Acoustic communication in laboratory mice is a relatively recent subject of experimental study, often yielding disparate findings. For example, researchers often manually place mouse ultrasonic vocalizations (USVs) into categories based on spectrotemporal characteristics, but the numbers and types of categories differ widely between laboratories. Here, we attempt to determine what cues CBA/CaJ mice use to discriminate between vocalizations by testing them in an operant conditioning paradigm. The mice were trained to discriminate a repeating background containing one USV from several target USVs. The targets were different call types used by Holmstrom et al. (2010) and manipulations of the background calls, such as removing the frequency modulation, shifting the entire call up or down in frequency, shortening or lengthening the call, or reversing the entire call. Results show that large frequency shifts were easy for the mice to discriminate, while reversing the calls and removing the frequency modulation were much more difficult. For most calls, similarity in spectrotemporal characteristics yielded poor discrimination performance. These results are the first to show that mice can discriminate between some vocalizations but not others, and that they may place different meaning to different call types, though not necessarily the call types designated by humans.

This content is only available via PDF.