Food access schedule and diet composition alter rhythmicity of serum melatonin and pineal NAT activity.

This study investigated the effect of dietary composition and food access schedule on the rhythmicity of serum melatonin and pineal N-acetyltransferase (NAT) activity. Wistar rats maintained on a 12:12 h light-dark cycle were assigned to two dietary groups: a group fed rat chow and a group fed a choice between a protein-rich and a carbohydrate-rich diet. Each dietary group was further divided based on feeding schedule, with food available between 0800 and 1600 h or ad lib access to food. Regardless of dietary condition, total food and carbohydrate intake of rats having free access to food was higher than under the restricted food access schedule. Protein intake of rats fed the dietary choice was lower with the restricted access than in the free access. In rats fed the dietary choice, melatonin levels and NAT activity were significantly decreased with restricted access compared to free access. Such results were not found in rats offered restricted chow. This study suggests that the rhythms of melatonin secretion and NAT activity can be altered by dietary composition.     

Food intake inhibition by melatonin in goldfish (Carassius auratus)

Feeding regulation by monoamines, neuropeptides and certain hormones has been studied in fish, but a possible role of melatonin is unknown. The purpose of the present study was to investigate the effects of melatonin on food intake in goldfish. Fishes were housed in 12L:12D and injected with different doses of either melatonin or 2-iodomelatonin. Two routes of administration, intracerebroventricular and intraperitoneal injections, and two times of the daily photocycle, midday and midnight, were tested. Food intake was measured at 2, 5 and 8 h postinjection. Melatonin and its analog, 2-iodomelatonin intracerebroventricularly injected had no effect on food intake at any time. However, intraperitoneal injections of both indoleamines significantly reduced food intake at different postinjection times. The inhibitory effect of melatonin was blocked by intraperitoneal administration of its antagonist, luzindole. These results demonstrate the in vivo efficiency of luzindole as melatonin antagonist, and thus provide a useful experimental tool to investigate melatonin functions. In conclusion, both melatonin and its agonist 2-iodomelatonin administered peripherally, inhibit food intake in goldfish, and this inhibitory effect appears to be mediated via luzindole-sensitive melatonin receptors. Our results strongly suggest that melatonin is involved in the peripheral satiety mechanisms in goldfish.     

Dietary influences on the acute effects of anorectic drugs

The effects of acute administration of d-amphetamine sulfate (0.0, 0.5, 1.0 and 2.0 mg/kg) and dl-fenfluramine hydrochloride (0.0, 1.5, 3.0 and 6.0 mg/kg) on food intake were examined in male Sprague-Dawley rats fed either a high-carbohydrate diet (carbohydrate equaled 65% of total calories) or a high-fat diet (fat equaled 65% of total calories). Animals were given ad lib access to the diets throughout the experiment. Drug injections were given at 0900 on experimental days and food intakes were measured at 1, 3 and 6 h postinjection. Amphetamine led to dose-related decreases in food intake for animals on both diets. The effects of amphetamine were most noticeable at 1 and 3 h postinjection. No differences in amphetamine's effects on food intake were found as a function of diet. Fenfluramine injections also led to dose-related reductions in food intake for animals in both dietary conditions. In contrast to amphetamine, however, fenfluramine led to greater reductions in food intake for rats fed the high-fat diet than for rats fed the high-carbohydrate diet. These data demonstrate that dietary variables must be considered when evaluating the anorectic actions of psychopharmacological agents.     

Preabsorptive factors are not the main determinants of intake depression induced by a high-protein diet in the rat

The factors involved in the depression of food intake produced by a high-protein diet are still poorly understood. The aim of this study was to assess the role of several preingestive or preabsorptive factors likely to influence food intake when rats were fed ad libitum. Food intake was measured after modifying the composition of the high-protein diet, i.e., the type of proteins, or carbohydrates. Moreover, correlations between high-protein diet intake and the quantity of fluid ingested or stomach volume were studied. By varying the carbohydrate composition (sucrose/cornstarch) and the protein source (soy or gluten or total milk protein) of high-protein diets, we modified the orosensory properties of these diets. However, no differences in food intake were observed between these groups of rats during the transition phase or after adaptation, except during the first day of soy- or gluten-based diets when the depression of food intake was intensified. The depression of high-protein diet intake was neither the consequence of any delay necessary to increase the fluid intake induced by eating a high-protein diet nor due to a marked increase in stomach volume, which might explain enhanced satiety and decreased food intake through the activation of vagal afferent fibers. Our experiments do not indicate a preponderant role for oropharyngeal or preabsorptive factors in the depression of food intake induced by a high-protein diet.     

Carbohydrate loading in human muscle: an improved 1 day protocol

It is generally acknowledged that even without a glycogen-depleting period of exercise, trained athletes can store maximal amounts of muscle glycogen if fed a carbohydrate-rich diet for 3 days. What has never been examined is whether under these conditions this many days are necessary for the content of muscle glycogen to attain these high levels. To examine this issue, eight endurance-trained male athletes were asked to eat 10 g·day^–1·kg^–1 body mass of high-carbohydrate foods having a high glycaemic index over 3 days, while remaining physically inactive. Muscle biopsies were taken prior to carbohydrate loading and after 1 and 3 days of eating the carbohydrate-rich diet. Muscle glycogen content increased significantly (P<0.05) from pre-loading levels of [mean (SE)] 95 (5) to 180 (15) mmol·kg^–1 wet mass after only 1 day, and remained stable afterwards despite another 2 days of carbohydrate-rich diet. Densitometric analyses of muscle sections stained with periodic acid-Schiff not only supported these findings, but also indicated that only 1 day of high carbohydrate intake was required for glycogen stores to reach maximal levels in types I, IIa, and IIb muscle fibres. In conclusion, these findings showed that combining physical inactivity with a high intake of carbohydrate enables trained athletes to attain maximal muscle glycogen contents within only 24 h. Electronic Publication     

Melatonin enhances NADPH-diaphorase activities in the hypothalamus of maternally-separated rats

Maternal separation or social isolation is a risk factor in the development of mammalian species affecting both physical and mental growth, and food intake regulation. Melatonin has been known to regulate body weight on various species including rodents. We investigated the effect of melatonin treatment on the expression of nitric oxide synthase, which may involved in food intake regulation, in the brain of maternally separated-rats using nicotinamide adenine dinucleotide phosphate-diaphorase (NADPH-d) histochemistry. Melatonin (10 mg/kg i.p.) was injected to 14-day-old maternally-separated rats for 7 days. Maternally-separated rats with melatonin administration showed significantly higher staining intensities of NADPH-d-positive neurons in the paraventricular nucleus (PVN) and in lateral hypothalamic area (LHA) than maternally-separated without melatonin administration (P < 0.05). Body weight of melatonin treated rats significantly increased at the 6th and 7th day compared to that of rats without melatonin treatment (P < 0.05). These results indicate that melatonin may be associated with increase body weight via NOS in the hypothalamic areas in maternally-separated or socially isolated rats.     

Bitter taste and blood glucose are not involved in the suppressive effect of dietary histidine on food intake

Histamine decreases food intake by activating histaminergic neurons in the hypothalamus. Histamine is synthesized by histidine decarboxylase (HDC) from histidine. The purpose of this three-part animal study was to clarify the mechanism underlying the suppressive effect of dietary histidine on food intake. In experiment 1, we attempted to distinguish palatability from a direct effect of dietary histidine because histidine tastes slightly bitter to humans. We measured food intake every hour for 24 h in rats fed with a histidine-enriched diet or one of various quinine diets (0.001–0.8% quinine), also bitter. In experiment 2, we measured changes in blood glucose levels in rats fed with a standard or histidine-enriched diet because blood glucose is known to decrease food intake. In experiment 3, we intraperitoneally injected fluoromethylhistidine (FMH), an antagonistic inhibitor of HDC, in rats fed with a histidine-enriched diet. In experiment 1, food intake was almost the same in rats fed with the histidine-enriched diet as that in rats fed with the 0.01% quinine diet until 6 h, but food intake was low in other groups compared with that in the histidine-enriched diet group. After 6 h, food intake did not increase in rats fed with the histidine-enriched diet. In experiment 2, the blood glucose level rose quickly and then began to decrease at approximately 2 h in both groups of rats. However, it decreased more dramatically in rats fed with the histamine-enriched diet and reaches a significant difference from the decrease in the standard-diet group by 6 h. In experiment 3, food intake increased significantly in FMH-injected rats fed with the histidine-enriched diet compared with in non-FMH injected rats. Our results suggest that dietary histidine suppresses food intake by activating histaminergic neurons in the hypothalamus, independently bitter taste and blood glucose level.     

The effect of hormone therapy on serum melatonin concentrations in premenopausal and postmenopausal women: A randomized,double-blind,placebo-controlled study

Objectives

Melatonin levels decrease physiologically with age, and possibly with the transition to menopause. The plausible influence of hormone therapy (HT) on melatonin is poorly understood. The aim of this randomized, placebo-controlled, double-blind trial was to investigate the effect of HT administration on serum melatonin concentrations in late premenopausal and postmenopausal women.

Study design

Analyses were carried out among 17 late premenopausal and 18 postmenopausal healthy women who participated in a prospective HT study in Finland. Serum melatonin was sampled at 20-min (21:00–24:00 h; 06:00–09:00 h) and 1-h (24:00–06:00 h) intervals at baseline and after 6 months with HT or placebo.

Main outcome measures

Melatonin levels and secretion profile after 6 months of HT compared to placebo.

Results

Mean melatonin levels, mean melatonin exposure level (area under curve, AUC) and mean duration of melatonin secretion did not differ after 6 months with HT vs. placebo, irrespectively of the reproductive state. However, in postmenopausal women the melatonin peak time (acrophase) was delayed by 2.4 h (2 h 21 min) on average after 6 months with HT vs. placebo (p < 0.05). No interaction between time and group was detected when melatonin level was modelled before or after treatment.

Conclusions

Administration of HT to postmenopausal women alters melatonin peak time, but not melatonin levels. Further research on larger clinical samples is needed to better understand the effects of HT on melatonin profile.     

Increased food intake and carbohydrate preference in the rat following treatment with the serotonin antagonist metergoline

The effect on feeding of antagonism of the serotonergic system was studied. Intraperitoneal injection of one of 8 doses of the serotonin antagonist metergoline (MET) (0.06-1.5 mg.kg-1) given prior to a 3 h food deprivation produced a significant dose-related increase in intake of a standard laboratory diet throughout a 2 h measurement period. A 1 mg/kg dose of MET administered prior to one of 3 isocaloric powdered diets a medium carbohydrate, a high carbohydrate, and a non-carbohydrate diet - produced a significant increase in the intake by rats of all 3 diets in the first hour after food presentation. However, in the second hour MET continued to increase consumption of the high carbohydrate diet, whereas intakes of the medium carbohydrate and non-carbohydrate diets were no longer affected by MET treatment. These data suggest that the serotonergic system is a tonic inhibitor of food intake and of carbohydrate preference.     

Responses to protein dilution in the adult female rat

Adult female rats were given ad lib access to 3 food cups containing casein, cornstarch, and a fat mixture, or a composite diet formulated from these nutrients. One-half of the animals in each diet condition were subjected to 25, 50, 150, and 500% sequential dilutions of the casein or the composite diet. Self-selecting rats decreased intake of the diluted casein, but increased consumption of cornstarch, so that total caloric intake remained constant. Composite-fed rats compensated for dilution by increasing the volume of diet consumed, but caloric regulation was less precise than in self-selectors.     

Maintenance on a high-fat diet impairs the anorexic response to glucagon-like-peptide-1 receptor activation

Previous data suggests that the adiposity signal leptin reduces food intake in part by enhancing sensitivity to short-term signals that promote meal termination, including glucagon-like peptide 1 (GLP-1). We hypothesized that maintenance on a high-fat (HF) diet, which causes resistance to leptin, would impair GLP-1's ability to reduce food intake. To test this hypothesis, we examined the anorexic responses to intraperitoneal injection of 100 μg/kg GLP-1 and 1 μg/kg exendin-4 (Ex-4), the potent, degradation resistant GLP-1 receptor agonist, in Wistar rats maintained on a low-fat (10%; LF) or HF (60%) diet for 4-6 weeks. Rats maintained on each of these diets were tested twice, once while consuming LF food and once while consuming HF food, to distinguish between effects of acute vs. chronic consumption of HF food. LF-maintained rats tested on LF diet reduced 60-min dark phase intake in response to GLP-1, but HF-maintained rats failed to respond to GLP-1 whether they were tested on HF or LF diet. LF-maintained rats tested on HF diet also showed no response, suggesting that even brief exposure to HF diet can impair sensitivity to GLP-1 receptor activation. Both LF- and HF-maintained rats showed significant anorexic responses to Ex4 at 4 h post-treatment, but only LF-maintained rats had significantly reduced intake and body weight 24 h after injections. To determine whether the ability of endogenous GLP-1 to promote satiation is impaired by HF maintenance, we examined the response to exendin 3 (9-39) (Ex9), a GLP-1 receptor antagonist. In LF-maintained rats, Ex9 increased intake significantly, but HF-maintained rats reduced food intake in response to Ex9. These data support the suggestion that maintenance on HF diet reduces the anorexic effects of GLP-1 receptor activation, and this phenomenon may contribute to overconsumption of high-fat foods.     

2-Deoxy-D-glucose and mercaptoacetate induce different patterns of macronutrient ingestion.

2-Deoxy-D-glucose (2DG) and mercaptoacetate (MA) are antimetabolic agents that reduce the metabolism of glucose and fatty acids, respectively, and stimulate feeding. The present study compared the effects of MA and 2DG on macronutrient self-selection. Because 2DG and MA have different metabolic actions and appear to activate different neural pathways, our hypothesis was that 2DG and MA would elicit different patterns of macronutrient selection. The first experiment examined macronutrient selection in response to 2DG, MA, and 0.9% saline in rats maintained on a three-macronutrient self-selection diet consisting of cornstarch, casein, and vegetable oil. Subsequently, one macronutrient source was replaced in each of three similar experiments with Polycose, albumin, or solid vegetable shortening. Finally, 2DG and MA tests were conducted in which only one macronutrient (cornstarch, casein, or oil) was available during the test. Results show that MA and 2DG elicit different macronutrient preferences. 2DG elicits intake of all three macronutrients in the same relative proportion consumed during spontaneous feeding across a number of dietary conditions, suggesting that glucoprivation activates interoceptive signals and neural pathways similar to those involved in normal hunger. MA elicits a selective intake of protein. Conditions in which carbohydrate palatability is enhanced or protein palatability is diminished lead to a relative increase in carbohydrate intake in response to MA. However, MA did not increase the intake of fat. Results suggest that intake of each macronutrient is subject to separate neural or endocrine control, and that these controls are linked to metabolic cues.     

Contrasting effects of systemic and central sibutramine administration on the intake of a palatable diet in the rat

Sibutramine hydrochloride monohydrate is the only centrally active weight-modifying agent currently approved by the FDA for long-term use in the treatment of obesity. Systemic sibutramine treatment has been shown to reduce food intake in humans and rodent models in a manner that is consistent with the enhancement of satiety mechanisms. Although it is generally assumed that the hypophagic effects of the drug are mediated by actions within the brain, the locus or loci of these effects remains unclear. These experiments compared the effects of systemic and intracranial injections of sibutramine on the intake of a palatable diet in non-deprived animals. Consistent with prior reports, systemic injections of sibutramine hydrochloride (at 0, 0.5, 1.0, or 3.0 mg/kg sibutramine i.p.) dose-dependently reduced feeding on a high fat/high sucrose diet across a 2-h feeding session, but did not alter water intake or locomotor activity. In contrast, bilateral injections of sibutramine (at 0.0, 2.0, 4.0 and 10.0 μg/0.5 μl/side) into either the paraventricular nucleus of the hypothalamus (PVN) or the medial nucleus accumbens shell (ACb) significantly and dose-dependently increased food intake of the sweetened fat diet. ACb treatment also modestly inhibited locomotor behavior; intracranial injections had no effect on water consumption. These experiments are the first to suggest that sibutramine treatment may have distinct actions upon separate neural circuits that modulate food intake behavior in the rat.     

Control of protein intake in golden hamsters

Experiments were performed to investigate the behavioural responses of golden hamsters to manipulations of dietary protein availability. In the first experiment, hamsters were maintained on a protein-free diet and a powdered diet containing 64.8% protein (P64.8). When the P64.8 diet was progressively diluted with cornstarch, hamsters increased their intake of this diet fraction, but protein intake nevertheless declined. When the protein content of the diet was 16.2%, animals derived only 6% of total calories from protein and lost weight despite normal intake of calories. In the remaining experiments, hamsters were maintained on a self-selection regimen of high-protein chow, pure carbohydrate (sugar cubes), and pure fat (vegetable shortening). When high-protein chow was removed for either 5 or 10 days, total caloric intake and body weight declined, and hamsters selectively increased protein intake for several days after high-protein chow was returned. Hamsters allowed access to high-protein chow for only one hour each day markedly increased the amount of high-protein chow they ate during this hour as protein-restriction continued, but still consumed only about 10% of their normal daily protein intake on this schedule and lost 20% of starting body weight in two weeks; when free access to high-protein chow was restored, these animals selectively increased their protein intake above pre-restriction levels. Hamsters given access to high-protein chow only on alternate days demonstrated a relatively modest and slowly developing increase in protein intake, perhaps because they incurred only a moderate protein deficit. The results suggest that when protein intake falls below normal minimum requirements, hamsters will demonstrate an adaptive protein hunger but make only a limited adjustment to the dilution of a protein-containing diet fraction.     

Activation of brain somatostatin 2 receptors stimulates feeding in mice: Analysis of food intake microstructure

We recently reported that the oligosomatostatin receptor agonist, ODT8-SST increases food intake in rats via the somatostatin 2 receptor (sst₂). We characterized ingestive behavior following intracerebroventricular (icv) injection of a selective sst₂ agonist in freely fed mice during the light phase. The sst₂ agonist (0.01, 0.03, 0.1, 0.3 or 1 μg/mouse) injected icv under short inhalation anesthesia dose-dependently increased cumulative light phase food intake over 4 h compared to vehicle with a 3.1-times increase at 1 μg/mouse (p < 0.05). Likewise, the sst_2,3,5 agonist octreotide (0.3 or 1 μg/mouse) dose-dependently increased 4-h food intake, whereas selective sst₁ or sst₄ agonists at 1 μg/mouse did not. In vehicle-treated mice, high fat diet increased caloric intake/4 h by 2.8-times compared to regular diet (p < 0.05) and values were further increased 1.4-times/4 h by the sst₂ agonist. Automated continuous assessment of food intake established a 6.6-times higher food intake during the dark phase due to increased number of meals, meal size, meal duration and rate of ingestion compared to non-treated mice during the light phase. During the first 4 h post icv sst₂ agonist injection, mice had a 57% increase in number of meals with a 60% higher rate of ingestion, and a 61% reduction in inter-meal intervals, whereas meal sizes were not altered compared to vehicle. These data indicate that the activation of brain sst₂ receptors potently stimulates the light phase ingestive behavior under basal or high fat diet-stimulated conditions in mice. The shortened inter-meal interval suggests an inhibitory effect of the sst₂ agonist on “satiety”, whereas “satiation” is not altered as indicated by normal meal size.     

Alpha-lactalbumin combined with a regular diet increases plasma Trp-LNAA ratio

Brain serotonin influences food intake and mood. It is synthesised from tryptophan (Trp) of which uptake in the brain is dependent on plasma ratio of tryptophan to the sum of other large neutral amino acids (Trp-LNAA). A carbohydrate-rich diet increases this ratio, whereas a protein-rich diet decreases it. Yet, if the protein source is alpha-lactalbumin the ratio increases. It is, however, unknown whether this also happens in the context of a regular diet (15% protein). We studied the effect of an alpha-lactalbumin supplement combined with regular diet on plasma Trp-LNAA ratio, serum prolactin (marker of serotonin synthesis), food intake, appetite, macronutrient preference and mood. Eighteen healthy males participated in a double-blind, randomised, placebo-controlled, crossover study. One hour after breakfast they received a drink containing alpha-lactalbumin and carbohydrates (AS) or carbohydrates (PS) only. Plasma Trp-LNAA ratio, serum prolactin, food intake, appetite, macronutrient preference and mood were assessed before and 90 min after consumption of the supplement. Changes of plasma Trp-LNAA ratio differed (P<.001) between both supplements, increasing by 16% after AS and decreasing by 17% after PS. Decrease of serum prolactin was slightly smaller after AS than after PS (P=.083). Appetite, food intake, macronutrient preference or mood did not differ between supplements. We conclude that an alpha-lactalbumin-enriched supplement combined with a regular diet increases plasma Trp-LNAA ratio and may influence serum prolactin, but we could not demonstrate effects on appetite, food intake, macronutrient preference and mood.     

Muscle ceramide content is similar after 3 weeks’ consumption of fat or carbohydrate diet in a crossover design in patients with type 2 diabetes

This study aimed at investigating the effect of prolonged adaptation to fat- or carbohydrate-rich diet on muscle ceramide in type 2 diabetes patients, using a longitudinal crossover study. Eleven type 2 diabetes patients consumed isocaloric fat- or carbohydrate-rich diet for 3 weeks in random order. After each dietary intervention period, muscle glycogen, triacylglycerol and ceramide content and plasma concentrations of insulin, adiponectin, glucose and FFA were determined. Insulin resistance was assessed by HOMA2 calculation. After the dietary period, plasma glucose and insulin, insulin sensitivity, muscle glycogen, triacylglycerol and ceramide content were similar. Plasma adiponectin concentration was significantly higher after fat compared with carbohydrate-rich diet. Results indicated that following fat-rich diet intake muscle ceramide and triacylglycerol concentrations were not different compared with that after carbohydrate-rich diet. Furthermore, plasma adiponectin concentration was higher after fat-rich compared with carbohydrate-rich diet, but insulin sensitivity remained similar despite the major difference in dietary macronutrient composition.     

Transdermal delivery of a pineal hormone: melatonin via elastic liposomes

Melatonin (MT) is a good candidate for transdermal delivery considering its short biological half-life, low molecular weight and a variable oral absorption. The objective of this work was to develop a novel formulation of melatonin for its efficient transdermal delivery. Melatonin loaded elastic liposomal formulation was prepared, characterized and the effect of this developed formulation on the in vitro permeation of melatonin across human cadaver skin was investigated, using a locally fabricated Franz diffusion cell. Skin permeation potential of the developed formulation was assessed using confocal laser scanning microscopy (CLSM), which revealed an enhanced permeation of the formulation to the deeper layers of the skin (up to 180 microm) following channel like pathways. Skin permeation profile of melatonin through elastic liposomal formulations was observed and the investigations revealed an enhanced transdermal flux (51.2+/-2.21 microg/cm(2)/h), decreased lag time (1.1h) and an optimum permeability coefficient (15.06+/-0.52 cm/h) for melatonin. The obtained flux was nearly 5 and 12.3 times higher than conventional liposomal and plain drug solution, respectively (P<0.005). Our result suggests the feasibility of elastic liposomal system for transdermal delivery of melatonin thereby eliminating the limitations of long lag time and poor skin permeation associated with the drug.     

Effect of sibutramine on macronutrient selection in male and female rats

Sibutramine, a serotonin-noradrenaline reuptake inhibitor (SNRI), has been shown to be a safe and effective weight-loss drug. The purpose of the present study was to examine whether sibutramine has an effect on macronutrient selection in both female and male rats in addition to total food intake. Wistar rats of both sexes were divided into three groups, and each group was offered a different set of three sensorily contrasting macronutrient-specific diets, each set including carbohydrate-, protein-, and fat-rich diets. Sibutramine (10 mg/kg) was shown to consistently decrease carbohydrate and fat intake at all data points regardless of gender and diet. Intake of carbohydrate differed between male and female rats at 2 h post administration with 2.5 and 5 mg/kg of sibutramine. The effect of sibutramine on protein intake was diet- and gender-specific. All doses of sibutramine decreased total food intake regardless of gender and diet group beginning at 6 h post administration. In conclusion, sibutramine affected macronutrient selection and emphasis on dietary recommendations, as well as appropriate dosage according to gender should be considered during therapy.     

Hyper-homeostatic learning of anticipatory hunger in rats

Anticipatory hunger is a learnt increase in intake of food having a flavour or texture that predicts a long fast. This learning was studied in rats trained on a single food or a choice between protein-rich and carbohydrate-rich foods, presented for 1.5 h after 3 h without maintenance food at the start of the dark phase. Eight training cycles provided a pseudo-random sequence of 3 h and 10 h post-prandial fasts with a day on maintenance food between each training fast. The measure of anticipatory hunger is the difference over one 4-day cycle between the intake of test food having an odour predictive of the longer fast (TL) and intake of food with an odour cuing to the shorter fast (TS). Previous experiments showed that conditioning of preference for the odour before the shorter fast competes with learning to avoid hunger during the longer fast (anticipatory hunger), generating a cubic or quartic contrast. TL minus TS showed a strong cubic trend over 8 training cycles with both single and choice meals. There was a switch from preference for the short-fast odour at cycle 2 (TL-TS=-0.86 g) to a peak of anticipatory hunger at cycle 6 (TL-TS=1.57 g). We conclude that anticipatory hunger is learnt when a choice is given between protein-rich and carbohydrate-rich foods as well as on a single food. In addition, since anticipatory hunger extinguishes itself, such learning improves on negative-feedback homeostasis with a feed-forward "hyper-homeostatic" mechanism.