 20-s interval. The first vertical dashed line marks the start of eating (A and B) or start of active interaction with the container (C and D). Second vertical dashed line marks the end of eating (A and B) or the moment of the container's removal from the case (C and D). Solid symbols show values significantly different with respect to each behavioral event. Because the duration of eating and food-seeking behavior differed in each trial, analyses were conducted for each behavioral event.

Rate of change (relative change for each subsequent 20 s) is another representation of temperature dynamics (see Fig. 5B). In the NAcc, a continuous temperature increase that was evident during the pre-eating food-seeking behavior (0.02–0.03°/20 s) was accelerated (0.05–0.06°/20 s) during eating but rapidly decreased within ∼40 s after the completion of eating. In the muscle, a slight temperature increase seen before consumption (∼0.01°C/20 s) was rapidly accelerated to very high values at the start of eating (∼0.1°C/20 s) and slightly declined during subsequent eating. When eating was finished, muscle temperature momentarily fell and continued to decline until returning to baseline. Skin temperature gradually increased immediately before and during the entire period of food intake (∼0.05°C/20 s) but rapidly dropped after eating was completed.

As shown in Fig. 5C, food-seeking activity with no eating was accompanied by a stable increase in NAcc and muscle temperatures, which was virtually identical to that occurring after presentation of the container and initiation of food consumption. In contrast to food consumption, both temperatures did not fall after the container was removed from the case. During this period, rats continued to search, and temperatures still slightly increased, although at a much slower pace than during preceding activity. Changes in skin temperature were more variable, but a gradual increase was evident during the entire period of food-seeking activity and abruptly stopped after the container was removed from the cage. The rate of temperature increase in the NAcc and muscle was similar (0.03–0.04°C/20 s) during the entire period of food-seeking activity, being slightly higher in the NAcc than in the muscle (see Fig. 5D).

Relationships between basal brain temperature and its stimulus-induced and behavior-related changes.

As shown in Fig. 6A, presentation of a container to satiated rats results in variable increases in NAcc temperature, ranging from weak (0.1°C) to large (1.3°C); they were independent of basal brain temperatures. Eating (Fig. 6B) also induced variable increases in NAcc temperature (0.7–2.2°C), but they were strongly dependent on basal levels (r = 0.77; P < 0.01). Increases were greater at low basal temperatures and weaker at high temperatures (mean 1.47°C). This correlation was absent for temperature increases associated with food-seeking behavior (Fig. 6C); these increases were independent of basal temperatures and, on average, were somewhat smaller than those seen during eating (1.28°C).

Fig. 6.

The relationships between basal NAcc temperature and its changes induced by stimulus presentation in satiated conditions (A), eating (B), and food-seeking activity (C). Each graph shows regression lines, a regression equation, and a coefficient of correlation, as well as mean values of temperature increase in each condition (bold horizontal hatched line). Correlation was significant and relatively strong (r = −0.77, P < 0.01) only for brain temperature response during eating.

DISCUSSION

The present study demonstrates that feeding behavior is accompanied by rapid and relatively large temperature fluctuations in both the brain and peripheral locations. These fluctuations differed in their latencies and amplitudes between these recording locations, had certain differences in food-deprived and nondeprived conditions, and were dependent on the ability to obtain food and the type of animal behavior.

Sources and mechanisms of brain, muscle, and skin temperature fluctuations.

Metabolism-related heat production appears to be a primary factor determining brain temperature increases occurring under behavioral conditions (22, 25, 39, 41). However, metabolic heat is continuously redistributed within brain tissue and removed from the brain by cerebral circulation to the lungs and then to the external environment. Although circulation is the primary means of heat dissipation from the brain and a contributor to brain temperature fluctuations, heat cannot be delivered to the brain from the periphery since brain tissue is always warmer than arterial blood arriving to the brain (6, 8, 10, 14, 22). The brain-arterial blood temperature gradients always remain positive during behavioral conditions (6, 16), intense physical exercise (29), and even general anesthesia (see 15), which results in marked metabolic inhibition, with significant brain and body hypothermia (18). Brain temperature increases induced by environmental challenges, moreover, have consistently shorter latencies and are stronger than those occurring in arterial blood (16) and nonlocomotor muscles (3; this study), suggesting brain metabolic activity as a primary factor of brain hyperthermia and a force behind a more delayed body hyperthermia.

Although most energy used for brain metabolism is spent restoring membrane potentials after electrical discharges (21, 33, 39), suggesting a basic relationship between electrical activity of neurons and metabolic neural activity, these are two different parameters. Metabolic activity involves both neurons, glial cells (which ratio in human brain is at least 1:10; see Ref. 20) and brain vascular endothelium, and significant energy is spent for neural functions not directly related to spike generation (i.e., synthesis of macromolecules, transport of protons across mitochondrial membranes counteracting the proton leak in the opposite direction; see Ref. 11). Although not as specific as other measures of neural activity (i.e., single-unit and multiunit electrical discharges, EEG) or brain metabolism (oxygen or glucose consumption), brain temperature provides a dynamic parameter that reflects both generalized and structure-specific changes in metabolic neural activity. This parameter could be valuable to quantitatively characterize arousal, a generalized physiological activation, which is usually graded based on alterations in EEG or other homeostatic parameters (i.e., cardiovascular, respiration, corticosterone). In contrast to true metabolic measures, which often require anesthesia and have a limited time window for experiments, temperatures may be recorded simultaneously from several brain and body sites in behaving animals during chronic experiments. This technical advantage was crucial for the goals of this study. Using the same animals recorded during several consecutive daily sessions, we were able to evaluate the pattern of central and peripheral temperature fluctuations induced by the same stimulus in food-deprived and nondeprived states and during different behaviors.

Because temperature fluctuations in skin and muscle depend upon two variables: the state of vessels (vasoconstriction/vasodilatation) and the temperature of arterial blood inflow, these two locations were important for evaluating the source of brain temperature fluctuations and its underlying mechanisms. By tracking temperature responses in different brain and body sites with a short collection interval, one is able to follow the dynamics of heat generation and flows within the organism.

Although metabolism-related heat production appears to be the major cause of brain temperature increases following exposure to salient stimuli and during motivated behavior, change in cerebral blood flow is another important variable affecting temperature fluctuations. It is unclear, however, how this parameter is regulated and what its relationships with brain metabolism and temperatures are. Although the direct action on brain vessels of certain products of enhanced brain metabolism is generally believed to be a primary force in inducing vascular dilation and increased cerebral blood flow (9, 35, 42), direct, linear correlations between temperatures and blood flow are well established in peripheral tissues [i.e., skin, intestine, muscle, liver (5, 27, 28, 36)] and appear to exist in the brain (24, 29). Therefore, metabolism-related increases in local brain temperature can be viewed as a factor that increased local blood flow. This factor could explain, at least partially, the well-known, but little-understood, phenomenon of blood flow increases that typically exceed metabolic activity of brain tissue (9). Using this mechanism, the brain is able to increase blood flow more and in advance of actual metabolic demands (“anticipatory” metabolic activation), thus providing a crucial advantage for the organism's adaptation to potential energetic demands. By increasing blood flow above the brain's current need, more oxygen and nutrients are delivered to the areas of potential demand and more potentially dangerous metabolic heat is removed from intensively working brain tissue. While special experiments with parallel measurements of temperature and blood flow (which is a quite complex task, especially in small animals) are necessary to clarify this issue, it appears that the brain could “regulate” its own temperature by temperature-dependent alteration of its blood flow.

Motivational state.

If the animal is deprived from food, it is assumed that it differs from a nondeprived animal by its motivational state, which should be reflected in physiological and behavioral parameters. We found that rats in a food-deprived state have lower spontaneous locomotion and lower temperatures in each recording point than in a nondeprived state. Although this may seem surprising from a behavioral point of view, this finding is consistent with other data, suggesting that food deprivation (starvation, fasting) decreases whole-body metabolism and body temperatures (7, 26, 30). Metabolism following food deprivation is at lower levels (resulting in low locomotion and low brain and muscle temperatures), and peripheral vessels are to some extent constricted (resulting in lower skin temperature) to retain heat inside of the body by limiting its dissipation to the external environment. However, food-deprived animals showed noticeably different behavioral responses to the presented food container than nondeprived animals. While the latter showed a lack of interest toward this container along with a weak temperature response, food-seeking activity associated with much stronger changes in temperature and locomotion was typical of food-deprived conditions. Therefore, presentation of the same “appetitive” stimulus to rats in deprived and nondeprived states differently affects behavior and temperatures, suggesting that specific environmental stimulation is essential to reveal (actualize) motivational state.

Appetitive or arousing stimulus.

While the food container, both open and closed, was an appetitive stimulus in food-deprived conditions and induced food-seeking behavior, the same stimulus also induced a locomotor and temperature response in satiated animals, which showed no specific interest in this container. While different in amplitude and duration, the initial phase of temperature and locomotor response induced by a container's presentation was similar in hungry and satiated conditions (see Fig. 4), mimicking the response induced by other salient somato-sensory stimuli (tail-pinch, presentation of male or female partner, sexually relevant stimuli, placement in a new cage, etc.). All these stimuli increase spontaneous locomotion, rapidly but transiently decrease skin temperature, and more slowly increase brain and muscle temperatures. Therefore, the same food container (a combination of visual and olfactory signals) used in nondeprived conditions could be better defined as an arousing stimulus, inducing an arousal response. A second peak of locomotion and temperature response induced by the removal of a container is consistent with this arousal mechanism. Previously, the same phenomenon—two activation peaks—was reported in rats after presentation of an opposite-sex partner and after its removal from the cage (23). Therefore, an increased arousal (or nonspecific activation) could be viewed as both a precondition for and a correlate of specific motivated behavior.

Decrease in skin temperature following a container's presentation is an obvious consequence of an acute peripheral vasoconstriction, a universal adaptive organism's response to limit heat dissipation. This response is very rapid (∼10 s), transient, and often followed by weak, rebound-like skin hyperthermia, which reflects vasodilatation and enhanced heat loss. Therefore, peripheral vasoconstriction contributes to increased body temperature, thus affecting intrabrain heat homeostasis. Although in this study, this centrally mediated response was triggered via activation of visual and olfactory pathways (sight, smell), it could also be induced via activation of somatic and visceral neural afferents by other somato-sensory stimuli and certain drugs (see 19). Despite different causes and underlying mechanisms, brain and body hyperthermia coupled with decreased skin temperatures is homologous to fever, when an organism actively promotes increased temperature by enhancing heat production and limiting heat dissipation (38; see, however, Ref. 34, for a review).

Food consumption (eating).

Eating was preceded and accompanied by a gradual increase in NAcc temperature, which peaked at the time when eating was finished. The rate of temperature increase rapidly grew after the start of eating, maintained during the entire period of consumption, and sharply dropped when eating was completed. While the temperature increase was more rapid and stronger in the NAcc than in the muscle after presentation of a container and during food seeking, muscle showed a stronger increase during eating. In this case, the rate of increase was maximal at the beginning of eating (0.12°C/20 s), slowly decreased during eating, and rapidly inverted into a decrease after eating was completed. Although temporal muscle is not involved in general motor behavior, it is activated by jaw closing and located in a close proximity to the masseter muscle, which is heavily involved with chewing. Therefore, a strong temperature increase in the muscle and an increase in skin temperature (due to increased arterial blood inflow to the head) could be viewed as a consequence of specific muscular activity associated with chewing. However, despite increased motor activity associated with eating, temperature increases were not related to general locomotion, which drastically reduced, when increases were maximal. On the other hand, temperatures began to fall after cessation of eating when the rat became hyperactive again, showing intense washing, grooming, and exploratory behavior.

Food-seeking activity.

Sustained food-seeking activity was associated with a robust increase in general locomotion and gradual increases in NAcc and muscle temperatures. Although these increases were somewhat weaker in amplitude than those during eating (see Table 1 and Fig. 3), they were more prolonged and variable. Temperature increases during food-seeking activity did not differ from the preceding arousal-associated increase that began after the container was presented to the rat but slowly decreased and disappeared after the container was removed from the case. In contrast to the relatively quick drop in temperatures after eating, NAcc and muscle temperatures always peaked after the container was removed from the case, and mean values were also significantly larger for ∼10 min (see Fig. 5). This extended hyperthermic response coupled with robust hyperlocomotion was the most important difference compared with eating.

Common features of temperature dynamics during different forms of motivated behavior.

Temperature changes associated with feeding has several important similarities to those occurring during sexual behavior (23), another basic motivated activity of living beings. Similarly to locomotor activation and gradual increases in NAcc temperature seen after presentation of a food container to hungry rats in this study, the same changes occurred in sexually experienced male rats after presentation of a female partner, which was seen and smelled but inaccessible for direct interaction. Similar to food-seeking activity, temperatures further increased when the rats were allowed free interaction and the male partner initiated copulatory behavior. Similar to eating, copulatory behavior was associated with gradual temperature increases, which were abruptly terminated immediately after ejaculation. While similar in its pattern, temperature fluctuations during sexual behavior were larger in amplitude, with both sharper temperature increases during copulation and stronger postejaculatory temperature decreases. In addition to different levels of metabolism, these quantitative differences in temperatures could also be related with specifics of animal behavior. While behavioral activation (grooming, washing, hyperlocomotion) was consistently seen after eating, prominent hypoactivity was typical of male rats immediately after completion of a copulatory cycle. Although cocaine is a drug reinforcer and intravenous self-administration is unnatural motivated behavior, the pattern of temperature fluctuations during this behavior (17) also had important similarities with feeding and sexual behaviors, suggesting some common behavior-associated changes in brain activational processes and metabolic neural activity. Within the cycle of repeated cocaine self-injections, NAcc temperatures gradually increased during cocaine-seeking activity, peaked at the moment of self-injection, and rapidly decreased after intravenous drug infusion.

Therefore, it is reasonable to speculate that motivated (seeking) behavior, independently of the reinforcer's nature (food in hungry conditions, sexual partner for a sexually experienced subject, drug for a drug-dependent subject), appears to be accompanied by sustained metabolic neural activation. This metabolic activation is triggered via somato-sensory pathways and maintained during consummatory behavior (eating, copulation, lever-pressing), until successful interaction with the reinforcer (food consumption, ejaculation resulting from copulation, drug self-injection) and its detection by the central nervous system. The time interval and mechanisms of this detection could be different for each type of motivated behavior, determining specifics of food, drug, and sex reward. While ejaculation per se has an immediate, neurally mediated “satisfying” effect, abruptly blocking further behavior (and inducing in humans acute euphoria followed by disphoria), some defined time is necessary for the self-administered drug to reach specific neural substrates and inhibit drug-seeking behavior. In the case of food, this delay between food consumption and true satiety could be the most prolonged, since in addition to rapid “eating-blocking” signals from the stomach, the brain needs to receive metabolic signals from consumed food to induce satiety and stop eating behavior. Despite all of these obvious differences, interaction with the reinforcer eventually results in termination of metabolic neural activation and disappearance of seeking behavior, correlating with satisfaction (reward).

Perspectives and Significance

This study demonstrates that brain temperature is a sensitive physiological parameter, which shows relatively rapid and large fluctuations following exposure to various arousing stimuli and during motivated behavior. Although brain temperature in this study was represented by the NAcc, a deep brain structure that is heavily implicated in regulation of motivational processes (31, 43), and our data support the involvement of this area in various aspects of feeding behavior (37), it could be reasonably assumed that generally similar temperature fluctuations occur in other brain structures. Although previous work revealed significant structural differences in basal temperatures, as well as minor but significant differences in temperature responses to various arousing stimuli and during motivated behavior, behavior-related temperature fluctuations were correlative, suggesting its generality for the brain as a whole (for a review, see Ref. 14). However, similar to this study, these fluctuations were highly specific with respect to the stimulus nature, its arousing potential, and key events of motivated behavior. Therefore, it appears that brain temperature provides a valuable measure of nonspecific arousing components common to different types of motivated behavior, but this measure is weaker in revealing structural specificity of neuronal activity.

GRANTS

This research was supported by the Intramural Research Program of the National Institutes of Health, National Institute on Drug Abuse.