Novelty at second glance: A critical appraisal of the novel-object paradigm based on meta-analysis

The study of consistent individual differences in behaviour has become an important focus in research on animal behaviour. Behavioural phenotypes are typically measured through standardized testing paradigms and one frequently used paradigm is the novel object test. In novel object tests, animals are exposed to new (unknown) objects and their reaction is quantified. When repeating trials to assess the temporal consistency of individual differences, researchers face the dilemma of whether to use the same or different ‘novel’ objects, since the same stimulus can result in habituation, while exposure to different objects can result in context-dependent responses. We performed a quantitative assessment of 254 effect sizes from 113 studies on novel-object trials to evaluate the properties of this testing paradigm, in particular the effect of object novelty and time interval between novel-object trials on estimates of individual consistency. We found an increase of sample sizes and an increase of estimates of repeatabilities with time. The vast majority of short-term studies (one month) used either the same or different novel objects about equally often. The average estimate for individual consistency was r = 0.47 (short-term r = 0.52, long-term r = 0.44). Novelty, time interval between trials and their interaction together explained only 3% of the total heterogeneity. Overall, novelobject trials reliably estimate individual differences in behaviour, but results were very heterogeneous even within the same study species, suggesting susceptibility to unknown details in testing conditions. Most studies that measure novel-object responses in association with food label the trait as neophobia, while novel-object trials in a neutral context are labelled variously as boldness/shyness, exploratory behaviour or neophobia/neophilia. Neophobia/neophilia is also the term most specific to novel object presentations. To avoid ambiguity, we suggest object neophobia/neophilia as the most specific label for novel-object responses.