Audio analysis

The playback timing was reviewed in the audio recordings and only those files in which the acoustic stimulus was triggered prior to the terminal buzz were included in the analysis. Bat attack phases were defined based on the inter-pulse interval (IPI) of the echolocation calls as: search phase (IPI longer than 50 ms); approach phase (IPI between 49 and 13 ms) and terminal buzz phase (IPI shorter than 13 ms) (Corcoran et al., 2009; Jones et al., 2018). We selected the first capture attempt of every trial for acoustic analysis. The duration of the terminal buzz phase was determined from the beginning of the first pulse to the end of the last one.

The SSGs were identified during the approach phase of each echolocation sequence and the number of calls per SSG was determined. For each interaction, we calculated the average number of calls produced per SSG and this value was used for further statistical analysis. Following the criteria of Kothari et al. (2014), we defined SSGs as clusters of two or more calls with a relatively constant IPI between them (within 5% error with respect to the mean IPI of the sound group) that are flanked by longer IPI (at least 1.2 times the mean interval within the call cluster). In this case, only those audio files in which the acoustic stimulus was triggered at the end of the echolocation search phase and continued during the approach phase were used for the analysis. We also classified each emitted SSG into one of three categories depending on their number of calls: doublets, triplets and multi-call. Doublets and triplets contained two and three pulses, respectively, within one strobe. The multi-call category included those SSGs with four or more calls in it. The number of SSGs produced per category for every trial, in response to different acoustic stimulations, was measured.

All audio recordings were analyzed in Avisoft SASLab Pro v5.2. We used the Automatic Parameter tool to automatically identify the acoustic signals in the audio recordings. The echoes of individual calls, the second and third harmonic and the stimuli were manually removed to facilitate automatic detection.