Hunger

Abstract

Hunger is a type of drive: a complex behavioural state that elicits intense food seeking and consumption. This article covers the classic peripheral (stomach-contraction), glucostatic and lipostatic accounts of what controls it, the hypothalamic lesions that tied it to the hypothalamus, the gut and hormonal signals that modulate food intake, the AgRP neurons whose activation evokes voracious feeding, the learned, remembered and hedonic influences on eating, and how hunger is measured. An important and unresolved question is what hunger neurons signal: one study found that their activity has negative valence, another that it is positively reinforcing in the presence of food.

Keywords: hunger, appetite, AgRP neurons, energy homeostasis

Hunger sits at the meeting point of physiology and psychology. Feeding behavior is a fundamental aspect of energy homeostasis, in which the amount of energy consumed must match the amount expended (Woods et al., 1998), yet people also eat for pleasure, not just for calories (Lowe & Butryn, 2007), and fasting-induced hunger has served as the model for social craving (Tomova et al., 2020).

What Hunger Is

Hunger is a complex behavioural state that elicits intense food seeking and consumption (Atasoy et al., 2012). Its place in physiology is homeostatic: feeding is a fundamental aspect of energy homeostasis, the process by which body fuel stored as adipose tissue is held constant over long intervals (Woods et al., 1998). That homeostatic framing has a rival. On an incentive account, hunger acts to enhance the incentive value of food; reduction of a hunger drive is usually not enough to reinforce behavior unless an incentive stimulus is also involved, and dogs given their full daily calories intravenously continued to eat customary amounts of food (Berridge, 2018). In current neuroscience, arcuate nucleus-based hypothalamic circuits link energy state to the motivational drive, hunger, and limbic and cognitive processes bring about hunger-mediated increases in the reward value and perception of food (Andermann & Lowell, 2017).

Hunger is not the whole of appetite. A review of appetite measurement separates satiation, the control of meal size, from satiety, the control of the post-eating period, and assesses a homeostatic toolkit for measuring subjectively perceived hunger alongside procedures for the influence of hedonic processes on food intake (Gibbons et al., 2019). Hunger is also used as a template for other needs: after 10 hours of fasting, midbrain regions responded selectively to food cues, and the same regions responded to social cues after isolation, supporting the idea that acute isolation causes social craving similar to the way fasting causes hunger (Tomova et al., 2020).

Types of Hunger

MeSH files hunger beneath drive and lists one narrower descriptor beneath it, appetite.

This is a classification for indexing literature, not a theory of how hunger divides. MeSH also files hunger under nutritional and digestive system physiological phenomena, so the same descriptor indexes physiological as well as psychological work.

Theories of Hunger

Early theories looked to the body. Before the twentieth century, as Mayer reviewed, theories of peripheral origin held that the taking of food resulted from the stimulation of afferent nerves by some change in the tissues or of a strictly local group of sensory nerves, mainly in the stomach, while a rival theory postulated a central origin (Mayer, 1953). Mayer, holding that the regulation of energy intake is fundamental to all homeostatic mechanisms, proposed a glucostatic mechanism of regulation of food intake (Mayer, 1953).

The best-known peripheral account located hunger in the stomach. Cannon and Washburn offered their own explanation of hunger (Cannon & Washburn, 1912): recording the stomach while Washburn reported his sensations, they found that when he stated that he was hungry, powerful contractions of the stomach were invariably being registered, and Cannon concluded that contractions are the immediate cause of hunger (Cannon, 1915). A patient whose entire stomach had been removed nonetheless became hungry frequently, although he could not recall ever having had, either before or after the operation, any sensation he would describe as a hunger pang; to him the sensation was one of emptiness in the stomach followed by a feeling of general weakness (Wangensteen & Carlson, 1931).

The lipostatic account located control in the hypothalamus. The young rat adjusts its food intake so precisely to its energy needs that its fat stores remain almost constant, and hypothalamic damage permits excessive intake and causes obesity (Kennedy, 1953). Kennedy suggested that the hypothalamic satiety mechanism is concerned only with preventing an overall surplus of energy intake over expenditure, and that the simplest way this lipostasis could be achieved is by sensitivity to the concentration of circulating metabolites (Kennedy, 1953). Hypothalamic lesions and adiposity in the rat had been reported earlier (Hetherington & Ranson, 1940). Early researchers found that lesions of the ventromedial hypothalamus (VMH) resulted in hyperphagia and obesity in a variety of species including humans, which led them to designate the VMH as the brain's 'satiety center' (King, 2006). Its counterpart in rats was a lateral area that produced eating when electrically stimulated and aphagia and weight loss when lesioned, which was dubbed the 'feeding center' (Berthoud & Münzberg, 2011). A later review concluded that VMH lesion-induced obesity is in large part a metabolic obesity independent of hyperphagia, that the VMH nonetheless plays a primary role in feeding behavior, and that the concept of hypothalamic feeding and satiety centers is outdated and unnecessary (King, 2006).

Gut and Hormonal Signals

Signals from the body modulate food intake. A review of these signals frames them as integrating the body's immediate and long-term energy needs (Woods et al., 1998). On the stopping side, the gut peptide cholecystokinin was reported to decrease food intake in rats (Gibbs et al., 1973). On the long-term side, mutation of the mouse obese (ob) gene results in profound obesity and type II diabetes, and the ob gene product may function as part of a signalling pathway from adipose tissue that acts to regulate the size of the body fat depot (Zhang et al., 1994).

Ghrelin has been proposed as a signal for starting a meal. It was purified from rat stomach as an endogenous ligand for the growth-hormone secretagogue receptor, a 28-amino-acid peptide that releases growth hormone (Kojima et al., 1999). In 10 healthy people given meals on a fixed schedule, plasma ghrelin levels increased nearly twofold immediately before each meal and fell to trough levels within 1 hour after eating, a pattern reciprocal to that of insulin, which supports the hypothesis that ghrelin plays a physiological role in meal initiation in humans (Cummings et al., 2001). Ghrelin is not critically required for appetite in mice, however: in ghrelin-null mice, size, growth rate, food intake and body composition were indistinguishable from wild-type littermates, responses to starvation were normal, and exogenous ghrelin still stimulated appetite, leading the authors to conclude that ghrelin is not critically required for appetite (Sun et al., 2003).

Sleep alters the same hormones. In a randomized crossover study of 12 healthy young men, two days of sleep restriction, compared with sleep extension, were associated with an 18% decrease in leptin, a 28% increase in ghrelin, a 24% increase in hunger and a 23% increase in appetite, especially for calorie-dense foods with high carbohydrate content (Spiegel et al., 2004).

Figure 1

Changes in Hormones and Ratings After Sleep Restriction Versus Sleep Extension

Hunger Neurons

Two intermingled hypothalamic populations, defined by agouti-related peptide (AGRP) or pro-opiomelanocortin (POMC), positively and negatively influence feeding (Aponte et al., 2011). Activation of only 800 AGRP neurons in mice evoked voracious feeding within minutes, the response increased with neuron number, stimulation frequency and duration, POMC neuron stimulation reduced food intake and body weight, and feeding was evoked selectively over drinking without training or prior exposure to the stimulus (Aponte et al., 2011). Mapping the circuit, an inhibitory connection to the paraventricular hypothalamus substantially accounted for AGRP neuron-evoked eating, and AGRP neurons target and inhibit oxytocin neurons there, a small population that is selectively lost in Prader-Willi syndrome, a condition involving insatiable hunger (Atasoy et al., 2012).

These neurons respond to food before it is eaten. Optical recording in awake mice showed that the sensory detection of food is sufficient to rapidly reverse the activation of AgRP and POMC neurons induced by energy deficit, before any food is consumed, modulated by palatability and nutritional state (Chen et al., 2015). Nutrients then sustain the change: AgRP neuron activity has a rapid but transient sensory-driven component and a slower, sustained calorie-dependent component, and nutrients are necessary and sufficient for the sustained reduction (Su et al., 2017). Intragastric nutrient infusion inhibited AgRP neurons in proportion to the calories infused, independent of macronutrient identity; serotonin, CCK and PYY were necessary or sufficient for these effects, whereas leptin had no acute effect but induced a slow modulation over hours that was required for inhibition of feeding (Beutler et al., 2017).

What the activity of these neurons signals is disputed. In one study, mice avoided activation of AGRP neurons, indicating that their activity has negative valence, and AGRP neuron inhibition conditioned preference for flavours and places, so that preference for cues associated with eating is learned through reduction of a negative-valence signal (Betley et al., 2015). In another, brief stimulation of AgRP neurons before food was available promoted intense appetitive and consummatory behavior that persisted for tens of minutes, mediated by a long-lasting potentiation of the rewarding properties of food, and AgRP neuron activity was positively reinforcing in the presence of food, although the authors did not rule out an additional negative-valence signal that functions primarily before food is discovered (Chen et al., 2016). The dispute echoes an older one: motivation was once thought of in terms of aversive drives and reward in terms of drive reduction, those concepts were largely replaced by incentive concepts, yet aversive drive concepts still occasionally surface in reward neuroscience today (Berridge, 2018).

Learning, Memory and Hedonic Hunger

Eating can start without a deficit. Pavlovian conditioning taught rats an association between an arbitrary external cue and food, and presenting the cue then elicited feeding in sated animals; the meal was about 20 percent of daily intake and was compensated for by a reduction of subsequent intake (Weingarten, 1983). Specific hungers have also been treated as adaptive specializations of learning (Rozin & Kalat, 1971).

Memory helps end meals. Two densely amnesic patients, who usually failed to remember that they had just eaten, readily consumed a second lunch offered 10 to 30 minutes after the first and usually began a third, suggesting that memory for what has recently been eaten substantially contributes to the onset or cessation of a meal (Rozin et al., 1998). In unrestrained eaters, being asked to think about that day's lunch reduced subsequent intake, although subjective ratings of hunger, fullness and desire to eat did not vary with the cue (Higgs, 2002). Social psychology took up the relation between obesity and eating (Schachter, 1968). Schachter proposed that obese individuals are especially, and perhaps exclusively, responsive to external food cues; that model was later modified and challenged, research on external cues declined and then resurged, and one review now distinguishes normative cues such as portion size, which affect everyone, from sensory cues such as palatability, which have a more powerful effect on some individuals, such as the obese, than on others (Herman & Polivy, 2008). The internal-external hypothesis for obesity was itself reassessed under the question of what went wrong (Rodin, 1981).

Hunger also has a hedonic side. An increasing proportion of human food consumption appears to be driven by pleasure, not just the need for calories; homeostatic and hedonic eating motives overlap but are dissociable, and the Power of Food Scale was developed to measure individual differences in appetitive responsiveness to the rewarding properties of the food environment (Lowe & Butryn, 2007). Some individuals seem to eat not only because of metabolic need but also because of an excessive hedonic drive, and neural work examines both how metabolic signals modulate cortico-limbic reward processing and how the cognitive and emotional brain may override homeostatic regulation (Berthoud, 2011).

Measuring Hunger

Hunger is measured by self-report. In 55 healthy men, visual analogue scales recorded hunger, satiety, fullness and prospective food consumption before breakfast and every 30 minutes for 4.5 hours; with a paired design and a power of 0.8, a 10 mm difference in fasting ratings and a 5 mm difference in mean 4.5-hour ratings could be detected with 18 subjects, mean 4.5-hour ratings gave the best correlations with subsequent energy intake (r = ±0.50 to 0.53), and the authors concluded that VAS scores are reliable for appetite research (Flint et al., 2000). A methodological report on appetite research gives detailed coverage of the extent to which a change in hunger can stand alone as a measure of appetite control, and distinguishes designs for satiation from designs for satiety (Blundell et al., 2010).

Hunger can also be measured as a trait. Factor analysis of an eating questionnaire administered to groups ranging from extreme dietary restraint to extreme lack of restraint yielded three stable factors, cognitive restraint of eating, disinhibition and hunger, which became the 51-item Three-Factor Eating Questionnaire (Stunkard & Messick, 1985).

Worked Example

Rating scales set the size of a hunger study. With a paired design and a power of 0.8, a 5 mm difference in mean 4.5-hour visual analogue ratings was detectable with 18 subjects (Flint et al., 2000). Because the required sample size of a paired comparison scales with the inverse square of the difference to be detected, halving that difference to 2.5 mm would, on the same variability, need about 18 × 2², or 72 subjects. The same 18 subjects could detect only a 10 mm difference in fasting ratings, consistent with the authors' finding that coefficients of repeatability were larger for fasting values than for mean 4.5-hour values (Flint et al., 2000).

Discussion

Three conclusions follow for cognitive psychology. First, eating is regulated by the body but not only by it. Ghrelin rises before meals (Cummings et al., 2001), yet a conditioned cue elicits eating in sated rats (Weingarten, 1983) and amnesic patients who fail to remember a meal readily eat a second one (Rozin et al., 1998).

Second, rated hunger and eating can come apart. A reminder of lunch reduced intake without changing hunger ratings (Higgs, 2002), and a methodological report asks how far a change in hunger can stand alone as a measure of appetite control (Blundell et al., 2010).

Third, the valence of hunger is unsettled. Mice avoided AGRP neuron activation (Betley et al., 2015), but AgRP neuron activity was positively reinforcing in the presence of food in another study (Chen et al., 2016), and aversive drive concepts still surface in reward neuroscience (Berridge, 2018).

Current Directions

Three lines of work are active. Circuits: three major neural circuits strongly and acutely influence appetite with notably different characteristics, forming what one review calls three pillars of appetite control (Sternson & Eiselt, 2017), and new tools have revealed unexpected findings such as the rapid regulation of homeostatic neurons by cues that predict future food consumption (Andermann & Lowell, 2017). Individual differences: there is increasing recognition of individual variability in the expression of appetite, and of the fact that the mean response conceals important information about appetite control (Gibbons et al., 2019). Other needs: fasting is the reference for social craving, with midbrain responses to food cues after fasting paralleling responses to social cues after isolation (Tomova et al., 2020).

Common Misconceptions

“People stop eating because they stop feeling hungry.”

A reminder of lunch reduced how much people ate without changing their hunger ratings (Higgs, 2002), and amnesic patients readily ate a second lunch soon after the first (Rozin et al., 1998).

“Animals eat only when they need energy.”

A conditioned cue elicited feeding in sated rats (Weingarten, 1983), and an increasing proportion of human eating appears to be driven by pleasure, not just by the need for calories (Lowe & Butryn, 2007).

“It is settled that hunger neurons produce an unpleasant state.”

Mice avoided AGRP neuron activation in one study (Betley et al., 2015), but AgRP neuron activity was positively reinforcing in the presence of food in another (Chen et al., 2016).

Glossary

AgRP neurons.

Hypothalamic neurons expressing agouti-related peptide, whose activation evokes feeding.

Cholecystokinin.

A gut peptide (CCK) reported to decrease food intake in rats.

Energy homeostasis.

The process by which body fuel stored as adipose tissue is held constant over long intervals.

Ghrelin.

An orexigenic peptide produced primarily by the stomach, whose plasma level rises before meals.

Glucostatic theory.

Mayer's proposal of a glucostatic mechanism of regulation of food intake.

Hedonic hunger.

Lowe and Butryn’s term for an appetitive counterpart, which the food environment in affluent societies may be creating, to the psychological effects of other hedonically-driven activities such as drug use and compulsive gambling.

Leptin.

An anorexigenic hormone, lowered by sleep restriction, that slowly modulates hunger neurons over hours.

Lipostasis.

Kennedy's term for the hypothalamic prevention of a surplus of energy intake that would deposit fat.

Meal initiation.

The start of a meal, in which ghrelin is hypothesized to play a physiological role.

Paraventricular hypothalamus.

The target of the inhibitory AGRP neuron circuit that substantially accounted for evoked eating.

POMC neurons.

Hypothalamic neurons expressing pro-opiomelanocortin, whose stimulation reduced food intake and body weight.

Power of Food Scale.

A measure of individual differences in appetitive responsiveness to the rewarding properties of the food environment.

Satiation.

The control of meal size.

Satiety.

The control of the post-eating period.

Three-Factor Eating Questionnaire.

A 51-item questionnaire measuring cognitive restraint of eating, disinhibition and hunger.

Visual analogue scale.

A rating scale used to record appetite sensations such as hunger, satiety and fullness, scored in millimetres.

Key Researchers

Kent C. Berridge

Professor in the Department of Psychology at the University of Michigan; reviewed how drive concepts of motivation gave way to incentive concepts. ORCID - Wikipedia - Faculty Page

John E. Blundell

Psychobiologist at the University of Leeds; co-authored work on visual analogue scales and on methods for measuring appetite and hunger. ORCID - Faculty Page - Google Scholar

Suzanne Higgs

Psychologist at the University of Birmingham; showed that memory for recent eating influences later food intake. ORCID - Wikipedia - Google Scholar

Bradford B. Lowell

Professor at Harvard Medical School and Beth Israel Deaconess Medical Center; co-authored a wiring-diagram review of appetite control. ORCID - Faculty Page - Lab

Stanley Schachter

(1922-1997). Social psychologist at Columbia University; proposed that obese individuals are especially responsive to external food cues. Wikipedia

Scott M. Sternson

Neuroscientist at the University of California, San Diego and the Howard Hughes Medical Institute; showed that AGRP neurons are sufficient to evoke feeding and mapped their circuits. ORCID - Faculty Page

Frequently Asked Questions

What is hunger in psychology?

Hunger is a complex behavioural state that elicits intense food seeking and consumption (Atasoy et al., 2012), linked by hypothalamic circuits to the body's energy state (Andermann & Lowell, 2017).

Which hormone signals hunger before meals?

Ghrelin, produced primarily by the stomach, rose nearly twofold in the blood immediately before each meal, supporting a role in meal initiation (Cummings et al., 2001), although mice lacking ghrelin ate normally (Sun et al., 2003).

Which brain cells control hunger?

Activating only 800 AGRP neurons in the hypothalamus of mice evoked voracious feeding within minutes (Aponte et al., 2011).

Does seeing food change hunger signals?

Yes: the sensory detection of food rapidly reversed the activation of AgRP neurons in hungry mice before any food was eaten (Chen et al., 2015), although sustained reductions required nutrients (Su et al., 2017).

Why do people eat when they are not hungry?

A conditioned cue elicited feeding in sated rats (Weingarten, 1983), and an increasing proportion of human eating appears to be driven by pleasure, not just by the need for calories (Lowe & Butryn, 2007).

Does memory affect eating?

Amnesic patients readily ate a second lunch 10 to 30 minutes after the first (Rozin et al., 1998), and recalling that day's lunch reduced later intake (Higgs, 2002).

Does lack of sleep increase hunger?

In 12 young men, sleep restriction was associated with lower leptin, higher ghrelin and 24% higher hunger (Spiegel et al., 2004).

How is hunger measured?

With visual analogue scales, which are reliable for appetite research and whose mean ratings correlated with later energy intake (Flint et al., 2000), and as a trait with the hunger factor of the Three-Factor Eating Questionnaire (Stunkard & Messick, 1985).

References