Edited by: Paul Koene, Wageningen University, Netherlands
Reviewed by: Birte L. Nielsen, INRA, France; Céline Tallet, INRA UMR PEGASE, France
This article was submitted to Animal Behavior and Welfare, a section of the journal Frontiers in Veterinary Science.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Spatial cognition in vertebrates is adversely affected by a lack of environmental complexity during early life. However, to our knowledge, no previous studies have tested the effect of early exposure to varying degrees of environmental complexity on specific components of spatial cognition in chickens. There are two main rearing systems for laying hens in the EU: aviaries and cages. These two systems differ from one another in environmental complexity. The aim of the present study was to test the hypothesis that rearing in a barren cage environment relative to a complex aviary environment causes long-lasting deficits in the ability to perform spatial tasks. For this purpose, 24 white Dekalb laying hens, half of which had been reared in an aviary system and the other half in a conventional cage system, were tested in a holeboard task. Birds from both treatment groups learnt the task; however, the cage-reared hens required more time to locate rewards and had poorer levels of working memory. The latter finding supports the hypothesis that rearing in a barren environment causes long-term impairment of short-term memory in chickens.
Animals must be able to perceive, store, and retrieve information in order to navigate their environment and maximize the ratio of benefits to costs. Birds should have good spatial cognition, allowing them to remember specific routes and landmarks so as to optimally utilize resources such as food, water, perches, and nests. They also need to use their knowledge of routes and landmarks effectively to escape potentially dangerous situations such as attacks that may have fatal consequences. Spatial learning and memory are, therefore, important for the fitness and survival of mobile species living in a complex environment. However, developing and maintaining cognitive ability is likely to be costly with regards to the energy required for neurogenesis and establishment of neural pathways (
Conventionally, laying hens in the EU are raised on specialized rearing farms to 15–18 weeks, at which time they are delivered to a specialized producer. They begin producing eggs between 18 and 22 weeks of age and are killed at 72–80 weeks of age, making space for a new set of birds. There are two major rearing systems for laying hens in the EU: aviary systems and cage systems. These two systems differ substantially from one another in, among other factors, environmental complexity. Rearing cages are barren environments containing 25 birds with access to food, water, and perches. Movement is restricted due to cage size. In the aviary-rearing system, at least 15,000 chickens are kept inside a large barn and are able to move both horizontally and vertically within it. Food, drinking nipples, and perches are available in specific locations on platforms elevated above the floor, and the chickens must navigate in order to access these. Differences in the early rearing environment have been shown to cause pronounced and long-lasting effects on spatial skill in domestic chickens (
To test whether rearing in these different environments influences spatial cognition in laying hens with the same genetic background, a holeboard task was used (
Both forms of working memory contain information that is trial dependent, and are thought to be forms of short-term memory (
To our knowledge, this is the first time a spatial holeboard task (
Non-beak trimmed, female white Dekalb chickens (
All the birds were exposed to the same light intensity, light schedule, and temperature, as recommended by the General Management Guide for White Dekalb Commercial Layers (
All the birds were housed in a single room in a Natura Primus 1600 system (Big Dutchman; Figure
Upon delivery to the rearing farm, immediately after hatching, all the chicks were initially placed in cages on the first and second tiers. At 4 weeks of age, the aviary-reared birds (half the birds in the house) were released from their cages by opening the doors, allowing them to move between the corridor floor and each aviary tier on either side of the corridor, until the end of the rearing phase at 16 weeks. Meanwhile, the cage-reared birds (the remainder) were kept in cages on the first and second tiers. The aviary-reared and the cage-reared birds were housed in separate corridors throughout the rearing phase.
The holeboard test can be used to quantify working memory, general working memory, and reference memory under different conditions following habituation to the test arena. An initial acquisition phase is used to test the birds’ ability to learn the location of baited cups without the provision of specific cues. A second cued phase involves the addition of novel cues associated with baited cups. Following the first trial during which neophobic responses may be observed, the cued phase introduces additional information that may improve cognitive performance relative to performance in the uncued task. This is followed by a third over-training phase in which cues are again removed in order to re-establish scores for baseline performance. The fourth reversal phase involves testing the birds’ ability to learn the location of rewarded cups after the introduction of a new uncued configuration. This last phase introduces a change that requires birds to replace previous information regarding the configuration of rewarded cups with information about the new configuration.
The holeboard test comprised a modification of methods described by Nordquist et al. (
The birds were habituated to the cups over 7 days, starting upon arrival at the experimental facilities at 16 weeks old. They were then habituated to the holeboard apparatus over 5 days starting at 17 weeks old. Following habituation, the birds were trained and tested individually in two trials per day over 28 working days (
All 48 birds (24 from the aviary-rearing treatment and 24 from the cage-rearing treatment) were habituated to the blue cups that contained rewards in the holeboard, and trained for a week to associate the cups with a mealworm reward by provision of live mealworms in the cups in the home pen three times daily. Next, these birds were allowed to habituate to the holeboard test arena for five consecutive days in daily sessions of 5 min. During the first habituation session, the birds were exposed to the apparatus in pairs of the same rearing treatment; for the remainder of the habituation, training, and testing sessions, the birds were exposed to the arena alone. During habituation sessions, all nine cups contained one mealworm. The habituation sessions were terminated when each bird had found and eaten all nine mealworms or 5 min had elapsed, whichever occurred first. According to a pilot study performed by our group, 33% of chickens fail to find any mealworms after several training days. A subset of 24 birds (
During holeboard training and testing, three cups were baited with worms and the remaining six left empty. The configuration refers to the spatial position of the baited cups in the arena. Six configurations were randomly chosen and four chickens were randomly assigned to each configuration. The order in which different birds were tested on a given day was randomized. The chickens were trained to find the three baited cups among the nine cups in the holeboard without any specific cues to guide them. The mealworms were only visible to birds after they had chosen a cup by entering the circle surrounding it. At the start of each test, the chicken was placed in the top left corner of the holeboard (Figure
For uncued acquisition training on test days 1–14 (trial 1–28), the same configuration of baited cups was used for each bird. During cued acquisition on test days 15–19 (trials 29–38), extra cues were added to the three baited cups in the form of a colored plywood base (red instead of the normal light wood color) without changing the configuration of baited cups. For over-training on test days 20–24 (trials 39–48), the baited cups were returned to their uncued form and the birds were further trained (over-training) in the uncued format of the holeboard. In the reversal phase on test days 25–28 (trials 49–56), the chickens were trained to find mealworms in a new configuration of baited, uncued, cups.
This experimental work was approved by the Institutional Animal Care and Use Committee at NMBU under ID number 6189.
The following measures were noted and/or calculated for each trial. Trial duration was defined as the total duration until all the mealworms had been eaten, or the maximum of 5 min had elapsed. Working memory was defined as the ratio of rewarded visits to the number of visits to the baited holes. This ratio reflects the chickens’ ability to avoid re-visits to the baited set of holes within the trial (
The effect of rearing environment (treatment) on the four parameters described above was tested in a repeated measures ANOVA, with bird as random factor nested in treatment, and treatment and phase as fixed factors. The interaction between treatment and phase was included in the model. Phase (uncued acquisition, cued acquisition, over-training, and reversal) was the repeated factor. The trial duration data did not fulfill all of the assumptions of ANOVA (equality of variance and normality of residuals), and was therefore transformed using a Box–Cox transformation. Where significant interactions were found, the data were subjected to a
Two chickens from the cage-reared treatment and one from the aviary-reared treatment did not search for bait in the holeboard, despite extensive training. Their data were excluded from the statistical analyses, reducing the number of individuals in the cage and aviary-reared treatments to 10 and 11, respectively. Mean values for trial duration, working memory, general working memory, and reference memory for each treatment during each holeboard phase are presented in Table
Aviary |
Cages |
|||
---|---|---|---|---|
Mean | ±SEM | Mean | ±SEM | |
Uncued acquisition | 187.23 | 6.89 | 172.57 | 7.24 |
Cued acquisition | 43.06 | 6.61 | 92.95 | 11.43 |
Over-training | 52.02 | 7.18 | 93.19 | 11.23 |
Reversal | 7.02 | 14.51 | ||
Uncued acquisition | 0.67 | 0.02 | 0.66 | 0.02 |
Cued acquisition | 0.90 | 0.016 | 0.79 | 0.027 |
Over-training | 0.85 | 0.019 | 0.87 | 0.019 |
Reversal | 0.02 | 0.045 | ||
Uncued acquisition | 0.77 | 0.014 | 0.74 | 0.016 |
Cued acquisition | 0.88 | 0.015 | 0.83 | 0.021 |
Over-training | 0.83 | 0.019 | 0.87 | 0.017 |
Reversal | 0.73 | 0.021 | 0.73 | 0.03 |
Uncued acquisition | 0.39 | 0.013 | 0.41 | 0.013 |
Cued acquisition | 0.62 | 0.022 | 0.59 | 0.024 |
Over-training | 0.50 | 0.018 | 0.54 | 0.017 |
Reversal | 0.36 | 0.014 | 0.27 | 0.018 |
Parameter | Statistics ( |
||
---|---|---|---|
Treatment | Phase | Treatment x phase | |
Trial duration | |||
Working memory | |||
General working memory | |||
Reference memory |
Trials | Working memory– general working memory |
Working memory– reference memory |
General working memory– reference memory |
|||
---|---|---|---|---|---|---|
1 | 0.1282 | 0.579 | 0.2791 | 0.220 | 0.1310 | 0.571 |
2 | 0.1801 | 0.435 | 0.1700 | 0.463 | ||
3 | 0.1569 | 0.497 | ||||
4 | ||||||
5 | 0.1075 | 0.643 | 0.1246 | 0.590 | ||
6 | ||||||
7 | 0.0747 | 0.748 | 0.2212 | 0.335 | ||
8 | 0.1700 | 0.461 | 0.1130 | 0.626 | ||
9 | 0.2430 | 0.288 | 0.1988 | 0.388 | ||
10 | ||||||
11 | −0.0438 | 0.850 | −0.2016 | 0.381 | ||
12 | 0.3346 | 0.138 | 0.2325 | 0.310 | ||
13 | 0.0197 | 0.932 | ||||
14 | 0.2856 | 0.209 | ||||
15 | 0.3597 | 0.109 | 0.3587 | 0.111 | ||
16 | 0.3488 | 0.121 | ||||
17 | 0.2133 | 0.353 | ||||
18 | ||||||
19 | 0.2221 | 0.333 | ||||
20 | ||||||
21 | ||||||
22 | 0.3237 | 0.152 | ||||
23 | ||||||
24 | 0.1380 | 0.551 | ||||
25 | ||||||
26 | −0.0720 | 0.756 | 0.2646 | 0.246 | ||
27 | 0.1707 | 0.459 | 0.3299 | 0.144 | ||
28 | ||||||
29 | ||||||
30 | ||||||
31 | 0.0673 | 0.771 | ||||
32 | 0.0596 | 0.797 | 0.2325 | 0.310 | ||
33 | 0.1087 | 0.039 | 0.0781 | 0.737 | ||
34 | 0.1964 | 0.393 | ||||
35 | ||||||
36 | ||||||
37 | ||||||
38 | ||||||
39 | 0.6320 | 0.002 | ||||
40 | ||||||
41 | ||||||
42 | ||||||
43 | 0.1750 | 0.448 | ||||
44 | 0.3039 | 0.180 | ||||
45 | ||||||
46 | ||||||
47 | 0.1844 | 0.424 | 0.3677 | 0.101 | ||
48 | ||||||
49 | ||||||
50 | 0.2342 | 0.307 | 0.1387 | 0.548 | ||
51 | ||||||
52 | 0.0524 | 0.821 | 0.1836 | 0.426 | ||
53 | 0.0304 | 0.896 | 0.2409 | 0.293 | ||
54 | 0.3604 | 0.108 | ||||
55 | ||||||
56 |
Trial | Working memory-trial duration (Box–Cox transformed) |
|
---|---|---|
49 | ||
50 | −0.3654 | 0.103 |
51 | ||
52 | ||
53 | ||
54 | ||
55 | ||
56 |
Trial duration and all memory parameters (Table
Correlations between working memory and general working memory were mostly weak to moderate during the first half of the acquisition phase, and most of the reversal phase, when performance was not at its peak (Table
There was no main effect of rearing treatment on any of the holeboard variables (Table
The results show the effects of the rearing environment on working memory and trial duration in a holeboard task, and support the hypothesis that rearing in a barren cage environment relative to a complex aviary environment causes long-lasting deficit in the ability of chickens to perform a spatial task. With the exception of the three birds that did not learn the task, the effects of phase indicated that the holeboard task was a valid approach to quantifying working memory, general working memory, and reference memory in laying hens. This was confirmed by higher average scores for working memory, general working memory, and reference memory during cued acquisition and over-training than during uncued acquisition. The present study also supports the previous finding with chickens (
Working memory is considered to be a form of short-term memory, while reference memory is considered to be a form of long-term memory (
Reference memory, as opposed to working memory, is memory of the general rules of the task, such as the fact that holes may or may not contain food rewards. It holds information that is relevant across several of the trials and is, therefore, trial independent (
In a typical aviary environment, chickens have ample opportunity to move in three-dimensional space and to perform a wide range of natural behaviors such as wing flapping, dust bathing, and flying. They also have both positive and negative contact with a large number of conspecifics. In the case of negative (antagonistic or aggressive) social interactions, a subordinate chicken has the option of moving away from the area to avoid or escape the attacker. The chickens must also be able to find food troughs, drinking nipples, nest boxes, and perches throughout the aviary.
In a furnished cage system, the chickens have very limited space in which to move. Some natural behaviors such as wing flapping and flying are difficult to perform. Vertical movement is limited to about 50 cm. Each hen normally has physical and social contact with 8–10 other hens, as for example, in the modified Victorsson T10 cages used in the present study. All resources available to each hen are within the cage and, therefore, the birds need not search for these. A caged laying hen’s environment may thus present her with cognitive challenges that are similar to the challenges met in the first three phases of the holeboard (uncued acquisition, cued acquisition, and over-training). The environment that a caged hen experiences is normally very stable. In this type of surroundings, reference memory is arguably the most relevant memory component, one that holds the general rules and facts about the environment on which the chickens can always depend. In the aviary system, however, each hen has the potential to find herself in a wider range of situations, both in terms of location and social interaction. In this case, working memory is likely to be valuable, as it allows the chicken to interpret stimuli based on each individual situation, and to navigate through a complex environment that may change depending on her location in the house and elevation above the floor. Overall, cognition is favored in environments with greater spatial variability than those that are stable (
No rearing treatment effects on general working memory were found. General working memory is the ratio of the number of unique holes visited to the total number of hole visits. During the reversal phase, where treatment effects were found in working memory, general working memory for both treatments showed a similar decrease. In this phase, the configuration of cups was changed, forcing the chickens to explore more cups to find the food rewards. The reduction during this phase in working memory and general working memory for cage-reared birds, but only in general working memory for aviary-reared birds, indicates that cage-reared chickens revisited both baited and unbaited cups. In contrast, aviary-reared birds revisited only the cups that may have been baited in previous phases, and are now unbaited, thus reducing general working memory but not working memory. This further underpins the suggestion that aviary-reared birds have better short-term memory than cage-reared birds.
The previous discussion partly rests on the assumption that spatial working and reference memory are psychologically distinct. Correlational data and a lack of correspondence between treatment effects influencing one indicator but not the other, from studies in rats and mice, support the idea that they are independent [reviewed by van der Staay et al. (
The corresponding adverse effects of cage rearing on trial duration and working memory are interesting in view of questions raised by van der Staay et al. (
An aviary-rearing system, with its complexity and higher opportunity for novel situations, seems to prepare chickens to cope with new tasks by increasing their ability to retain short-term spatial and temporal information about the environment. This may be possible through increased neuroplasticity (
General working memory levels were variable during the uncued acquisition phase of the holeboard task. As previously mentioned, working memory is a ratio of rewarded visits to the number of visits to the baited set of holes, while general working memory is the ratio of all holes visited to the total number of visits to any hole. This high degree of variability observed for general working memory may therefore indicate that chickens are better at remembering events when they have recent memory of both successful and failed attempts to find the food rewards. This suggests that the information the hen acquires from visiting an empty hole is as informative as the information acquired from a visit to a baited hole. During trials where the chickens performed well and visited only baited cups, the trial was terminated as soon as all mealworms had been found, so the chickens had no chance to explore the other cups. This could then result in poorer performance in the following trial as chickens had to explore other cups as well as the ones with mealworms. Indeed, a similar phenomenon has been observed in studies of spatial memory in other non-caching bird species, with authors suggesting a distinct effect of proactive interference, that is, the information about a rewarded site is influenced by the exploration of other sites prior to finding the reward (
The results of this study support the hypothesis that rearing in a barren cage environment relative to a complex aviary environment causes long-lasting deficit in the ability to perform a spatial task, as indicated by effects on chickens’ working memory. Exposure to varying degrees of early environmental complexity thus influences how well birds remember the type of stimulus presented, when it was presented, and where this happened. Furthermore, the effects documented in the present study were rather long-term, as the last treatment effects were found over two months after birds were removed from the rearing environment.
FT participated in the design of the study, carried out data collection and data analysis, and drafted the manuscript; RN participated in the design of the study and drafted the manuscript; JN participated in the design of the study, carried out data analysis, and drafted the manuscript; AJ participated in the design of the study and drafted the manuscript. All authors approved the final manuscript.
No conflicts of interests exist in regards to this study. The funding organizations the Foundation for Research Levy on Agricultural Products (FFL), the Agricultural Agreement Research Fund (JA), and Animalia (Norwegian Meat and Poultry Research Centre) finance applied agricultural research in collaboration with the private and public sectors. These parties’ sole interest in the present study was to support publication of the unbiased results in order to provide advice to poultry rearers.
This work was funded by the Foundation for Research Levy on Agricultural Products (FFL), the Agricultural Agreement Research Fund (JA), and Animalia (Norwegian Meat and Poultry Research Centre) through the Research Council of Norway, grant number 207739. Anne Lene Hovland provided helpful comments to the manuscript.
1
2