The way in which the data are combined affects the interpretation of short-term feeding behavior

Novel phenotypes of feeding and social behaviour and their relationship with individual rabbit growth and feed efficiency

Due to the lack of a recording system for individual consumption of group-housed rabbits, published studies about feeding behaviour are based on information recorded at the group- and not at the individual level and periods covering only a few days or, in some cases, only part of a day. Such information could be used to inform rabbit management systems but cannot be used for genetic selection. We aimed to generate and use information from a novel automated feeder for group-housed rabbits to identify new phenotypes for individual animals that could be incorporated into breeding programs to improve feed efficiency and social behaviour under different feeding regimens. At 39 d of age, rabbits from 15 batches were placed in cages and fed ad libitum to become used to the electronic feeder. From 42 to 58-59 d, one group of 1 086 rabbits was fed ad libitum (AL), while another group of 1 134 rabbits was fed on a restricted feeding schedule (R) by limiting the feeding time to the period between 1800 and 0600 h of the following day. We implemented a reliable multivariate method to remove anomalous feeding behaviour records. We then defined novel traits for feeding behaviour that apply to both types of feeding regimes, and for social behaviour that indicates an animal's rank within the cage hierarchy. We based these traits on feeder records and a biologically sound definition of a meal. Finally, we estimated the phenotypic correlations of those traits with growth and feed efficiency traits. Our findings demonstrate that variables about resource distribution among cage mates and an animal's priority for feed access were found to be good indicators of an animal's dominant or subordinate status within the cage. Based on results obtained in R animals (results were similar in AL animals), the most efficient animals were those that ate less frequently (phenotypic correlation with feed conversion ratio, rho = 0.6), and consumed smaller amounts per meal (rho = 0.7), spent less time at the feeder (rho = 0.4), and appeared to be subordinate, as they did not have priority access to the feeder (rho = -0.3), and had the smallest share of resources (range of rho = 0.2-0.6). We conclude that quantifying feeding and social behaviour traits can enhance the understanding of the mechanisms through which individuals exert their effects on the performance of their cage mates.

Consequences of variation in feeding behaviour for the probability of animals starting a meal as estimated from pooled data

Better understanding of how animals regulate their intake may be gained by analysis of feeding behaviour. This is often recorded in terms of feeding events, e.g. visits to feeders, which can be clustered into meals. This enables calculation of the probability of animals starting a meal in relation to time since the last meal, which is thought to give insight into food intake regulation. Starting probabilities are often calculated with pooled data but recent work suggests that pooling may strongly affect conclusions.

In this study we analysed feeding behaviour of cows to investigate how previous conclusions about feeding behaviour may have been affected by pooling. Using parameters derived from experimental data, we constructed simulation models to further explore under what circumstances pooling, either across day and night or across individuals, could affect the interpretation of starting probabilities. Data were simulated to explore the consequences of pooling as either the proportion of meals occurring during the day or the individual variation in their mean number of meals per 24 h changed. Simulation allowed us to extend the analysis of the consequences of pooling for the interpretation of starting probabilities.

Analysis of experimental data, collected with 16 dairy cows, showed that they ate a mean of six meals per 24 h. Individual variation resulted in a proportional CV of the individual mean number of meals per 24 h of 0·14. Cows ate a mean proportion of 0·59 of their meals during the day. Analysis of experimental data suggested that pooling, conducted in previous studies, has probably led to a quantitative underestimation of the increase in starting probability with time since the last meal but not a qualitative misinterpretation of the direction of change in the starting probability.

Simulation studies showed that pooling had no serious consequences when the mean number of meals per 24 h, or the variation about this mean, was low. However, as the number of meals per 24 h and variation increased, pooling led to conclusions that may wholly misrepresent both magnitude and direction of the change in starting probabilities calculated separately for the individuals or for day and night. This may explain why the results of some published studies seem not to agree with biological principles of food intake regulation.

Review: Quantifying animal feeding behaviour with a focus on pigs

The study of animal feeding behaviour is of interest to understand feeding, to investigate the effect of treatments and conditions or to predict illness. This paper reviews the different steps to undertake when studying animal feeding behaviour, with illustrations for group-housed pigs. First, one must be aware of the mechanisms that control feeding and the various influences that can change feeding behaviour. Satiety is shown to largely influence free feeding (ad libitum and without an operant condition) in animals, but 'free' feeding seems a very fragile process, given the many factors that can influence feeding behaviour. Second, a measurement method must be chosen that is compatible with the goal of the research. Several measurement methods exist, which lead to different experimental set-ups and measurement data. Sensors are available for lab conditions, for research on group-housed pigs and also for on-farm use. Most of these methods result in a record of feeding visits. However, these feeding visits are often found to be clustered into meals. Thus, the third step is to choose which unit of feeding behaviour to use for analysis. Depending on the situation, either meals, feeding visits, other raw data, or a combination thereof can be suitable. Meals are more appropriate for analysing short-term feeding behaviour, but this may not be true for disease detection. Further research is therefore needed. To cluster visits into meals, an appropriate analysis method has to be selected. The last part of this paper provides a review and discussion of the existing methods for meal determination. A variety of methods exist, with the most recent methods based on the influence of satiety on feeding. More thorough validation of the recent methods, including validation from a behavioural point of view and uniformity in the applied methods is therefore necessary.

The temporal structure of feeding behavior

Meals have long been considered relevant units of feeding behavior. Large data sets of feeding behavior of cattle, pigs, chickens, ducks, turkeys, dolphins, and rats were analyzed with the aims of 1) describing the temporal structure of feeding behavior and 2) developing appropriate methods for estimating meal criteria. Longer (between-meal) intervals were never distributed as the negative exponential assumed by traditional methods, such as log-survivorship analysis, but as a skewed Gaussian, which can be (almost) normalized by log-transformation of interval lengths. Log-transformation can also normalize frequency distributions of within-meal intervals. Meal criteria, i.e., the longest interval considered to occur within meals, can be estimated after fitting models consisting of Gaussian functions alone or of one Weibull and one or more Gaussian functions to the distribution of log-transformed interval lengths. Nonuniform data sets may require disaggregation before this can be achieved. Observations from all species were in conflict with assumptions of random behavior that underlie traditional methods for criteria estimation. Instead, the observed structure of feeding behavior is consistent with 1) a decrease in satiety associated with an increase in the probability of animals starting a meal with time since the last meal and 2) an increase in satiation associated with an increase in the probability of animals ending a meal with the amount of food already consumed. The novel methodology proposed here will avoid biased conclusions from analyses of feeding behavior associated with previous methods and, as demonstrated, can be applied across a range of species to address questions relevant to the control of food intake.

Measuring meals: structure of prandial food and water intake of rats

Attempts to understand ingestion have sought to understand the control of meals. The present study evaluated a meal definition that included prandial drinking (drinking-explicit meals). The spontaneous nocturnal intake of male Wistar rats was studied. The meal breakpoint was defined as the interval between feeding or drinking events providing the most stable estimate of meal structure. Alternative breakpoints derived from prevailing methodology, log-survivorship, or frequency histogram analysis of interfeeding intervals without respect to drinking were compared (drinking-naive meals). Threshold interfeeding intervals that accounted for drinking indirectly were evaluated as surrogate breakpoints (drinking-implicit meals). Definitions were compared by determining which criterion better conformed to predictions of satiety. Microstructural differences resulting from the definitions were also studied. Under the drinking-explicit definition, rats averaged nine or ten 13-min meals/night, during which they consumed food and water equally in duration (5–6 min) and quantity (2.3 g). Individual differences were observed in microstructure measures. Meals defined by drinking-informed, but not drinking-naive, methods were followed by the behavioral satiety sequence and by an initially low likelihood of resuming feeding that monotonically increased with time. The drinking-explicit definition uniquely revealed preprandial and postprandial correlations of similar magnitude. Under drinking-informed definitions, food restriction increased meal size, whereas drinking-naive definitions increased meal frequency. Drinking-implicit definitions provided workable approximations of meal frequency and size but inferior estimates of feeding duration, eating rate, and the preprandial correlation. Thus log-survivorship analysis is not appropriate for identifying meal breakpoints, and the consideration of drinking in meal definitions can provide a better estimate of meal structure.

The relationship between meal composition and long-term diet choice

When animals are offered a choice of feeds that are nutritionally complementary, they are able to select a consistent combination of these feeds over long periods of time. Analysis of how such consistent diet choice is achieved, in terms of short-term feeding behavior, may further our knowledge of how animals regulate nutrient intake. Previous work, on meal pattern analysis and on nutrient synchronization, led us to hypothesize that animals may select a consistent diet within a meal. In three experiments cows were offered a choice between high- (H) and low- (L) protein feeds and short-term feeding behavior data were collected using computerized feeders. Feeding behavior was first analyzed in terms of visit characteristics. A greater average daily intake of H, relative to L, was more closely related to the ratio of H visits to L visits than to differences in the intake per visit to feeders supplying H or L. Individual meal criteria were estimated using a mixed-distribution model, and visits were clustered into meals. Cows typically had approximately six meals per day. The observed frequency distribution of meal composition, in terms of the proportion of visits to H feeders, was determined. Subsequently, the observed visits were randomly reclustered into bouts consisting of the same number of visits as were observed in meals, and the frequency distribution of random bout composition was calculated. If frequency distributions of meals and random bouts coincide, then this is evidence that cows do not regulate diet choice within a meal. Comparison of the frequency distributions of meals and random bouts provided no evidence that cows attempted to achieve their long-term average diet composition within a meal. We also investigated whether cows tried to achieve a consistent diet choice within a meal by adjusting their intake per visit, depending on the feed type visited and the proportion of visits to H feeders in a meal. There was no evidence that this occurred. In conclusion, our analyses have shown that cows did not attempt to select within a meal a consistent diet in terms of protein to energy ratio. Indeed, our data and the literature suggest that the timeframe over which the intake of energy and protein is regulated must be greater than a meal in a number of animal species.

The short-term feeding behavior of growing pigs fed foods differing in bulk content

We investigated the effects of foods of different bulk on the short-term feeding behavior (STFB) of 16 individually housed pigs. The three foods used had different bulk contents [low-control (C), medium-70% wheat bran (WB), and high-70% sugar beet pulp (SBP)]. We expected the different intakes of the foods to be reflected in differences in STFB. Three hypotheses were developed based on ideas about the way in which a physical constraint to intake could arise. H(1): there would be less diurnal variation in feeding on high-bulk foods that limit food intake. H(2): feeding patterns on the bulky foods would be less flexible than those on C. H(3): a change in food type would result in food intake and STFB being rapidly altered to become appropriate to the new food. There were significant differences in food intake and STFB between the foods as intended. Pigs fed WB and SBP spent longer eating and had a slower feeding rate (FR) than pigs fed C. H(1) was rejected, as there was no difference in diurnal variation in intake between the foods. Feeding was not extended into the night on WB and SBP and the proportion of feeding that occurred during the night was the same for all three foods. H(2) was supported, as pigs fed WB and SBP were unable to maintain food intake and performance when time of access to the feeder was reduced. There was no adaptive change in STFB. H(3) was supported as a change from WB or SBP to C, or vice versa, caused a rapid change in STFB so that it became appropriate to the new food. It is concluded that physical constraints to food intake, caused by food bulk, may bring about changes in STFB and that they are important for the regulation of intake of such foods.

The use of mixed distribution models to determine bout criteria for analysis of animal behaviour

Feeding behaviour consists of feeding events, separated by non-feeding intervals. Feeding events are often clustered into bouts, which may be called meals. Grouping feeding events into meals requires the determination of a bout or meal criterion, that is, the longest interval accepted as part of a meal. Tolkamp & Kyriazakis (1999a) proposed a three-Gaussian model to estimate meal criteria. The three Gaussians each described the frequency distribution of the log(e)-transformed lengths of a population of intervals. These populations were thought to be: (1) short intervals within meals; (2) intervals within meals during which animals drink; (3) intervals between meals. This model predicted that the probability of an animal starting a meal would first increase, and then decrease with time since the last meal. This contrasts with expectations based on the satiety concept, which predicts that the probability of an animal starting a meal will increase with time since the last meal. This discrepancy is related to the symmetrical nature of the Gaussian distribution. Alternatively, the two-parameter Weibull distribution can take a skewed form and perhaps is more suitable to describe the different populations of intervals. In this study, models consisting of combinations of Gaussian and Weibull distributions were examined for their suitability to describe the observed feeding behaviour of cows. Weibulls did not improve the description of the populations of within-meal intervals, compared to Gaussians. However, the Weibull distribution was found to describe the between-meal population of intervals statistically better than the Gaussian. Additionally, this inclusion of a Weibull, as opposed to a Gaussian, resulted in predictions that were in better agreement with the satiety concept over the entire range of interval lengths observed. A model based on Gaussians to describe the within-meal populations of intervals and a Weibull to describe the population of between-meal intervals is, therefore, proposed. This model leads to biologically more satisfactory estimates of bout criteria than previous models and is likely to be applicable both across species and behaviours.