INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE PREOPTIC ANTERIOR hypothalamic area (POAH) is considered the thermointegrative center of the mammalian brain. Cooling the POAH elicits appropriate heat-gain responses, and, conversely, heat loss responses are activated when the POAH is heated. Changes in ambient, hence skin, temperature (T_s) alter the threshold POAH temperatures (T_POAH) for thermoregulatory responses or alter the gain of the responses. Simple neuronal models have been proposed as hypotheses for how POAH neurons could interact and respond to thermoafferent input to produce these system characteristics that have been described in a number of mammalian species (for reviews, see Refs. 3, 14-16). If any of these models for POAH thermointegration, or even their basic assumptions, are correct, then there should be POAH cells that respond both to local temperature (thermosensitive) and to T_s (thermoresponsive). There have been many single-unit studies of POAH neurons on anesthetized and on unanesthetized animals. These studies have demonstrated POAH units that are thermosensitive, either warm or cold sensitive. In addition, studies have reported units that respond to changes in T_s (for reviews, see Refs. 3, 5). These data seem to support models that place the integration of thermoafferent information in the POAH.

Recent investigations of putative thermoafferent pathways, however, have revealed a possible problem with the assumption that changes in firing rates of POAH cells that correlate with changes in peripheral and/or POAH temperatures are reflecting thermointegrative processes (12, 13). As reviewed in those papers, urethan-anesthetized animals show electroencephalographic (EEG) changes similar to those seen in unanesthetized animals changing arousal states. In addition, EEG state changes in anesthetized animals can be induced by thermal and other stimuli. Units in many areas of the brain, including the hypothalamus, are EEG state selective: they change their firing rates with changes in the EEG (12, 20, 22, 23, 28, 30). Therefore, changes in POAH unit activity recorded in anesthetized preparations in response to thermal stimuli may be reflecting changes in cortical EEG state rather than thermoregulatory integrative activities. To control for this possible confounding variable, it is necessary to monitor EEG in single-unit studies, and this has not been done in most studies of the thermosensitivity and thermoresponsiveness of POAH cells.

If the POAH is a thermointegrative area, there should be cells in this area that change firing rate because of changes in local temperature, changes in peripheral temperature, and changes in both temperatures within EEG-defined arousal states. Although it has been shown that some hypothalamic units are locally thermosensitive within an arousal state (10, 11, 26, 27), there is no such unequivocal demonstration that POAH units respond specifically to peripheral thermal stimulation independently of EEG changes. Such a demonstration was the purpose of this study, and to that end we searched the POAH for cells responding to changes in both local and peripheral temperature in urethan-anesthetized rats while monitoring the EEG states of the animals.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals and surgical procedures. Experiments were conducted on 28 male Wistar rats weighing 250-350 g. Animals were housed in a temperature-controlled room (22-24°C) on a 12:12-h light-dark photoperiod. They had access to food and water ad libitum.

Each animal was anesthetized with 4% halothane (Halocarbon Laboratories) followed by an intraperitoneal injection of 1.0 g urethan/kg body wt. The anesthetic effect of the urethan injection lasted for the duration of the experiment. The animal was shaved, and the scalp and body trunk were treated with a depilatory cream. The animal was placed in a Kopf stereotaxic instrument. A midline incision was made, the skin was deflected, the top of skull was scraped, bleeding vessels were cauterized, and the skull surface was treated with hydrogen peroxide (3%). Five EEG electrodes were implanted into the skull. Four electrodes recorded frontal-occipital and occipito-occipital EEG activity, and one served as the common electrode for single-unit recordings. A pair of thermodes, each consisting of stainless steel concentric inner and outer cannulas (the outer one closed at the bottom) allowing one-way circulation of water, were implanted +1.5-2.0 mm from bregma, ±2.0 mm lateral of midline, and 9.0 mm deep. The thermodes and a thermocouple reentrant tube were glued into a Plexiglas block that was drilled to allow unidirectional perfusion of the thermodes. The block was anchored to the skull with bone screws and dental acrylic. The POAH was made accessible to electrode penetration through a 4.0-mm hole in the skull centered 1.0 mm off midline and 1.0 mm behind bregma, and the dura was removed. A reentrant tube was situated on the opposite side of midline so that the center of the recording area and the reentrant tube were equidistant from the thermodes.

Experimental protocol. Immediately after surgery, the animal, still in the stereotaxic device, was placed in the experimental setup. A thermocouple was inserted into the hypothalamic reentrant tube to measure T_POAH. An integrated T_s was obtained from six thermocouples, wired in parallel, attached to various skin areas about the animal's trunk. A thermocouple was inserted 2.0-3.0 cm into the rectum to monitor core temperature. The animal was wrapped in a water-perfused blanket to control skin and core temperatures. Because changes in scrotal temperature are known to have strong effects on the EEG (19, 20), we were careful that the water-perfused blanket did not extend to the scrotal area. The rate of change of both T_POAH and T_s was ~0.5-1.0°C/min. T_POAH was maintained at ~37-38°C during T_s manipulations, and T_s was maintained at ~36-38°C during T_POAH manipulations.

A glass microelectrode, previously pulled to a resistance of 10-15 M using a model P-87 Flaming Brown Micropipette Puller (Sutter Instrument) and filled with a 3.0 M NaCl solution, was stereotaxically advanced through the target area with the use of a Trent Wells hydraulic microdrive until a cell was detected. Single-unit firing detected by the electrode was passed to a Grass high-impedance probe. The signal was amplified and filtered by a Grass preamplifier and viewed on a Tektronix S113 oscilloscope. Action potentials with a signal-to-noise ratio greater than 3:1 were passed through a window discriminator (our design). The discriminator output was digitized and stored on a personal computer as firing rate frequency averaged over 10-s epochs. The raw single-unit firing data were also stored on polygraph paper. After a cell was discriminated it was recorded for up to 1 h while T_s and T_POAH were manipulated. The microelectrode was then advanced until another cell was located. A recording session lasted several hours, during which the microelectrode was moved to different POAH sites.

Data recording and EEG analysis. EEG, single-unit activity (SUA), and four temperatures were recorded continuously throughout the experiments. The EEG signal was recorded on paper by a Grass model 7 polygraph at a chart speed of 5 mm/s. EEG state analysis was done manually by scoring the polygraph records in 10-s epochs (13). In urethan-anesthetized rats, five EEG states can be distinguished as previously described (13). State 1 EEG is a low-amplitude, high-frequency pattern similar to the EEG of an awake animal. State 3 EEG is a high-amplitude, low-frequency pattern similar to the EEG of non-rapid eye movement sleep. State 2 EEG consists of rapid alternations between states 1 and 3 EEG patterns, and it can vary from being almost all state 1 to being almost all state 3. States 4 and 5 are mostly seen in animals that are hypothermic. State 5 EEG is characterized by alternating high spikes and very low amplitude waves. State 4, like state 2, is a transition state and consists of alternations between state 3 and state 5 activity in varying proportions.

Temperatures stored on the computer were T_POAH, T_s, rectal temperature, and the water-perfused blanket temperature. Inputs from thermocouples were processed by a custom-made signal conditioner, which converted the temperature signal to a voltage input for the computer. A time code generator was synchronized with the computer time clock, and 10-s intervals were recorded on the polygraph paper.

Histology. At the end of an experiment the animal was euthanized by an intracardial injection of a general anesthetic (50 mg/kg ketamine, Parke-Davis; 10 mg/kg acepromazine, Tech-American; and 5 mg/kg xylazine, Miles Laboratories). The brain was removed, quickly frozen in 2-methylbutane at 40°C, and mounted for sectioning. The brain was sectioned into 30-µm sections on a cryostat (Hacker Instruments), and the sections were dried and stained with cresyl violet. The anterior/posterior and lateral position of the electrode track was easily distinguishable from the slides. The recorded depth of the electrode, as determined stereotaxically during the recording session, was used to pinpoint the actual recording site.

Data analysis. One-way analysis of variance (ANOVA) was used to determine whether there was a significant (P  0.01) effect of EEG state on the firing rates of the individual POAH units. Scheffé's F test with an error rate set at P  0.01 was used following a significant overall ANOVA to make pairwise comparisons between EEG state groups.

To determine the T_POAH sensitivity of a cell, its thermal coefficient (T_c; impulses per second per degree Celsius) was determined by linear regression (frequency vs. T_POAH) over the temperature range of maximum slope (7, 8). A minimum temperature range of 2°C was used for calculating the slope. As in previous studies, a cell was classified as warm sensitive if the slope was +0.80 and cold sensitive if the slope was 0.60 (for review, see Ref. 5). All others were classified as insensitive. Responsivity of POAH cells to changes in T_s was also determined by linear regression (frequency vs. T_s), and the same criteria were used for classifying the cells as were used for T_POAH sensitivity.

In the cases of cells with T_c values that classified them as T_POAH sensitive or T_s responsive, further analyses were performed to determine whether each cell was T_POAH sensitive or T_s responsive independent of EEG changes. These additional analyses were threefold. 1) The T_c of each cell was determined within the states characterized by uniform EEG patterns (states 1 and 3; Ref. 13). If the T_c within EEG state 1 or 3 met the criteria as outlined above, the cell was classified as T_POAH sensitive or T_s responsive. However, if the T_c within both EEG states 1 and 3 did not meet with the above criteria, the cell was classified as appearing to be T_POAH sensitive or T_s responsive. EEG state 2 was not used in this analysis because it consists of varying percentages of state 1 and state 3 activity. States 4 and 5 were not used for this analysis because they are associated with body temperatures outside of a normal range (13). 2) For a cell to be classified as only appearing to be T_POAH sensitive or T_s responsive, the cell also had to show statistically significant changes in firing rate with changes in EEG according to criteria as stated above. 3) If a cell only appeared to be T_POAH sensitive or T_s responsive by the first two tests stated above, a third analysis was performed. This analysis was a determination of the effect of T_POAH or T_s on the EEG state of the animal. It was done by a one-way ANOVA for each cell to determine whether there was a significant (P  0.01) effect of T_POAH or T_s on the EEG state of the animal for data used to determine the T_c of the cell. Scheffé's F test with an error rate set at P  0.01 was used following a significant overall ANOVA to make pairwise comparisons between EEG state groups. Although a positive correlation was not considered necessary for classification of a cell as only appearing to be T_POAH sensitive or T_s responsive, it was considered to be further evidence that the changes in EEG were brought about by the changes in either T_POAH or T_s and that the firing rate of the cell was primarily reflecting these EEG changes rather than a thermoregulatory process.

Because of the nature and results of the study, it was deemed necessary to compare our data with data from previously published studies. The question to be asked was whether our data set was equivalent to the data sets on which previous studies based their interpretations. To make this possible, we analyzed our data using the most stringent criteria used in previous studies (5). Statistical comparisons of the numbers of cells within each classification group in the present study with those from previous studies were done by a power analysis using confidence intervals of 99%. This resulted in the ability to determine whether or not the numbers of cells within each classification group in the present study were significantly different (P  0.01) from those of previous studies.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

A total of 66 cells from 28 urethan-anesthetized rats was characterized in terms of responsivity to changes in EEG state and T_s and sensitivity to T_POAH. Figure 1 summarizes the anatomic distribution of these cells. Most of the cells in this study were recorded in the lateral and medial preoptic areas.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Anatomic distribution of cells recorded. Dots indicate approximate locations of cells recorded. AC, anterior commissure; AHA, anterior hypothalamic area; AVPO, anterioventral preoptic nucleus; BST, bed nucleus of the stria terminalis; HLDBB, horizontal limb of the diagonal band of Broca; LAH, lateral anterior hypothalamic nucleus; LHA, lateral hypothalamic area; LPOA, lateral preoptic area; MPOA, medial preoptic area; MPON, medial preoptic nucleus; PVN, paraventricular nucleus; StH, striohypothalamic nucleus; 3V, third ventricle; ZI, zona incerta.

SUA responses to EEG states. Changes in EEG state were elicited while recording from 55 of the 66 cells. Of these 55 cells tested, 41 (75%) showed significant (P  0.01) changes in firing rate with changes in the cortical EEG state of the animal.

SUA responses to T_POAH manipulations. T_POAH was manipulated while recordings from 27 of the 66 cells were made. Of the 27 cells tested, 5 cells (18%) were warm sensitive, 1 (4%) was cold sensitive, and 21 (78%) were temperature insensitive. The warm- and cold-sensitive cells were thermosensitive within EEG states 1 and/or 3. Figure 2 shows one cell (cell 1911) that had T_c values +0.80 for all EEG states during which it was recorded. Because these EEG states include 1 and 3, this cell was classified as T_POAH sensitive. In addition, this cell did not show significant changes in firing rate with changes in EEG states, supporting the conclusion that this cell was indeed responding to T_POAH, not to changes in EEG states brought about by changes in T_POAH.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Preoptic anterior hypothalamus (POAH) warm-sensitive cell (cell 1911). Each point is average firing rate of the cell over a 10-s epoch. Symbols denote electroencephalographic (EEG) state of the animal during each 10-s recording epoch. (For details of analysis see Table 1.)

Six of the 21 temperature-insensitive cells appeared to be T_POAH sensitive if EEG state changes were ignored. Five appeared warm sensitive, and one appeared cold sensitive. Within EEG states 1 and 3, all of these cells were insensitive to changes in T_POAH or there were not enough data to make that determination (either the cell was not recorded during EEG state 1 and/or 3, or data within state 1 and/or 3 did not span 2°C). It is important to note that whenever state 1 and/or 3 T_c determinations were not available, apparent thermosensitivity was dependent on data from states 2 and/or 4. Because states 2 and 4 consist of continuous, temperature-dependent mixtures of EEG activity characteristic of the other states, the assumption of EEG dependence of the apparent thermosensitivity was warranted. For example, Fig. 3 presents data from a cell that appeared to be warm sensitive to T_POAH. This apparent warm sensitivity was due to T_c values +0.80 during EEG states 2 and 4. This cell had significantly different firing rates in states 1 and 3 at a T_POAH of ~38°C. Because state 2 in this cell consisted of predominantly state 3 activity at the highest T_POAH and mostly state 1 activity at the lowest T_POAH, the cell appeared to be thermosensitive in state 2. Similarly, state 4 in this cell consisted predominantly of state 3 activity at 38°C and the amount of state 3 activity declined with temperature. The conclusion was that this cell was responding to EEG changes and was not intrinsically thermosensitive.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   POAH cell (cell 612) that appears warm sensitive. Each point is average firing rate of the cell over a 10-s epoch. Symbols denote EEG state of the animal during each 10-s recording epoch. (For details of analysis see Table 1.)

Table 1 shows the data for the 12 cells that were or appeared to be T_POAH sensitive (had T_c values of 0.60 or +0.80 for T_POAH). This table gives the overall T_c for the cell, the T_c for the cell in each EEG state during which it was recorded, the EEG responsivity, the influence of T_POAH on EEG state for each cell, and the resulting classification of each cell. Because the T_c for each cell was determined over the temperature range of maximum slope, the T_POAH in each EEG state was determined for those data used to determine the T_c. The EEG state responsivity, on the other hand, was determined for the entire time the cell was being recorded.

Closer inspection of Table 1 shows that our methodology not only enabled us to detect cells that appeared to be T_POAH sensitive when they were actually EEG state responsive (e.g., cell 612, Fig. 3), but we were also able to detect T_POAH-sensitive cells that would not have been so classified if data from all EEG states were combined. One such cell was cell 1041. The overall T_c of this cell was 0.18. However, in state 1 its T_c was +0.92, and within this state it was clearly T_POAH warm sensitive.

SUA responses to T_s manipulations. T_s was manipulated while recordings from 58 of the 66 POAH were made. When EEG was not taken into account, 11 of these 58 cells (19%) appeared to be warm responsive, 6 (10%) appeared to be cold responsive, 2 (3%) appeared to be both warm and cold responsive, and 39 appeared nonresponsive. However, when EEG was taken into account, none of these cells were thermoresponsive in the uniform EEG states 1 or 3. All of the cells that appeared to be responsive to T_s were also responsive (P  0.01) to EEG state changes, possibly indicating that these cells could have been responding not to T_s, but to the EEG state changes (Table 2). Most of these cells also showed significant effects of T_s on EEG state (Table 2). There were no cells recorded in this study that responded to changes in T_s without a concomitant change in EEG.

Figure 4 shows data for one cell that appeared to be warm responsive to T_s. During the recording of this cell, the EEG state of the animal was strongly related to T_s. The firing rate of the cell was significantly higher in state 3 than in state 1. Once again, state 2 consists of a temperature-dependent mixture of state 1 and state 3 activity, with state 3 activity predominant at the highest temperature and state 1 activity predominant at the lowest temperature. Therefore, the apparent thermoresponsiveness of this EEG state-selective cell is due to the temperature dependence of the EEG state of the animal.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   POAH cell (cell 3031) that appears responsive to increases in skin temperature (Ts). Each point is average firing rate of the cell over a 10-s epoch. Symbols denote EEG state of the animal during each 10-s recording epoch. (For details of analysis see Table 2.)

Figure 5 shows an interesting case in which a cell could appear to be both warm sensitive and cold sensitive. Cell 1044 had a higher firing rate in state 1 than in state 3. State 1 could be stimulated by either high or low T_s. Thus, in Fig. 5A, a 2.5-min segment of recording begins with a neutral blanket and skin temperature, and the EEG is state 2 dominated by state 3-like activity. In response to the lowering of blanket temperature, T_s falls and the EEG shows an increasing amount of state 1-like activity until it is finally all state 1. The firing rate of the cell increases as T_s falls and EEG state 1 activity increases. Thus the cell appears to be cold responsive with a T_c of 4.4 (Table 2). In contrast, Fig. 5B is a 2.0-min recording of the same cell, which begins with blanket and skin temperatures at high levels. Under these thermal conditions the animal is in EEG state 1 and the cell has a high firing rate. As blanket temperature is returned to a neutral temperature, T_s falls, and the EEG progresses from state 1 to state 2 to state 3, and the firing rate of the cell declines. Now the cell appears to be warm responsive with a T_c of 9.24 (Table 2).

View larger version (62K): [in this window] [in a new window]   Fig. 5.   Recordings of firing of a POAH cell (cell 1044) and the simultaneous recordings of the EEG, Ts (in °C), and water-perfused blanket temperature (Tbl; in °C). A: 2.5-min segment of the recording during which Tbl and Ts began at a neutral level and were lowered. As Ts falls, EEG state changes from being mostly state 3-like activity to being all state 1 activity. Firing rate of this cell was higher in state 1 than in state 3, and therefore it appears to be cold sensitive. B: 2.0-min segment of the recording during which Tbl and Ts begin at a high level and are lowered to neutral. At the beginning of this recording the EEG state is 1 and the cell has a high firing rate. As temperatures fall, EEG state changes through 2 to 3 and the cell has a lower firing rate. During this segment the cell, therefore, appears to be warm sensitive. F-F, frontal-frontal; F-O, frontal-occipital; SUA, single-unit activity.

Table 2 shows the data for all 19 of the 58 cells that had T_c values of 0.60 or +0.80 for T_s. This table gives the overall T_c for each cell, the T_c for each cell in each EEG state during which it was recorded, the EEG responsivity, the influence of T_s on EEG state for each cell, and the resulting classification of each cell. Because the T_c for each cell was determined over the temperature range of maximum slope, the T_s in each EEG state was determined for those data used to determine the T_c. The EEG state responsivity, on the other hand, was determined for the entire time the cell was being recorded.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

If the POAH is the thermointegrative center of the brain, it should be possible to record from POAH cells that respond to changes in both T_POAH and T_s. A number of investigators have undertaken such studies in urethan-anesthetized animals, and these studies have yielded results that seem to support the basic premise that peripheral temperature information is integrated in the POAH (4, 17, 18, 31, 34; see Ref. 3 for a review). There was an unanticipated confounding variable, however, which undermines interpretations of the data obtained in these excellent studies. This confound has three components: 1) urethan-anesthetized rats show EEG state changes, some of which are similar to those which characterize changes in arousal states in unanesthetized animals (12, 13, 19, 20, 22-24, 28, 30); 2) these EEG state changes can be spontaneous (22-24, 30), induced by thermal stimulation (12, 19, 20, 22) or a variety of other sensory modalities (22, 24, 28, 30); and 3) most (50-76%) POAH neurons are arousal state selective and vary their firing rates with changes in EEG activity (12, 13, 20, 22, 23, 28, 30). Therefore, unless EEG state is monitored, it is not possible to determine whether or not POAH SUA reflects a specific stimulus modality independently of changes in EEG activity, which in turn are driven by the applied stimulus. Therefore, our goal was to test whether there were POAH cells that were thermosensitive and/or thermoresponsive independent of EEG states. Such cells would be candidates for playing roles in thermoregulatory integration.

Our results are in agreement with previous studies on two counts. First, there are POAH cells that have firing rates that are a function of EEG state. Of the 55 POAH cells we recorded in different EEG states, 75% were EEG state selective. This is within the range (50 76%) of, and is not significantly different from, the numbers of EEG state-responsive neurons recorded in previous studies (12, 19, 20, 22-24, 28-30). Second, our study also shows POAH neurons that are sensitive to local temperature changes within uniform, EEG-defined arousal states. Previous studies show a range of 8-70% of POAH neurons as T_POAH sensitive, depending on the arousal state of the animal, with more neurons temperature sensitive during wake than non-rapid eye movement sleep (10, 11, 26, 27). The number of T_POAH-sensitive cells found in this study (22%) is not significantly different from the numbers of T_POAH-sensitive neurons recorded in previous studies.

The present study is not in agreement with previous studies that reported T_s-responsive cells in the POAH. These various studies on a number of species using different methods of thermal stimulation have found an average of 22% of the POAH cells responsive to changes in peripheral temperature, with a range of 3-39% (for a review, see Ref. 3). In the present study done on rats, using a water-perfused blanket for manipulating T_s, 33% of the POAH cells recorded appeared to be T_s responsive if EEG state changes were ignored. This 33% is not significantly different from the 22% average in previous reports. However, of the 19 cells recorded in this study that appeared to respond to changes in T_s with changes in firing rate, none responded to changes in T_s independently of EEG state changes. All of these cells were EEG state selective, and none showed T_c values of 0.60 or +0.80 within the uniform EEG states 1 and 3. Most showed significant effects of T_s on EEG state, which indicates that the changes in T_s determined the EEG state, which in turn determined the firing rates of the cells. Although all of these cells showed significant differences in firing rates with a change from at least one EEG state to another, in some cases the mean T_s between the corresponding EEG states were not significantly different (e.g., cell 2473), indicating again that firing rate was correlated with EEG state changes rather than changes in T_s. Therefore, there were no POAH cells found in the present study that were unequivocally T_s responsive.

There is the possibility that cells that respond to T_s were simply missed in this study. However, statistical analysis showed that enough cells were recorded (given the average number of purported T_s-responsive cells found in previous studies), so that there is only a 1% chance that actual T_s-responsive cells were missed. It is also possible that some of the cells we recorded may have been found to be T_s responsive in EEG states 1 or 3 if we had been able to manipulate T_s over a broader range without disrupting the state. It is very difficult to change T_s of an anesthetized animal significantly without producing EEG state changes. Presumably, this would have been true of previous studies as well. Apparent thermoresponsiveness was associated in virtually all cases with EEG state 2 or 4. These states consist of temperature-dependent, continuously variable proportions of states 1 and 3 or states 3 and 5 EEG activity, respectively. Close inspection of the EEG and unit activity recordings reveal a very close association of moment-to-moment changes in EEG activity and cell firing rates, indicating that the primary effect of T_s in these experiments was to alter EEG activity, which in turn determined firing rates.

An important conclusion from this study is that it is essential to monitor EEG when recording the activity of neurons in the POAH to determine their responsiveness to various modalities of stimulation. This is not a new conclusion; the point was made 30 years ago. In 1967, while investigating the effects of progesterone on neural activity, Komisaruk and co-workers (22) wrote, "...we were impressed by the striking temporal correlation between changes in activity of single neurons and alterations in the arousal level of the cortical EEG. Since elevated activity of the majority of the neurons we observed was closely correlated with cortical arousal, it was imperative to distinguish the effects of progesterone that might be produced indirectly by an induced change in arousal from the effects on particular neurons independent of changes in brain arousal." This conclusion has been reiterated by other investigators (10, 23, 24).

Other studies of thermoregulatory integration have also called attention to the EEG state selectivity of neurons as being a confounding variable that has led to probably false conclusions that cells are involved in thermoregulatory processing of information. Grahn and colleagues (12, 13) investigated the thermoresponsiveness of cells in the rostral ventromedial medulla and in the subceruleus area that were purported to be involved in thermoafferent information processing. The conclusion of those studies was that virtually none of these cells responded to T_s within an EEG-defined state (one rostral ventromedial medulla cell was thermoresponsive without a change in EEG activity). Rather, most cells were arousal state selective, and thermal stimulation of the skin altered arousal states. Kanosue and colleagues (19, 20) examined the responses of diencephalic cells to thermal stimulation of the scrotum, and they also concluded that the cell responses were not specific to the thermal stimulus but reflected EEG activation caused by the thermal stimulation as well as other modalities of stimulation.

Perspectives

Yet, we know that peripheral temperature information is used in thermoregulation. The threshold T_POAH for activating thermoregulatory responses shift in response to changes in T_s. This observation, however, involves thermal stimulation in the POAH and in the periphery while systemic thermoregulatory responses are measured. Thus the integration of central and peripheral temperature information could occur at any location between the POAH and the ultimate motor output neuron.

Our interpretation is supported by a number of experiments that have examined thermoregulatory responses to T_s after partial or complete transections of the spinal cord. Many older studies showed that after complete transection of the spinal cord at the cervical or thoracic level, mammals regain the ability to respond to cooling of the skin with vasoconstriction and shivering (for a review, see Ref. 33). Clinical observations of paraplegic patients (9) report that spinal cord injuries do not prevent thermoregulatory responses (vasodilation and sweating) in areas innervated by the spinal cord below the level of the lesion. Cats (1, 6), monkeys (25), and rats (2) have been shown to regain the ability to respond to peripheral cooling eliminated by high-level transection (at the level of the superior colliculus to the mammillary bodies) by subsequent low-level transection (at the inferior colliculus to the lower one-third of the pons) in the same animal. Although these responses are much reduced in comparison with intact animals and are not sufficient to maintain body temperature, it is important to recognize that all descending paths in these preparations have been severed, and even if those paths are not directly involved in temperature regulation, they may play a role in modulating the general level of excitability of spinal circuits. In a difficult series of experiments, Klussmann (21) selectively cut the ventrolateral funiculi of the spinal cord at the cervical level in rabbits and demonstrated no deficiencies in metabolic heat production responses to skin cooling. Because most ascending thermosensitive fibers are found in the ventrolateral funiculi, it must be concluded that the peripheral thermoafferent information need not be communicated to the hypothalamus to stimulate thermoregulatory responses.

The results reported in this paper, combined with the results of spinal transection studies, support a model of the thermoregulatory system in which hypothalamic thermosensitivity results in descending commands that modulate communication between peripheral thermosensors and thermoregulatory effectors at lower levels on the neural axis. This conclusion supports the earlier view of Satinoff (32) who wrote, "I suggest that the hypothalamus is not the sole integrator of body temperature. Rather, it is the most important among many in that it coordinates the activity of other integrating mechanisms at lower levels of the neuroaxis."