The perceptions of the thirst for water or a hunger for specific minerals or macronutrients that result from homeostatic deficits are the result of unique patterns of neural activity in the central nervous system (CNS). A useful heuristic for understanding the biological underpinnings of states of thirst and various hungers is to consider them to be similar to the perceptions generated by sensory modalities such as vision, audition, olfaction, taste, or the somatic senses (i.e., the classic senses). The neurobiology underlying thirst and related appetites is similar to that of many sensory systems. However, unlike primary sensory systems, thirst and specific hungers do not employ unique sensory organs (eyes, ears, tongue, nose, and skin) or a single type of dedicated receptor (rods, cones, taste buds, hair cells) sensing the information necessary to generate the central states associated with the respective perception. To produce the neural states associated with thirst or the hunger for a particular substance, the brain receives inputs from multiple sources that are related to the physiological status of the body. This information is conveyed to the brain in the form of various chemical and neural signals over multiple modalities or input pathways and then is processed by the brain in an extensive neural network. In effect, thirst and specific hungers are the synthetic products of the CNS-they are created in the brain.
This review will discuss the biological nature of the psychological states of thirst and salt appetite from the perspective of the signals that generate them and the brain neuronal and chemical systems responsible for creating them. By considering thirst and similar motivated states from the perspective of sensory and perceptual systems, it is easier to appreciate that these behavioral controls of fluid homeostasis operate under unique sets of biological constraints. Understanding the sensory and perceptual neurobiology of thirst and specific hungers allows an appreciation of the probable limits of these processes and the restrictions they have in restoring homeostasis under conditions of environmental and physical challenges.
TOTAL BODY WATER AND BODY FLUID COMPARTMENTS
Fluid homeostasis requires mechanisms that maintain overall sodium and water balance as well as the appropriate distributions of these substances within the intracellular (ICF) and extracellular fluid (ECF) compartments of the body. Consistency of the fluid matrix of organisms is achieved by multiple control systems with physiological (i.e., autonomic nervous system; hormonal) and behavioral effectors that are coordinated by the CNS. The level of activation of each of these mediators participates in the process of maintaining body fluid homeostasis and the patterned output of each response system is adjusted by the brain. When separate sources of sodium and water are available, it is possible to establish rates of sodium and water intake and loss that are largely independent of one another.
Although the importance of water for maintaining fluid homeostasis seems obvious, the contribution of sodium (or related extracellular ions and molecules) to overall fluid balance and distribution in the body may not be so apparent. Sodium ions do not readily cross cell membranes and are the key to determining the partitioning of water between the ICF and the ECF compartments. That is, the sodium concentration of the ECF, in large part, determines the osmotic gradients and forces that move water in and out of cells.
Animals drink water and ingest impermeable ions and molecules intermittently. Between the postabsorptive phase, a time when the hydromineral milieu is theoretically closest to balance and the conclusion of the next drinking episode, physiological and behavioral control systems work together to optimize the distribution of available fluid resources and to restore lost body water and sodium. When animals are challenged by physiological (e.g., exercise), environmental (e.g., increased ambient temperature), or pathophysiological (e.g., emesis; diarrhea) conditions, afferent, neural, and humoral signals directed to the CNS are generated and serve as independent indices reflecting the degree of fluid loss from the ICF and ECF compartments (Fig. 1).
Body fluid balance is determined by interacting physiological and behavioral effector systems. Behavioral mechanisms participate in the intake of sodium and water, and hormonal and autonomic mediators establish the rate of loss of these substances from the body. Both autonomic and endocrine mechanisms target the kidney and the sweat and salivary glands to adjust the rates of sodium and water loss from the body. In the kidney, the antidiuretic hormone, vasopressin, is a well-defined endocrine factor acting on renal collecting ducts to increase water permeability and water reabsorption. Aldosterone, a mineralocorticoid, acts on renal tubules to promote sodium reabsorption. These actions of hormones are complemented by renal tubule sympathetic innervation that, when activated, increases the reabsorption of sodium and water.
Hormonal and autonomic mechanisms act on the kidneys and on sweat and salivary glands to reduce the rate of sodium and water loss in the face of accruing fluid deficits. However, despite renal- and endocrine-targeted defenses, it is only by engaging motivational mechanisms for the acquisition and consumption of water (thirst) and of sodium (sodium or salt appetite) that restoration of sodium and water lost to the environment can be attained.
SIGNALS REFLECTING DEHYDRATION OF BODY FLUID COMPARTMENTS GENERATE THIRST AND SODIUM APPETITE
In classic studies in dogs conducted in the 1930s, Gilman (34) demonstrated that intravenous infusions of hypertonic saline, but not hypertonic urea, induce thirst. These findings indicated that substances excluded from cells generate an effective osmotic pressure and, thereby, produce drinking for the simple reason that they dehydrate the ICF compartment. The additional discovery that water intake was evoked after experimental procedures in animals that allowed investigators to selectively deplete either ICF or ECF led to the formulation of the so-called double depletion hypothesis of thirst in the early 1970s (23,31).
Hypovolemia, the depletion of ECF produced by experimental procedures that generate a localized edema or remove isotonic fluid from the body (e.g., hemorrhage or intraperitoneal dialysis), is also an effective condition for inducing thirst. For example, Fitzsimons (30) produced drinking in rats by intraperitoneal injections of large molecular weight substances (hyperoncotic colloids, such as polyethylene glycol and gum acacia) that reverse the normal Starling forces present at the capillaries. Increased oncotic pressure in the extracellular space at the injection site of such colloids progressively sequesters isotonic ECF to dehydrate interstitial fluid and reduce blood volume. With hemorrhage, intraperitoneal dialysis or hyperoncotic colloid-induced dehydration, the osmotic equilibrium across cell membranes is not affected. Therefore there is no movement of water out of the cellular compartment and hence no reduction of ICF.
Although thirst can be induced by selective dehydration of either ECF or ICF (i.e., the double depletion hypothesis), under the majority of physiological conditions thirst arises as a result of a simultaneous depletion of the two fluid compartments. Conjoint dehydration of the ICF and ECF by coadministered hypertonic saline and polyethylene glycol produces a total water intake that equals the sum of what either dipsogenic treatment would produce when administered independently (13).
Under physiological or pathophysiological conditions, various hydration challenges generate different degrees of dehydration of the ICF and ECF compartments. For example, transpiration of water through the skin or loss by evaporation from the respiratory system or oral mucosa results in a depletion of only water and, therefore, a pure dehydration of the ICF compartment; loss of hypotonic fluid as sweat or saliva produces loss of both ECF and ICF; fluid depletion as a result of bleeding (e.g., hemorrhage due to tissue damage; menstruation) produces a selective ECF loss (i.e., hypovolemia).
Disproportionate dehydration in one body compartment can become translated into a loss from the other compartment. For example, an initial loss of hypotonic fluid results in a relative hypernatremia that, in turn, dehydrates the intracellular compartment. Ultimately, further changes to apportion the distribution of dehydration to each of the body fluid compartments can occur as a result of renal mechanisms, because the kidney has the capacity to excrete either hypotonic or hypertonic urine.
Total water loss can be differentially shared by the ICF or ECF compartments over time and under different physiological tasks or environmental conditions. Classic demonstrations of thirst being maintained after the administration of hypertonic saline while water was withheld were made by Adolph and colleagues (2) and Holmes and Gregersen (36,37). After delays of 3-8 h, rats and dogs drank amounts of water that were similar to those under experimental conditions where they were given immediate access to water. During the time when water was not accessible, sufficient amounts of solute were excreted so that plasma osmolarity was nearly normalized. In other words, by excreting Na+, the animals exchanged the prevailing stimulus for thirst, an ICF depletion, for an ECF dehydration.
Not only are the processes underlying water intake induced by depletion of ICF versus ECF compartments different, so are the mechanisms of satiety. Fitzsimons (29) demonstrated that nephrectomized rats infused with effective osmotic agents drink sufficient water to restore body fluids to isotonicity. Anephric rats that cannot use renal mechanisms to excrete Na+ behave as near-perfect osmometers. In contrast, animals with pure hypovolemia do not restore fluid balance by ingesting only water. Studying rats made hypovolemic by subcutaneous administration of polyethylene glycol, Stricker (68,69) demonstrated that rats do not maintain high rates of water intake in the face of persisting hypovolemia. That is, water intake slows beginning 6-9 h after polyethylene glycol-induced hypovolemia. However, if the hypovolemic animals are given access to isotonic saline rather than water, they continue to ingest the salty fluid at high rates (69). This indicates that sodium, in addition to water, is required to correct hypovolemia.
In fact, Stricker's experiments (68,69) model what had been described in man by Adolph (1) as voluntary dehydration. Later, the term involuntary dehydration was suggested to be more appropriate to describe this phenomenon (35). Dehydrated humans with access only to water stop drinking before the fluid deficit is repaired. Hypovolemic rats will not only restore fluid balance by ingesting isotonic saline but will consume nonpreferred, unpalatable concentrations of hypertonic saline if they are offered in conjunction with water (68,70,71) (Fig. 2). A significant increase in the intake of a concentrated, normally rejected or unpreferred solution of NaCl is a commonly employed operational definition of sodium appetite (i.e., salt appetite).
The experimental study of sodium appetite dates from Curt Richter's (58) pioneering demonstration that adrenalectomy in the rat results in the ingestion of significant volumes of sodium solution. Adrenalectomy removes the body's primary source of aldosterone and therefore causes a loss of sodium in the urine, saliva, and feces. This absence of sodium-retaining hormone causes compensatory sodium ingestion to replace the lost sodium. Numerous other experimental manipulations have been shown to enhance the ingestion of concentrated solutions of NaCl in animals. These treatments range from pharmacological doses of aldosterone to the induction of sodium depletion by any of several methods including dietary restriction (Table 1). It is interesting to note that the majority of the experimental manipulations used to induce sodium appetite require substantial periods of time to elapse, often on the order of several hours to days, before a significant increase in concentrated NaCl becomes apparent. This is in contrast to the shorter latencies for the onset of thirst (i.e., significant water intake) after experimental depletion of the ICF or ECF compartments.
The first time animals are made sodium deficient, they will consume large quantities of NaCl on being given access to salty fluids even though they have had no previous experience with such solutions (25). Display of this appetite for sodium by naïve animals provides evidence that the appetitive and consummatory behaviors associated with sodium appetite are innate. However, this native behavior is also modifiable as a result of prior experience. A single sodium depletion, regardless of whether the animal does or does not orally consume sodium, produces an enhanced NaCl intake response on subsequent tests in response to sodium depletion (26,62). The phenomenon of enhanced NaCl intake after an initial treatment can be mimicked by administration of exogenous angiotensin II (ANG II) and aldosterone (62).
Many of the treatments that induce sodium appetite also induce water intake; however, the converse does not hold. For example, depletion of ICF with hypertonic saline induces water intake but little, if any, intake of hypertonic NaCl (57). Such observations led Peck (57) to suggest that ICF depletion-induced water drinking and ECF depletion-induced water drinking and sodium ingestion may be the reflection of different types of thirst. In experimental animals, it is impossible to test whether a thirst induced by ICF depletion is qualitatively different than a thirst induced by ECF depletion. It is probably best to use objective operational definitions of (the state of) thirst as measured by a significant increase in water intake and (the state of) sodium appetite as indexed by a significant increase in hypertonic saline solution intake. Although in principle both thirst and sodium appetite can exist simultaneously, under common experimental conditions only one behavior can be expressed and measured at a time. The investigation of the patterns of water and NaCl intake after dehydration proves instructive and has increased the understanding of the neural and endocrine mechanisms that control the generation of the hydromineral-related ingestive behaviors involved in maintaining body fluid homeostasis.
Stricker and colleagues (70,71) characterized the time course of the onset of thirst (water intake) and sodium appetite (NaCl intake) after the induction of hypovolemia induced by subcutaneous polyethylene glycol treatment. The general result from such studies indicates that although significant increases in water intake are apparent within an hour or two after polyethylene glycol treatment, it usually takes on the order of 5-7 h before significant increases in the consumption of concentrated (i.e., 2 or 5%) NaCl solutions are observed (Fig. 2). The reasons for this relative delay in onset are probably because of the need for an increase in one or more facilitory factors or a decrease in inhibitory mechanisms. Such processes may be related to the degree or duration of hypovolemia, the consequences of first drinking water (i.e., to produce a decrease in osmolarity of ECF), increased hormonal levels (e.g., ANG II; aldosterone), the development of hypotension, or some combination of these factors (44,45,72).
After NaCl intake commences, animals alternate between drinking water and ingesting concentrated sodium solution (Fig. 2). By adjusting the volumes of water and hypertonic saline consumed over time, rats end up consuming the equivalent of a near-isotonic (0.9-1.2%) mixture (70). Thus, two distant behaviors driven by two different motivational states provide the optimal strategy for the hypovolemic animal to correct an ECF deficit. Sodium appetite as demonstrated by a significant increase in hypertonic NaCl solution intake has been demonstrated in many avian and mammalian species. This paradigm has not been applied to study the recovery from dehydration in humans.
The classical studies of human sodium deficiency were conducted by McCance (50), who used himself and three volunteers as subjects. Severe sodium deficits were induced by low-sodium diet and episodes of sweating. During the sodium-depleted state, which was maintained for 11 d, the subjects reported excessive fatigue, lethargy, and a general feeling of exhaustion. The sense of taste was generally affected. Food and cigarettes were described as tasteless. The subjects differed in their desire for sodium. One subject described a distinct longing for salt and often went to sleep thinking about it, but McCance reported having no specific craving for sodium.
More recently, better-controlled experiments employing more objective methods have been used to assess sodium appetite in humans after sodium depletion. Beauchamp and colleagues (6) studied the effects of a very-low-sodium diet and diuretic treatment on taste in normal subjects for a 10 d period. As assessed after 5 and 10 d of depletion, thresholds for the taste of salt decreased in the majority of subjects, and the preference judgments for salt in foods tended to be greater. It is interesting to note that during the depletion period, despite increased plasma aldosterone and renin activity, supine and upright blood pressure were slightly decreased. Together, the data are interpreted as indicating that sodium depletion in humans produces moderate sensory (salt taste) changes and increased preference for salty foods.
The time course and development of sodium appetite in humans rehydrating without sodium replacement has been studied by Takamata and colleagues (74). These investigators induced H2O and Na+ depletion by light exercise (eight bouts during 7 h), which was followed by a rehydration period providing only ad libitum water. After a delay of several hours, the subjects began to express a significant increase in palatability ratings for hypertonic saline solutions. The increased acceptance over the hypertensive range of solutions became most apparent 17-23 h after the start of rehydration (Fig. 3). It is unclear whether factors such as reduced plasma osmolarity or the presence of hypotension were responsible for increasing perceived sodium palatability during the course of rehydration.
VISCERAL AND BRAIN SENSORY RECEPTORS DETECT THE STATUS OF BODY FLUID BALANCE
Both chemo- and mechanoreceptors sense the status of body fluids. These receptors are located in the systemic viscera and in the brain. Increasing the osmolarity of brain ECF by injection of hypertonic cerebrospinal fluid into the cerebral ventricles activates vasopressin release, sympathetic outflow, and thirst. These physiological and behavioral responses are biologically consistent with one another and cooperatively produce an increase in total body water and a tendency to elevate blood pressure. Such responses have been proposed to be mediated by neuron-like cells, which sense their own volume and are referred to as osmoreceptors (80,81).
Although there are many osmosensitive regions in the CNS and body, including the hepatoportal region, one of the key areas housing osmoreceptors for thirst and vasopressin release lies within the periventricular tissue surrounding the anteroventral portion of the third cerebral ventricle (AV3V) (40,44,45). AV3V-associated structures that play particularly important roles in osmoreception and/or processing are the ventral median preoptic nucleus and the organum vasculosum of the lamina terminals (OVLT). The OVLT along with two other anatomically and functionally unique structures, the subfornical organ (SFO) and the area postrema (AP), are referred to as sensory circumventricular organs (43). Sensory circumventricular organs like the other similar structures are devoid of a blood-brain barrier, but whereas most of the other circumventricular organs function as sites of secretion, sensory circumventricular organs have been implicated as structures sensing humoral signals in blood or interstitial fluid. Cells with stretch-sensitive ion channels have been identified electrophysiologically within the OVLT (7) and are, therefore, primary candidates as osmoreceptors.
A second key humoral signal that acts on the brain to effect responses consistent with restoration or expansion of body fluids and elevation of blood pressure is circulating ANG II. Systemic administration of ANG II elevates arterial blood pressure, releases vasopressin, and increases the ingestion of water and sodium (44,45). These ingestive, pressor, and secretory responses are also produced by ANG II administered directly into the brain ventricles.
There are several regions in the CNS sensitive to ANG II. The SFO in particular has been implicated in the activation of thirst, salt appetite and vasopressin release by circulating ANG II (44,45). It has been hypothesized that the brain renin-angiotensin system (i.e., a system with its components synthesized de novo in the brain) is coupled with the systemic (i.e., renal) renin-angiotensin system through the SFO (38). Circulating angiotensin in the mode of a hormone acts on sensory circumventricular organs to activate brain neural pathways. Brain interneurons, in turn, release angiotensin from their axon terminals, which then functions like a traditional neurotransmitter (38,39,43).
Considerable attention has been paid to the contribution of brain steroid receptors in the generation of sodium appetite (24). Mineralocorticoids (e.g., aldosterone), when given in pharmacological doses (82) or in more physiological concentrations in conjunction with ANG II (32), produce significant NaCl ingestion. The facilitory interaction that can be achieved with systemic injections of very low doses of aldosterone and small amounts of ANG II injected into the brain has been referred to as the synergy hypothesis for sodium appetite (24).
Mechanoreceptors identified as key for sensing blood volume and arterial blood pressure are located in the periphery. The great veins and atria (i.e., the low-pressure side of the circulation) contain what are often referred to as cardiopulmonary receptors (a misleading terminology, because there are many different types of visceral sensory receptors associated with the heart and lungs) that are responsive to stretch. These sensors generate afferent nerve traffic to the CNS in proportion to distention of the walls of the great veins (e.g., superior vena cava) and atria. Similar stretch receptors located in the aortic arch and carotid sinus sense changes in arterial blood pressure. Output from these receptors change in proportion to arterial blood pressure.
Information from low- and high-pressure baroreceptors as well as from cardiac ventricular mechano-/chemoreceptors reaches the brain by the IXth and Xth cranial nerves. When blood pressure and/or blood volume falls below a "normal" set point, nerve activity declines and results in triggering the reflex release of vasopressin, increased sympathetic outflow, and, ultimately, thirst and sodium appetite. A moderate reduction in blood pressure (induced by antihypertensive agents) in the presence of hypovolemia synergizes to rapidly generate sodium appetite and thirst, which are dependent on a renin-angiotensin system that is intrinsic to the brain (76). Maintaining blood pressure or cutting neural afferents from high-pressure baroreceptors (sinoaortic denervation) blocks this type of experimentally induced sodium appetite (77). Blood pressure at hypertensive or normotensive levels is likely to inhibit thirst and sodium appetite (44,45,59).
A BRAIN NEURAL NETWORK PROCESSES FLUID BALANCE-RELATED INPUT TO MAINTAIN BODY FLUID HOMEOSTASIS
The generation of the appetitive and consummatory behaviors necessary to acquire and ingest water and sodium requires the integration of information derived from several types of sensory systems. Visceral afferent signals must communicate the body's need for water and sodium to the brain. In the process of seeking water and sodium, sensory systems are engaged for identification of potential sources in the external environment. When in contact with food and fluids, gustatory mechanisms assess the taste qualities of the ingesta to identify the commodity as an acceptable source of water or sodium for consumption.
The VIIth, IXth, and Xth cranial nerves carrying taste and visceral information enter the brain stem and synapse primarily in the nucleus of the tractus solitarius (NTS). Axons from cells in the NTS, in turn, project to other brain nuclei. Similarly, input detected by sensory circumventricular organs (SFO, OVLT, AP) is also carried into the brain by nerves emanating from these structures. Many of the brain pathways and regions carrying and processing input from the systemic chemo- and mechanoreceptors and from sensory circumventricular organs have been defined and characterized by anatomical, neurochemical, and functional techniques. The result has been the identification of a neural network that both carries and integrates information critical for the regulation of fluid balance. Pathways and information initially processed by the AP and NTS ascend the neuraxis to meet information deriving from the SFO and OVLT. Between the forebrain and hindbrain portals of input, there are several nodes (brain nuclei) where information from such various sources converges.
In principle, activity in any pathway and nucleus in the brain can influence information handling (processing and storage) within the neural network that controls the effectors maintaining body fluid and cardiovascular homeostasis. However, there are brain areas that have been found to be especially critical in these processes. Most notable among these brain regions are 1) the sensory circumventricular organs (i.e., SFO, OVLT, and AP), 2) the NTS, 3) caudal and rostral ventrolateral medulla, 4) parabrachial nucleus (PBN), 5) the median preoptic nucleus, 6) parvocellular hypothalamic paraventricular nucleus, 7) magnocellular paraventricular nucleus, 8) supraoptic nucleus, 9) amygdala (particularly the central and medial nuclei of the amygdala), 10) bed nucleus of the stria terminalis, 11) lateral hypothalamus, 12) ventrolateral medulla, and 13) the spinal cord intermediolateral cell column (42,44,45). Reciprocal pathways connect most of these structures with one another and provide feedforward and feedback neural circuits that are necessary for achieving the processing of information within the CNS. Information handling and processing within this hydromineral-neural network is necessary for both acute and chronic adjustments involved in regulating body fluid balance and distribution (e.g., hemodynamics). There are many neurotransmitter systems associated with this body fluid-related brain network. Two of the most important of the central neurochemical mediators are norepinephrine and angiotensin. Pathways containing these putative neurotransmitters are distributed throughout the CNS and are associated with most, if not all, of the key structures involved in fluid balance and cardiovascular control. Evidence indicates that these two brain neurochemical systems are important for mobilizing effector systems that expand ECF volume (39,44,45).
In addition to facilitory mechanisms, there are inhibitory counterparts represented in the central fluid regulatory network that retard overexpansion of the ECF compartment. Evidence indicates that such inhibition may involve brain oxytocin (72), serotonin (44,45,51-53), tachykinins (49), cholecystokinin (54), and bombesin (15,16).
In recent years, significant progress has been made in several laboratories in identifying forebrain and hindbrain components of the hydromineral neural network subserving thirst and sodium appetite as well as those for physiological control systems (i.e., sympathetic and endocrine). Systemically derived information has immediate access to the forebrain via actions on rostral sensory circumventricular organs associated with the lamina terminalis (Fig. 4). The lamina terminalis is comprised of a layer of ependymal cells that during early development covers the rostral end of the neural tube. In mature animals, this sheet of cells forms the rostral wall of the third ventricle, and four midline structures, the SFO, the median preoptic nucleus, the anterior commissure and the OVLT, lie adjacent to the lamina terminalis. The SFO is situated in the dorsal aspect of the third ventricle and the OVLT is located along the ventral portion of the lamina terminalis, both of these sense plasma ANG II levels and extracellular osmolarity.
Both the SFO and OVLT project to the median preoptic nucleus that lies between them and is located inside the blood-brain barrier. The median preoptic nucleus is an important integrative node, processing both angiotensin-derived and osmotic information. Neural activity in the median preoptic nucleus partly determines the degree and pattern of activation of the physiological and behavioral control systems for body fluid homeostasis (41). Numerous studies have implicated the SFO (66,75), the ventral median preoptic nucleus (27,28), and the combined ventral median preoptic nucleus and OVLT region in the control of thirst and sodium appetite (5,8,19,40). Several pathways descend from lamina terminalis structures to innervate key forebrain nuclei that have been implicated in thirst (e.g., lateral hypothalamic/perifornical area), sodium appetite (central nucleus of the amygdala), anterior pituitary hormone control (parvocellular paraventricular nucleus; PVN), posterior pituitary function (magnocellular PVN and supraoptic nuclei), and sympathetic outflow (parvocellular PVN) (Fig. 4).
In contrast to the forebrain facilitory actions of ANG II, hyperosmolarity, and mineralocorticoid binding, other signals act through a hindbrain system to keep water and sodium intake in check. This inhibitory system involves the neural circuitry of the AP, the medial portion of the NTS (AP/mNTS), and the lateral PBN (44,45).
Ablation of the AP/mNTS dramatically increases the intake of concentrated sodium chloride solution both in chronic (11) and acute (21) experimental tests. An important component of a neural pathway projecting from the AP/mNTS to the lateral PBN (14,64,79) contains serotonin (48). Bilateral injections of serotonin antagonists directly into the lateral PBN markedly enhance both thirst and sodium appetite induced by many dipsogenic and natriorexigenic challenges (44,45,53). Several recent studies have also implicated the actions of cholecystokinin, GABA, corticotropin-releasing hormone, and norepinephrine in the lateral PBN on sodium and water intake (3,9,17,18,54). In many respects the LPBN seems to be a switch to determine whether water intake versus sodium intake becomes the predominant behavior used to correct body fluid deficits.
The lateral PBN (Fig. 5) is an integrative region involved in processing viscerally derived inhibitory input. An example of the importance of the lateral PBN in the inhibition of behaviors related to volume expansion has recently been demonstrated. Inflating a small balloon at the junction of the vena cava and right atrium in rats inhibits thirst (46) and sodium appetite (78). The expansion of such a balloon in rats with lesions of the lateral PBN is no longer effective in inhibiting isoproterenol-induced drinking (56). This observation is consistent with other findings (22,54) indicating that the lateral PBN is a key brain region controlling behaviors that influence ECF volume. Information processed within the lateral PBN is carried deeper into the neural network for hydromineral balance where it interacts with other types of input, particularly those from sites receiving humorally derived signals from the lamina terminalis (Fig. 5).
Although much of the neural circuitry of thirst and sodium appetite is contained within the brain stem and limbic system, we probably become aware of our thirst and hunger for salty substances because of activation of parts of the cerebral cortex. This is a reasonable expectation (hypothesis) because the highest levels of information processing for perception in other sensory systems occur at various cortical loci. In recent studies using positron emission tomography, Derek Denton and his colleagues (20) have identified in human volunteers several brain regions that are activated in conjunction with a strong sensation of thirst produced by intravenous infusions of hypertonic saline. Among the brain regions showing changes in activity associated with the desire to drink were the anterior and posterior portions of the cingulate cortex. This is a brain region associated with motivational and affective processes.
EXERCISE, CONDITIONING, AND BIOLOGICAL CONSTRAINTS ON THIRST AND SODIUM APPETITE
There is a substantial body of experimental evidence indicating that maintaining adequate hydration is critical for health and well-being (4,63). The excessive loss of water and sodium in extreme environments or while engaging in athletics not only impairs performance but also increases the likelihood of thermal injury (73). Significant water and sodium losses can occur very quickly in both fit and untrained individuals, especially if exercising in a hot environment. The speed of onset of severe dehydration under extreme conditions can exceed the onset of thirst and sodium appetite. Thirst drive resulting from dehydration may occur on the order of minutes to tens of minutes, but the onset of sodium appetite is likely to require many hours before a significant hunger develops (70,74). Impaired thirst and sodium appetite mechanisms that accompany the normal aging process (47) are likely to only complicate this problem. Although behavioral mechanisms of drinking and salt ingestion can act in concert with the kidney to restore body fluid homeostasis over the long-term, the mobilization of thirst- and especially sodium appetite-related behaviors do not have sufficiently rapid response times to prevent immediate consequences of severe water and sodium losses.
Beyond concerns of the extended lag times for the onset of thirst and sodium appetite that constrain the efficiency in converting exercise-induced fluid deficits into corrective behaviors, there are exercise-induced fluid shifts and hemodynamic changes that are likely to also compromise the reliability of rehydration-related behaviors for deficit correction. The nature of circulatory changes and of fluid movement among lymphatic, intravascular and interstitial spaces, as well as the movement between ICF and ECF compartments (60) that accompany exercise, needs to be considered. Unfortunately, little is known about how changes in fluid distribution produced either as a result of chronic exercise and physical conditioning (e.g., expansion of blood volume (12)) or by acute exercise will alter afferent signaling mechanisms. At present, one can only speculate about the effects of exercise and conditioning on thirst and sodium appetite based on hydromineral-related hormonal or autonomic responses under such conditions. For example, Nose and colleagues (55) have demonstrated that during graded exercise the threshold for the release of vasopressin is increased. Therefore, one might also suspect that increased inhibition of water or sodium intake would accompany exercise. Perhaps increases in arterial pressure and venous return associated with exercise would contribute to such inhibition, but this has not been studied in relation to thirst or sodium appetite.
The degree of loading of cardiopulmonary baroreceptors have been demonstrated to affect both thirst and vasopressin secretion. Head-out water immersion produces a shift of blood volume toward the thorax stretching the large capacitance vessels and the atria produces a reduction of thirst, water intake, and vasopressin secretion in dehydrated subjects (61,67). Interestingly, this inhibitory effect on thirst and vasopressin release does not occur in older subjects (67). It has been suggested by Stachenfeld and colleagues (67) that the attenuated inhibitory mechanisms may be due to impaired cardiopulmonary reflexes (10). From this perspective, one might consider the possibility that thirst in athletes with cardiac hypertrophy and diminished cardiovascular reflex function (33) may have impaired fluid-related behavior and endocrine responses to altered stretch of vessels containing cardiopulmonary receptors.
Despite the limitations imposed by the sensory nature of thirst and sodium appetite, it is important to recognize that the cognitive capacities of humans provide a means of compensating for limitations inherent in intrinsic biological motivational processes. That is, cognitive mechanisms can activate behaviors to drink or consume salt and override an absence of thirst and sodium appetite. However, fluid intake resulting from only cognitive processes without the activation of thirst- or sodium appetite-related mechanisms may lead to inappropriate water and salt intake. In this context it is important to note that not only is hypohydration caused by inadequate thirst and sodium appetite a concern, but under conditions of extreme athletic challenges there is also the risk of cognitively overdriven fluid intake. The result can be exercise-induced hyponatremia, a dilutional hyponatremia associated with positive fluid balance. This disorder is produced in part by the excess secretion of antidiuretic hormone (i.e., arginine vasopressin (65)) and can be life threatening. Excessive ingestion of fluid prompted only by a cognitively based formula clearly carries a risk of untoward consequences produced by hyponatremia.
By viewing thirst and sodium appetite from the perspective of sensory psychobiology, it becomes easier to appreciate that there are biological limits on the functional capacities of motivational systems. That is, just as there are limitations in the classic senses in terms of thresholds for detection and discrimination, there are restrictions as to how fast and how reliably the receptors and neural systems subserving thirst and salt appetite can sense and respond to changes in hydrational status. Just as we employ microscopes, telescopes, pressure transducers, audio amplifiers, and other instrumentation to enhance and resolve physical stimuli, which normally serve to excite the classic senses, it is not unreasonable to expect that it may be necessary to employ instrumentation (scales, osmometers, sodium analyzers) to assess hydrational status under extreme and rapidly changing conditions.
SUMMARY AND CONCLUSIONS
Maintaining body fluid and cardiovascular homeostasis requires the integrative capacity of the CNS. A thirst for water or a hunger for sodium are perceptions that result from synthetic processes carried out in the brain. The CNS receives visceral and somatic sensory inputs and integrates this information. Through this process, calculated decisions are made by the brain and the best pattern of reflexes and behaviors are generated to optimize available fluid distribution between body compartments and to restore hydromineral balance. The key behaviors that are necessary for the maintenance and recovery of normal fluid balance are those that lead to the acquisition and consumption of both water (thirst) and sodium (sodium appetite).
Figure 6 summarizes several of the major points identifying the afferent stimuli and central circuitry relevant to the behavioral controls of ECF volume. To the left of the box, one of the primary sites for osmoreception related to thirst and vasopressin release is identified as a component of the lamina terminalis. To the right and above the large box representing the brain are the neural and humoral inputs that communicate the status of body fluid and cardiovascular homeostasis to the CNS. Under a range of physiological conditions, circulating ANG II is a participant in the generation of thirst and sodium appetite and its actions are complemented by additional visceral afferent inputs. ANG II accesses the CNS through brain circumventricular organs, particularly the SFO.
Systemic input carried over nerves from the viscera originates in vascular baroreceptors. Hypovolemia and/or hypotension result in input to the CNS that enhances activity in central facilitory systems. Conversely, hypervolemia and/or hypertension activate inhibitory mechanisms. Other types of facilitory and inhibitory input from the viscerally derived gastrointestinal tract, splanchnic and hepatoportal circulation and the kidneys also are likely to feed into the neural network for fluid balance.
In the brain, angiotensin acting within the SFO has been proposed to activate several descending projections to key forebrain structures such as the median preoptic nucleus, hypothalamic perifornical area, central nucleus of the amygdala, supraoptic nucleus, and paraventricular nucleus. Some of the fibers from the SFO and other lamina terminalis structures are angiotensinergic. These pathways and their termination sites are associated with the control of the behaviors and reflexes that maintain body fluid balance and cardiovascular homeostasis (Fig. 4). Ascending facilitory and inhibitory inputs project into many of these same forebrain regions. For example, norepinephrine is released at key forebrain sites and acts to reinforce the actions of angiotensin to generate thirst and sodium appetite.
Several recent studies have described the role of inhibitory influences arising in the periphery and activating neural circuitry in the hindbrain. Central structures implicated in this inhibitory system are the AP/mNTS and the lateral PBN. The inhibitory actions of serotonin within the lateral PBN have been associated with a hindbrain projection from the AP/mNTS to the lateral PBN that has been proposed to protect from overexpansion of the ECF compartment.
The integration of information derived from facilitory and inhibitory humoral and visceral neural inputs is conducted within many different nodes (nuclei) in a sensory network that regulates hydromineral balance. The outputs of this network both controls motor-pattern generators responsible for the appetitive and consummatory behaviors and give rise to a sensory experience perceived as thirst and sodium appetite.
Taking into consideration the basic neurobiology of the brain systems maintaining fluid balance, it becomes apparent that generation of thirst or sodium appetite results from a set of neural computations based on a complex of facilitory and inhibitory afferent signals. Phenomena such as voluntary (or involuntary) dehydration arise from failure to consume adequate sodium in conjunction with water.
It is important to recognize that there are limits to the operating capabilities of the behavioral controls of body fluid homeostasis. Like any sensory system, there are thresholds for detection and discrimination of the pattern of stimuli that are perceived as thirst or sodium appetite. It should be recognized that cognitive instructions derived from the cortex can supplement the mechanisms of thirst and sodium appetite which arise from phylogenetically old parts of the nervous system. Appropriately informed, cognitive mechanisms can be useful in maintaining adequate hydration. However, the strength of cognitive mechanisms can also lead to either inappropriate hypo- or hyperhydration, conditions with their own attendant risks.
Under extreme conditions all sensory systems fail. Technology to enhance sensory capacity is appropriate in many circumstances. As more is learned about the nature of controls affecting thirst and sodium appetite and how these are influenced by the environment and exercise, a body of knowledge will be developed to provide the best strategies for maintenance and restoration of body fluid homeostasis for health and well-being.
The work was supported in part by grants from the National Heart, Lung, and Blood Institute HL 14388 and HL-57472, and from the National Institute of Diabetes and Digestive and Kidney Diseases DK-066086.
1. Adolph, E. F. Physiology of Man in the Desert
, New York, NY: Interscience Publishers, pp. 254-270, 1947.
2. Adolph, E. F., J. P. Barker, and P. A. Hoy. Multiple factors in thirst. Am. J. Physiol.
3. Andrade, C. A., S. P. Barbosa, L. A. De Luca JR., and J. V. Menani. Activation of alpha2-adrenergic receptors into the lateral parabrachial nucleus enhances NaCl intake in rats. Neuroscience
4. Barr, S. I. Effects of dehydration on exercise performance. Can. J. Appl. Physiol.
5. Bealer, S. L., and A. K. Johnson. Sodium consumption following lesions surrounding the anteroventral third ventricle. Brain Res. Bull.
6. Beauchamp, G. K., M. Bertino, D. Burke, and K. Engelman. Experimental sodium depletion and salt taste in normal human volunteers. Am. J. Clin. Nutr.
7. Bourque, C. W., S. H. Oliet, and D. Richard. Osmoreceptors, osmoreception, and osmoregulation. Front. Neuroendocrinol.
8. Buggy, J., and A. K. Johnson. Preoptic-hypothalamic periventricular lesions: thirst deficits and hypernatremia. Am. J. Physiol. Regul. Integr. Comp. Physiol.
9. Callera, J. C., L. B. Oliveira, S. P. Barbosa, D. S Colombari, L.A. De Luca JR., and J. V. Menani. GABA(A) receptor activation in the lateral parabrachial nucleus induces water and hypertonic NaCl intake. Neuroscience
10. Cleroux, J., C. Giannattasio, G. Bolla, et al. Decreased cardiopulmonary reflexes with aging in normotensive humans. Am. J. Physiol. Heart Circ. Physiol.
11. Contreras, R. J., and P. W. Stetson. Changes in salt intake lesions of the area postrema and the nucleus of the solitary tract in rats. Brain Res.
12. Convertino, V. A. Blood volume: its adaptation to endurance training. Med. Sci. Sports Exerc.
13. Corbit, J. D. Cellular dehydration and hypovolaemia are additive in producing thirst. Nature
14. Cunningham, E. T. JR., R. R. Miselis, and P. E. Sawchenko. The relationship of efferent projections from the area postrema to vagal motor and brain stem catecholamine-containing cell groups: an axonal transport and immunohistochemical study in the rat. Neuroscience
15. De Caro, G., C. Polidori, T. G. Beltz, and A. K. Johnson. Area postrema, lateral parabrachial nucleus, and the antinatriorexic effect of bombesin. Peptides
16. De Caro, G., C. Polidori, J. V. Menani, and A. K. Johnson. Bombesin affects the central nervous system to produce sodium intake inhibition in rats. Physiol. Behav.
17. De Castro e Silva, E., J. B. Fregoneze, and A. K. Johnson. Corticotropin-releasing hormone in the lateral parabrachial nucleus inhibits sodium appetite in rats. Am. J. Physiol. Regul. Integr. Comp. Physiol.
18. De Gobbi, J. I. F., L. A. De Luca JR., A. K. Johnson, and J. V. Menani. Interaction of serotonin and cholecystokinin in the lateral parabrachial nucleus to control sodium intake. Am. J. Physiol. Regul. Integr. Comp. Physiol.
19. De Luca, L. A. JR., O. Galaverna, J. Schulkin, S. Z. Yao, and A.N. Epstein. The anteroventral wall of the third ventricle and the angiotensinergic component of need-induced sodium intake in the rat. Brain Res. Bull.
20. Denton, D., R. Shade, F. Zamarippa, et al. Neuroimaging of genesis and satiation of thirst and an interoceptor-driven theory of origins of primary consciousness. Proc. Natl. Acad. Sci. U. S. A.
21. Edwards, G. L., T. G. Beltz, J. D. Power, and A. K. Johnson. Rapid-onset "need-free" sodium appetite after lesions of the dorsomedial medulla. Am. J. Physiol. Regul. Integr. Comp. Physiol.
22. Edwards, G. L., and A. K. Johnson. Enhanced drinking after excitotoxic lesions of the parabrachial nucleus in the rat. Am. J. Physiol. Regul. Integr. Comp. Physiol.
23. Epstein, A. N. Epilogue: retrospect and prognosis. In: The Neuropsychology of Thirst: New Findings and Advances in Concepts
, A. N. Epstein, H. R. Kissileff, and E. Stellar (Eds.). Washington, DC: V.H. Winston, pp. 315-332, 1973.
24. Epstein, A. N. Mineralocorticoids and cerebral angiotensin may act together to produce sodium appetite. Peptides
25. Epstein, A. N., and E. Stellar. The control of salt preference in the adrenalectomized rat. J. Comp. Physiol. Psychol.
26. Falk, J. L. Serial sodium depletion and NaCl solution. Physiol. Behav.
27. Fitts, D. A. Effects of lesions of the ventral ventral median preoptic nucleus or subfornical organ on drinking and salt appetite after deoxycorticosterone acetate or yohimbine. Behav. Neurosci.
28. Fitts, D. A., D. S. Tjepkes, and R. O. Bright. Salt appetite and lesions of the ventral part of the ventral median preoptic nucleus. Behav. Neurosci.
29. Fitzsimons, J. T. Drinking by nephrectomized rats injected with various substances. J. Physiol. (Lond.)
30. Fitzsimons, J. T. Drinking by rats depleted of body fluid without increase in osmotic pressure. J. Physiol. (Lond.)
31. Fitzsimons, J. T. Some historical perspectives in the physiology of thirst. In: The Neuropsychology of Thirst: New Findings and Advances in Concepts
, A. N. Epstein, H. R. Kissileff, and E. Stellar (Eds.). Washington, DC: V.H. Winston, pp. 3-33, 1973.
32. Fluharty, S. J., and A. N. Epstein. Sodium appetite elicited by intracerebroventricular infusion of angiotensin II in the rat: II. Synergistic interaction with systemic mineralocorticoids. Behav. Neurosci.
33. Giannattasio, C., G. Seravalle, G. B. Bolla, et al. Cardiopulmonary receptor reflexes in normotensive athletes with cardiac hypertrophy. Circulation
34. Gilman, A. The relation between blood osmotic pressure, fluid distribution and voluntary water intake. Am. J. Physiol.
35. Greenleaf, J. E. Problem: thirst, drinking behavior, and involuntary dehydration. Med. Sci. Sports Exerc.
36. Holmes, J. H., and M. I. Gregersen. Observations on drinking induced by hypertonic solutions. Am. J. Physiol.
37. Holmes, J. H., and M. I. Gregersen. Role of sodium and chloride in thirst. Am. J. Physiol.
38. Johnson, A. K. The periventricular anteroventral third ventricle (AV3V): its relationship with the subfornical organ and neural systems involved in maintaining body fluid homeostasis. Brain Res. Bull.
39. Johnson, A. K. Brain mechanisms in the control of body fluid homeostasis. In: Perspectives in Exercise Science and Sports Medicine, Volume 3: Fluid Homeostasis During Exercise
, C. V. Gisolfi and D. R. Lamb (Eds.). Carmel, IN: Benchmark Press, pp. 347-419, 1990.
40. Johnson, A. K., and J. Buggy. Periventricular preoptic-hypothalamus is vital for thirst and normal water economy. Am. J. Physiol. Regul. Integr. Comp. Physiol.
41. Johnson, A. K., J. T. Cunningham, and R. L. Thunhorst. Integrative role of the lamina terminalis in the regulation of cardiovascular and body fluid homeostasis. Clin. Exp. Pharmacol. Physiol.
42. Johnson, A. K., J. De Olmos, C. V. Pastuskovas, A. M. Zardetto-Smith, and L. Vivas. The extended amygdala and salt appetite. Ann. N. Y. Acad. Sci.
43. Johnson, A. K., and P. M. Gross. Sensory circumventricular organs and brain homeostatic pathways. FASEB J.
44. Johnson, A. K., and R. L. Thunhorst. The neuroendocrinology of thirst and salt appetite: visceral sensory signals and mechanisms of central integration. Front. Neuroendocrinol.
45. Johnson, A. K., and R. L. Thunhorst. The neuroendocrinology, neurochemistry and molecular biology of thirst and salt appetite. In: Handbook of Neurochemistry and Molecular Neurobiology: Behavioral Neurochemistry, Neuroendocrinology and Molecular Neurobiology
, 3rd ed., A. Lajtha and J. D. Blaustein (Eds.). New York, NY: Springer, pp. 641-688, 2007.
46. Kaufman, S. Role of right atrial receptors in the control of drinking in the rat. J. Physiol. (Lond.)
47. Kenney, W. L., and P. Chiu. Influence of age on thirst and fluid intake. Med. Sci. Sports Exerc.
48. Lança, A. J., and D. van der Kooy. A serotonin-containing pathway from the area postrema to the parabrachial nucleus in the rat. Neuroscience
49. Massi, M., L. Gentili, M. Perfumi, G. De Caro, and J. Schulkin. Inhibition of salt appetite in the rat following injection of tachykinins into the medial amygdala. Brain Res.
50. McCance, R. A. Medical problems in mineral metabolism: III. Experimental human salt deficiency. Lancet
51. Menani, J. V., L. A. De Luca JR., and A. K. Johnson. Lateral parabrachial nucleus serotonergic mechanisms and salt appetite induced by sodium depletion. Am. J. Physiol. Regul. Integr. Comp. Physiol.
52. Menani, J. V., L. A. De Luca JR., R. L. Thunhorst, and A. K. Johnson. Hindbrain serotonin and the rapid induction of sodium appetite. Am. J. Physiol. Regul. Integr. Comp. Physiol.
53. Menani, J. V., and A. K. Johnson. Lateral parabrachial serotonergic mechanisms: angiotensin-induced pressor and drinking responses. Am. J. Physiol. Regul. Integr. Comp. Physiol.
54. Menani, J. V., and A. K. Johnson. Cholecystokinin actions in the parabrachial nucleus: effects on thirst and salt appetite. Am. J. Physiol. Regul. Integr. Comp. Physiol.
55. Nose, H., A. Takamata, G. W. Mack, et al. Water and electrolyte balance in the vascular space during graded exercise in humans. J.Appl. Physiol.
56. Ohman, L. E., and A. K. Johnson. Role of lateral parabrachial nucleus in the inhibition of water intake produced by right atrial stretch. Brain Res.
57. Peck, J. W. Discussion: Thirst(s) resulting from bodily water imbalances. In: The Neuropsychology of Thirst: New Findings and Advances in Concepts
, A. N. Epstein, H. R. Kissileff, and E. Stellar (Eds.). Washington, DC: V.H. Winston & Sons, pp. 99-112, 1973.
58. Richter, C. P. Increased salt appetite in adrenalectomized rats. Am. J. Physiol.
59. Robinson, M. M., and M. D. Evered. Pressor action of intravenous angiotensin II reduces drinking response in rats. Am. J. Physiol. Regul. Integr. Comp. Physiol.
60. Rowell, L. B. Human Cardiovascular Control
. New York, NY: Oxford University Press, 1993.
61. Sagawa, S., K. Miki, F. Tajima, et al. Effect of dehydration on thirst and drinking during immersion in men. J. Appl. Physiol.
62. Sakai, R. R., W. B. Fine, A. N. Epstein, and S. P. Frankmann. Salt appetite is enhanced by one prior episode of sodium depletion in the rat. Behav. Neurosci.
63. Sawka, M. N., and E. F. Coyle. Influence of body water and blood volume on thermoregulation and exercise performance in the heat. Exerc. Sport Sci. Rev.
64. Shapiro, R. E., and R. R. Miselis. The central neural connections of the area postrema of the rat. J. Comp. Neurol.
65. Siegel, A. J. Exercise-associated hyponatremia: role of cytokines. Am. J. Med.
66. Simpson, J. B., and A. Routtenberg. Subfornical organ: site of drinking elicitation by angiotensin II. Science
67. Stachenfeld, N. S., L. DiPietro, E. R. Nadel, and G. W. Mack. Mechanisms of attenuated thirst in aging: role of central volume receptors. Am. J. Physiol. Regul. Integr. Comp. Physiol.
68. Stricker, E. M. Osmoregulation and volume regulation in rats: inhibition of hypovolemic thirst by water. Am. J. Physiol.
69. Stricker, E. M. Inhibition of thirst in rats following hypovolemia and/or caval ligation. Physiol. Behav.
70. Stricker, E. M., K. S. Gannon, and J. C. Smith. Thirst and salt appetite induced by hypovolemia in rats: analysis of drinking behavior. Physiol. Behav.
71. Stricker, E. M., and J. E. Jalowiec. Restoration of intravascular fluid volume following acute hypovolemia in rats. Am. J. Physiol.
72. Stricker, E. M., and J. G. Verbalis. Sodium appetite. In: Handbook of Behavioral Neurobiology, Vol. 10, Neurobiology of Food and Fluid Intake
, E. M. Stricker (Ed.). New York, NY: Plenum Press, pp. 387-419, 1990.
73. Sutton, J. R. Clinical implications of fluid imbalance. In: Perspectives in Exercise Science and Sports Medicine, Volume 3: Fluid Homeostasis During Exercise
, C. V. Gisolfi and D. R. Lamb (Eds.). Carmel, IN: Benchmark, pp. 425-444, 1990.
74. Takamata, A., G. W. Mack, C. M. Gillen, and E. R. Nadel. Sodium appetite, thirst, and body fluid regulation in humans during rehydration without sodium replacement. Am. J. Physiol. Regul. Integr. Comp. Physiol.
75. Thunhorst, R. L., T. G. Beltz, and A. K. Johnson. Effects of subfornical organ lesions on acutely induced thirst and salt appetite. Am. J. Physiol. Regul. Integr. Comp. Physiol.
76. Thunhorst, R. L., and A. K. Johnson. Renin-angiotensin, arterial blood pressure, and salt appetite in rats. Am. J. Physiol.
77. Thunhorst, R. L., S. J. Lewis, and A. K. Johnson. Effects of sinoaortic baroreceptor denervation on depletion-induced salt appetite. Am. J. Physiol.
78. Toth, E., J. Stelfox, and S. Kaufman. Cardiac control of salt appetite. Am. J. Physiol. Regul. Integr. Comp. Physiol.
79. van der Kooy, D., and L. Y. Koda. Organization of the projectionsof a circumventricular organ: the area postrema in the rat. J. Comp. Neurol
. 219:328-338, 1983.
80. Verney, E. B. The antidiuretic hormone and the factor which determines its release. Proc. R. Soc. Lond. B Biol. Sci.
81. Wolf, A. V. Osmometric analysis of thirst in man and dog. Am. J. Physiol.
82. Wolf, G. Sodium appetite elicited by aldosterone. Psychon. Sci.