Physiology & Behavior 181 (2017) 10–15
Contents lists available at ScienceDirect
Physiology & Behavior journal homepage: www.elsevier.com/locate/physbeh
Circadian and economic factors affect food acquisition in rats restricted to discrete feeding opportunities
MARK
Dulce M. Minaya, Kimberly L. Robertson, Neil E. Rowland⁎ Department of Psychology, University of Florida, Gainesville, FL 32611-2250, USA
A R T I C L E I N F O
A B S T R A C T
Keywords: Closed economy Unit price Operant behavior Zeitgeber Satiation
The purpose of this study is to examine aspects of operant behavior-modeled economic choice for food in rats in closed economy protocols in which food is available for only a few discrete times per daily 23-h session, designed to emulate clustering of human food intake into meals. In the first experiment, rats performed lever press responses for food pellets in an ascending series of ratios or fixed unit prices (FUP) when food was available for four 40-min food opportunities (FO) per day. Daily intake at low FUP was comparable to ad libitum intakes. Intake declined as FUP increased and was not distributed equally among the four FOs. In particular, the last FO of a session (occurring at about lights on in a 12:12 cycle) was the smallest, even when total intake was low due to the response requirement at high FUP. Within FOs, satiation was evident at low FUPs by a decrease in rate of intake across a 40 min FO; at high FUPs responding was evenly distributed. In the second experiment, rats had a choice of responding on two levers for either intermittent inexpensive (II; low FUP according to a four FO schedule) or costly continuous (CC; 20-fold higher FUP but available throughout 23-h sessions) food. Most (73%) of the rats consistently chose almost all of their food from the II source. Further, as the timing of the four II FOs were changed relative to the light: dark Zeitgeber, the time of the smallest meal changed such that the smallest meal (s) were during the light period regardless of ordinal position within a session. These data are discussed in terms of economic and Zeitgeber effects on consumption when food is available intermittently, and are contrasted with results from comparable protocols in mice.
1. Introduction Starting with the work of Curt Richter [1], rats (Rattus norvegicus) have been the subjects of an overwhelming majority of scientific studies on mechanisms of food intake. Early studies showed that, with free access to food, rats eat in ~10 well-defined episodes or meals per day separated by substantial periods of no eating (intermeal interval, IMI). Further, food intake is higher at night, with shorter IMIs, than during the daytime [1–3]. These studies led to the belief that the meal is the fundamental neurobiological unit of eating [4] and numerous studies have investigated possible physiological mechanisms. This meal pattern is conserved when rats have to perform an operant task for small food pellets in a closed economy using ratio or fixed unit price (FUP) schedules [5–7] although, consistent with economic demand theory, total intakes decline at the highest FUPs [5–10]. In contrast, when the operant cost is instead imposed at the point of obtaining access to food, rats adopt a cost-minimizing or global economic strategy of fewer but larger meals as access cost increases [5,11]. The mechanism behind this change in meal strategy is unknown.
⁎
Corresponding author. E-mail address: nrowland@ufl.edu (N.E. Rowland).
http://dx.doi.org/10.1016/j.physbeh.2017.09.003 Received 3 July 2017; Received in revised form 22 August 2017; Accepted 3 September 2017 Available online 05 September 2017 0031-9384/ © 2017 Elsevier Inc. All rights reserved.
Several reports from our laboratory have described the behavior of mice (Mus musculus) in operant protocols for food acquisition. As in rats, intake of mice declines as FUP increases, often more sharply because mice seem to emit or tolerate a lower maximum number of operant responses per day than rats [12]. In contrast to rats, mice spend at least the first half of the night engaged in almost continuous locomotor activity during which many small but frequent eating episodes or grazing occur [6,14]. Meals, as such, are poorly defined during this period but tend to be better defined in the late night and daytime [6]. These observations led us to investigate what would happen to regulation of intake if mice were forced to take discrete meals, in the limit emulating the typical pattern of humans [15], by enabling them to work for food only during discrete intervals or food opportunities (FOs). Two important findings emerged from that work. First, when food was available in four 40-min FOs spaced 4-h apart, intake was not the same at each FO but showed a characteristic variation that appeared in part to be due to the light-dark cycle and/or an endogenous circadian oscillator [13,16,17]. Second, within an FO, the rate of intake declined monotonically and most individuals stopped responding and eating
Physiology & Behavior 181 (2017) 10–15
D.M. Minaya et al.
5TUM) were dropped into the trough from a dispenser outside the cage. Water was available ad libitum from a sipper spout. A computer running custom programs in Med-PC IV software recorded responses and controlled pellet deliveries in daily 23-h sessions. In experiment 2, the front wall of each chamber was equipped with two levers, each with a cue light above that was illuminated when the lever became operational, located on either side of a the food trough. Other details were as in experiment 1.
before the end of the 40 min [13]. This within-FO behavioral satiation occurred at all FUPs, including at higher costs at which total intake was substantially reduced and weight loss was rapid and would have resulted in death if the animals had not been removed from the experiment. In contrast, when a fixed-interval schedule for earning food was imposed to slow (at higher values) the rate at which food could be procured, pellets were earned evenly across a FO with no evidence of satiation [13]. Thus, a characteristic of higher ratio (FUP) schedules in mice is to produce a maladaptive pattern of early satiation despite inadequate intake. The purpose of the first experiment in this paper is to examine whether rats will exhibit within-FO satiation similar to those shown by mice using ratio schedules. This will address the question of whether there is or is not a significant difference in mechanism or behavioral strategy of satiation between rats and mice in this situation and whether rats, like mice, show circadian variation in FO size. The second experiment examines choice strategy and meal structure of rats as a function of economic factors. To approach this question, rats were allowed continuous access to costly food (i.e., high FUP) intercalated with intermittent short periods of inexpensive food (i.e., low FUP; effectively, cheap FOs). Further, to investigate the role of timing relative to the light: dark (Zeitgeber) cycle on size of successive FOs, this experiment was performed in three segments using standard, phase advanced, or phase delayed timing of the II episodes. In a prior study, we showed that in mice exposed to this experimental protocol, light exerted a suppressive effect on food intake and had a greater modulatory effect than economic choice [17]. We have been surprised by both early satiation and circadian sensitivity findings in mice, given the context of a large and diverse literature in rats. It is thus important to validate this potential species difference using exactly analogous protocols. If the species difference holds, then rats should show less early satiation (exp. 1) and greater economic sensitivity (exp. 2) than mice.
2.3. Experimental procedures 2.3.1. Experiment 1 These animals had previously served in a continuous access operant food study [6] and re-adaptation to the chambers was rapid, achieved in one day during which one food pellet was delivered for every two lever presses (FUP2). At the start of the experimental phase, males and females weighed means of 402 g and 264 g, respectively. Food access was then tapered to 16-h/day for three days, then to four 40-min feeding opportunities (FOs) each 23-h session for six days. FOs started at 1800, 2200, 0200, and 0600-h (FOs #1–4, respectively and ZTs 12, 16, 20, and 00), during which time food pellets were delivered according to the prevailing FUP schedule. The animals were removed from the testing chambers for one hour during the middle of each day (11:30–12:30-h) during which they were weighed and the chambers serviced. During the formal experimental phase, rats were tested for four contiguous daily sessions at a given FUP after which the number of responses required to deliver a food pellet was increased sequentially to 2, 5, 10, 25, 50, 100, and 200. Some animals did not complete the higher FUPs due to weight loss > 15%. 2.3.2. Experiment 2 The goal of this experiment was to examine whether the timing of cheap food relative to the ambient light Zeitgeber (lights out ZT 12-00) affects economic choice. Following a 30-day rest period in their home cage, rats were allowed to re-adapt to the test chambers and the presence of two levers. For initial training, animals were exposed to FUP2 per 45-mg food pellet for two consecutive days followed by an incrementing FUP series (5, 10, 25, and 50 for 1–2 days each) to again adapt the animals to higher food costs. The animals were then tested in three successive segments that differed in the timing of four inexpensive FOs, a protocol previously described by Minaya et al. [17]. Briefly, in each segment, one lever (continuous costly, CC) was operational throughout each 23-h session and delivered one 45-mg food pellet upon completion of 100 responses. The other lever (intermittent inexpensive, II) was operational only during four 15-min FOs and delivered a food pellet upon completion of 5 responses; the shorter inexpensive FOs in this experiment were determined on the basis of results at FUP5 in experiment 1. The relative position of the levers was switched every 2 days to account for and avoid side preference. In segment 1, the inexpensive 15-min FOs (#1–4, respectively) were available on the II lever starting at ZT 13, 17, 21, and 01. In segment 2, the inexpensive FOs were 4-h earlier relative to segment 1 with the II lever operational starting at ZT 09, 13, 17, and 21. In segment 3, the inexpensive FOs were delayed 4-h relative to segment 1, with the II lever operational starting at ZT 17, 21, 01, and 05. Data were collected for eight consecutive days in each segment.
2. Methods 2.1. Subjects The same 6 male and 6 female Sprague-Dawley rats (Harlan/Envigo Labs, Indianapolis IN), initially 4 months old were used in both experiments. They had previously served in an operant study that examined the pattern of food pellet acquisition in daily sessions as a function of price, using an increasing fixed unit price schedule of reinforcement [6]. Between studies, rats were housed individually in polycarbonate cages with free access to standard pellets of Purina 5001 and autoclaved water. Fluorescent vivarium lights were on 06–18 h, ambient temperature was ~24 °C, and relative humidity was ~50%. We did not assess the estrous cycle of females. All procedures were approved by the Institutional Animal Care and Use Committee of University of Florida with the stipulation that rats were removed from the study as soon as body weight loss exceeded 15% of initial. 2.2. Behavior test chambers During experiments, rats were housed individually in operant conditioning chambers (Med Associates, St. Albans, VT) enclosed in ventilated, sound-attenuating cubicles (relative humidity ~ 45%; ambient temperature ~ 23 °C). Indirect light from a 7-W bulb inside the cubicle provided the same 12:12-h light-dark cycle as the vivarium with lights on 0600–1800 h local time, equivalent to Zeitgeber time (ZT) 00–12. Chambers measured 30 × 24 × 21 (L, W, H) cm and were made of Plexiglas with aluminum front and rear panels, and standard rat stainless steel rod floor (~ 1 cm between rods). A paper-lined stainless steel pan was placed below the floor. In experiment 1, the front wall of each chamber was equipped with a single fixed lever located on the left side of a food trough. A cue light was illuminated above the lever when food was available. Grain-based 45-mg food pellets (Purina Test Diet
2.4. Statistical analysis Data are presented as group means with statistical comparison using two-way RM ANOVA and Holm-Sidak post hoc contrast (Sigma Plot); P < 0.05 was used as the criterion for statistical significance. Intakes were not adjusted for spillage from the pan because this was minimal and, in any event, we could not determine when any food spillage occurred within a session. In the first experiment, to ensure analysis of the most stable behavior, data from the last two days at each FUP were 11
Physiology & Behavior 181 (2017) 10–15
D.M. Minaya et al.
Table 1 Food intake and body weight of rats at each fixed unit price (FUP) in experiment 1. FUP
N
Males
N
Pellets/day 2 5 10 25 50 100 200
6 6 6 6 6 5 3
471 484 468 417 341 225 124
± ± ± ± ± ± ±
24⁎ 22⁎ 28⁎ 32⁎ 38 32 33
Body weight (g) 413 417 420 418 403 390 371
± ± ± ± ± ± ±
10 7⁎ 8⁎ 11 13 11 11
Females Pellets/day
5 5 5 5 5 4 2
363 353 323 327 265 195 126
± ± ± ± ± ± ±
22⁎ 17⁎ 12⁎ 24⁎ 37 33 30
Body weight (g) 261 263 261 259 251 239 234
± ± ± ± ± ± ±
5 5⁎ 4⁎ 5 7 11 13
Shown are means ( ± SE) number of 45-mg pellets earned/day averaged for the last two days at each FUP and body weight on the last day of each FUP. As indicated by the Ns, some animals were removed due to weight loss. For the FUPs completed by all animals. ⁎ Indicates value at given FUP > than corresponding value at FUP50 (Ps < 0.05).
averaged and used for analyses. In the second experiment, to allow adaptation to changes in timing of FOs, only data from the last four days of each segment were used for analysis. During segment 1, one female inexplicably stopped responding and those data were not included in analysis. In all three segments of experiment 2, three animals (2 of the 5 females and 1 of the 6 males) showed atypical behavior with regard to amount consumed on the CC lever and/or they exhibited pronounced side preference. Their data were also excluded from the statistical analyses but are presented separately. 3. Results 3.1. Experiment 1 Mean body weights at the end of each 4-day FUP segment are shown in Table 1. All animals were studied through the end of FUP50 but thereafter several were removed due to weight loss > 15%. Females weighed an average of 37% less than males throughout the study and rats of both sexes lost weight as FUP increased. Two-way RM ANOVA on body weight through FUP50 showed a main effect of sex (F(1, 36) = 199.6, P < 0.001) and FUP (F(4, 36) = 3.9, P = 0.01), but no interaction. Post hoc contrasts showed that body weight was significantly higher at the end of FUPs 5 and 10 compared to FUP50. Total daily intakes (i.e., sum of the four FOs) at each FUP are shown in Table 1. Consistent with their higher body weight and an assumed (0.67 exponent) effect on metabolic rate, males consumed ~20% more than females. Intake declined similarly in males and females as FUP increased. The mean intakes at the lowest two FUPs, equivalent to 21.5 g in males and 16.1 g in females, were comparable to those of rats fed ad libitum. Two-way RM ANOVA on total food intake through FUP50 showed a main effect of sex (F (1, 36) = 15.2, P = 0.004) and FUP (F (1, 36) = 10.2, P < 0.001), but no significant interaction. Post hoc contrasts showed that intake was significantly higher at FUPs 2, 5, 10, and 25 compared with FUP50 (Ps < 0.05). Average food pellets consumed in each FO, for the last two days at FUPs 5, 25, 50, and 100, are shown in Fig. 1. Given the sex differences in intake reported above, data for males and females are shown separately. For males, two-way RM ANOVA showed a main effect of FUP (F (4, 60) = 6.6, P < 0.01), FO (F(3, 60) = 26.6, P < 0.001), and a significant interaction (F(12, 60) = 4.0, P < 0.001). During FO #1, intake at FUP5 was significantly higher than at FUP50 (P < 0.05). During FO #2, intake was significantly higher at FUP5 compared to FUP25 and 50 (Ps < 0.05), and at FUP10 compared to FUP50 (P < 0.001). There were no significant differences in intakes at FOs #3 and 4. Within FUP comparisons showed that intakes at FOs #1, 2 and 3 were significantly higher than at FO #4 (Ps < 0.05) for all FUPs. Similar results were observed in female rats with a few exceptions. At FO #2, intakes were significantly higher at FUPs 5 and 10 compared to FUP50 (Ps < 0.05). There were no differences in intakes at FOs #1, 3,
Fig. 1. Number of 45-mg pellets earned at each feeding opportunity in Experiment 1. Shown are mean ( ± SE) for males (top panel) and females (bottom panel). As FUP increased, several animals were removed from the study due to weight loss. Data for FUPs 5–50 and FUP100 is for 6, 5 males and 5, 4 females, respectively. Within an FUP, bars denoted with a symbol differ significantly from FO #4. Between FUPs, bars denoted with a letter differ significantly from FUP50.
and 4. Comparisons within FUP showed that at FUP50, intakes during FOs #2 and 3 were significantly higher than during FO #4 (Ps < 0.001). At all other FUP levels, the profile of intake of females was similar to that of males. Fig. 2 shows mean intakes as a function of contiguous 5-min epochs within each FO and averaged for the last two days at FUPs 5, 25, 50, and 100. At FUP5, during FOs #1, 2, and 3 intake was rapid for the first 10 min and declined gradually thereafter. Intake at FO #4 was very low throughout the 40-min FO. At FUP25, pellet acquisition remained at a constant rate throughout for FOs #1, 2 and 3. Intake at FO #4 was again very low. At higher FUPs, maximum pellet acquisition during FOs #1, 2, and 3 was apparently constrained by the greater time needed to emit the response requirement, yielding a maximum of 14 and 9 pellets per 5 min throughout the session at FUPs 50 and 100, respectively. During FO #4, pellet acquisition was low but with considerable individual variability; some animals earned consistently more pellets than others. A similar pattern was observed in females, albeit with somewhat lower maximum rates than males at the higher FUPs: 9 and 8 pellets per 5 min for FUPs 50 and 100, respectively. 3.2. Experiment 2 Mean body weights at the end of each segment are shown in Table 2. Females weighed an average of 40% less and consumed 30% less than males (Table 1). However, due to the relatively small Ns in this analysis (see below) and individual variability, two-way RM ANOVA on 12
Physiology & Behavior 181 (2017) 10–15
D.M. Minaya et al.
Fig. 2. Mean ( ± SE) number of 45-mg pellets earned in 5min epochs of 40-min feeding opportunities #1–4 at FUPs 5 (top left), 25 (top right), 50 (bottom left), and 100 (bottom right) for male rats in Experiment 1. Data are for 6 and 5 animals in FUPs 5–50 and 100, respectively. A similar acquisition pattern was observed in female rats (data not shown).
Segment
1 2 3
Body weight (g)
Total number of pellets earned/day
II (%)
CC (%)
M
M
M+F
M+F
436 ± 13 441 ± 18 454 ± 19
F 262 ± 7 262 ± 5 264 ± 5
383 ± 40 418 ± 39 380 ± 39
F 298 ± 25 293 ± 11 272 ± 18
93 ± 2 92 ± 5 85 ± 7
Mean % daily intake at each feeding opportunity
Table 2 Body weight, total food intake, and fraction of intake on the intermittent inexpensive (II) and continuous costly (CC) levers for the three Zeitgeber-differing segments in experiment 2.
7 ± 2 8 ± 5 15 ± 7
Shown is mean ( ± SE) body weight on the last day of each segment and intakes averaged over the last 4 days of each segment for 5 males and 3 females.
total food intake showed no significant effect of sex (F(1, 12) = 4.7) or segment (F(2, 12) = 1.0, Ps > 0.05). The fraction of daily intake from the II and CC levers or food sources for the 5 males and 3 females who exhibited consistent behavior (see below) are shown in Table 2. Two-way RM ANOVA on fraction of daily intake with segment and food source as factors showed a main effect of food source (F(1, 14) = 90.3, P < 0.001) but no significant interaction. In all three segments, rats procured a significantly higher fraction of their daily intake from the II lever than on the CC lever. Mean fractions of daily intake at each FO from the II level are shown in Fig. 3. Two-way RM ANOVA showed a main effect of FO (F(3, 42) = 5.8, P < 0.01) and a significant interaction between FO and segment (F(6, 42) = 9.1, P < 0.001). In all three segments, intake was highest during the FOs at ZT 1700 and 2100. In segment 1, intakes were significantly higher at FOs #1, 2, and 3 than at FO #4 (Ps < 0.05). In segment 2, the animals ate significantly more at FOs #3 and 4 than at FOs #1 and 2. Segment 3 was almost a mirror image of segment 2, with intakes at FOs #1 and 2 being significantly higher than at FOs #3 and 4. The fraction of daily intake obtained on the CC lever was relatively small during all segments and pellet acquisition on this lever typically occurred during the dark phase of the light: dark cycle. Individual data for rats exhibiting atypical or inconsistent behavior (1 male, 2 females) are shown in Table 3. Rats F1 and M1 (female and male, respectively) exhibited a marked side preference, earning food
40 a a * # * 30
Segment 1 Segment 2 Segment 3
a * a #
b * b
20
b
c b
10
0 0900
1300
1700
2100
0100
0500
Zeitgeber time Fig. 3. Mean ( ± SE) fraction of daily food intake on the intermittent inexpensive (II) lever in Experiment 2. Data are shown for 5 males and 3 females. In segment 1, inexpensive food was available at ZT 1300, 1700, 2100, and 0100. In segment 2, inexpensive food was available at ZT 0900, 1300, 1700, and 2100. In segment 3, inexpensive food was available at ZT 1700, 2100, 0100, and 0500. The horizontal line indicates the dark period (ZT12–24). Within a segment, bars denoted with a symbol differ significantly, * > FOs 3, 4, # > FO1. Between segments, FOs denoted with different letters differs significantly. Ps < 0.05.
pellets on only one of the two levers regardless of the prevailing FUP on that lever. Rat F2, female, earned pellets only on the CC lever except on the last day of segment 3. These three rats account for 27% of the cohort studied and their choice behavior was not logically controlled by the economic conditions; it is important to document a substantial incidence of such off-schedule behavior in rats that were otherwise wellaccustomed to one lever economic conditions (e.g., experiment 1) and during which their behavior was not atypical. 4. Discussion The principal aim of Experiment 1 was to describe the behavior of 13
Physiology & Behavior 181 (2017) 10–15
D.M. Minaya et al.
first 3 FOs. However, this does not explain the small FO #4 (~ 30 pellets or 1.3 g) at FUPs 50 or 100 when intake from FOs #1–3 totaled only ~300 and 200 pellets or 13.5 and 9.0 g, respectively. Homeostatic explanations would predict that, in an attempt to conserve intake and obtund weight loss, intakes at FO #4 should have been as high as allowed by the ambient FUP. Instead, since FO #4 was the only FO to occur during the light phase (at lights on), a reasonable interpretation is that light and/or endogenous periodicity is responsible for low responding and intake, a position that is supported in part by the results of experiment 2. A similar conclusion about light inhibition was reached in a somewhat different food operant protocol in rats [22]. It is informative to compare these results in rats with corresponding data we have reported using mice in a temporally similar 4 FO protocol [13]. In those studies, mice performed nose poke responses for 20-mg pellets. A normal daily complement of such pellets for a mouse is ~200, compared with more than double that number of 45-mg pellets in rats (Table 1), meaning that rats must work for ~2-fold more food pellets than mice to achieve a proportional intake. Mice show greater apparent cost sensitivity: average daily intake fell to ~50% of that at FUP2 near FUP100 in rats (Table 1) compared with near FUP50 in three mouse studies using a similar 4FO protocol [13,16,17]. To compare within-FO eating between rats and mice in these protocols, we took data from FUP5 and 50 in rats (Fig. 2) and from FUP5 and 25 in mice [13]. We chose these values because, in both species, daily intake was maximal at FUP5 and had dropped to 70–80% of that maximum at the higher FUP. Pellets earned in each 5 min segment of the 3 largest FOs (#1–3) were then expressed as % of daily intake at that FUP i.e., 340 and 480 for FUP5 and 50 in (male) rats; 200 and 160 for FUP5 and 25 in (male and female) mice. The results show that at FUP5 (Fig. 4, left panel) both rats and mice earned ~7% of their daily intake in the first 5 min of a FO. That fraction declined quite rapidly in mice, to a low asymptote after 15–20 min, whereas in rats the rate declined more gradually and without a clear inflection point. At the higher FUP (Fig. 4, right panel), mice again showed a marked drop in response or eating rate after 15–20 min, whereas rats showed no such decline. For a given reduction in food intake, relative weight loss in mice exceeded that of rats: for the proportionally matched intake data used in Fig. 4, rats lost ~1% body weight by the end of FUP50 whereas mice lost ~ 10% by the end of FUP25 [13]. It might be expected that this more rapid weight loss, a consequence of higher mass-specific metabolic rate of smaller animals [23], would lead to a more aggressive defense of food intake as cost increases in mice than rats, but the opposite was observed. Mice appear to be particularly vulnerable to
Table 3 Number of 45-mg pellets earned by subjects that exhibited atypical behavior in Experiment 2. Subject
F1
F2
M1
Segment
1 2 3 1 2 3 1 2 3
Day 6
Day 8
Total intake
CC/Total
Total intake
CC/Total
298 434 2 277 315 397 423 541 544
100 0 100 100 100 100 0 100 100
365 209 350 294 362 305 521 558 482
0 100 0 100 100 20 100 0 0
Shown are total daily intake and percent consumed on the CC lever. Days 6 and 8 represent the last testing day on which CC food was available on the left and right lever, respectively.
rats in the 4FO protocol. Many studies have shown that rats adapt well to quite severe schedules of daily food restriction (e.g. [18].) and, as expected at low FUPs, rats in the present study restricted to a total of 160 min access per day had intakes typical of ad libitum fed rats. In this operant protocol, the minimum time between pellets is the time to complete the FUP plus the time spent handling and consuming the pellet. Assuming that the maximum rates observed in Fig. 2 represent continuous food-directed behavior for an entire 5 min epoch, then simultaneous equations of the form.
(N × FUP × Tr) + (N × Th) = 300 (N = number of pellets in epoch, Tr = time per response, and Th = time to handle/consume pellet) can be solved for pairs of FUPs. For FUPs 10 and 25, these solve to Tr = 0.32 s/response and Th = 6.3 s, while FUPs 50 and 100 solve to yield Tr = 0.24 s/response and Th = 1.5 s. These estimates of response time are comparable to those reported by others [10] and show a modest decrease (i.e., increased rate of responding) across FUP. In comparison, the 4-fold decrease in handling time is large. Higher temporal resolution data collection and/ or video assessment would be needed to confirm the underlying assumption of continuous on-task behavior but, especially at high FUP when these rates are sustained for the entire 40 min, it is clear that intakes become limited by time available. Given this analysis, two aspects of the data are of particular note. First, at FUP5 and 25 the mean intake within each of FOs #1–3 decreased with elapsed time. This was most pronounced at FUP5 when intake in the first 20 min was ~100 pellets (4.5 g) while intake in the last 20 min was about half this. This observation is consistent with descriptions of the behavioral satiety sequence of rats or mice eating solid food [19,20] in which behaviors such as grooming or other activity occur with higher frequency as the session proceeds until, in the end, resting behavior prevails. Satiation behavior in our operant situation reflects longer response and food handling time and/or prolonged post-pellet pauses in responding. The 5-min resolution of our data does not allow us to distinguish these alternatives but the latter interpretation seems more likely given that mice almost invariably complete response runs without interruption [21]. Also, comparing published behavioral satiety sequence data for rats [19] and mice [20] suggests that mice show a faster decline in time spent eating and relatively more in grooming or other activity quite early in a test meal. In contrast to the satiation profile for rats at low FUP, at FUP 100 a maximum rate of food acquisition was sustained and yielded ~3 g over an entire 40-min FO. The second noteworthy aspect is the characteristic profile of size in successive FOs. FO #2 was slightly larger than FOs #1 and 3, but FO #4 was always smaller. This was particularly marked at FUP5 and may be because sufficient daily food (~ 20 g in males) was taken during the
% of total daily intake in each 5-min epoch
8
Low unit price 7
High unit price
Mouse FUP 5 Rat FUP 5
6
Mouse FUP 25 Rat FUP 50
5 4 3 2 1 0 5
10 15 20 25 30 35 40
5
10 15 20 25 30 35 40
Time (min) within feeding opportunity
Fig. 4. Group mean pellet intakes (Ns = 5–8) in successive 5 min epochs within a 40-min feeding opportunity of rats (derived from present data) and mice (derived from [13]). The left panel shows data at low fixed unit price (5 responses per 20- or 45-mg pellet in mice and rats, respectively) and the right panel shows data at high fixed unit price (25 and 50 responses per pellet in mice and rats, respectively). Data are averages from the largest 3 FOs within a day and are expressed as % of the total daily intake at that unit price.
14
Physiology & Behavior 181 (2017) 10–15
D.M. Minaya et al.
until sufficient responses can no longer be emitted in the time available. In a choice situation, rats elect to compress almost all of their food intake into inexpensive but brief opportunities. These findings have modest differences from those in mice, whose intake and choice behavior does not meet theoretical maxima or apparent optimal choice.
serious and eventually fatal weight loss by exhibiting what we have termed elective or cost-based anorexia [12]. In the second experiment, rats were presented four 15-min FOs at low cost (II) concurrent with continuous access to the same commodity at 20-fold the cost (CC). As noted in the discussion of experiment 1, rats earned ~100 pellets in the first 15 min of each of FOs #1–3 at FUP5 (Fig. 2), so we anticipated that rats could earn up to 400 pellets from II in the present experiment. The data in Table 2 indicate that, averaged across segments, males and females took 354 and 259 II pellets, respectively, although these were not evenly distributed across FOs (Fig. 3). Regardless of whether these amounts do or do not reflect ceilings imposed by the experimental conditions, it can be stated that most (8/11) rats elected to compress their daily intake into the four II episodes, and take relatively little (~ 10% on average) from the constantly-available but 20-fold more expensive food source (CC). This result is consistent with results of operant food acquisition in rats when cost was imposed to gain access to the food [5]: in those studies, even a small access or procurement cost caused rats to shift from a pattern of small and frequent meals to one of large and infrequent meals, thereby reducing the number of access response strings they would have to emit per day. Thus, the present and Collier's [5,11] results are consistent with a strategy in rats of maximizing energy gain relative to foraging expenditure. We recognize, however, that if the experimental conditions had been less favorable (e.g., higher FUP on the II lever), rats may have been forced to take a larger fraction from the CC lever. Intake profile on the II lever (Fig. 3) was similar to the profile of FO size observed in experiment 1 at FUP5 (Fig. 1). In segment 1, II FOs that occurred in the middle of the night were larger than those that occurred during the light phase. In segment 2, FO #1 was in the light and was the smallest II meal. In segment 3, FOs #3 and 4 were in the light and were the smallest II meals. These data indicate that light and/or the consequences of endogenously entrained circadian physiological oscillations exert a partial suppressive effect on operant food intake. These data are consistent with results examining Zeitgeber shifts in relation to II food in mice [17] except that the suppressive effect in mice seems to be much more pronounced than in rats. Also, the fraction of food earned on the II lever was an average of only 46% (compared with ~ 90% in the present rats). This appears to be due to the fact that, aside from light-related factors, mice do not sustain high rates of responding for > 15–20 min even at low FUP (Fig. 4) and thus cannot or do not eat more than ~50% of their intake from the II lever in the time it is available. This could be due to lower maximum rates of responding than in rats and/or to longer post-pellet pauses which presumably include other behaviors such as grooming or activity. Concurrent measures of activity and food in free-feeding mice support the latter interpretation [13,14]. In summary, rats adapt operant foraging and consumption behavior to restricted access times in which they eat adequate daily amounts
References [1] C.P. Richter, Animal behavior and internal drives, Q. Rev. Biol. 2 (1927) 307–323. [2] J. Le Magnen, Hunger, Cambridge University Press, 1985. [3] S. Balagura, S.D.V. Coscina, Periodicity of food intake in the rat as measured by an operant response, Physiol. Behav. 3 (1968) 641–643. [4] G.P. Smith, The controls of eating: a shift from nutritional homeostasis to behavioral neuroscience, Nutrition 16 (2000) 814–820. [5] G. Collier, E. Hirsch, P.H. Hamlin, The ecological determinants of reinforcement in the rat, Physiol. Behav. 9 (1972) 705–716. [6] N.E. Rowland, D.M. Minaya, M.R. Cervantez, V. Minervini, K.L. Robertson, Differences in temporal aspects of food acquisition between rats and two strains of mice in a closed operant economy, Am. J. Phys. Regul. Integr. Comp. Phys. 309 (2015) R93–108. [7] R.B. Kanarek, G. Collier, Patterns of eating as a function of the cost of the meal, Physiol. Behav. 23 (1979) 141–145. [8] S.E.G. Lea, The psychology and economics of demand, Psychol. Bull. 85 (1978) 441–466. [9] S.R. Hursh, Behavioral economics, J. Exp. Anal. Behav. 42 (1984) 435–452. [10] R. Bauman, An experimental analysis of the cost of food in a closed economy, J. Exp. Anal. Behav. 56 (1991) 33–50. [11] G. Collier, A functional analysis of feeding, Adv. Study Behav. 35 (2005) 63–103. [12] N.E. Rowland, D.M. Minaya, K.L. Robertson, Restricted temporal access to food and anorexia: modelling systems, in: V.R. Preedy, V.B. Patel (Eds.), Handbook of Famine, Starvation, and Nutrient Deprivation, Springer International Publishing, 2017, , http://dx.doi.org/10.1007/978-3-319-40007-5_53-1. [13] N.E. Rowland, M. Cervantez, K.L. Robertson, Restricted temporal access to food and anorexia in mice: microstructure of eating within feeding opportunities, Appetite 96 (2016) 621–627. [14] E.H. Goulding, A.K. Schenk, P. Juneja, A.W. MacKay, J.M. Wade, L.H. Tecott, A robust automated system elucidates mouse home cage behavioral structure, Proc. Natl. Adad. Sci. USA 105 (2008) 20575–20582. [15] J.M. DeCastro, The control of food intake in free-living humans: putting the pieces back together, Physiol. Behav. 100 (2010) 446–453. [16] D. Atalayer, N.E. Rowland, Effects of meal frequency and snacking on food demand in mice, Appetite 58 (2012) 117–123. [17] D.M. Minaya, N.E. Rowland, K.L. Robertson, Effect of day-night cycle on distribution of food intake and economic choice among imposed food opportunities in mice, Physiol. Behav. 164 (2016) 395–399. [18] N. Rowland, Feeding patterns in rats on restricted access schedules: palatability, bulk, and other determinants of intake, Bull. Psychon. Soc. 5 (1975) 306–308. [19] J.C.G. Halford, S.C.D. Wanninayake, J.E. Blundell, Behavioral satiety sequence (BSS) for the diagnosis of drug action on food intake, Pharmacol. Biochem. Behav. 61 (1998) 159–168. [20] B.C. Finger, T.G. Dinan, J.F. Cryan, Behavioral satiety sequence in a genetic mouse model of obesity: effects of ghrelin receptor ligands, Behav. Pharmacol. 22 (2011) 624–632. [21] N.E. Rowland, A.M. Giddings, V. Minervini, K.L. Robertson, Economics of food intake in mice: energy yield of the reinforcer, Physiol. Behav. 136 (2014) 104–110. [22] R.A. Bauman, G.J. Kant, Time cost of alternation reduced demand for food in a closed economy, Physiol. Behav. 57 (1995) 1187–1193. [23] A.J. Hulbert, P.L. Else, Basal metabolic rate: history, composition, regulation, and usefulness, Physiol. Biochem. Zool. 77 (2004) 869–876.
15