Intra- and inter-test consistency of fear tests for adult rabbits

Validating fear tests by measuring intra-test correlation (i.e., repeatability analysis) is disputed because fear tests are often based on novelty (which decreases with re-testing). However, if test-retest scores correlate positively, re-testing could be used to decrease experimental noise due to disturbances or day-to-day mood fluctuations. Alternatively, testing can be repeated by applying different tests measuring the same underlying construct (i.e., fearfulness), which is evaluated by inter-test correlations. In search of a robust set of indicators of rabbit fear, we studied intra- and inter-test Spearman correlations, using 23 adult females. Per rabbit we performed 4 open-field tests (1 test/day, expected to cause fear due to novelty, social isolation and lack of hiding places), 1 novel-object test (directly after the 4th open-field test) and 4 heart rate recordings (directly after the behavioural tests). The number of open-field squares entered was decreased upon ret esting (signed rank test, median day 1: 83(74–95), day 2: 56(35–75), P<0.001), suggesting that decreased novelty during re-testing affected the rabbits’ response. Still, these daily values were highly correlated (rs=0.73, P<0.001). The same was observed for post-test heart rate (day 1: 258(252–267), day 2: 240(225–249) bpm, P<0.001; rs=0.50, P=0.02). Although the number of squares entered and post-test heart rate did not correlate on any day (rs=0.1-0.2, P=0.5-0.8), both correlated with novel-object contact latency. These correlations were stronger when using a 4 day average than when using day 1 values only (squares: rs=-0.61, P=0.002 vs. rs=-0.50, P<0.015; heart rate: rs=-0.45, P=0.03 vs. rs=-0.40, P<0.06), suggesting that re-testing improved the open-field test’s reliability. The results suggest that increased open-field locomotion and heart rate are uncorrelated indicators of decreased fearfulness in adult rabbits. However, the observed correlation between locomotion and novel-object contact latency contrasts with previous results for juvenile rabbits, suggesting that such results are age or context specific.