Fig 1.
Experimental setup for visual preference and cross-modal tests.
The cow was positioned centrally between two screens. Each screen showed a video of a person’s face: one familiar and one unfamiliar to the cow. During cross-modal tests, a speaker placed between the screens played the voice of one of the two individuals. Cameras recorded the cow’s behavioural responses throughout the test.
Fig 2.
Following the familiarisation phases, the cows were subjected to a test phase consisting of two trials. Each trial included a visual preference test lasting 8 seconds, during which two faces were simultaneously presented on the two screens: one familiar face and one unfamiliar face. This test was followed by a cross-modal test, also lasting 8 seconds, during which the voice of one of the two individuals was played. The video was considered congruent when the displayed face corresponded to the played voice, and incongruent when the voice did not match the displayed face. A, B, C and D represent different men.
Table 1.
Summary of the results of the generalised mixed models. Gaze durations were analysed with familiarity of the person (visual preference tests) and video congruence (cross-modal tests). Heart rate (HR) measures were analysed with the familiarity of the voice. Significant results are shown in bold.
Fig 3.
First gaze (A, C) and total gaze (B, D) durations.
Figures A and B show results from the visual preference test, according to the familiarity of the human face. Figures C and D show results from the cross-modal test, according to the congruence of the video. Boxplots indicate the median (central line) and the first and third quartiles (box limits). Red crosses represent means. Circular points represent individual values for trial 1, and triangular points represent individual values for trial 2. *: p ≤ 0.05, **: p ≤ 0.01, GLMMs.