INTRODUCTION

Top Abstract Introduction Methods Discussion References

THE PREFERENCE FOR, and hyperphagia- and obesity-promoting effects of, high-fat (HF) diets relative to high-carbohydrate (HC) diets are the subject of considerable interest. Several features of HF foods are thought to contribute to the overeating response: their palatable flavor, high energy density, and reduced satiating action, although this remains an unresolved issue (18, 19, 24). In addition, HF diets may produce obesity even in the absence of hyperphagia because of increased feed efficiency; dietary fat requires less energy to be converted into body fat than does dietary carbohydrate or protein (24).

In a recent "behavioral dissection" of HF hyperphagia, Warwick and Weingarten (26) fed rats HF and HC liquid diets that were equal in caloric density and protein content. Over a 16-day period the rats consumed more energy and gained more weight on the HF diet than on the HC diet. A second experiment assessed the palatability response to the flavors of the HF and HC diets using two-bottle, sham-feeding tests to minimize postingestive influences. The rats preferred the HF to the HC diet, indicating that the palatable flavor of the HF may have contributed to the hyperphagic response. The feeding effects of the two diets were then compared in the absence of flavor differences by using a self-regulated intragastric infusion procedure. Rats were infused intragastrically with HF or HC diet as their sole source of nutrition whenever they drank a saccharin solution available ad libitum 23 h/day. The rats in the HF group drank more saccharin and thus infused themselves with more diet and gained more weight than the rats in the HC group. These results demonstrate that when flavor differences are eliminated and caloric density is equated, the postingestive effects of the HF diet were sufficient to induce overeating relative to the HC diet. The authors (26) hypothesized that this occurred because the HF diet is less satiating than the HC diet, i.e., produced less suppression of ongoing eating (reduced satiation) and less suppression of eating during the postmeal interval (reduced satiety).

The postingestive actions of nutrients on food intake are usually considered to be inhibitory (satiation and satiety), but recent work demonstrates that nutrients can have positive postingestive effects that can increase food preferences and acceptance. This is demonstrated by reports that pairing the intake of a novel flavored solution with intragastric nutrient infusions increases the relative intake of that flavor in two-bottle preference tests and the absolute intake of the flavor in one-bottle acceptance tests (20). Furthermore, separate physiological systems appear to mediate the postingestive positive and negative feedback so that the reinforcing and satiating actions of nutrients are not necessarily correlated (20). In theory, at least, the hyperphagia induced by the postingestive actions of HF diets may result from increased positive feedback as well as, or instead of, decreased negative feedback. Also, the preference for HF over HC foods in choice tests may occur, at least in part, because of differences in the positive postingestive actions of the foods. Evidence to support these possibilities is lacking, and what data are available suggest that dietary fat has weaker, not stronger, postingestive stimulatory effects than dietary carbohydrates (13, 17). However, these data are based on studies of pure nutrients, and the positive postingestive effects of mixed diets high in fat or carbohydrate have not been investigated.

The aims of the present study, therefore, were to replicate the finding that intragastric self-infusions of an HF liquid diet promote hyperphagia relative to an HC liquid diet and to compare the effectiveness of the two diets to condition flavor preferences. This was accomplished by training rats to drink different flavored saccharin solutions paired with intragastric infusions of isocaloric HF and HC diets. The HF diet contained 60% fat (in kcal) and 34% carbohydrate, and the HC diet contained 15% fat and 79% carbohydrate; the diets were equivalent in protein content. Amounts consumed during one-bottle training and preferences in two-bottle choice tests were evaluated. In addition, drinking patterns were recorded to determine the effects of the different diet infusions on bout size and frequency. Variations of daily intake may result from changes in one or both of these bout parameters, which may reflect different components of ingestive behavior related to postingestive satiation, satiety, and reinforcement.

	GENERAL METHODS

Top Abstract Introduction Methods Discussion References

Subjects. Adult female Sprague-Dawley rats (260-350 g at the beginning of the experiments) obtained from Charles River Laboratories (Wilmington, MA) or bred in our laboratory from this stock were used. The animals were individually housed in wire-mesh cages in a room maintained at 21°C under a 12:12-h light-dark cycle (lights on at 0800).

Surgery. The rats were surgically implanted with stainless steel gastric cannulas as described by Elizalde and Sclafani (7). The animals were allowed to recover from surgery for at least 1 wk before the start of training. When not in use, the cannulas were kept closed with stainless steel screws.

Apparatus. Intragastric infusions were accomplished using an "electronic esophagus" apparatus described in detail elsewhere (7). In brief, the rats were housed in standard stainless steel cages modified so that powdered chow was available from a food cup accessible through a hole in the back wall of the cage. Drinking fluids were available through stainless steel sipper tubes available though two 19-mm holes in the front of the cage. Trays below the sipper tubes collected spillage. A slot in the cage floor permitted a stainless steel spring containing two catheters to connect the rat's gastric cannula to outputs of a dual-channel infusion swivel located below the cage. Plastic tubing connected the swivel's inputs to two peristaltic infusion pumps. The infusion pumps were operated automatically by drinkometer circuits and a microcomputer whenever the rat licked one of the two sipper tubes; the computer stored the number of licks per minute for offline drinking pattern analysis. The rate of infusion was 1.5 ml/min, and ~1 ml of fluid was infused intragastrically for each 1 ml of fluid orally consumed. With this infusion system, the rats were relatively unrestrained and free to eat and drink normally 22 h/day. The equipment was serviced during the remaining 2 h/day.

Test solutions and diets. The oral test solutions [conditioned stimuli (CSs)] consisted of 0.2% sodium saccharin (Sigma, St. Louis, MO) solutions flavored with 0.05% cherry, grape, or orange unsweetened Kool-Aid (General Foods, White Plains, NY). These flavors are equally preferred by naive rats (unpublished observations). The rats were infused with tap water or an HF or HC liquid diet. The liquid diets used for the intragastric infusions were based on the evaporated milk diets described by Warwick and Weingarten (26). The HF and HC versions were prepared by adding extra carbohydrate (sucrose in expt 1 or maltodextrin in expts 2-4) and fat emulsion (corn oil) to the milk (see Table 1). The HF and HC diets were equal in caloric density (2.1 kcal/ml) except in expt 4, in which the HC diet was diluted with water to 1.4 kcal/ml. The diet caloric densities and macronutrient compositions differ slightly from those reported by Warwick and Weingarten (26), presumably because of small differences in caloric values used for the individual diet ingredients. The following values were used here: sucrose = 3.85 kcal/g, maltodextrin = 3.8 kcal/g, corn oil = 9.0 kcal/g, and evaporated milk = 1.19 kcal/g. Note that the net caloric density of the liquid diets in the stomach was reduced by one-half as they were mixed with the orally consumed CS solutions. In addition to the liquid diet infusions, powdered Purina chow (no. 5001) was available ad libitum; the chow was low in fat (12.1% by calories) and high in carbohydrate (59.4%) and had a caloric density of 3.3 kcal/g. Chow intakes were recorded by weighing to the nearest 0.1 g. Intakes of CS solutions and intragastric diets were recorded in grams (±0.1 g) and converted to volumes and energy, respectively.

Conditioning procedure. The rats were trained to associate flavors (the CSs) with intragastric infusions during one-bottle training sessions (22 h/day). The flavors paired with HF, HC, and water infusions are referred to as CS+HF, CS+HC, and CS, respectively. The specific flavor (orange, cherry, grape) paired with each infusion was counterbalanced across subjects. Flavor preferences were assessed in two-bottle choice tests (22 h/day) with all possible pairs of CS flavors, still paired with their associated intragastric infusions. Chow was available ad libitum throughout training and testing. Water was available during one-bottle CS training days, but intakes were negligible (<4 ml/day) and are not reported.

Data analysis. Intakes during training and preference testing were averaged over sessions within the same condition (i.e., 2- to 6-day means) before being submitted to repeated-measures analysis of variance. Individual comparisons were evaluated by simple main effects or t-tests when appropriate. Intakes of the CS during two-bottle tests were also expressed as a percentage of total intake. Drinking patterns (bout frequency and size) were analyzed using the licks per minute and 22-h intake data. A bout was defined as a period of drinking containing a minimum of 10 licks and interlick intervals no larger than 10 min. Bout frequency was expressed as number of bouts per 22-h session. Mean bout size was calculated by dividing total 22-h intake by bout frequency.

	EXPERIMENT 1

Experiment 1 compared the effects of HF and HC diets developed by Warwick and Weingarten (26) on food intake and flavor preference conditioning. In their study, separate groups of rats were infused intragastrically with HF or HC diet whenever they drank a plain saccharin solution. In the present experiment, all rats were infused with HF and HC diets when they drank different flavored saccharin solutions; a third flavor was paired with intragastric water infusions. Chow was also available so that the rats were not food deprived on water-infusion days.

Originally the rats were tested with the different flavored solutions in short-term (30 min/day) sessions, but intakes were rather low, presumably because the rats were not food deprived and were tested at midday. The rats were then trained during long-term (22 h/day) sessions, and only these data are reported below.

Methods. Ten rats fitted with an intragastric cannula were adapted to the infusion apparatus, where they were maintained on ad libitum powdered chow and water. As noted above, they were initially trained with the CS solutions for 30 min/day but were then switched to 22 h/day sessions. Before the start of long-term training, the rats were given ad libitum access to only chow and water for 3 days. On training day 1, one-half the rats were given the CS+HF paired with intragastric infusion of the HF diet, followed on day 2 by the CS paired with intragastric water and on day 3 by the CS+HC paired with intragastric infusion of the HC diet. The remaining rats were trained with the CS+HC on day 1 and CS+HF on day 3. This 3-day pattern was repeated twice for a total of nine training sessions. Two-bottle preference tests were then conducted, with one-half the rats being offered the CS+HF vs. CS for 2 days, followed by the CS+HC vs. CS for 2 days; the remaining rats were given the CS+HF and CS+HC in the reverse order. All rats were then given the choice between the CS+HF and the CS+HC for 6 days. During the first 4 days of this choice, the CS+HF and CS+HC were paired with their corresponding liquid diets (reinforced test); during the next 2 days, the CS+HF and CS+HC were paired with intragastric water (extinction test).

Results. Table 2 summarizes daily CS intakes, intragastric diet intakes, and drinking patterns during one-bottle training days. The rats differed in their total daily intakes of the three CS solutions [F(2,18) = 30.34, P < 0.001]: CS intake (paired with intragastric water) exceeded (P < 0.05) CS+HF intake (paired with intragastric HF diet), which, in turn, exceeded (P < 0.05) CS+HC intake (paired with intragastric HC diet). These differences in total intakes were due primarily to differences in bout frequency [F(2,18) = 44.48, P < 0.001]. The highest bout frequencies were obtained with the CS, followed (P < 0.05) by the CS+HF, followed (P < 0.05) by the CS+HC. Mean bout sizes did not vary reliably, although CS+HF bouts tended to be larger than CS+HC bouts. Because they drank more CS+HF than CS+HC, the rats also infused themselves with more HF diet than HC diet [t(9) = 7.65, P < 0.001]. Table 2 also includes chow intake data from the pretraining baseline period and during one-bottle training. Chow intake varied as a function of training condition [F(3,27) = 100.7, P < 0.001]. Chow intakes did not differ on CS+HF and CS+HC training days but were suppressed (P values < 0.05) compared with CS days; in addition, CS chow intakes were reduced relative to the baseline period. Total energy intake (chow and diet intake) was also influenced by training condition [F(3,27) = 127.1, P < 0.001]. The rats consumed more (P < 0.05) energy on CS+HF days than CS+HC days, and more (P < 0.05) with both diet types than during baseline or CS days.

View larger version (39K): [in this window] [in a new window]   Fig. 1.   Intakes (means ± SE) of conditioned stimulus (CS) solutions during 22-h, 2-bottle tests of expt 1. Rats were given choice of CS solution paired with high-carbohydrate diet (CS+HC) vs. CS paired with water infusion (CS) (left), CS paired with high-fat diet (CS+HF) vs. CS (middle), and CS+HC vs. CS+HF (right) (see text for order counterbalancing). Intakes of CS, CS+HC, and CS+HF were paired with intragastric infusions of water, HC diet, and HF diet, respectively. Diets were prepared using sucrose. Numbers atop bars represent percentage intake of corresponding solution.

Discussion. During the one-bottle training period, the rats consumed more CS+HF than CS+HC and thus infused themselves with more HF diet than HC diet. Although they reduced their chow intake when the liquid diet infusions were available, this did not compensate fully for the energy provided by the diets. Consequently, total caloric intakes were increased above chow baseline when the rats had the CS+HF and CS+HC with diet infusions available, and the degree of hyperphagia was greater with the CS+HF and HF diet than with the CS+HC and HC diet. Chow intake was suppressed on CS days, relative to baseline, and this presumably represents a compensatory response to the elevated caloric intakes on preceding CS+HF and CS+HC days. Rats trained with CS+HF and CS+HC flavors paired with dilute nutrient infusions do not suppress their chow intake on CS days, which indicates that the CS flavor and water infusions do not have direct inhibitory effects on chow intake (2).

	EXPERIMENT 2

Following Warwick and Weingarten (26), the carbohydrate used to prepare the HC and HF diets in expt 1 was sucrose. However, we have found that the postingestive effects of fructose-containing carbohydrates (fructose, sucrose) are less reinforcing than those of glucose-based carbohydrates (glucose, maltose, maltodextrins) (1, 21). Thus the stronger preference conditioned by the HF diet relative to the HC diet in expt 1 may have occurred because the diets was prepared with a "suboptimal" carbohydrate, not because HF diets are, in general, more reinforcing than HC diets. Experiment 2 tested this possibility by comparing the preferences conditioned by HF and HC diets prepared with maltodextrin rather than sucrose.

Methods. New rats (n = 12) fitted with intragastric cannulas were trained and tested using the procedures of expt 1 except that the HC and HF diets were prepared with maltodextrin instead of sucrose (Table 1) and there were no preliminary short-term sessions.

Results. The one-bottle intake data are summarized in Table 3. Daily intake of the CS was greater (P < 0.05) than that of the CS+HF, which, in turn, was greater (P < 0.05) than that of the CS+HC [F(2,22) = 31.22, P < 0.001]. The differences in total intakes were due exclusively to differences in bout frequencies [F(2,22) = 28.21, P < 0.001]. The highest bout frequency was seen with the CS, followed (P < 0.05) by the CS+HF, followed (P < 0.05) by the CS+HC. CS bout sizes did not reliably differ. The rats infused themselves with more liquid diet on CS+HF days than on CS+HC days [t(11) = 7.10, P < 0.001]. Chow intakes did not reliably differ on CS+HF and CS+HC training days but were suppressed (P < 0.05) compared with CS and baseline days; chow intake on CS days was also less (P < 0.05) than baseline intake [F(3,27) = 120.8, P < 0.001]. Total energy intake also varied as a function of training condition [F(3,27) = 162.7, P < 0.001]; intake on CS+HF days exceeded (P < 0.05) that on CS+HC days, which, in turn, was greater (P < 0.05) than the intakes on baseline and CS days.

View larger version (36K): [in this window] [in a new window]   Fig. 2.   Intakes (means ± SE) of CS solutions during 22-h, 2-bottle tests of expt 2. Rats were given choice of CS+HC vs. CS (left), CS+HF vs. CS (middle), and CS+HC vs. CS+HF (right) (see text for order counterbalancing). Intakes of CS, CS+HC, and CS+HF were paired with intragastric infusions of water, HC diet, and HF diet, respectively. Diets were prepared using maltodextrin.

Discussion. The results of this experiment are very similar to those of expt 1. Most importantly, the CS+HF was significantly preferred to the CS+HC, and this was due to the rats taking almost twice as many CS+HF bouts than CS+HC bouts in the two-bottle tests. The CS+HF preference was somewhat weaker than that observed in expt 1 (64 vs. 72%), which is consistent with other findings showing that maltodextrin, the carbohydrate used in this experiment, has a more potent reinforcing effect than sucrose, the carbohydrate used in expt 1. A pilot study also revealed that rats trained with the two different HC diets developed a preference for the CS flavor paired with the maltodextrin-based diet over the CS flavor paired with the sucrose-based diet (unpublished findings).

	EXPERIMENT 3

Experiments 1 and 2 revealed that, although the magnitude of the preference may differ as a function of the carbohydrate used, the HF diet conditioned a stronger preference relative to both HC diets. Conceivably, the rats may have developed the CS+HF preference not because the HF diet is inherently more reinforcing than the HC diets but because they consumed more CS+HF than CS+HC and were infused with more HF than HC diet because the HF diet was less satiating than the HC diet. Experiment 3 determined whether the amount of exposure to the CS+HF and CS+HC solutions and/or amount of intragastric diet infused is responsible for the CS+HF preference. This was accomplished by limiting the amount of CS consumed and diet infused during one-bottle training. The maltodextrin-based HF and HC diets of expt 2 were used.

Methods. A new group of rats (n = 11) was adapted to the infusion cages. The cages were modified in that the drinking bottles were mounted on retractable bottle holders controlled by the computer. The rats were first given ad libitum access to chow and water for 2 days. Next, they were given an unflavored 0.2% saccharin solution to drink paired with intragastric water infusions during 10 30-min access periods per day. At specified times, the retractable bottle holder brought the drinking spout within reach of the rat, and 30 min later it was retracted out of reach. The first access period was in the light phase (1810-1840), and the remaining nine periods occurred during the night phase, starting at 2010 and then every 80 min. This approximated the rats' natural nocturnal drinking pattern. During the first four saccharin "pretraining" days, saccharin intake during the 30-min access periods was unlimited. For the next 3 days, the rats could only drink a maximum of 3 ml per access period; the amount of water infused intragastrically approximately matched the amount of saccharin consumed. Intake was limited individually for each rat by the computer retracting that rat's drinking bottle after a fixed number of licks (equivalent to 3 ml of intake) had been emitted. The maximum number of access periods (10/day) and maximum amount of total intake (30 ml/day) were selected on the basis of observation that most rats in expt 2 would take at least this many bouts and consumed at least 30 ml/day of the CS+HF and CS+HC solutions. Chow was available ad libitum, but water was not provided to increase the likelihood that rats would drink during the limited-access periods.

Results. The one-bottle training data are summarized in Table 4. Despite the limited-access schedule, the rats differed slightly in their CS intakes [F(2,22) = 4.88, P < 0.05]: CS and CS+HF intakes were similar and modestly greater (P < 0.05) than that of the CS+HC. This small intake difference resulted because the rats took 1.3 fewer bouts per day of the CS+HC than of the CS+HF and CS [F(2,20) = 11.92, P < 0.001 (with all CSs the rats did not take every bout that was available)]. Bout sizes did not vary, and the rats usually reached the limit imposed by the access schedule. The amounts of HF and HC diets infused per day did not differ because of the adjustment in the intake-to-infusion ratios. Chow intakes varied as a function of training condition [F(3,30) = 66.6, P < 0.001]. Relative to the water infusions, the diet infusions reduced chow intakes, although the HF diet reduced chow consumption less (P < 0.05) than did the HC diet. Total caloric intakes varied as a function of training condition [F(3,30) = 73.2, P < 0.001]; caloric intake was greater (P < 0.05) on CS+HF days than CS+HC days, and greater on CS+HF and CS+HC days than CS days. In addition, caloric intakes on CS+HF and CS+HC days exceeded (P < 0.01) the baseline level, whereas intake on CS days was less (P < 0.05) than baseline.

View larger version (47K): [in this window] [in a new window]   Fig. 3.   Intakes (means ± SE) of CS solutions during 22-h, 2-bottle tests of expt 3. Rats were first tested with limited access to CSs as in training (i.e., 10 30-min presentations/22 h): they were given choice of CS+HC vs. CS, CS+HF vs. CS, and CS+HC vs. CS+HF (see text for order counterbalancing). In a final test (right), rats had a choice of CS+HC vs. CS+HF with unlimited access (Ad Lib). Intakes of CS, CS+HC, and CS+HF were paired with intragastric infusions of water, HC diet, and HF diet, respectively. Rats also had ad libitum access to standard chow.

Discussion. The limited-access training procedure greatly reduced the intake differences of the CS+HF and CS+HC (to only 3 ml/day) and eliminated the difference in HF and HC diet infusions. In addition, the bout sizes were similar, and bout frequencies were less different than in expt 2. Yet the rats developed a preference for the CS+HF over the CS+HC that was nearly identical to that seen in expt 2 (63 vs. 64%). This indicates that the CS+HF preferences obtained in expts 1 and 2 were not due to the rats overconsuming the CS+HF and HF diet relative to the CS+HC and HC diet during one-bottle training. Rather, the data suggest that the HF diet is more reinforcing per amount infused than the HC diet. This issue is addressed further in expt 4.

	EXPERIMENT 4

Experiment 4 investigated the possibility that the conditioned preference for the CS+HF resulted from the different satiety potencies of the HF and HC diets. Although it might seem that the food that produces less satiety should also be less reinforcing, recent data argue that just the opposite may be true: too much satiety may limit postingestive reinforcement (12, 27). We therefore hypothesized that reducing the satiating potency of the HC diet, by reducing its caloric density, might enhance its postingestive reinforcing action such that the CS+HC and CS+HF were equally preferred. The rats in expt 4, therefore, were trained with the HF diet used in expt 2 and a diluted version of the HC diet.

Methods. New rats (n = 11) were trained and tested as in expt 2, except that the HC diet (now referred to as Hc) was diluted with water such that its caloric density was 1.4 kcal/ml rather than 2.1 kcal/ml (see Table 1). This amount of dilution was designed to equate the rats' volume intakes of the CS+Hc and CS+HF and of the Hc and HF diets during one-bottle training. Plain water was not available during one-bottle training because expts 1 and 2 revealed that the rats consumed very little water.

Results. The one-bottle intake data are presented in Table 5. Daily CS intakes differed [F(2,20) = 3.63, P < 0.05], in that the rats consumed more (P < 0.05) CS than CS+HF. The intake of the CS+Hc was slightly, but not reliably, greater than that of the CS+HF. Bout frequencies of the CS+HF and CS+Hc were virtually identical, and both were lower (P < 0.05) than that of the CS [F(2,20) = 9.19, P < 0.01]. Mean bout sizes (in volume) of the different CSs did not reliably differ. The rats infused themselves with similar volumes of the HF and Hc diets (46 and 47 ml/day, respectively). Because the HF diet was calorically denser than the Hc diet, more energy was infused on CS+HF than CS+Hc days [t(10) = 10.51, P < 0.001]. Chow intakes did not reliably differ on CS+HF and CS+Hc days and were reduced (P values < 0.001) on CS+HF and CS+Hc days relative to CS and baseline days [F(3,30) = 150.1, P < 0.001]. Total energy intake was greatest on CS+HF days, followed by CS+Hc days, then baseline days, and finally CS days, all differences P < 0.05 [F(3,30) = 263.1, P < 0.001].

View larger version (38K): [in this window] [in a new window]   Fig. 4.   Intakes (means ± SE) of CS solutions during 22-h, 2-bottle tests of expt 4. Rats were given the choice of CS + reduced HC (CS+Hc) vs. CS (left), CS+HF vs. CS (middle), and CS+Hc vs. CS+HF (right) (see text for order counterbalancing). Intakes of CS, CS+Hc, and CS+HF were paired with intragastric infusions of water, Hc diet, and HF diet, respectively; Hc diet was a diluted version of the HC diet of expts 2 and 3.

Discussion. The CS+HF was not reliably preferred to the CS+Hc, which contrasts with the significant CS+HF preferences obtained in the preceding experiments. Admittedly, the nonreliable 58% CS+HF preference obtained here was not that much less than the significant 63-71% CS+HF preferences obtained with the maltodextrin-based diets in expts 2 and 3. Nevertheless, the fact that calorically diluting the HC diet enhanced the preference for its paired flavor relative to the CS+HF flavor is noteworthy. Diluting the HC diet also increased CS+Hc total intake and bout frequency to the levels of the CS+HF; mean bout size was similar between the two diets. Thus on a volume basis the CS+Hc and CS+HF and their associated infusions had similar satiating and reinforcing effects even though the HF infusions provided 50% more energy than did the Hc infusions. Taken together, these findings support the idea that the HF diet has a stronger reinforcing effect than the full-strength HC diet, not because it stimulates greater energy intake but because it produces less satiety.

	GENERAL DISCUSSION

Top Abstract Introduction Methods Discussion References

The present study compares the hyperphagia-promoting and flavor-conditioning effects of HF and HC liquid diets and also provides data related to the satiating actions of the diets. These findings will be discussed in turn.

HF and HC diet-induced hyperphagia. HF diets have often been found to produce excessive energy intake compared with HC diets (24). Typically, the HF and HC diets have differed in flavor, caloric density, and nutrient composition so that the cause(s) of the overeating are not certain. Following the lead of Warwick and Weingarten (26), the present study eliminated the inherent flavor and caloric density differences between HF and HC diets by having the rats self-infuse isocaloric liquid diets into the stomach. Although our design and focus were different (e.g., use of CS flavors, within-group design, availability of chow), we confirmed Warwick and Weingarten's (26) finding that rats self-infused more HF diet than HC diet. In expt 1, given sucrose-based diets, the rats self-infused 36% more HF than HC, which is similar to the 30% difference reported by Warwick and Weingarten using the same diets. An even greater difference (50%) in HF and HC infusions was obtained in expt 2 using maltodextrin-based diets. In terms of total caloric intake (liquid diet and chow), the rats consumed 22-26% more energy on HF days than HC days in expts 1 and 2.

HF- and HC-conditioned flavor preferences. Consistent with many earlier studies using different types of nutrient infusions, the rats in expts 1-4 acquired preferences for the flavors paired with the diet infusions (HF and HC diets) over the flavor paired with water. The new finding here is that in three of the experiments the rats also developed a preference for the flavor paired with the HF diet infusions over the flavor paired with the HC diet infusions. The magnitude of the preference was consistent across experiments (63-72%) and, although moderate, it is significant given that the diets had the same caloric density and differed only in their fat-to-carbohydrate ratio. Furthermore, prior experiments using pure macronutrient infusions indicate that fat is less reinforcing than carbohydrate. In long-term studies, carbohydrate (maltodextrin solution) infusions conditioned stronger flavor preferences and increased flavor acceptance more than did isocaloric fat (corn oil emulsion) infusions (13, 17). Taken together, these data suggest that the relative reinforcing effects of fat and carbohydrate depend on whether the nutrients are presented in pure form or as components of mixed diets.

Hyperphagic, reinforcing, and satiating effects of HF and HC diets. The finding that HF infusions conditioned a stronger preference than did isocaloric HC diet infusions raises the possibility that HF hyperphagia occurs, at least in part, because HF diets have stronger positive postingestive actions than do HC diets. That the postingestive reinforcing effects of nutrients can actually stimulate consumption is demonstrated by reports that rats increase their daily intake of flavored solutions paired with intragastric carbohydrate infusions rather than water (16). Note though that dilute fat infusions are less effective than isocaloric carbohydrate infusions in stimulating flavor acceptance (17), which is consistent with their reduced ability to condition a flavor preference. More concentrated infusions do not increase total intake, but they may have a stimulatory effect that is obscured by the satiating actions of the nutrient infusions (16). In theory, the nutrient's reinforcing action may counteract its satiating action such that the net inhibitory effect on intake is reduced. The present results, however, indicate that the reinforcing effect of the HF diet was not highly correlated with the degree of hyperphagia seen with the diets. This is illustrated in Fig. 5, which plots HF diet overconsumption in one-bottle training as a function of CS+HF preference in two-bottle tests, relative to the HC diet and CS+HC, respectively. The overall correlation using the data from expts 1, 2, and 4 was only 0.057. The r² values obtained in the individual experiments were 0.090, 0.031, and 0.279, none of which were significant. (Data from expt 3 were not included because the diet infusions were restricted.) This analysis suggests that the hyperphagic response to the HF diet was not primarily a result of the diet's postingestive reinforcing action.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Relationship between HF diet overconsumption and CS+HF preference. Percent HF diet overconsumption represents percent increase in energy self-infused on HF 1-bottle days relative to HC (or Hc) 1-bottle days. Percent CS+HF preference represents percent of total intake consumed as CS+HF on CS+HF vs. CS+HC or CS+Hc 2-bottle tests. Data points represent data from individual rats from expts 1, 2, and 4.

Influence of flavor on food intake and preference. In the present study and prior work (26), flavor differences inherent to HF and HC diets were eliminated by having the rats self-infuse the diets when they drank either identical saccharin solutions or flavored saccharin solutions that did not differ in their initial palatability. In both cases, the rats overconsumed the solution paired with the HF diet infusions, which would seem to indicate that flavor stimuli are not important in diet-induced overeating. However, other findings document the importance of flavor in food intake and choice. Warwick and Weingarten (26) observed that rats did not learn to self-infuse liquid diet if plain water rather than saccharin was used as the oral cue. This is consistent with earlier reports that rats do not readily self-infuse liquid diets when required to bar press in the absence of oral cues (10). Thus, flavor cues are essential to maintain ingestion and allow for the expression of diet-induced overeating.

Perspectives