Identification of specific call types produced by pre-weaning gilts in response to isolation

Abstract

Monitoring the vocalizations produced by livestock in commercial facilities can be used to remotely assess the arousal, and perhaps welfare, of the animals. Domestic swine is an ideal candidate for this type of assessment as it is a highly social and vocal mammal, however, a standardized description of the vocal repertoire of domestic swine has yet to be established. The objectives of this study were (1) to identify specific vocalizations produced by pre-weaning gilts in response to brief 5 min isolation from dam and natal farrowing pen using advanced acoustic analysis, and, (2) examine the relationship between such vocalizations and behaviors displayed during isolation in order to examine the potential for specific vocalizations, or “call types”, to represent different arousal levels in the animal. Vocalizations were classified according to acoustic and spectral characteristics such as peak frequency, duration, frequency range, and harmonics. Over 14,000 vocalizations across 40 gilts were analyzed and categorized into seven call types (A–G). The most frequent vocalizations identified were type E calls (squeals, 28.9 %), C calls (croaks, 21.9 %), D calls (high frequency grunts, 18.4 %), and A calls (low frequency grunts, 18 %). Call type was found to co-vary with several behavioral responses to isolation. C calls (croaks) were positively correlated (P < 0.05) to behaviors interpreted as pacing in the novel environment, while G calls (low frequency grunt variant) were positively correlated (P < 0.05) to the behaviors related to active exploration of novel environment. There were no effects of litter or sow parity on the proportion of call types produced or behaviors displayed in isolation, however, there were noticeable individual differences in vocal and behavioral responses. This study demonstrates that humans may perceive multiple call types as the same vocalization produced by gilts, and therefore, more advanced acoustic analysis is needed to create an audiogram for swine objectively. Given that gilts were observed to produce the same call types, while displaying different behavioral responses to isolation (freeze, explore, or, escape), it is difficult to narrow down the valence or arousal strength of a gilt’s response to this stress-inducing context. Future research may include physiological and/or cognitive indicators of affective state in order to refine the meaning behind each call type expressed by pre-weaning gilts in isolation.