Cerebellum size is related to fear memory and domestication of chickens

Abstract

Red Junglefowl (Gallus gallus) were selected for divergent levels of fear of humans during eight generations, causing the selection lines to differ in fear levels as well as in the proportional brain and cerebellum masses. Birds from the two lines were then crossed to obtain an F3 intercross in order to study the correlations between brain mass and fear learning. We exposed 105 F3-animals individually to a fear habituation and memory test at 8 days of age, where the reactions to repeated light flashes were assessed on 2 consecutive days. After culling, the absolute and relative sizes of each of four brain regions were measured. Stepwise regression was used to analyse the effects of the size of each brain region on habituation and memory. There were no effects of any brain region on the habituation on day one. However, birds with a larger absolute size of cerebellum had significantly reduced reactions to the fearful stimuli on day two, indicating a better memory of the stimuli. No other regions had significant effects. We conclude that increased cerebellum size may have been important in facilitating chicken domestication, allowing them to adapt to a life with humans.

1. Introduction

A defining trait of the so-called domestication syndrome in animals is a reduced size of the brain relative to body mass [1]. However, different brain regions are affected differently by domestication [2], and in chickens, we previously found that the genetic architecture governing the size of the regions differs, allowing parts of the brain to respond to selection independently [3]. In line with this, domesticated chickens have smaller brains relative to body size compared to the wild ancestor, the Red Junglefowl (Gallus gallus), while the cerebellum is in fact considerably larger both in absolute and proportional measures [3]. In Red Junglefowl selected for either high or low fear of humans, we found similar effects: after only few generations of selection, the tamer birds had overall smaller brains relative to body size, but larger cerebellum than the fearful ones [4]. This indicates that the cerebellum may be involved in handling fearful stimuli in chickens, and that this may have been important during domestication when the normally very shy ancestral Red Junglefowl adapted to a life among humans.

The cerebellum was for a long time considered to be concerned mainly with locomotor control, but recent research shows that it is involved in a variety of cognitive processes [5–7]. Both positive and negative emotions appear to be processed by cerebellar circuits, with a predominance of the negative ones [8]. However, most knowledge about cerebellar involvement in emotional processing comes from mammals. For example, in humans and other primates, it is involved in cognitive development and social interactions [9], and in rats, it plays a role in fear-conditioning memory consolidation [10]. In the light of this, and our previous observations in chickens, we therefore hypothesized that cerebellum size may be related to central behavioural aspects relevant for domestication, such as the ability to habituate to fearful stimuli.

Here, we studied fear reactions in an F3-intercross between Red Junglefowl selected for high versus low fear of humans and measured the size and composition of their brains. The aim was to analyse how birds with different brain compositions habituate to and memorize a startling but harmless stimulus, as a measure of fear memory.

2. Material and methods

The complete dataset is available as electronic supplementary material, table S1.

Starting from an outbred population of Red Junglefowl, we selected separate lines depending on their responses in a standardized fear-of-human test. Details about the breeding and housing of the birds, as well as the selection scheme, have been reported previously [11]. A range of correlated selection responses was observed, including a relative decrease in brain mass and increase in relative cerebellum mass in the low fear line [4].

Using birds from the eighth selected generation, we bred an intercross by crossing two males and two females from each selection line with a reciprocal design. In the F1- and F2-intercross generations, birds were randomly bred to generate an F3-intercross consisting of a total of 105 animals, constituting our test population. Further details about the breeding scheme can be found in [12].

At 8 days of age, the chicks were tested in fear habituation and memory test, consisting of two test occasions interspersed by 24–28 h. On each of the two test occasions, one chick was placed on its own in a closed, evenly illuminated test arena, measuring 25 × 25 × 30 cm. Integrated into the floor was a LED light source emitting blue light, and after 30 s in the arena, the chicks were exposed to a series of five brief (1 s) light flashes with a 30 s interval in-between. Thirty seconds after the last flash, the chick was removed and replaced by the next animal.

The tests were recorded on video and the immediate startle responses of the chicks to each of the consecutive light flashes were scored using the Observer software (Noldus Inc) from the recordings according to a five-degree scale, where the lowest score (1) was recorded when no reaction to the flash could be observed, and the highest (5) for the maximal fear reaction. To validate the scoring, a random sub-sample (10%) of the videos were scored by two independent observers, and the observer score correlation was r_s = 0.91 (p < 0.001). Details and ethogram for the scores can be found in [12].

To obtain an overall measure of the intensity and waning of the fear response, we calculated the area under the curve (AUC) for each chick on each of the test days. The AUC was calculated by adding the areas under the curve for each of the five 30 s intervals. A lower AUC means lower overall reaction scores.

When the birds were 32 weeks old, all were weighed and culled by rapid decapitation. The brains were removed and dissected into four parts according to the protocol used by Henriksen et al. [3]: cerebral hemispheres, optic lobes, cerebellum and the remaining part containing, e.g. brainstem and hypothalamus. Immediately after dissection, each brain region was weighed separately, and the wet mass obtained using a scale with a precision of 0.001 g. Previous research has found a high correlation between wet mass and volume [3].

The measured masses of the brain regions had a distribution that did not significantly deviate from normality, as determined by inspection of Q-Q-plots (Shapiro–Wilks test p > 0.05 for all regions). To analyse the relationship between brain composition and fear reactions, we performed a stepwise regression analysis. The model consisted of AUC as response and included body weight and absolute weight of the total brain and of each of the four brain regions. The regression was performed separately for the first and second test occasion.

All deviations are given as standard errors of the mean.

3. Results

Males (N = 59) were significantly larger than females (N = 46) (1087.1 ± 10.4 g versus 791.9 ± 9.1 g; p = 0.031, t-test) and had larger absolute brain masses (males: 2.77 ± 0.02 g, females: 2.50 ± 0.02 g; p < 0.001, t-test). Consistent with the overall larger brain, the absolute masses of each of the separate brain regions were also significantly larger in males (p < 0.001, t-test). There were no sex effects on AUC either on the first test occasion (9.4 ± 0.4 versus 10.1 ± 0.6; p = 0.33, t-test), or on the second test (males: 8.1 ± 0.4; females: 8.8 ± 0.5; p = 0.32, t-test).

In the first fear test, most chicks reacted strongly to the first light flash and then gradually showed a less intense startle response to the following light flashes (figure 1), consistent with a fast habituation response. On the second test, the average startling reactions were less intense already at the first light flash (figure 1), indicating that the chicks remembered the stimulus from the first test and again showed evidence of within-test habituation. The difference between the intensity of the reactions is reflected in the AUC, which was significantly lower on the second test occasion (figure 1; t = 3.45 p < 0.01; paired sample t-test), indicating a consolidation of memory from the first test.

Figure 1.

Fear reaction scores (1 = lowest, 5 = highest startle reaction) to each of five consecutive light flashes (time) on two separate tests, separated by 24–28 h (day 1 and day 2, respectively).

In the stepwise regression model, no variables were retained as significant for the AUC at the first test occasion. For the second test, the absolute size of the cerebellum was the only significant predictor of AUC (R² = 0.041, F = 4.25, p = 0.042), where a larger cerebellum was associated with a smaller AUC (figure 2). The effect appeared more pronounced in males (figure 2), although the association did not reach significance in either of the sexes separately.

Figure 2.

Cerebellar mass plotted against area under the curve (AUC) for the consecutive fear startle reactions in the fear test on day 2. Higher AUC means more fearful reactions and slower habituation.

4. Discussion

We found that, in an F3-intercross between Red Junglefowl selected for high versus low fear of humans, birds with a larger cerebellum were significantly better at remembering and habituating to a fearful but harmless stimulus. This indicates that cerebellum size may have been an important factor for the successful domestication of chickens.

A larger cerebellum has previously been reported in domesticated White Leghorn egg layers compared to ancestral Red Junglefowl [3], and likewise in Red Junglefowl selected for reduced fear of humans [4], suggesting that the cerebellum may have important functions in relation to tameness and domestication in chickens. Our present results help corroborate this, since the ability to live and thrive close to humans must have been central to early domesticates [13], and this would undoubtedly have been facilitated by an increased ability to habituate to harmless stimuli that may initially have been perceived as frightening.

The brain is a costly organ. To optimize energy use, adaptive changes in brain size and composition therefore have evolved during domestication [1,2,14,15–17]. Consistent with the mosaic brain theory, different regions can evolve independently, as evidenced by the different genetic architectures underlying size and growth of different brain parts [2,3]. The cerebellum is a part of the vertebrate brain that has long been considered mostly a centre for locomotor control, but research during the past decades has shown that it has much wider functions. In humans, it is involved in social cognition and learning [5,18], and in chickens and quail, it affects reproductive behaviour [3,6]. In rats, the cerebellum plays an important role in the consolidation of fear memories, and our present results are consistent with this [10].

Our results show a significant relationship between cerebellar mass and fear habituation in chickens but do not allow any conclusions on what it is that causes this, and the effect is relatively weak with an R² of 4.1%. The cerebellum of birds has a different anatomy from in mammals, and its composition and connectivity differs even between breeds of the same species [2]. The effects we observe could, for example, be a result of an increased number of neurons or other cell types or increased neuronal density. Future research should focus on elucidating which aspects of the increased mass underlie the observed behavioural effects.

It is noteworthy that no other brain region size had a significant association with the learning process measured in the present experiment. However, with the relatively gross measurements used, we cannot rule out the involvement of other parts of the brain, such as, e.g. the amygdala or hypothalamus, and we were not able to assess the connectivity between the cerebellum and other areas.

Chickens were domesticated from Red Junglefowl at least 8000 years ago [19] and the ancestor in the wild is extremely shy and fearful towards humans [20]. Hence, a first and necessary step towards successful domestication must have been a reduction in fear, and it has been suggested that this was in fact a major driving force behind many typical domesticated traits [13]. In rabbits, brain structure modifications during domestication are consistent with altered fear processing [17] and domesticated chickens have an overall reduced brain mass relative to body size [3]. However, in domesticated chickens, the cerebellum makes up a larger proportion of the brain than in Red Junglefowl [3]. The same effects were observed after only a few generations of selection on Red Junglefowl for reduced fear of humans: the tamer selection line evolved a proportionally smaller brain as a whole and a proportionally smaller telencephalon, but a proportionally larger cerebellum [4]. This was associated with changes in fear memory [21]. The differences previously observed between the two selected Red Junglefowl-lines can be caused by many different factors, including genetic drift. However, the F3-intercross approach used here clearly demonstrates a genetic correlation between cerebellum mass and fear memory, suggesting that cerebellum size in chickens may have conferred adaptive value during domestication.

In conclusion, our results show that cerebellum mass in Red Junglefowl is associated with more efficient habituation and memory of a fearful but harmless stimulus. In the light of previous findings that the cerebellum is larger in domesticated chickens as well as in Red Junglefowl selected for reduced fear of humans, the present findings suggest that larger cerebellum mass may have facilitated the adaptation of wild Red Junglefowl to a life with humans, and hence may have been important for the successful domestication of chickens.

Ethics

The experiments were approved by the Linköping Animal Ethics committee, licence no 14916–2018.

Data accessibility

The complete dataset is available as electronic supplementary material, table S1.

Authors' contributions

R.K., D.W., R.H. and P.J. acquired data; R.K. and P.J. analysed data with substantial input from D.W. and R.H.; P.J. conceived and coordinated the study and drafted the paper with critical input from R.K, D.W. and R.H. All authors gave final approvement of the version to be published and agree to be held accountable for the contents therein.

Competing interests

We declare we have no competing interests

Funding

This study was funded by H2020 European Research Council (322206) and Vetenskapsrådet (2015-05444).