Time horizons in rats: the effect of operant control of access to future food.

The primary goal of this experiment was to determine whether the addition of an operant requirement for access to a less costly (continuous reinforcement) patch of future food increased the time horizon over which that future patch decreased intake in a currently available depleting (progressive-ratio) patch. Three groups of 4 rats were tested. Each member of the earned-time group was required to cumulate a fixed-time outside the progressive-ratio patch to obtain access to food in the less costly patch; the fixed-time requirement ranged from 2 to 64 min. Rats in the matched-time group received response-independent access to less costly food at the average delay shown by the earned-time group. Rats in the matched-time no-food group were removed from the chamber at the same average delay without receiving access to less costly food. Two of the earned-time rats showed an increased time horizon relative to that shown by the matched-time rats (approaching 40 min for 1 rat). The other 2 earned-time rats markedly increased instrumental responding but showed suppression of intake only when food was less than 20 min away. The matched-time group showed less suppression of intake over a similar range of delay intervals. Surprisingly, the matched-time no-food animals also showed suppression of intake concentrated at the end of the session, possibly reflecting the receipt of their entire daily ration 30 min after the session. The potential importance of time horizons to the foraging process is clear, but experimenters are still working out paradigms for investigation of these horizons.     

A temporal limit on the effect of future food on current performance in an analogue of foraging and welfare

Rats obtained access to food twice each 24-hour period. The first session was a work session in which food was available on a progressive-ratio schedule. During the second session, which occurred between 1 and 23 hours after the work session, food was freely available up to a fixed total intake each 24 hours. The situation resembled elements of several real world circumstances, including the choice between continuing to forage in a rapidly depleting patch and waiting for a better patch, and between working now and receiving a guaranteed income later. The purpose of the experiment was to explore the time period over which future access to reward could affect current responding. Contrary to what might be expected from recent theorizing, anticipation of future food delayed by an hour or more after the start of the work session had no effect on current performance. Food intake was high and constant during work sessions except for a prefeeding effect that occurred when the free session closely preceded the next day's work session. Also, an increase in the difficulty of the work schedule increased the amount of work and the maximum price paid for food as if the work session were the only time food was available. The results indicate the importance of considering temporal limits in theories that require animals to integrate input over time to determine the allocation of resources among alternatives.     

Foraging in a simulated natural environment: There's a rat loose in the lab

Rats were required to earn their food in a large room having nine boxes placed in it, each of which contained food buried in sand. In different phases of the experiment the amount of time allowed for foraging, the amount of food available in each food patch, and the location of the different available amounts were varied. The rats exhaustively sampled all patches each session but seemed to have fairly strong preferences for certain locations over others. If position preferences were for patches containing small amounts of food, the sensitivity to amount available was increased so that when location was compensated for, a pattern of optimal foraging was evident. The importance of environmental constraints in producing optimal behavior and the relation of the observed behavior to laboratory findings are discussed.     

An experimental analysis of optimal foraging behaviour in patchy environments

A series of experiments was designed to explore the cognitive mechanisms involved in optimal foraging models by using the behavioural controls of operant methodology. Rats were trained to press one of two levers to obtain reinforcement on a progressive variable-interval schedule, which modelled food patch depletion; the schedule was reset by pressing the other lever. Thus both duration (residence time in a patch) and rate-related (interval before and after the final reward) measures were obtained. Experiment 1, which manipulated environmental stability and quality, and Experiment 2, which varied travel time between patches, found results that supported the marginal value theorem (Charnov, 1976) and suggested that rats adjust capture rate to the environment average by monitoring the length of the interval between rewards. Experiment 3 modelled the clumping of food items and found that capture rate was now adjusted by adoption of a fixed giving-up time. Finally, Experiments 4a and 4b ruled out a time expectancy hypothesis by manipulating the number of food clumps and the series of inter-reinforcement intervals. Overall the experiments demonstrate the value of modelling foraging strategies in operant apparatus, and suggest that rats adopt rate predictive strategies when deciding to switch patches.     

Drinking in a patchy environment: the effect of the price of water.

Rats in a laboratory foraging paradigm searched for sequential opportunities to drink in two water patches that differed in the bar-press price of each "sip" (20 licks) of water within a bout of drinking (Experiment 1) or the price and size (10, 20, or 40 licks) of each sip (Experiment 2). Total daily water intake was not affected by these variables. The rats responded faster at the patch where water was more costly. However, they accepted fewer opportunities to drink, and thus had fewer drinking bouts, and drinking bouts were smaller at the more costly patch than at the other patch. This resulted in the rats consuming a smaller proportion of their daily water from the more costly patch. The size of the differences in bout frequency and size between the patches appears to be based on the relative cost of water at the patches. The profitability of each patch was calculated in terms of the return (in milliliters) on either effort (bar presses) or time spent there. Although both measures were correlated with the relative total intake, bout size, and acceptance of opportunities at each patch, the time-based profitability was the better predictor of these intake measures. The rats did not minimize bar-press output; however, their choice between the patches and their bout sizes within patches varied in a way that reduced costs compared to what would have been expended drinking randomly. These data accord well with similar findings for choices among patches of food, suggesting that foraging for water and food occurs on the basis of comparable benefit-cost functions: In each case, the amount consumed is related to the time spent consuming.     

Timing and Foraging: Gibbon's Scalar Expectancy Theory and Optimal Patch Exploitation

We present a study that links optimal foraging theory (OFT) to behavioral timing. OFT's distinguishing feature is the use of models that compute the most advantageous behavior for a particular foraging problem and compare the optimal solution to empirical data with little reference to psychological processes. The study of behavioral timing, in contrast, emphasizes performance in relation to time, most often without strategic or functional considerations. In three experiments, reinforcer-maximizing behavior and timing performance are identified and related to each other. In all three experiments starlings work in a setting that simulates food patches separated by a flying distance between the two perches. The patches contain a variable and unpredictable number of reinforcers and deplete suddenly without signal. Before depletion, patches deliver food at fixed intervals (FI). Our main dependent variables are the times of occurrence of three behaviors: the “peak” in pecking rate (Peak), the time of the last peck before “giving in” (GIT), and the time for “moving on” to a new patch (MOT). We manipulate travel requirement (Experiment 1), level of deprivation and FI (Experiment 2), and size of reinforcers (Experiment 3). For OFT, Peak should equal the FI in all conditions while GIT and MOT should just exceed it. Behavioral timing and Scalar Expectancy Theory (SET) in particular predict a Peak at around the FI and a longer (unspecified) GIT, and make no prediction for MOT. We found that Peak was close to the FI and GIT was approximately 1.5 times longer, neither being affected by travel, hunger, or reinforcer size manipulations. MOT varied between 1.5 and just over 3 times the FI, was responsive to both travel time and the FI, and did not change when the reinforcer rate was manipulated. These results support the practice of producing models that explicitly separate information available to the subject from strategic use of this information.     

Comparing Locomotion With Lever-press Travel In An Operant Simulation Of Foraging

An operant model of foraging was studied. Rats searched for food by pressing on the left lever, the patch, which provided one, two, or eight reinforcers before extinction (i.e., zero reinforcers). Obtaining each reinforcer lowered the probability of receiving another reinforcer, simulating patch depletion. Rats traveled to another patch by pressing the right lever, which restored reinforcer availability to the left lever. Travel requirement changed by varying the probability of reset for presses on the right lever; in one condition, additional locomotion was required. That is, rats ran 260 cm from the left to the right lever, made one response on the right lever, and ran back to a fresh patch on the left lever. Another condition added three hurdles to the 260-cm path. The lever-pressing and simple locomotion conditions generated equivalent travel times. Adding the hurdles produced longer times in patches than did the lever-pressing and simple locomotion requirements. The results contradict some models of optimal foraging but are in keeping with McNair's (1982) optimal giving-up time model and add to the growing body of evidence that different environments may produce different foraging strategies.     

The relationship between feeding rate and patch choice.

Rats in a laboratory foraging simulation searched for sequential opportunities to feed in two patches that differed in the rate at which food pellets were delivered (controlled by fixed-interval schedules) and in the size of the pellets. The profitability of feeding in each patch was calculated in terms of time (grams per minute) and in terms of effort (grams per bar press). These values were the result of the imposed fixed interval, the size of the pellets, and the rate at which the rats pressed the bar in each condition. The rats ate more food and larger meals, but not more frequent meals, at the patch offering the higher rate of food consumption, calculated as grams per minute. The relative intake at any patch was a function of the relative rate of intake during meals at that patch compared to the other patch. Rats respond to explicit manipulations of feeding time in the same manner as they respond to manipulations of feeding effort.     

Changes in feeding and foraging patterns as an antipredator defensive strategy: a laboratory simulation using aversive stimulation in a closed economy.

The effects of the risk of electric shock on the meal patterns of rats living in an operant chamber were investigated. Rats could obtain food by working on a response lever that provided reinforcement according to chained fixed-ratio continuous reinforcement schedules that allowed the animals control over meal size. Using a two-compartment operant chamber with a safe nesting area and manipulanda area with a grid floor, shock could be correlated with responding on the schedule. Shocks (less than or equal to 1.25 per hour) were scheduled to occur randomly throughout the day, independent of the rat's behavior. Shock caused a reorganization of meal patterns such that the animals took less frequent but larger meals. This pattern reduced the time the animals spent at risk without compromising caloric balance. Similar changes in feeding pattern were obtained in both hooded and albino rats. Exposure to shock in a separate chamber did not produce these behavioral modifications. The magnitude of shock-induced alterations of meal patterns was greater with chained fixed-ratio 90 continuous reinforcement than with chained fixed-ratio 10 continuous reinforcement. Additionally, the rats seemed to be able to reduce food intake but increase caloric efficiency, such that the reduced food intake did not have deleterious effects on maintenance of body weight. These behavioral modifications reduced the number of shocks received from that which would have been expected if meal pattern changes had not occurred. We suggest that this technique may provide a useful laboratory simulation of the impact that the risk of predation has on foraging behavior.     

Fishing for the right words: decision rules for human foraging behavior in internal search tasks

Animals depleting one patch of resources must decide when to leave and switch to a fresh patch. Foraging theory has predicted various decision mechanisms; which is best depends on environmental variation in patch quality. Previously we tested whether these mechanisms underlie human decision making when foraging for external resources; here we test whether humans behave similarly in a cognitive task seeking internally generated solutions. Subjects searched for meaningful words made from random letter sequences, and as their success rate declined, they could opt to switch to a fresh sequence. As in the external foraging context, time since the previous success and the interval preceding it had a major influence on when subjects switched. Subjects also used the commonness of sequence letters as a proximal cue to patch quality that influenced when to switch. Contrary to optimality predictions, switching decisions were independent of whether sequences differed little or widely in quality.     

Rats (Rattus norvegicus) modulate eating speed and vigilance to optimize food consumption: effects of cover, circadian rhythm, food deprivation, and individual differences.

The eating behavior of rats (Rattus norvegicus) given food pellets of specified size was examined as a function of environmental, circadian, and experiential influences. Eating times were shorter in lighted, exposed environments than in dark, covered environments, even though in novel, exposed conditions the rats made many scanning movements as they ate. Eating time also varied as a function of the circadian cycle in that eating times were shorter in the night portion of the day-night cycle. Finally, eating times decreased if rats were food deprived, and deprivation had a small but enduring influence. Within the tests there were differences in the eating times of individual rats that were not attributable to the experimental manipulations. That rats can optimize food intake by varying eating speed is discussed in relation to physiological regulation of feeding and to optimal foraging theory.     

Conspecific aggression influences food carrying: Studies on a wild population of Rattus norvegicus

Laboratory rats, like many other animal species, transport food. Their behavior is influenced by factors such as food size, the time required to eat, travel distance, travel difficulty, and the availability of cover, etc. Recent versions of optimal foraging theory suggest that species that display such behavioral patterns do so in order to minimize risk to predation while at the same time maximizing food gain. Nevertheless, it is not clear that this explanation applies to rats, nor is it easy to investigate this problem in a laboratory. The results of the present study on urban feral rats show that their food handling behavior is similar to that of domestic rats. The results also suggest that food carrying can serve defensive as well as communicative functions. Aggression around food sources was high. Smaller rats always carried food and some large rats infrequently carried food, suggesting that the food carrying by smaller, subordinate rats may help them avoid conspecific aggression. The rats also vigorously attempted to steal food that was carried home by conspecifics. This result suggests that food carrying can redistribute food resources and inform conspecifics about food availability. The results demonstrate the utility of multilevel behavioral analysis and also demonstrate that for rats, food transport has functions other than predator avoidance, including avoidance of conspecific aggression and communication about food availability. © 1996 Wiley-Liss, Inc.     

Postprandial scanning by the rat (Rattus norvegicus): the importance of eating time and an application of "warm-up" movements

We observed the movements of rats (Rattus norvegicus) after they had eaten food pellets of various size or hardness. With rooted hindlegs, they made head scans, with vibrissae in contact with the substrate, that began over the area below where they had eaten and then expanded to include almost the entire area surrounding their body. Scanning was not contingent on the presence of dropped food. It occurred when rats ate on a screen through which any dropped crumbs could fall. It also occurred when rats were trained to find food at a location distant to where they ate. Although the duration of scanning increased in proportion to the size of food consumed, when eating time was varied, using food items of similar size but different hardness, scanning increased in proportion to eating time. Postprandial scans resemble the exploratory (warm-up) movements that bridge transitions from immobility to locomotion. We propose that a subset of the movements of warm-up are co-opted in this postprandial period. It is likely that in natural foraging situations they are useful for food searching. The results suggest that although the motor system may be conservative in the number of actions that it can produce, diversity is achieved by applying fundamental patterns to many uses.     

Residence time in concurrent foraging with fixed times to prey arrival

Five pigeons were trained in a concurrent foraging procedure in which reinforcers were occasionally available after fixed times in two discriminated patches. In Part 1 of the experiment, the fixed times summed to 10 s, and were individually varied between 1 and 9 s over five conditions, with the probability of a reinforcer being delivered at the fixed times always .5. In Part 2, both fixed times were 5 s, and the probabilities of food delivery were varied over conditions, always summing to 1.0. In Parts 3 and 4, one fixed time was kept constant (Part 3, 3 s; Part 4, 7 s) while the other fixed time was varied from 1 s to 15 s. Median residence times in both patches increased with increases in the food-arrival times in either patch, but increased considerably more strongly in the patch in which the arrival time was increased. However, when arrival times were very different in the two patches, residence time in the longer arrival-time patch often decreased. Patch residence also increased with increasing probability of reinforcement, but again tended to fall when one probability was much larger than the other. A detailed analysis of residence times showed that these comprised two distributions, one around a shorter mode that remained constant with changes in arrival times, and one around a longer mode that monotonically increased with increasing arrival time. The frequency of shorter residence times appeared to be controlled by the probability of, and arrival time of, reinforcers in the alternative patch. The frequency of longer residence times was controlled directly by the arrival time of reinforcers in a patch, but not by the probability of reinforcers in a patch. The environmental variables that control both staying in a patch and exiting from a patch need to be understood in the study both of timing processes and of foraging.     

Evidence for future cognition in animals

Evidence concerning the possibility of mental time travel into the future by animals was reviewed. Both experimental laboratory studies and field observations were considered. Paradigms for the study of future anticipation and planning included inhibition of consumption of current food contingent on future receipt of either a larger quantity or more preferred food, choice between quantities of food contingent on future pilfering or replenishment of food, carrying foods to different locations contingent on future access to those locations, and selection of tools for use to obtain food in the future. Studies of non-human primates, rats, black-capped chickadees, scrub-jays, and tayras were considered. It was concluded that current evidence favors future cognition in animals, and some theoretical issues concerning this ability were discussed.     

Leaving patches: effects of economy, deprivation, and session duration

Three pigeons pecked keys for food reinforcers in a laboratory analogue of foraging in patches. Half the patches contained food (were prey patches). In prey patches, pecks to one key occasionally produced a reinforcer, followed by a fixed travel time and then the start of a new patch. Pecks to another key were exit responses, and immediately produced travel time and then a new patch. Travel time was varied from 0.25 to 16 s at each of three session durations: 1, 4, and 23.5 hr. This part of the experiment arranged a closed economy, in that the only source of food was reinforcers obtained in prey patches. In another part, food deprivation was manipulated by varying postsession feeding so as to maintain the subjects' body weights at percentages ranging from 85% to 95% of their ad lib weights, in 1-hr sessions with a travel time of 12 s. This was an open economy. Patch residence time, defined as the time between the start of a patch and an exit response, increased with increasing travel time, and consistently exceeded times predicted by an optimal foraging model, supporting previously published results. However, residence times also increased with increasing session duration and, in longer sessions, consistently exceeded previously reported residence times in comparable open-economy conditions. Residence times were not systematically affected by deprivation levels. In sum, the results show that the long residence times obtained in long closed-economy sessions should probably be attributed to session duration rather than to economy or deprivation. This conclusion is hard to reconcile with previous interpretations of longer-than-optimal residence times but is consistent with, in economic terms, a predicted shift in consumption towards a preferred commodity when income is increased.     

Discrimination learning in a foraging situation

Rats were allowed to forage in a simulated natural environment made up of eight food sources (patches) each containing a fixed number of pellets. Two of the eight contained an extra supply of peanuts. The peanut patches were signaled by an olfactory/visual cue located at the bottom of the ladder leading to the patch. In successive phases the number of sessions per day, height of the patches, and availability of peanuts were manipulated. Subjects showed evidence of discrimination learning under these conditions, although the degree of discriminatory behavior varied as a function of environmental manipulations. Assessment of behavior within foraging sessions showed that subjects systematically changed their patterns of utilization of patches across time. Sampling or exploration, as well as food reinforcement, seem implicated in these results.     

Experimental evidence for route integration and strategic planning in wild capuchin monkeys

Both in captivity and the wild, primates are found to travel mostly to the nearest available resource, but they may skip over the closest resource and travel to more distant resources, which are often found to be more productive. This study examines the tradeoff between distance and reward in the foraging choices of one group of wild capuchin monkeys (Cebus apella nigritus) using feeding platforms in large-scale foraging experiments conducted over four years. Three feeding sites were arrayed in an oblique triangle, such that once the monkey group had chosen one site to feed, they had a choice between two remaining sites, a close one with less food and the other up to 2.3 times as far away but with more food. Sites were provisioned once per day. The capuchins generally chose the closer feeding site, even when the more distant site offered up to 12 times as much food. The distances to, rewards of, or various profitability measures applied to each alternative site individually did not explain the group’s choices in ways consistent with foraging theory or principles of operant psychology. The group’s site choices were predicted only by comparing efficiency measures of entire foraging pathways: (1) direct travel to the more rewarding distant site, versus (2) the longer ‘detour’ through the closer site on the way to the more distant one. The group chose the detour more often when the reward was larger and the added detour distance shorter. They appeared to be more sensitive to the absolute increase in detour distance than to the relative increase compared to the straight route. The qualitative and quantitative results agree with a simple rule: do not use the detour unless the energy gain from extra food outweighs the energy cost of extra travel. These results suggest that members of this group integrate information on spatial location, reward, and perhaps potential food competition in their choice of multi-site foraging routes, with important implications for social foraging. This contribution is part of the special issue “ A Socioecological Perspective on Primate Cognition” (Cunningham and Janson 2007b).     

Patch choice as a function of procurement cost and encounter rate.

The effects of patch encounter rate on patch choice and meal patterns were studied in rats foraging in a laboratory environment offering two patch types that were encountered sequentially and randomly. The cost of procuring access to one patch was greater than the other. Patches were either encountered equally often or the high-cost patch was encountered more frequently. As expected, rats exploited the low-cost patch on almost 100% of encounters and exploited the high-cost patch on a percentage of encounters that was inversely proportional to its cost. Meal size was the same at both patches. Surprisingly, when low-cost patches were rare, the rats did not increase their use of high-cost patches. This resulted in spending more time and energy searching for patches and a higher average cost per meal. The rats responded to this increased cost by reducing the frequency and increasing the size of meals at both patches and thereby limited total daily foraging cost and conserved total intake.     

A neural network model of foraging decisions made under predation risk

This article develops the cognitive—emotional forager (CEF) model, a novel application of a neural network to dynamical processes in foraging behavior. The CEF is based on a neural network known as the gated dipole, introduced by Grossberg, which is capable of representing short-term affective reactions in a manner similar to Solomon and Corbit’s (1974) opponent process theory. The model incorporates a trade-off between approach toward food and avoidance of predation under varying levels of motivation induced by hunger. The results of simulations in a simple patch selection paradigm, using a lifetime fitness criterion for comparison, indicate that the CEF model is capable of nearly optimal foraging and outperforms a run-of-luck rule-of-thumb model. Models such as the one presented here can illuminate the underlying cognitive and motivational components of animal decision making.