Mood induction alters attention toward negative-positive stimulus pairs in sheep

Mood is a lasting affective state that influences motivation and decision-making by pre-shaping a subject’s expectations (pessimism/optimism). Mood states affect biases in judgment, memory, and attention. Due to a lack of verbal report, assessing mood in non-human animals is challenging and is often compromised by intense training sessions. Measuring mood using attentional biases can circumvent this problem, as it takes advantage of observing a spontaneous reaction. As in humans, we expected that negative mood will heighten attention toward negative compared to positive stimuli. Here, we validate measures of attention toward acoustic stimuli in sheep (N = 64) and assess sheep’s differential attention toward acoustic stimuli before and after mood induction (N = 32). Mood was induced by manipulating the environment. We used animal vocalizations (dog barking and sheep bleating as negative and positive stimuli, respectively) varying in intensity and played simultaneously from one side each, and measured lateral attention based on the sheep’s behavior. Overall results were somewhat ambiguous. Yet, negative mood sheep seemed to shift their attention more toward dog vocalizations when the stimulus pair was well balanced at baseline. Though some adaptations are still needed, our approach could be a promising alternative to measure animals’ mood without prior training.


Preparation of the acoustic stimuli
We used two sets of stimuli. We used white noise (i.e. random noise with a uniform frequency spectrum over a wide range of frequencies: 20 Hz to 20 kHz) as a stimulus, for which we assumed no emotional content that depended on innate reactions or reactions learnt throughout ontogeny (apart from its novelty and the suddenness of its presentation). The direction of attention to one or the other side from where the noise was played back was measured. We used dog vocalizations as presumed negative and sheep vocalizations as presumed positive stimuli. These stimuli were played back simultaneously to assess the relative attention of the sheep toward the two types of stimuli. The amplitude of all these stimuli when not reduced was set to the same level. For the sheep vocalizations, we recorded the habituated sheep bleating while they anticipated the delivery of fresh hay and concentrates (a mixture of UFA 763 ProRumin COMBI QM, Herzogenbuchsee, Switzerland) at a time-point before the present experiment (approximately 12 long bleats per 10 s stimulus). As in Maigrot, et al. 2 and de la Torre, et al. 3 , the positive valence experienced by the bleating sheep was inferred from the context, considering that food enhances fitness and triggers approach behavior. The approach behavior could also be observed while recording the vocalizations used in this experiment. Sheep calls were recorded with the same devices as the dog calls from outside the home pens, at distances of 2 to 7 m. Calls amplitude was equalized and we built 15 sheep vocalizations' stimuli, i.e. direct successions of sheep bleating, lasting 10 s each.
Each dog and sheep vocalization stimulus was manipulated to obtain two intensities, i.e. one reduced and one full, mimicking two distances of the sound source, i.e. one close and the other further away. For the habituated sheep, both the dog and sheep stimuli were used at their full amplitude for the high intensity, and the general stimulus amplitude was halved using Praat v.6.0.37 DSP Package 1 for the low intensity. However, after the habituated sheep were tested and their behavior assessed, we decided to modify the amplitudes of the stimuli used for the naïve sheep to make the salience of the two types of stimuli even more comparable. In order to have a more similar loudness for the high intensity stimuli, the sheep stimuli were used at their full amplitude while the general dog stimuli amplitude was reduced (amplitude at 0.8) using Praat. In order to have a stronger difference between the two intensities, we adapted the method used by Naguib, et al. 4

in songbirds.
The low intensity stimuli were attenuated in a frequency-dependent manner, i.e. the relative intensities of high frequencies decreased more with increasing propagation distance. To apply this strong reduction of the stimuli amplitude, we filtered animal vocalizations at 2 dB/octave using a graphic frequency filter in Cool Edit Pro 2.1 (Syntrillium Software Cooperation, Phoenix, Arizona, U.S.A.). During a test, the sheep could hear four different barking dogs. Each vocal stimulus (at low or high intensity) was used only once per sheep and approximately the same number of times across all the sheep.

Definition of sheep's attention based on behavioural measurements
Attention was defined based on the sheep's behavior (ethogram in Supplementary Tab. S1). This ethogram was developed in a pilot test involving four sheep that were subjected to stimulus sequences close to those used in the final testing. Sheep behavior might have been accompanied by sounds, e.g. a writhing sheep moved around quickly and the feet on the floor could be heard as well as the body moving along the wire mesh of the box, that help us to confirm a specific behavior. In principle, we were interested in the overall attention that was directed to either side. Nevertheless, strong attention may be a clearer sign of where attention is focused to and, therefore, strong attention was additionally evaluated (Supplementary Tab. S1).
The same experimenter (CR) first assessed sheep's attention within each half-second interval of the videos with the sound turned off to be blinded in respect of the exact time when the stimuli were played. Without the sound, the scorer had no information on the side from which the white noise or the dog/sheep vocalizations originated nor about their intensity. In addition, the full sequence of stimuli was scored without specific knowledge of the exact time of the stimuli presentation. All these aspects assured that the observer was biased in scoring as little as possible. The behavior was then re-scored directly with the sound on to ascertain specific behaviors such as writhing, sniffing, eating, and nibbling the table. We consider assessing these half-second intervals as being very close to continuous scoring of the videos and therefore also address the respective data as durations. Intra-observer agreement for sheep's attention was assessed in a total of 30 test sessions. To do so, ten sheep per habituated and naïve batch in the validation, and five sheep per mood group in the reassessment after mood induction were randomly chosen and their attention assessed a second time 3-5 days later by the same experimenter (CR). To quantify intra-observer agreement on weak and strong attention to to the 5° and 30° angles used to define the side (left or right) and intensity (weak or strong) of the attention.   3 min / sheep Increased heart rate and cortisol concentration 9,10,15 , bleating N17 Unfamiliar object (an umbrella was opened once in front of the all sheep of each home pen) 1 time Flight response, increased eye white [16][17][18][19] (b)

N° Positive events (description) Duration
Evidence of positivity P1 P2 Objects regularly added (blue ball, big pink ball, tires) 20,21 Foraging substrates (branches, hay net) 20,21 3 times 3 times Exploration, occupation 21,22 , sniffed the object, ate the foraging materials P3 Gently groomed with a brush or by hand 5 min Stayed near the stockperson, increased heart rate 23-25 , relaxed spine posture, passive or backward ears and semi-closed eyes P4 A familiar human in green clothes gently handled animals (positive tactile contacts, food reward) 20 15 min Stayed calm near the experimenter 25 P5 Sheep on pasture 21 All day long Optimistic-like judgement in horses 21 , run to the pasture, anticipated (alert and bleating) P6 New type of food reward as treats on the passage to and from pasture (apples/carrots, spinach, pasta or butter cakes) 21 4 times Sniffed and ate the food P7 Novel relaxing odor (plastic box containing a compress soaked with essential oil of lavender) 21 All day long Sniffed the plastic box, forward ears 14 (c) Week