Contents lists available at ScienceDirect Applied Animal Behaviour Science journal homepage: www.elsevier.com/locate/applanim The Bovine Zero Maze: Development of a novel fear test for cattle Amanda J. Hubbard, Gordon C. Carstens, Lydia Forehand, Courtney L. Daigle ⁎ Department of Animal Science, Texas A&M University, College Station, TX 77843, United States ARTICLE INFO Keywords: Cattle Behavior Fear Anxiety Zero Maze ABSTRACT Fear tests for cattle are criticized as being 1) designed for animals uncomfortable in open spaces, 2) influenced by previous experience, 3) influenced by context, or 4) subjective. The Elevated Zero Maze (EZM) quantifies fear and anxiety in rodents by exploiting their fear of open spaces and propensity for dark, enclosed spaces. Inversely interpreting the EZM, the Bovine Zero Maze (BZM) was developed to quantify cattle fear and anxiety. Frequency, duration of, and latency to perform steps, escape attempts (EA), direction changes (DC), vocalizations, elim- inations, entry into closed/open sections in the BZM were decoded from video recordings. Heifers (n = 14) were re-tested four weeks later. Repeatability was analyzed for both latency to perform and total number of steps, DC, EA, vocalizations, eliminations, standing bouts, and total time spent standing using Pearson correlation coeffi- cients (PROC CORR). Test re-test reliability was evaluated using intraclass correlation coefficients (ICC). Relationships among ADG, EV, and total number and latency to perform steps, EA, DC, vocalizations, elim- inations, and standing bouts were analyzed using Pearson correlation coefficients (PROC CORR). Relationships between behavior, minute, and test were analyzed using a Generalized Linear Mixed Model (PROC MIXED) and post hoc analyses were conducted with a Bonferroni adjustment. Total number of steps was the only metric that was repeatable across both tests (R 2 = 0.71; P = 0.004) and had moderate reliability (ICC = 0.54). The longer heifers were in the BZM, the frequency of entries into the open and closed sections of the maze and eliminations decreased (all P < 0.01). Heifers vocalized less during the BZM re-test than during the novel BZM test (P = 0.01). Total number of steps made by heifers was impacted by minute (P = 0.006) and heifers tended to take more steps during the second minute of testing (60–120 s) than the seventh minute (P = 0.09). Test × minute affected total number of DC (P < 0.001), however no differences were detected between tests for each minute. Data from this pilot study suggest that behaviors during retesting differ from the novel test and that behaviors change as the test progresses. In addition, the lack of association between behavior in the BZM and EV indicate that the temperament traits evaluated by the BZM may not be related to productivity and may be evaluating different components of temperament than EV. 1. Introduction Quantifying an animal’saffective state is difficult because 1) there are no direct measures of emotion in animals, 2) the animal’s discrete emotions influence their behavior towards an immediate or acute sti- mulus (e.g., predator) that promote immediate survival, and 3) the discrete emotions that motivate the performance of survival behaviors have an underlying valence, or core affect (Mendl and Paul, 2004). Furthermore, unlike humans, animals are unable to verbally self-report emotional experiences (LeDoux and Hofmann, 2018). A fundamental component of good animal welfare, and a primary principle of animal welfare science and philosophy, is the freedom from fear and distress (Brambell, 1965). Thus, to quantify animal welfare, objectively evalu- ating emotions, including fear and anxiety, is of paramount importance. Fear and anxiety are different: fear is the discrete emotional reac- tion to the perception of actual danger, while anxiety is the core affec- tive state to potential danger (Forkman et al., 2007). Animals that per- ceive danger or are in a dangerous situation will express changes both physiologically (e.g., elevated heart rates, increased glucocorticoid le- vels) and behaviorally (e.g., active or defensive avoidance behaviors). Nevertheless, these changes are only indicators that the animal is re- sponding to a stimulus, and cannot serve as a direct measure exclusive to a fearful or anxious state. The seven core affective brain systems are located in the highly integrated regions of the brain are not mutually exclusive, use similar brain structures, and influence one another (Panksepp, 2011). For example, the SEEKING affective system is located in the ventromedial frontal lobe of the brain and the RAGE and FEAR affective systems are located in the lateral and medial temporal lobes of https://doi.org/10.1016/j.applanim.2019.104865 Received 18 February 2019; Received in revised form 20 August 2019; Accepted 27 August 2019 ⁎ Corresponding author. E-mail address: cdaigle@tamu.edu (C.L. Daigle). Applied Animal Behaviour Science 221 (2019) 104865 Available online 31 August 2019 0168-1591/ © 2019 Elsevier B.V. All rights reserved. T