Festina lente: late-night thoughts on high-throughput screening of mouse behavior.

A recent perspective discussed high-throughput behavioral analysis using mice, giving the overall impression that this area is lagging behind in neuroscience and biomedical research. Not only are we more optimistic about the current state of the art in behavioral neuroscience and its promise, but we also have reservations about whether high-throughput analysis is always an appropriate goal for most behavioral studies. We argue that behavioral studies should be carried out with clear goals and more regard to the intellectual context in which they have developed. In addition, behavioral studies can be performed quite easily, but this does not ensure the required validity or reliability of the particular tests used. Finally, high throughput may not always be an appropriate goal. We discuss the role of automated data collection and unique data-mining algorithms, and the question of the ethological relevance of behavioral tests.