The utility of behavioral studies for aquatic toxicology testing: a meta-analysis

Abstract

Behavioral responses have been applied for decades as tools for aquatic toxicity testing, but have received far less attention than studies assessing lethality, development or reproduction. With improved visual and non-visual assessment tools and increased knowledge of the importance of behavior for organism health and fitness, interest in behavioral analysis has increased in recent years. However, to our knowledge there has never been a quantitative assessment of the available techniques for organismal toxicity testing, so it is not clear whether behavioral studies represent valuable additions to environmental monitoring. We performed a meta-analysis comparing the relative sensitivities and average durations of behavioral studies to those assessing acute lethality, development and reproduction. Results demonstrate that the average duration of behavioral studies is consistently less than developmental or reproductive studies, and that behavioral endpoints are generally more sensitive than those assessing development or reproduction. We found effect sizes to be lower but power to be higher in behavioral and reproductive studies compared to studies assessing development, which likely relates to low sample sizes commonly used in developmental studies. Overall, we conclude that behavioral studies are comparatively fast and sensitive, and therefore warrant further attention as tools for assessing the toxicological effects of environmental contaminants. We suggest that research aimed at developing and optimizing techniques for behavioral analysis could prove extremely useful to the field of toxicology, but that future work must be directed at determining what specific behaviors are most sensitive to various classes of contaminants, and at understanding the relevance of changes to discrete behaviors for influencing organismal and population-level health and fitness.

Keywords: Aquatic toxicology; Behavioral analyses; Experimental methodology; Meta-analysis; Sensitivity.