Barry Keverne

Facial and vocal discrimination in sheep

The ability of sheep, Ovis aries, to discriminate between sheep, humans and other animals on the basis of facial and vocal cues was tested in an enclosed Y-maze. Pairs of faces or voices were presented which had a clear differential significance for the sheep (i.e. sheep versus human, dog or an unfamiliar breed or species). Some Clun Forest and Dalesbred sheep could actively distinguish (>75% choice) between different faces, with the best performances being seen for sheep versus human faces (the sheep face being preferred to that of the human). Other animals showed pronounced position preferences in the maze whatever faces were shown although they approached sheep faster than other faces, suggesting that they could discriminate between them. Dalesbred, but not Clun Forest, sheep could also discriminate between sheep and human vocalizations alone in the maze. For both breeds, combining appropriate sight and sound stimuli did not significantly enhance performance although mismatching them reduced it. In a further experiment on animals that performed at more than 75% choice criterion in the maze, inverting the faces, turning them to profile or masking the eyes all significantly reduced performance in discriminating between sheep and human faces. In another experiment, Clun Forest ewes could also distinguish between the faces of male and female breed members. During anoestrus they preferred the faces of females and when they were in oestrus they preferred those of rams. Overall, these results suggest that sheep can use facial cues to discriminate between different species, breeds and male and female members of the same breed. They also show that discriminatory performance is influenced by orientation, the presence of eyes and, to some extent, vocal cues.

Are faces special for sheep ? Evidence from facial and object discrimination learning tests showing effects of inversion and social familiarity

We have previously shown that sheep, like monkeys, have neural circuits within the temporal lobe that respond preferentially to faces. They can also discriminate between sheep, humans and other animals on the basis of facial cues using an enclosed Y-maze. In the present study we investigated the speed with which Clun Forest sheep learn to discriminate between familiar and unfamiliar faces, as opposed to symbols, in order to gain a food reward using the same Y-maze apparatus. Animals (n = 10) received 1 day of training where projected images of the pairs of faces or symbols were paired for 20 trials with a picture of either an empty or full bucket of food (which indicated which choice of face or symbol would result in the animal receiving a food reward) and on the next 4 days they were given a further 20 trials a day with the faces or symbols alone. Results showed that sheep learned significantly faster (by day 1 or 2 post training) to recognise sheep faces of a familiar breed compared to geometrical symbols (3-4 days post training). Learning using faces of animals of another unfamiliar breed was also significantly better than for symbols but was significantly worse than that seen using faces of a familiar breed. Inverting the faces significantly reduced learning speed for faces of a familiar breed but not for that of an unfamiliar one. Inverting familiar objects, food buckets, also did not impair discriminatory performance. In a further set of trials where discrimination learning was made more difficult by excluding cued trials and reducing the number of daily trials to eight, social familiarity was found to further improve the animals ability to learn to discriminate between the faces of a familiar breed. Finally, while discriminatory performance for adult sheep faces was very good, that for young lamb faces was poor, with only one animal learning to choose the face associated with food. It was confirmed in maternal ewes that they were also slow to learn to recognise the faces of their lambs (2-3 weeks). Overall these results show that sheep can learn to distinguish between individual adult sheep faces but that breed and social familiarity influence the level of performance. Further, discrimination learning of familiar and unfamiliar facial stimuli is better than between simple geometrical symbols, indicating that faces may be preferentially processed by the brain compared to other objects suggesting that faces are indeed special in this species as has been claimed for human and non-human primates.

Monoallelic gene expression and mammalian evolution

Monoallelic gene expression has played a significant role in the evolution of mammals enabling the expansion of a vast repertoire of olfactory receptor types and providing increased sensitivity and diversity. Monoallelic expression of immune receptor genes has also increased diversity for antigen recognition, while the same mechanism that marks a single allele for preferential rearrangement also provides a distinguishing feature for directing hypermutations. Random monoallelic expression of the X chromosome is necessary to balance gene dosage across sexes. In marsupials only the maternal X chromosome is expressed, while in eutherian mammals the paternal X genes are silenced in the developing placenta and early blastocyst. These examples of epigenetic gene regulation commonly employ asynchrony of replication, the binding of polycomb proteins and antisense RNA, and histone modifications to chromatin structure. The same is true for genomic imprinting which among vertebrates is unique to mammals and represents a special kind of epigenetic modification that is heritable according to parent of origin. Genomic imprinting pervades many aspects of mammalian growth and evolution but in particular has played a significant role in the co‐adaptive evolution of the mother and foetus.

Differences in blood levels of androgens in female talapoin monkeys related to their social status

Serum testosterone and androstenedione levels were lower in the subordinate female talapoin monkeys of four social groups than either dominant or intermediate-ranking females. This was found in both intact or ovariectomized (oestrogen-treated) animals, which suggests that androgen from the adrenals contributed to this rank-related endocrine effect. These differences disappeared when the females were housed singly, levels in all animals becoming similar to those in subordinates in the group cage. There were no rank-related differences in progesterone levels during either the follicular or luteal phase of the cycle in intact females, or in those of ovariectomized females of different rank, but cortisol was highest in dominant group-living animals in these experiments. Significant correlations were found between androgen levels in group-living females and the amount of sexual interest shown in them by males; the amount of aggressive interaction involving each female did not correlate with her androgen levels. Social rank is defined according to the direction, not the amount, of aggression. These findings suggest that the social hierarchy regulates androgen levels in these female monkeys; there may also be effects on the ability of females to respond to their own, or to administered, androgen. Similar findings have been made previously in male talapoins. Since androgens fill a critical role in the sexual behaviour of both sexes in primates, this may be a neuroendocrine mechanism of general significance relating behaviour to social rank.

Importance of Genomic Imprinting in the Evolution and Development of the Maternal Brain

It was the French reproductive biologist, Alfred Jost (1970), who proposed that mammalian sexual differentiation is biased in a female direction and masculine characteristics are imposed on an essentially female life plan. The reproductive success of mammals places a considerable burden of time and energy on the matriline, with some 95 % of female adult life committed to pregnancy, lactation and maternal care. Viviparity has thus provided a major selection pressure on the matriline in the evolution of these events and with particular emphasis on the placenta and hypothalamus. Increased maternal feeding, maternal care, suspension of fertility and sexual behaviour, parturition and milk provision are all integral to hypothalamic function and have evolved under the influence of the placental hormones to meet the demands of the developing infant (Keverne 2006). Viviparity has also introduced a new dimension to evolutionary genetics in providing the co-existence and continuity for three generations of matrilineal genomes (i.e., mother, developing offspring and developing oocytes) in one individual (Keverne 2011). Also unique to the mammalian matriline has been the evolution of epigenetic marks (imprint control regions) which are heritable and undergo reprogramming to regulate gene expression according to parent of origin. This imprinting of autosomal genes (genomic imprinting) plays a significant role in mammalian development, particularly development of the placenta and hypothalamus (Keverne 2009). Indeed, a number of imprinted genes are co-expressed in the placenta and hypothalamus and are important for the co-adapted functioning of these structures. Such transgenerational co-adaptation ensures the foetal hypothalamus is genetically and epigenetically programmed for ensuring optimal maternal care and nurturing (Broad and Keverne 2011). In this way the foetus not only controls its own destiny via the placenta but also that of the next generation via the developing hypothalamus.

Human evolution: brain, birthweight and the immune system

The large size of the human brain at birth is one of the defining features of our species, and yet comes at a price. Among the primates, humans have particularly difficult births, with high rates of maternal and fetal morbidity and mortality. Approximately 287 000 maternal deaths occurred in 2010