“Tasting and wasting” behavior in non-human primates: Aberrant behavior or normal behavior in “times of plenty”

Physiology & Behavior 89 (2006) 587 – 597

“Tasting and wasting” behavior in non-human primates: Aberrant behavior or normal behavior in “times of plenty” Richard W. Foltin ⁎ Division on Substance Abuse, New York State Psychiatric Institute, 1051 Riverside Drive, Unit 120, New York, NY 10032, USA Department of Psychiatry, College of Physicians and Surgeons of Columbia University, 1051 Riverside Drive, Unit 120, New York, NY 10032, USA Received 7 March 2006; received in revised form 17 July 2006; accepted 18 July 2006

Abstract The purpose of this study was to develop a foraging model that engenders large meals. Eight free-feeding baboons were first given periodic access to a chocolate sugar-coated candy (M & Ms®) and then a jelly sugar-coated candy (Skittles®). Baboons had access to food 24 h each day, but they had to complete a two-phase operant procedure in order to eat. Responding on one lever during a 30-min appetitive phase was required before animals could start a consumption phase, where responding on another lever led to food delivery, i.e., a meal. 3 days a week for 8 or 9 weeks baboons received candy during the first meal and then food pellets were available: a 2 month interval when only pellets were available separated periods of candy access. All baboons ate as much candy in the single candy meal as they did pellets throughout the remainder of the day. Beginning week 5 of M & M® access, five baboons began to waste a large number of M & Ms® by spitting them out. Baboons wasted few Skittles® or pellets. Pellet intake was less, but total caloric intake was greater on days that animals had access to either candy. Pellet, but not candy eating varied between males and females: males began eating pellets sooner in the day, ate more pellet meals and more pellets. Periodic access to a preferred candy food engendered large amounts of candy consumption in all baboons, and periodic access to M & Ms® engendered food tasting and wasting behavior in 5 of 8 baboons. © 2006 Elsevier Inc. All rights reserved. Keywords: Motivation; Foraging; Food intake; Sugar; Preferred food; Sex differences; Non-human primate; Baboon

1. Introduction In the wild, non-human primates spend a considerable amount of time foraging for food [1]. Foraging consists of time spent moving from one food source to another, time spent acquiring food within a food source, and a comparably small amount of time actually eating [2]. In a pioneering set of studies, Collier developed models for studying foraging behavior of a wide range of species in a controlled laboratory setting [3,4]. We have adapted these procedures in order to study foraging behavior of a large non-human primate, the baboon [5]. Previous studies have focused on how environmental and pharmacological manipulations affect feeding behavior related to a standardized laboratory diet [6]. In the natural ecology, however,

⁎ New York State Psychiatric Institute 1051 Riverside Drive, Unit 120 New York, NY 10032, USA. Tel.: +1 212 543 5717; fax: +1 212 543 5991. E-mail address: rwf [email protected]. 0031-9384/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.physbeh.2006.07.018

both human and non-human primates do not eat the same food at every meal. The first purpose of this study was to determine how periodic access to a highly-preferred food i.e., a patch of calorically dense food, affected the distribution and size of meals of the preferred item and the standard food pellet diet. In humans, the occasional consumption of a large meal characterizes several eating disorders. For some, the consumption of preferred foods becomes a ritualized pattern of excessive intake [7] known as a “binge.” Consumption of a binge can be followed by “inappropriate” compensatory behavior such as vomiting, or not. The presence of such compensatory behavior is characteristic of bulimia nervosa [8]. The absence of such compensatory behaviors is characteristic of binge-eating disorder [8]. Binge-eating disorder is behaviorally defined by the recurrent consumption of large amounts of food in a brief time frame, and psychologically characterized by feelings of loss of control over eating, and feeling distressed after overeating [8]. Approximately 2% of adult Americans have binge-eating disorder [9], with similar rates in males and females [10].

588

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597

Approximately, 1 to 2% of adult female Americans have bulimia [11]; bulimia rarely occurs in males. The second purpose of this study was to develop a model of foraging behavior that resulted in an occasional large meal. Once a stable pattern of large meals is observed in the laboratory it will be possible to use the model to evaluate potential behavioral and pharmacological interventions for decreasing “binge” eating behavior in laboratory animals [12]. Corwin and colleagues have developed a model of excessive eating of a single food in the rat based on limiting access to a preferred food, but not limiting access to chow [13]. Rats given access to fat for 2 h/day (2 h before the dark cycle) on only 3 days per week develop a bingetype eating pattern of fat intake during those 2 h [14,15]. Human children with restricted access to preferred foods also overeat when preferred foods are available [17]. Thus, restricted access to a preferred food leads to large intake of that food in a short time frame. In this study, we limited access to preferred food to a single meal in the morning 3 days a week for 8 or 9 weeks. Baboons were not food deprived and had access to food pellets for the remainder of the day. We had originally planned to assess eating only when baboons had access to chocolate sugar-coated candy (M & Ms®). M & Ms® were chosen because they contain both fat and sugar (38% of calories derived from fat). Because five animals developed an unexpected eating behavior when they had access to M & Ms®, we then decided to assess a jelly sugar-coated candy (Skittles®; 10% of calories derived from fat). The third purpose of this study was to determine if a food's reinforcing efficacy predicted how much of the food would be consumed during a single meal. We used a progressive ratio (PR) schedule [18], known to be sensitive to reinforcer magnitude, to assess the reinforcing efficacy of the preferred foods and the maintenance pellets before and after access to the preferred food. A PR procedure increases the response requirement for reinforcement after each reinforcer delivery [19]; it is assumed that a larger PR breakpoint (i.e., last completed ratio value) indicates a more efficacious reinforcer [20]. For example, increasing the concentration of sucrose increases the reinforcing efficacy of sucrose in rodents as defined by an increase in PR breakpoint [21,22]. We hypothesized that the initial reinforcing efficacy would predict subsequent intake. The fourth purpose of this study was to compare how male and female baboons behaved in response to having preferred food available on a limited basis. Although eating disorders are more common in human females than males [23,24,14,16], many studies have reported excessive food intake by both male and female rats. For this reason, we hypothesized that sex would not be a factor in determining eating of the preferred candy. In contrast, we hypothesized that males, based on their larger size, would eat more food pellets than females. 2. Material and methods 2.1. Animals and housing Four male baboons (Papio cynocephalus anubis), weighing 9.5 to 13.5 kg, and four female baboons, weighing 7.9 to 10.8 kg,

were individually housed in standard non-human primate cages (0.94 × 1.21 × 1.52 m high) at The New York State Psychiatric Institute. All baboons were sexually mature, i.e., females had menstrual cycles, but not full grown. The baboons had 13 months experience responding for food under the operant schedule used in this study. The room was illuminated with fluorescent lighting from 7:00 AM to 7:00 PM daily. The only food that baboons received daily, other than pellets delivered during the session, were two chewable vitamins (“Kiddy Chews,” Schein Pharmaceutical, Inc., Port Washington, NY), two pieces of fresh fruit (80–100 kcal, each), and a dog biscuit (150 kcal, Old Mother Hubbard, Inc., Lowell, MA). Water was available ad libitum from a spout located at the back of each cage. All aspects of animal maintenance and experimental procedures complied with the U.S. National Institutes of Health Guide for Care and Use of Laboratory Animals, and were approved by the New York State Psychiatric Institute Animal Care and Use Committee. 2.2. Apparatus A response panel holding, from bottom to top, a food hopper, two Lindsley levers spaced 0.30 m apart (Gerbrands, Arlington, MA), two stimulus lights above each lever, two stimulus lights above the food hopper, and a pellet dispenser (BRS-LVE model PDC-005, Beltsville, MD) was attached to the front of each cage. All schedule contingencies were programmed using Pascal on Macintosh (Cupertino, CA) computers located, along with the interface, in an adjacent room. 2.3. Schedule of reinforcement Responding under each phase of a two-phase chain schedule of reinforcement was on a separate response manipulandum. The session began with the illumination of a single light above the appetitive lever. The first response on the appetitive lever began a 30-min timer and illuminated a second light over the appetitive lever, i.e., the 30-min appetitive phase was indicated by the illumination of two lights above the appetitive lever. The appetitive phase was a fixed-interval (FI) 30 min schedule, with a FR 10 second-order phase [FI 30′ (FR 10:S)]. Thus, after every 10th response during the FI phase, the stimuli associated with reinforcer delivery during the second phase were presented. There was a 10 min limited hold for the appetitive phase, such that after the expiry of the 30 min FI, the next FR 10 had to be completed within 10 min. Failure to complete a FR 10 within 10 min canceled that appetitive phase, and extinguished one light over the appetitive lever such that only a single light was illuminated over the appetitive lever. The baboon received no indication that the 30-min interval had elapsed. The first FR 10 completed after 30 min resulted in the two lights above the left lever being extinguished and a single light above the right lever being illuminated, signalling the availability of food under the FR consumption phase of the chain schedule. The consumption phase of the chain schedule was maintained under a FR 10 schedule of food reinforcement. After a 10-min interval in which no responses occurred, the consumption phase terminated, i.e., meal size was determined by each baboon. The single light

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597

above the right consumption lever was then extinguished, and the single light above the left appetitive lever was again illuminated. In order to initiate another meal, the baboon was required to start another 30-min appetitive phase by pulling on the left lever. This schedule was in effect 24 h/day beginning at 9:00 AM, with the exception of a brief period during which the data were backed up and printed (∼ 5 min), which occurred at 8:55 AM each morning. During each regular-diet meal, baboons received 1 food pellet (banana-flavored 1-g food pellets containing 3.3 kcal/g: 0.55 g carbohydrate, 0.03 g fat, 0.2 g protein; Bio-Serv, Frenchtown, NJ). Pellet delivery was accompanied by the illumination of all 4 stimulus lights above the 2 levers for 8 s. The illumination of the 4 lights for 8 s also occurred upon completion of each FR 10 during appetitive phases when food was available. During the M & M® candy meal, baboons received 1 plain chocolate M & M® (Mars Corp., Hackettstown, NJ; 4.4 kcal: 0.6 g carbohydrate, 0.2 g fat, 0.04 g protein); 38% kcal derived from fat. During the Skittle® meal, baboons received 1 Skittle® (Mars Corp., Hackettstown, NJ; 4.3 kcal: 0.9 g carbohydrate, 0.04 g fat, 0 g protein); 10% kcal derived from fat. While the M & Ms® come in 6 colors, there is only 1 flavor. In contrast, there are 5 flavors of Skittles®. Candy delivery was accompanied by the flashing (1 s on:1 s off) of 2 white stimulus lights located above the food hopper for 8 s. The flashing of the 2 white lights for 8 s also occurred upon completion of each FR 10 during the appetitive phase when candy was available. Under these behavioral requirements when pellets are available, baboons typically eat a meal in the morning after the room lights are illuminated at 7:00 AM. A second meal usually occurs about 11:00 AM, a third meal usually occurs in the late afternoon, and some animals have a fourth meal after 5:00 PM. Unlike rodents, baboons rarely eat when the room is dark. 2.4. Daily sessions 4 days a week (Tuesday, Thursday, Saturday, and Sunday), only food pellets were available. On the other 3 days each week (Monday, Wednesday, and Friday), daily sessions began with a single candy meal. M & M® candies were tested first: M & M® candy-before-pellets and pellet-only sessions continued for 9 weeks. Baboons were free to start responding for pellets or candy beginning at 9:00 AM. Completion of the first 10 responses on the appetitive lever started the appetitive phase, which lasted a minimum of 30 min and a maximum of 40 min. After completion of the appetitive phase, baboons could earn as many pellets or pieces of candy as they wanted (1 delivered after every 10 responses) with the meal ending when the baboon stopped pulling the lever for 10 min. On pellet days baboons could have as many meals as they wanted over 24 h, but they had to complete a 30 min appetitive phase before each consumption phase. On candy days, after the end of the candy meal, the baboons then could work for as many pellet meals as they wanted until 9:00 AM the following morning. If a baboon did not have a candy meal in 4 h, then access to candy terminated and access to food pellets began. There were no stimuli indicating if the first meal of the day would be candy or pellets until the first

589

stimulus presentation during the appetitive phase, i.e., light flashes indicated a candy meal would occur after completion of the appetitive phase, and prolonged illumination of a different set of lights indicated a pellet meal would occur after completion of the appetitive phase. Approximately 2 months after completion of testing with M & Ms®, baboons were then tested with Skittles®: Skittle® candy-before-pellets and pellet-only sessions continued for 9 weeks. Finally, baboons were then given access to M & Ms® for 2 weeks in order determine if eating Skittles® altered the response to M & Ms®. This retest of M & Ms® occurred 6 months after M & Ms® were originally tested. 2.5. Assessment of reinforcing efficacy Reinforcing efficacy of pellets and candy was assessed using PR test sessions, during which the PR breakpoint, i.e., the largest completed ratio, for candy or pellets was determined. PR test sessions started at 9:00 AM, but rather than starting the acquisition interval timer, completion of the first 10 responses resulted in the delivery of a single reinforcer, i.e., 1 piece of candy or 1 pellet. Another reinforcer was available after a 15-s time out, but that piece of candy or pellet required 10% more lever pulls, and so on (e.g., 10, 11, 12, 13, 14, 16, 17, 19, 21, 23, 25, 28...). The session continued for 3 h (long enough for all animals to have reached a breakpoint), and was followed by a 21 h pellet session. Because the data collection program would only work with 4 baboons at a time, PR test sessions occurred on Monday and Thursday for 4 baboons and Tuesday and Friday for the other 4 baboons. PR breakpoints for candy were determined approximately 2 weeks before, and 2 weeks after the 8 or 9 weeks of access to candy 3 days a week. During the PR tests, candy was available only during the 2 PR test sessions each week. PR breakpoints for a pellet were determined approximately 4 weeks before and 3 weeks after the 8 or 9 weeks of candy access. Data from the 2 sessions each week were averaged to provide a single breakpoint value for each animal for each week. 2.6. Statistical analysis 2.6.1. Measures of candy and food pellet intake The total number of stimulus presentations, e.g., lights during appetitive phases and total number of pellets or candies earned (accompanied by stimulus presentations), number of pellet meals (consumption phases), the latency to the first candy and pellet meal (including the time required to complete the first meal appetitive phase), and the running rate (response/s timed from the first response after reinforcer delivery to the next reinforcer delivery) during the first appetitive and consumption phases for candy and pellets of each session were calculated. Total calories consumed each day were estimated based on the caloric content of the pellets earned that day plus the caloric content of the candies eaten that day (on days when candy was available). 2.6.2. Wasters vs. eaters During the initial 3 1/2 weeks of M & M® candy access some baboons began to waste small numbers of candy each day. When

590

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597

it became clear that this was a pattern of some animals, not a random wasting of a few candies, the number of uneaten candies were counted each day after the candy meal ended. The total number of candies earned, the total number of candies wasted and total number of candies eaten (earned minus wasted) during the candy meal were determined each day. Beginning in week 6 a natural dichotomy became obvious such that 3 baboons wasted few candies (b 25) and 5 baboons wasted a large number of candies (N 100). For this reason the data were analyzed using the amount wasted as a grouping variable with 5 animals in one group and 3 in the other group. Three of the 5 “wasters” were males. There was 1 male and 2 female “eaters”. This tasting and wasting was not observed with Skittles®. The primary data analyses compared candy and pellet consumption between the wasters and eaters separately for M & Ms® and Skittles®. Candy data were summarized using analyses of variance (ANOVA) with 1 between-groups factor (wasters vs. eaters) and 2 within-animal factors. The first within-animal factor was the number of weeks of access to candy (8 for Skittles®, 9 for M & Ms®), and the second was day within week (3). The number of M & M® candies wasted data was only analyzed for weeks 5–9 when accurate counts of wasted candy were obtained. Pellet data were summarized using ANOVAs with 1 between-groups factor (wasters vs. eaters) and 3 within-animal factors. The first withinanimal factor was the number of weeks of access to candy (8 or 9), the second factor was presence of candy (Yes, No), and the third factor was day within week (3; Saturday pellet data were not used in the analyses as the animal cages were cleaned on Saturdays). When a data point was missing, for example, due to mechanical problem, data from an adjacent day under that condition was substituted for the missing data point. Each baboon had 4 to 8 missing data points during the 9 weeks of M & M® access (out of a total of 54); with the exception of a TB test day, data were collected on most of these days, but no summaries were available due to computer or human error. Between weeks 4 and 5 of Skittles® access, there were 10 days when data were not collected due to electrical problems in the laboratory. During this time baboons were given food pellets daily and on Mondays, Wednesdays and Fridays they were given Skittles® in amounts similar to those earned during the previous week. We treated this interval as equivalent to 1 week of M & M® access. Progressive ratio breakpoints were summarized using ANOVAs with 1 between-groups factor (wasters vs. eaters) and 2 within-animal factors. The first within-animal factor was type of food (Skittle®, M & M®, food pellet), and the second was the time of observation (before vs. after candy access). Correlations between PR breakpoint and intake of candy or food pellets during the first meal of the day were calculated using simple regression analyses: one data point was classified as an outlier, i.e., made a greater contribution than the other data points to the regression, and not used in the analyses. 2.6.3. Females vs. males The secondary data analyses compared candy and pellet consumption between males and females during the last week of access to candy. Candy data were summarized using analyses of variance (ANOVA) with 1 between-groups factor (males vs.

females) and 2 within-animal factors. The first within-animal factor was the type of candy (Skittles®, M & Ms®), and the second was day within week (3). Pellet data were summarized using ANOVAs with 1 between-groups factor (males vs. females) and 3 within-animal factors. The first within-animal factor was the type of candy (Skittles®, M & Ms®), the second was presence of candy (Yes, No), and the third factor was day within week (3). Results for all ANOVAs were considered significantly different at P b 0.05, using Huynh–Feldt corrections. 3. Results 3.1. M and M® wasters compared to M and M® eaters 3.1.1. Candy deliveries The top panels of Fig. 1 compare the number of candies earned between the M & M® wasters and eaters as a function of week of access to candy. As shown in the upper left panel, M and M® intake nearly doubled between the first and second day of candy access then remained stable for the first 4 weeks. The wasters increased the number of candies earned during week 5, while the number of candies earned by the other 3 baboons remained stable throughout the 9 weeks. There was a significant effect of group with wasters earning more candies than the eaters each day [F (1, 6) = 6.1, P b 0.048]. There was also a significant effect of week with the number of M & Ms® earned increasing from week 1 to week 9 [F (8, 48) = 5.6, P b 0.0001]. A significant week by group interaction confirmed that the increase in candies earned was accounted for by increased candy earnings in the group that wasted candy [F (8, 48) = 7.4, P b 0.0001]. We had originally planned for 8 weeks of M & M® access, but because there was a decrease in the number of M & Ms® earned during week 8, we extended the period of M & M® access for another week. As shown in the upper right panel of Fig. 1, Skittle® intake nearly doubled between the first and third day of candy access, but then Skittle® intake remained stable for the remainder of the 8 weeks of candy access. There was no significant difference between the M & M® wasters and eaters in the number of Skittles® earned (P b 0.12). The lower panels of Fig. 1 compare the number of candies wasted between the 2 groups of animals as a function of week of access to candy. As shown in the lower left panel, beginning in week 5 there was an increase in the number of M & Ms® thrown away by the group that wasted M & Ms®, but not in the other 3 baboons. The 5 baboons who wasted M & Ms® did so by throwing some untouched candy on the floor, but most were initially placed in the mouth, then spat out, i.e, baboons sucked off the candy coating and threw the chocolate part away. There was a significant effect of group with wasters throwing more M & Ms® on the floor than eaters [F (1, 6) = 8.2, P b 0.028]. There was also a significant effect of week with the number of M & Ms® wasted increasing from week 5 to week 9 [F (8, 48) = 3.2, P b 0.03]. A significant week by group interaction confirmed that the increase in M & Ms® wasted was accounted for by the group that wasted M & Ms® [F (8, 48) = 5.6, P b 0.002]. As

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597

591

Fig. 1. Upper panels: Mean total daily number of candies earned each candy session for the 5 animals who wasted M & Ms® and the 3 animals who ate nearly all of the M & Ms® earned as a function of week of access to candy. Lower panels: Mean total daily number of candies wasted each candy session for the M & M® wasters and eaters as a function of week of access to candy. Although small amounts of M & Ms® were wasted during the first 3 1/2 weeks of M & M® access, the number wasted were not accurately counted until week 4. Although shown sequentially, candy was available on Mondays, Wednesdays and Fridays. Error bars on the data points represent ±1 SEM.

shown in the lower right panel of Fig. 1, all baboons wasted few Skittles®. No baboon wasted pellets. The pattern of results for M & Ms® consumed (earned minus wasted) differs from that of M & Ms® earned and wasted. Both groups ate a similar number of M & Ms® and Skittles® which remained consistent across the 8–9 weeks of candy access. During the last week of candy access, baboons began their Skittle® meal 38 + 2 min after the start of the session and their M & M® meal 48 + 6 min after the start of the session with no differences between groups in the latency to start eating either candy. The candy meal lasted significantly longer for the baboons who wasted M & Ms® when M & Ms® [F (1, 6) = 6.8, P b 0.04: 127 + 5 min vs. 66 + 4 min] or Skittles® [F (1, 6) = 8.3, P b 0.03: 108 + 4 min vs. 65 + 3 min] were available. The duration of the M & M® meal [F (8, 48) = 3.6, P b 0.008: 97 + 10.5 min during week 1 to 122 + 14.5 min during week 9], but not the Skittle® meal increased over weeks of candy access. Finally, all baboons responded rapidly during candy meals (4.5 responses/second). During the appetitive phase prior to a candy meal, the number of light flashes decreased over weeks of M & M® access [F (8, 48) = 2.8, P b 0.02: 43.2 + 3.1 during week 1 to 25.9 + 2.6 during

week 9], but not Skittle® access. All baboons responded slowly during candy appetitive phases (1.5 responses/second). In order to determine if the wasting behavior observed with M & Ms® would occur after baboons had access to Skittles®, baboons had access to M & Ms® using the same schedule 6 months after the first M & M access period and 1 month after access to Skittles. The 5 wasters continued to waste M & Ms® 6 months later: 181 + 32 vs. 191 + 34. One of the female eaters wasted M & Ms® 6 months later (280 vs. 0), while the other 2 eaters failed to waste M & Ms® 6 months later (22 vs. 15). 3.1.2. Pellet deliveries The upper panels of Fig. 2 compare the total number of pellets consumed each day during the first week and last week of candy access between the M & M® wasters and eaters. As shown in the upper panels, both groups consumed similar numbers of pellets across all 9 weeks of M & M® access and all 8 weeks of Skittle® access on the days that only food was available. Both groups consumed fewer pellets on the days that M & Ms® [F (1, 6) = 35.1, P b 0.001] or Skittles® [F (1, 6) = 106.3, P b 0.0001] were available across all 8–9 weeks of candy access. The decrease in pellet intake in both groups was due to a decrease in the number

592

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597

Fig. 2. Upper panels: Mean total daily number of candies and pellets consumed for the 3 days that candy was available and 3 of the 4 days that candy was unavailable during the first and last week of access to candy for the 5 animals who wasted M & Ms® and the 3 animals who ate nearly all of the M & Ms®. Lower panels: Mean total daily caloric intake for the 3 days that candy was available and 3 of the 4 days that candy was unavailable during the first and last week of access to candy for the M & M® wasters and eaters. Although shown sequentially, candy was available before pellets on Mondays, Wednesdays and Fridays, and only pellets were available on Saturdays, Sundays, Tuesdays and Thursdays. Error bars on the data points represent ±1 SEM.

of pellet meals per day from 3.8 + 0.1 on days when M & Ms® were not available to 2.9 + 0.1 on days that M & Ms® were available [F (1, 6) = 24.6, P b 0.003], and from 3.9 + 0.1 on days when Skittles® were not available to 2.5 + 0.1 on days that Skittles® were available [F (1, 6) = 161.1, P b 0.0001], i.e., the morning M & M® meal substituted for the morning pellet meal. Finally, all baboons responded rapidly during the first pellet meal (4.6 responses/second). In order to determine how pellet intake during the period of M & M® access compared to pellet intake when M & Ms® were not available, the number of pellets consumed during the 3 test days when M & Ms® were not available during week 9 were compared to pellet intake on 3 days 1 month later (2 weeks after the last M & M® PR test session). Total pellet intake during the last week of M & M® access (211 + 15) did not differ from pellet intake 1 month later (217 + 10). Thus, consuming candy on one day did not affect total pellet con-

sumption the next day. The size of the first pellet meal during the last week of M & M® access was, however, significantly smaller [F (1, 6) = 11.7, P b 0.014] than the size of the first pellet meal 1 month later (64 + 7 vs. 85 + 6). Baboons earned significantly fewer pellet appetitive reinforcers when M & Ms® [F (1, 6) = 11.9, P b 0.013: 53 + 3 vs. 73 + 3] or Skittles® were available [F (1, 6) = 48.7, P b 0.0004: 53 + 4 vs. 74 + 3]. Animals who wasted M & Ms® earned significantly fewer stimulus deliveries (13.6 + 0.8) than the animals who did not waste M & Ms® [24.9 + 1.7; F (1, 6) = 9.9, P b 0.012]. No differences in appetitive behavior prior to the first pellet meal between M & M® wasters and eaters were observed when Skittles® were available. All baboons responded slowly during the first pellet appetitive phase (1.4 responses/second). The lower panels of Fig. 2 compare the total caloric intake each day during the first week and last week of candy access

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597

593

between the M & M® wasters and eaters. Both groups had a similar caloric intake on days that pellets only were available across all 8–9 weeks of candy access. All baboons consumed more calories on the days that M & Ms® [F (1, 6) = 71.1, P b 0.0002: 1101 + 29 kcal/session vs. 654 + 15 kcal/session] or Skittles® [F (1, 6) = 22.7, P b 0.003]: 1148 + 25 kcal/session vs. 739 + 17 kcal/session] were available across all 8–9 weeks of candy access. Although baboons ate fewer pellets on the days that candy was available, this decrease was not enough to maintain stable caloric intake across days.

Fig. 3. Mean candy and food pellet intake during the first meal of the session as a function of progressive ratio breakpoint for 1 piece of candy or 1 food pellet for males and females.

3.1.3. Reinforcing efficacy There were no significant differences between the M & M® wasters and eaters in PR breakpoints for 1 pellet, 1 M & M® or 1 Skittle®, and breakpoints obtained before and after access to candy did not differ. PR breakpoints did vary across the 3 commodities [F (2, 12) = 11.8, P b 0.001], with the mean breakpoint, based all 8 baboons, for 1 M & M® (175 + 20) and 1 Skittle® (196 + 25) being larger than the mean breakpoint for 1 pellet (72 + 9). Fig. 3 presents the mean number of pellets and pieces of candy eaten during the first meal of the day during

Fig. 4. Mean number of candies eaten (earned–wasted; upper left panel), pellets eaten (upper right panel), latency to the first pellet meal (lower left panel), and number of pellet meals (lower right panel) for males and females during the last week of candy access as a function of type of candy. Error bars on the data points represent 1 SEM.

594

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597

the last week of access to candy as a function of PR breakpoint (each animal contributed 1 observation each for M & Ms®, Skittles® and pellets). A significant positive correlation between breakpoint and intake (adjusted r2 = 0.66) exists for the baboons as a group [F (1, 21) = 43.7, P b 0.000002]. Females [r2 = 0.65; F (1, 10) = 21.4, P b 0.0009] and males [r2 = 0.64; F (1, 9) = 18.5, P b 0.002] had nearly identical adjusted r2 s. PR breakpoints for a commodity were highly predictive of single meal intake. 3.2. Males compared to females The top panels of Fig. 4 compare the total number of candies and pellets eaten each day during the last week of candy access between males and females. There were no significant differences between males and females in the number of M & Ms® or Skittles® consumed (upper left panel), or any other measure of candy consumption. In contrast to candy consumption, sex and type of candy influenced pellet intake. As shown in the upper right panel of Fig. 4, males consumed more pellets than females [F (1, 6) = 5.8, P b 0.052]. There was also a significant interaction between type of candy and sex [F (1, 6) = 8.2, P b 0.029] with females eating similar numbers of pellets under both candy conditions, and males eating more pellets when Skittles® were available compared to when M & Ms® were available. There were significant sex differences in the latency to the first pellet meal of the day and the number of pellet meals each day. As shown in the lower left panel of Fig. 4, the average latency to the first pellet meal was significantly shorter [F (1, 6) = 9.5, P b 0.021] for males than females. As shown in the lower right panel of Fig. 4, the number of pellet meals was significantly greater [F (1, 6) = 14.1, P b 0.009] for males than females. With respect to reinforcing efficacy of the 3 commodities, there were no differences between males and females in PR breakpoints for either an M & M® or Skittle®. The lack of difference in breakpoint parallels the lack of significant differences in either M & M® or Skittle® consumption between males and females. The breakpoint for a single food pellet was significantly larger [F (1, 6) = 17.4, P b 0.006] for males (99 + 9) than females (52 + 7), however, which corresponds to the greater consumption of pellets by males than females. Males weighed more than females. 2 weeks before access to M & Ms®, the mean body weight of the males was 11.2 + 1.7 kg, and the females was 8.5 + 1.1 kg. All baboons gained weight. 2 weeks after access to M & Ms® the mean weight gain of the males was 2.1 + 0.7 kg, and the females was 1.1 + 0.5 kg. These weight gains did not differ, however, from weight gains observed during the 3 months prior to the access to M & Ms® when the mean weight gain of the males was 1.8 + 0.4 kg, and the females was 1.1 + 0.6 kg. 2 weeks before access to Skittles®, the mean body weight of the males was 13.5 + 2.1 kg, and the females was 11.3 + 1.9 kg. All baboons gained weight. 2 weeks after access to Skittles® the mean weight gain of the males was 1.3 + 0.5 kg, and the females was 0.7 + 0.7 kg. Again, the changes in weight were not statistically different from changes observed prior to starting this study.

4. Discussion 4.1. Candy consumption The results of the present experiment clearly indicate that free-feeding non-human primates given periodic opportunities to “forage” for a preferred food item, i.e., candy, will work for, and consume, a large quantity of the preferred item. When candy was available, the number of candies consumed in the first meal of the day was similar to, or larger than, the total number of food pellets consumed over the rest of the day. Due to the greater caloric content of candy, baboons also consumed more calories on days that candy was available. Candy intake increased during the first week of candy access, but then remained stable across the period of access to candy. Access to candy engendered a significant amount of appetitive and consummatory behavior and provided enrichment for the baboons. The present procedure was based on that used by Corwin and colleagues to engender excessive eating in free-feeding rodents [14–16]. Their studies have demonstrated that providing rats access to preferred fat or sucrose for 2 h/day (2 h before the dark cycle) 3 days per week results in caloric intakes during that interval that can be as much as ten times the caloric intake when the preferred item is not available. Hoebel and colleagues have used a different procedure for engendering excessive intake of a preferred food item (sugar) in rodents [23,25]. Their studies have demonstrated that limiting food intake to 12 h each day (8 h dark cycle, 4 h light cycle), and providing a sucrose solution and chow, leads to increased consumption of both sugar and chow: intake of chow and sucrose during the first 3 h of the session increase 3 to 4 fold after 8 days on this regime [24]. The consumption of larger amount of food in a discrete time frame by non-human primates replicates the findings described above with rodents. Candy meals, based on a priori definition, lasted longer than pellet meals at the same time of day. When instructed to binge, humans with binge-eating disorder increased their caloric intake by having a longer meal [26,27]. All 8 baboons ate 1 less pellet meal on the days candy was available thereby reducing total pellet intake on the days when candy was available. Laboratorybased and self-report data indicate that when humans eat a large meal, this meal substitutes for a small meal, and is not an additional meal that day [28–30]. These similarities between human and non-human primates suggest that the present procedures may provide a model for large meals. Over the course of M & M® or Skittle® access, however, baboons gained weight at a rate similar to that observed over a similar time frame before access to that candy, suggesting that some compensatory change in behavior occurred so that growth remained stable. Based on published norms [31,32] the male baboons were at the age when they would be expected to gain, if they were living in the wild, 0.65 to 0.75 kg in a 3 month interval, while the female baboons would be expected to gain 0.25 to 0.35 kg in a 3 month interval. Thus, weight gains in the present study were about twice that observed in the wild. Wild baboons that received supplemental food, however, gained weight significantly faster than non-supplemented wild-living baboons [33], which parallels the weight gains observed here.

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597

Corwin [15] demonstrated that rats given periodic access to a preferred food show a compensatory decrease in caloric intake on the day after the large caloric intake engendered by access to the preferred food. In the present study there was no evidence for compensatory decreases in pellet intake on days that candy was not available. Although the size of the first pellet meal on noncandy days was smaller during the last week of candy access than the first pellet meal when candy was no longer available, the total number of pellets on non-candy days during the period of access to candy did not differ from pellet intake when candy was no longer available. This decrease on non-candy days may reflect an effort to compensate for the large caloric intake on the previous day [5], or a contrast effect to the candy meal. Jen [34] reported that although male rhesus monkeys given access to diets known to produce obesity in rodents increased their caloric intake, their body weight did not change, and concluded that monkeys must have adjusted their activity levels to expend more energy. The difference between data obtained in rodents and non-human primates may be due to the much greater size of nonhuman primates which makes non-human primates less sensitive to acute caloric manipulations [5]. There were two instances where appetitive behavior was predictive of consummatory behavior. First, responding during the appetitive component for candy was greater than responding during the appetitive component for the first pellet meal of the day. Second, the M & M® wasters responded about half as much as the eaters during the appetitive component for the first pellet meal of the day when M & Ms® were available. These differences in appetitive behavior between candy and pellets were quite subtle compared to the differences in candy and pellet consumption. These findings confirm previous data from this laboratory [35,6] and elsewhere [3,4,36] showing that the appetitive and consummatory phases of eating can be behaviorally and pharmacologically differentiated. 4.2. Tasting and wasting behavior Five (3 males and 2 females) of the 8 baboons developed an unexpected behavior when they had access to M & Ms®. These animals did not eat all of the M & Ms® they earned. They placed M & Ms® in their mouths and then spat them out with the candy coating removed, i.e., they did not eat the chocolate center. This wasting pattern generally occurred after they had eaten most of the M & Ms® that they were going to eat that day. Of note is the finding that minimal tasting and wasting behavior occurred when baboons had access to Skittles®. The limited wasting of Skittles® may have been due to the fact that Skittles® have 5 flavors which provided greater variety than the M & Ms®, which have only 1 flavor (6 colors). It might also be due to the fact that baboons, who are excellent at discriminating sugar concentrations [37], have a preference for sweet flavored diets over a chocolate-flavored diet [38]. When baboons had access to M & Ms® again after access to Skittles® all 5 of the M & M® wasters, who had not wasted Skittles®, again wasted M & Ms®, while 2 of the 3 eaters, continued to eat the M & Ms®. This indicates that the behavior is stable, limited to M & Ms®, and not related to animal age.

595

Because of its similarity to the “chewing and spitting” behavior that has been identified in eating-disordered patients [39,40], it is tempting to conclude that the wasting behavior observed here is an aberrant behavior. In the natural ecology baboons are omnivorous and will eat almost all foods [41]. Baboons can be “fussy” eaters, however, in that when in rich food patches or when food is plentiful they will strip off the tough exterior of grasses and eat the tender shoots, or pull up grass and eat only the tender roots [42], which is a naturalistic type of wasting behavior. Perhaps the wasting behavior, which only occurred in the laboratory when sugar-coated M & Ms®, were available, reflects normal behavior observed when baboons are in a dense patch of preferred food. Of course, Skittles® are also coated in sugar, but they have a fruit-flavored, not chocolate-flavored interior. In general the M & M® wasters ate more candy, and the wasting behavior may reflect differences between the groups on some dimension other than the tendency to spit out M & Ms®. In addition to possible differences in satiety, fullness, and preference for sugar, the groups could have varied in their response to novelty or the willingness to forage (see next section). For example, under certain situations, zoo-housed baboons will forage for food treats even when those treats do not fill any nutritional need [43]. Because only 3 baboons did not waste food, there is limited power to the comparisons between groups. In additional to the small number of M & M® eaters, another limitation to this study is that the order of testing candies was not balanced: the decision to test Skittles® was made based upon the results obtained with M & Ms®. 4.3. Reinforcing efficacy The relative reinforcing efficacy of candy and pellets was assessed before and after access to candy using a progressive ratio schedule [18], which increased the response requirement for reinforcement after each reinforcer delivery [19]. The PR breakpoint for 1 M & M® or Skittle® was more than double the breakpoint for 1 food pellet. This confirmed our hypothesis that candy would be more reinforcing than pellets, and predicted the larger intake of candy during the candy meal than intake of pellets during a single meal. There was no change in the PR breakpoints for either candy or pellets as a consequence of access to candy. Although the breakpoint for a commodity predicted single meal intake of that commodity, the breakpoint for 1 M & M® was not predictive of M & M® wasting behavior. Animals did not waste candy during PR sessions. The failure of the M & M® breakpoints to predict M & M® wasting behavior is not surprising because the breakpoint was obtained when animals were eating, not wasting the M & Ms®. Candy intake was quite low during PR sessions (25–35 candies) such that the wasters did not consume the amount of M & Ms® (150 candies) that they would normally have eaten before wasting occurred. It would be interesting to provide animals a meal of M & Ms® equivalent to what was normally eaten, then conduct a PR test session to determine if the wasters would work hard to earn candy that they then tasted and wasted. There may be multiple reinforcing aspects of the M & M® candy including the eating of a single candy, licking the sugar coating off the candy, and or the act of foraging itself.

596

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597

4.4. Sex differences

References

Although there were only 4 males and 4 females in this study, there were still significant differences between the sexes in terms of pellet intake, but not candy intake. Males had larger PR breakpoints for pellets, and had a shorter latency to the first pellet meal of the day and had more pellet meals and ate more pellets than females on days when only food was available. Finally males ate more pellets when Skittle® were available, while females at a similar number of pellets under both candy conditions. The greater pellet intake of males was expected based on their larger size, but the similar intake of candy between the sexes was unexpected. Several studies have reported that male and female rats have similar PR breakpoints for standard rodent chow [44,45], but at least 1 study has reported that male rats consumed more chow diet and had greater PR breakpoints for the diet than female rats [46]. The results of present study, in combination with Merali et al. [46], indicate that sex differences in reinforcing efficacy will be observed in situations where one sex consumes more of the commodity.

[1] Rodman PS, Cant JGH. Adaptations for foraging in nonhuman primates. New York: Columbia University Press; 1984. [2] Owen J. Feeding strategy. Chicago: The University of Chicago Press; 1980. [3] Collier GH. Life in a closed economy: The ecology of learning and motivation. In: Zeiler MD, Harzem P, editors. Advances in analysis of behaviour. John Wiley & Sons, Ltd; 1983. p. 223–74. [4] Collier GH. Satiety: an ecological perspective. Brain Res Bull 1985;14:693–700. [5] Foltin RW, Fischman MW. Effects of caloric manipulations on food intake in baboons. Appetite 1990;15:135–49. [6] Foltin RW. Effects of dietary and pharmacological manipulations on appetitive and consummatory aspects of feeding in non-human primates. Appetite 2005;45:110–20. [7] Klein DA, Walsh BT. Eating disorders: clinical features and pathophysiology. Physiol Behav 2004;81:359–74. [8] American Psychiatric Association, S.Z.. Diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC: American Psychiatric Press; 1994. [9] Dingemans AE, Bruna MJ, van Furth EF. Binge eating disorder: a review. Int J Obes Relat Metab Disord 2002;26:299–307. [10] Mitchell KS, Mazzeo SE. Binge eating and psychological distress in ethnically diverse undergraduate men and women. Eat Behav 2004;5:157–69. [11] Hoek HW, van Hoeken D. Review of the prevalence and incidence of eating disorders. Int J Eat Disord 2003;34:383–96. [12] Carter WP, Hudson JI, LaLonde JK, Pindyck L, McElroy SL, Pope Jr HG. Pharmacological treatment of binge eating disorder. Int J Eat Disord 2003;34:S74–8. [13] Corwin RL, Buda-Levin A. Behavioral models of binge-type eating. Physiol Behav 2004;82:123–30. [14] Corwin RL, Wojnicki FH, Fisher JO, Dimitriou SG, Rice HB, Young MA. Limited access to a dietary fat option affects ingestive behavior but not body composition in male rats. Physiol Behav 1998;65:545–53. [15] Corwin RL. Binge-type eating induced by limited access in rats does not require energy restriction on the previous day. Appetite 2004;42:139–42. [16] Dimitriou SG, Rice HB, Corwin RL. Effects of limited access to a fat option on food intake and body composition in female rats. Int J Eat Disord 2000;28:436–45. [17] Fisher JO, Birch LL. Restricting access to palatable food affects children's behavioral response, food selection, and intake. Am J Clin Nutr 1999;69:1264–72. [18] Hodos W. Progressive ratio as a measure of reward strength. Science 1961;134:943–4. [19] Richardson NR, Roberts DCS. Progressive ratio schedules in drug selfadministration studies in rats: a method to evaluate reinforcing efficacy. J Neurosci Methods 1996;66:1–11. [20] Arnold JM, Roberts DC. A critique of fixed and progressive ratio schedules used to examine the neural substrates of drug reinforcement. Pharmacol Biochem Behav 1997;57:441–7. [21] Sclafani A, Ackroff K. Reinforcement value of sucrose measured by progressive ratio operant licking in the rat. Physiol Behav 2003;79: 663–70. [22] Reilly S. Reinforcement value of gustatory stimuli determined by progressive ratio performance. Pharmacol Biochem Behav 1999;63: 301–11. [23] Avena NM, Hoebel BG. A diet promoting sugar dependency causes behavioral cross-sensitization to a low dose of amphetamine. Neuroscience 2003;122:17–20. [24] Colantuoni C, Rada P, McCarthy J, Patten C, Avena NM, Chadeayne A, et al. Evidence that intermittent, excessive sugar intake causes endogenous opioid dependence. Obes Res 2002;10:478–88. [25] Avena NM, Hoebel BG. Amphetamine-sensitized rats show sugar-induced hyperactivity (cross-sensitization) and sugar hyperphagia. Pharmacol Biochem Behav 2003;68:1–5. [26] Goldfein JA, Walsh BT, LaChaussee JL, Kissileff HR, Devlin MJ. Eating behavior in binge eating disorder. Int J Eat Disord 1993;14:427–31.

5. Conclusions In summary, periodic access to a preferred candy food engendered 1) significantly greater intake of candy during a single meal than observed with pellets in all baboons; and 2) wasting behavior, i.e., sucking and spitting out of M & M®, but not Skittle® candy in 5 of 8 baboons. Reinforcing efficacy, estimated using PR breakpoints, predicted greater intake of candy within a single meal than food pellets in both males and females. Finally, males consumed significantly more food pellets than females, but there were no significant differences in intake of either candy between males and females. The large candy meals in the current procedure parallels the consumption of excessive amounts of food within a single meal that is a defining characteristic of disordered eating in patients with bulimia nervosa and binge-eating disorder. The current procedure may provide a behavioral baseline for evaluating potential pharmacological interventions [47,48] for these two disorders. The specificity of an intervention on eating of a large meal of preferred items compared to “normal” eating can be evaluated by comparing the effects of drugs on days when a preferred food is available, and eaten during a large meal, to days when only food pellets are available, and eaten in multiple smaller meals. Finally, providing periodic opportunities to forage for candy will provide enrichment for non-human primates in captivity. Acknowledgements This research was supported by DA-04130 from The National Institute on Drug Abuse, MH-65024 from the National Institute on Mental Health, and approved by the New York State Psychiatric Institute Animal Care and Use Committee. The assistance of Jean Willi, April Modranowski, Angel Ramirez, Jana Colley and Drs. Suzette Evans, Margaret Haney, Diane Klein and Mohamed Osman is gratefully acknowledged.

R.W. Foltin / Physiology & Behavior 89 (2006) 587–597 [27] Yanovski SZ, Leet M, Yanovski JA, Flood M, Gold PW, Kissileff HR, et al. Food selection and intake of obese women with binge-eating disorder. Am J Clin Nutr 1992;56:975–80. [28] Guertin TL, Conger AJ. Mood and forbidden foods' influence on perceptions of binge eating. Addict Behav 1999;24:175–93. [29] Kaye WH, Weltzin TE, McKee M, McConaha C, Hansen D, Hsu GLK. Laboratory assessment of feeding behavior in bulimia nervosa and healthy women: methods for developing a human-feeding laboratory. Am J Clin Nutr 1992:372–80. [30] Weltzin TE, Hsu GLK, Pollice C, Kaye WH. Feeding patterns in bulimia nervosa. Biol Psychiatry 1991;30:1093–110. [31] Mahaney MC, Leland MM, Williams-Blangero S, Marinez YN. Crosssectional growth standards for captive baboons: II. Organ weight by body weight. J Med Primatol 1993;22:415–27. [32] Strum SC. Weight and age in wild olive baboons. Am J Primatol 1991;25:219–37. [33] Altmann J, Alberts SC. Growth rates in a wild primate population: ecological influences and maternal effects. Behav Ecol Sociobiol 2005;57:490–501. [34] Jen K-LC. Failure to induce weight with palatable diets in monkeys (Macaca mulatta). Primates 1987;28:61–9. [35] Foltin RW. Effects of amphetamine, dexfenfluramine, diazepam, and dietary manipulations on responding reinforced by stimuli paired with food in non-human primates. Pharmacol Biochem Behav 2004;77:471–9. [36] Doucet E, St-Pierre S, Almeras N, Tremblay A. Relation between appetite ratings before and after a standard meal and estimates of daily energy intake in obese and reduced obese individuals. Appetite 2003;40:137–43. [37] Laska M, Schüll E, Scheuber H-P. Taste preference thresholds for foodassociated sugars in baboons (Papio hamadrayas anubis). Int J Primatol 1999;20:25–34. [38] Wene JD, Barnwell GM, Mitchell DS. Flavor preferences, food intake, and weight gain in baboons (papio sp.). Physiol Behav 1982;28:569–73.

597

[39] Guarda AS, Coughlin JW, Cummings M, Marinilli A, Haug N, Boucher M, et al. Chewing and spitting in eating disorders and its relationship to binge eating. Eat Behav 2004;5:231–9. [40] Kovacs D, Mahon J, Palmer RL. Chewing and spitting out food among eating-disordered patients. Int J Eat Disord 2002;32:112–5. [41] Richard AF. Primates in nature. New York: W.H. Freeman and Company; 1985. [42] Dunbar RIM. Feeding ecology of Gelada baboons: a preliminary report. In: Clutton-Brock TH, editor. Primate ecology: studies of feeding and ranging behaviour in lemurs, monkeys and apes. New York: Academic Press; 1977. [43] Jones M, Pillay N. Foraging in captive hamadrayas baboons: implications for enrichment. Appl Anim Behav Sci 2004;88:101–10. [44] Van Haaren F, Haworth SC, Bennett SM, Cody BA. The effects of pyridostigmine bromide on progressive ratio performance in male and female rats. Pharmacol Biochem Behav 2001;68:81–5. [45] Van Hest A, van Haaren F, van de Poll NE. The behavior of male and female wistar rats pressing a lever for food is not affected by sex differences in food motivation. Behav Brain Res 1988;27:215–21. [46] Merali Z, Brennan K, Brau P, Anisman H. Dissociating anorexia and anhedonia elicited by interleukin-1β: antidepressant and gender effects on responding for “free chow” and “earned” sucrose intake. Psychopharmacology 2003;165:413–8. [47] Grilo CM, Masheb RM, Wilson T. Efficacy of cognitive behavioral therapy and fluoxetine for the treatment of binge-eating disorder: a randomized double-blind placebo-controlled comparison. Biol Psychiatry 2005;57:301–9. [48] Appolinario JC, Bacaltchuk J, Sichieri R, Claudino AM, Godoy-Matos A, Morgan C, et al. A randomized, double blind, placebo-controlled study of sibutramine in the treatment of binge-eating disorder. Arch Gen Psychiatry 2003;60:1109–16.